Summary of findings
Pruritus, derived from the Latin word prurire, which means to itch, is defined as "an unpleasant sensation associated with the desire to scratch". This definition of pruritus was introduced in 1660 by the German physician Samuel Hafenreffer (Haffenreffer 1660; Misery 2010; Proske and Hartschuh 2010). In modern medicine, the term pruritus is generally used to refer to a pathological condition in which the sensations of itch are intense and often generalized and trigger repeated scratching in an attempt to relieve the discomfort. Pruritus is not a disease, rather it is a common and still poorly understood symptom of both localized and systemic disorders that may accompany many conditions (Bernhard 2005; Summey 2009; Zylicz 2004). Pruritus is a prevalent symptom in many skin conditions. However, much less is known about pruritus that is not associated with primary skin disease. This latter problem is of major relevance to many medical specialities, and notably palliative care. Pruritus or itch is not the most prevalent but is one of the most puzzling symptoms in advanced incurable diseases and can cause considerable discomfort in patients receiving treatment for cancer and patients with non-malignant terminal illnesses. In addition to a social embarrassment, the itch-scratch-itch cycle damages skin integrity, decreases resistance to infections, and impairs quality of life in a similar way to pain.
Prevalence of pruritus
Pruritus and malignant diseases
Pruritus may be associated with virtually any malignancy (Chiang 2011). Some neoplasms, particularly haematologic malignancies, are frequently associated with pruritus. Among patients with polycythaemia vera, 48% to 70% of patients have aquagenic pruritus. Pruritus is seen in about 30% of patients with Hodgkin's disease (Krajnik 2001a). The incidence and significance of pruritus in other lymphomas and leukaemia are unknown, but it has been reported in approximately 3% of patients with non-Hodgkin's lymphoma (Lober 1988). Solid tumours can be associated with paraneoplastic pruritus, which in fact might be a presenting symptom that precedes the diagnosis by months or years. The pathophysiology is not well understood, but it appears to involve an immunologic reaction to tumour-specific antigens (Seccareccia 2011). Pruritus is also frequent in cutaneous lymphomas (Ahern 2012). Additionally, it is a common symptom in malignancies of the biliary tract. Retrospective studies have revealed that malignant diseases are present in 2% to 11% of chronic itch cases (Weisshaar 2009).
Pruritus and nonmalignant internal diseases
Many internal diseases other than cancer may be associated with pruritus. Pruritus has been reported to herald the onset of thyroid disease, renal insufficiency, liver disease, iron deficiency, diabetes mellitus, paraproteinaemia, Sjögren's syndrome, and other conditions. Pruritus is a frequent symptom in end-stage renal disease (ESRD). In internal diseases, itch has been best studied in hepatic diseases and uraemic diseases (Metz 2010; Wang 2010; Weisshaar 2009). About one third of uraemic patients treated without dialysis exhibit uraemic pruritus, and on maintenance haemodialysis, the incidence of uraemic itching increases to 70% to 80% (Manenti 2009; Narita 2006). Hepatogenic or cholestatic pruritus affects 100% of patients with biliary cirrhosis and is the initial symptom in almost half of the patients with this disease (Bergasa 2008). Furthermore, the prevalence of pruritus in patients with end-stage HIV is over 20% (Smith 1997a; Uthayakumar 1997).
Pruritus in palliative care in general
In advanced diseases, as seen in palliative care units, the prevalence of severe pruritus is not too high, but pruritus is a distressing symptom for palliative care patients and may be difficult to manage. A specific problem in palliative medicine involves systemic pruritus in terminal illnesses (Twycross 2001; Twycross 2004) because pruritus is often a result of changing organ functions in this phase of illness. In this case, the itch is multifactorial, associated with both liver and kidney function deterioration and increased anxiety (Yosipovitch 2003). Additionally, in the field of palliative care, pruritus is a well-known adverse effect of opioid administration. Even though the incidence is low (approximately 1% after systemic administration), pruritus as an adverse effect must be kept in mind (Krajnik 2001).
Description of the condition
In the field of palliative care, pruritus is a symptom occurring in patients with disparate underlying diseases that is based on different pathologic mechanisms but ending in the same phenomenon. The pathogenesis of pruritus is complex and not fully elucidated, but it is known that central and peripheral nerves and specific brain regions are involved (Langner 2009). For a long time, itch was regarded as a variant of pain; however, the neural transmission associated with pruritus follows distinct neuronal pathways and causes unique sensations (Ikoma 2006; Schmelz 1997). The pathogenesis of itch is diverse and involves a complex network of cutaneous and neuronal cells. Mediators of pruritus presumably act on nerve fibres or lead to a cascade of mediator release, resulting in nerve stimulation and the sensation of pruritus. The group of potential chemical mediators is large and is steadily increasing. It contains amines (for example histamine, serotonin), proteases (for example tryptases), neuropeptides (for example substance P (SP), calcitonin gene-related peptide (CGRP), bradykinin), opioids (for example morphine, beta-, met-, leu-enkephalin), eicosanoids, growth factors, and cytokines (Weisshaar 2003). In recent years, more and more itch-specific mediators and receptors, such as interleukin-31, gastrin-releasing peptide receptor or histamine H
Description of the intervention
Due to the complex physiology of pruritus, many of the underlying mechanisms are still poorly understood (Krajnik 2001a). Modulations by serotonergic and enkephalinergic systems take place on all levels of the 'pruritus tract'. Additionally, opioid receptors seem to be of particular importance, which is not surprising considering the involvement of almost identical mediators for pain and itch (Moore 2009). This implies that various completely different pathologic mechanisms may form the basis of pruritus, making it difficult to find an effective medication for managing pruritus. Until now, no universally valid therapy concept has been developed. For the treatment of pruritus, the efficacy of several different substance classes have been postulated and researched. Treatment options researched and included in this review comprise antiphlogistic substances, psychotropic drugs, antagonistic drugs, anaesthetics, adsorbent substances, and topical treatments.
How the intervention might work
Of the mediators that trigger pruritus, histamine is the best known and most thoroughly researched. Preformed histamine is present in large amounts in mast cell granules. For this reason, after cell activation, it can be immediately released into the surrounding area where it can induce pruritus via H
Opioid receptor antagonists and (partial) agonists
Opioid receptor antagonists were originally developed for the treatment of heroin addiction and for symptom reversal of postanaesthetic depression, narcotic overdose, and opioid intoxication (Gowing 2010; Rösner 2010). Clinical and experimental observations have demonstrated that pruritus can be evoked or intensified by endogenous or exogenous opioids (Metze 1999; Metze 1999a). This phenomenon can be explained by the activation of spinal opioid receptors, mainly μ-opioid receptors on pain transmitting neurons, which often induce analgesia in combination with pruritus. Thus, reversing this effect through μ-opioid antagonists results in the inhibition of pruritus (Ständer 2008).
Pruritus sensations may arise from the superficial layers of the skin, which contain clustered nerve endings at 'itch points' close to the dermoepidermal junction, as well as the mucous membranes and conjunctiva (Krajnik 2001; Yosipovitch 2003). These receptors may be acted upon directly by physical or chemical stimuli, or indirectly via histamine release. Itch impulses are transmitted through the C fibres of polymodal nociceptors to the dorsal root ganglia, where a synapse occurs with secondary neurons. Efferents traverse to the contralateral spinothalamic tract and pass to the posterolateral spinothalamic tract, the posterolateral ventral thalamic nucleus, and then to the somatosensory cortex of the postcentral gyrus (Mela 2003). 5-Hydroxytryptamine (5-HT; serotonin) is an important neurotransmitter in these pathways (O'Donohue 2005). Ondansetron is one of a group of drugs that act as antagonists at 5-HT
Serotonin reuptake inhibitors and antidepressants
Selective serotonin reuptake inhibitors (SSRI) like sertraline and paroxetine play an increasingly important role in the management of pruritus (Balaskas 1998; Larijani 1996; Raap 2012; Schworer 1995; Tandon 2007; Tennyson 2001; Wilde 1996; Ye 2001; Zylicz 1998). They are believed to increase the extracellular level of the neurotransmitter serotonin by inhibiting its reuptake into the presynaptic cell and increasing the level of serotonin in the synaptic cleft that is available to bind to the postsynaptic receptor. They have varying degrees of selectivity for the other monoamine transporters, with pure SSRIs having only weak affinity for the noradrenaline and dopamine transporters.
Antiepileptics are used to prevent or reduce the severity and frequency of seizures (Duley 2010; Ratilal 2005; Wiffen 2011). Gabapentin, a γ-aminobutyric acid analogue, was originally developed as an antiepileptic and is speculated to hinder the transmission of nociceptive sensations to the brain and thereby also suppress pruritus (Ständer 2008).
Rifampicin is an antibiotic that induces detoxicating hepatic enzymes and competitively inhibits the reuptake of bile acids by hepatocytic transporters (Trauner 2005). Some hypothesize that rifampicin might influence pruritus by changing the bacterial growth in the intestines, which can influence the reabsorption of pruritrogens.
Thalidomide is a drug that modifies or regulates the immune system and has anti-inflammatory properties. It is used as an immunomodulator to treat graft versus host reactions. It suppresses tumour necrosis factor alpha (TNF-α) production and leads to a predominant differentiation of Th2 lymphocytes with suppression of interleukin-2 (IL-2) producing Th1 cells (McHugh 1995; Mettang 2010). The antipruritic action of this drug may be secondary to inhibition of TNF-α. Another possibility is that thalidomide can act as both a peripheral and a central nerve depressant (Moretti 2010).
Flumecinol (3-trifluoromethyl-alpha-ethylbenzhydrol), a benzhydrol derivative, has been reported to induce microsomal drug metabolising enzymes (Turner 1990). Flumecinol also lowers serum bilirubin in Gilbert's syndrome, possibly by inducing bilirubin undine diphosphate (UDP)-glucuronyltransferase. It may, therefore, induce a range of enzymes similar to phenobarbitone and rifampicin.
Colestyramine is an intestinally active anion exchange resin. It interrupts the enterohepatic circulation of bile acids and has been used for many years to relieve pruritus in cholestatic disorders (Datta 1966; Sharp 1967). Another bile acid sequestrant also used for the treatment of pruritus is colesevelam.
Cromolyn sodium is a drug that blocks mast cell degranulation in response to antigens, leading to decreased release of histamine, leukotrienes, and other inflammatory mast cell products. It is hypothesized that mediators released from mast cells are most likely to be responsible for chronic kidney disease-associated pruritus. Another hypothesis is that cromolyn sodium may decrease the severity of pruritus via decreasing serum tryptase levels.
Leukotriene antagonists prevent the inflammatory response produced by leukotrienes (Watts 2012).
Erythropoietin is a hormone produced naturally by the kidneys that stimulates the production of red blood cells by the bone marrow. It is hypothesized that erythropoietin may have an antipruritic effect related to a lowering effect of the hormone on plasma histamine concentrations (Bohlius 2009).
Activated charcoal is an agent that can bind many poisons in the stomach and therefore prevent them from being absorbed. Charcoal has also been shown to be effective in uraemic pruritus (Giovanetti 1995; Yatzidis 1972).
Capsaicin is the prototype of topical antipruritic agents that target the transient receptor potential (TRP) gene family of ion channels that respond to physical activation (heat, cold), protons (pH changes), or biological mediators (for example prostanoids) and counteract itch via activating pain neurons (Derry 2012; Derry 2013; Steinhoff 2011). Capsaicin (trans-8-methyl-N-vanillyl-6-nonenamide) is an alkaloid naturally found in many botanical species of the nightshade plant family (Solanacea).
Tacrolimus is an immunosuppressant used for the prevention of rejection in transplant patients. It suppresses the differentiation of Th1 lymphocytes and the ensuing IL-2 production (Suthanthiran 1994; Webster 2005).
Photodynamic therapy, transcutaneous electrical nerve stimulation (TENS), and other non-pharmaceutical therapies for pruritus will be assessed in a systematic review conducted separately by our research group.
In conclusion,: pruritus is a frequent and distressing symptom. The medical literature is full of recommendations for the management of pruritus, but it contains only a few clinical trials and evidence-based data. There are some reviews about pruritus in general (Winkelmann 1964; Winkelmann 1982), about dermatologic causes of pruritus (Fransway 1988), pruritus in systemic diseases (Kantor 1983; Summey 2005), and pruritus related to specific causes like cholestasis or uraemia (Khandelwal 1994; Szepietowski 2004). There is also some literature on the management of pruritus in palliative care patients (Krajnik 2001a). However, until now, there is no literature systematically reviewing the effectiveness of different medical interventions on pruritus in the field of palliative care.
Why it is important to do this review
Pruritus or itch is one of the most puzzling symptoms in advanced incurable diseases, and can cause considerable discomfort in patients. Nevertheless, it is a kind of Cinderella symptom, tucked away and hidden behind more fashionable symptoms such as pain. As already explained, pruritus is multifactorial in origin and can be a symptom of diverse pathophysiologies. Particulary over the last decade, clinical observation and controlled trials have done much to aid the understanding and treatment of pruritus, especially in liver disease and uraemia and also in other kinds of chronic pruritus. Therefore, this review aimed to systematically collect and review the evidence for adequate treatment of pruritus in the field of palliative care, to put this symptom into perspective, and to make new therapy strategies accessible for clinicians and patients (Wee 2008).
The objective was to determine the efficacy of pharmacological treatment (systemic and topical) for preventing or treating pruritus in adult palliative care patients. The focus was on patients who were in an advanced stage of their disease, including cancer, hepatic and renal failure, HIV/AIDS, and other non-malignant terminal illnesses.
Criteria for considering studies for this review
Types of studies
We have considered full reports concerning pruritus in patients with advanced diseases with a focus on pharmacological treatment. The primary outcome of the studies had to be pruritus. Only randomised controlled trials (RCTs) in adult humans were included. We defined 'randomised' as studies which were described as 'randomised' by the authors anywhere in the manuscripts. Both published and unpublished studies were eligible for inclusion.
Contrary to our first considerations in the protocol, we did not include controlled clinical trials (CCTs).
Types of participants
Defining the population for systematic reviews in palliative care has been identified as a problem in previous reviews. Therefore, we have drawn upon the definition "adult patients in any setting, receiving palliative care or suffering an incurable progressive medical condition", which has been used previously in other Cochrane reviews (Dorman 2010; Perkins 2009). Studies eligible for this review included participants:
- suffering from pruritus combined with an incurable advanced malignant or non-malignant disease such as advanced cancer, HIV/AIDS, renal failure, liver failure, or others;
- aged 18 years or older; and
- of both sexes.
Since many of the studies that came into consideration also included patients who were not necessarily in advanced stages of their disease, and were not palliative care patients, we decided to define comprehensible criteria for the patients included in this systematic review.
- We included all patients who were described as palliative care patients or as patients in advanced stages of malignant or non-malignant diseases.
If no detailed information on the stages of the underlying disease was given, the following patients were considered as patients with palliative care needs.
- Chronic kidney disease-associated, renal, or uraemic pruritus: all patients suffering from end-stage renal disease and patients in need of haemodialysis.
- Cholestatic or hepatogenic pruritus: all patients suffering from primary biliary cholestasis or primary sclerosing cholestasis and all patients who were described as being in an advanced stage of the disease. If patients with different kinds of cholestatic pruritus were included in the studies, only studies which included more than 75% of patients with primary biliary cholestasis, primary sclerosing cholangitis, or patients described as being in advanced stages of their disease were eligible for the systematic review.
- HIV-associated pruritus: all patients with pruritus associated with HIV were included.
- Pruritus associated with malignancies: all patients in advanced stages of cancer (with metastases or described as in an advanced stage of the disease) were included.
We excluded studies researching patients with pruritus related to acute or chronic cholestasis, acute or chronic dermatological diseases, or acute medical or surgical interventions. Furthermore, primarily dermatological diseases or infections were not included.
Types of interventions
We only included studies that measured pruritus as a primary outcome. Identified studies had to evaluate and report the effect of a pharmacological treatment on pruritus compared with a placebo, no treatment, or an alternative treatment. We included studies where any pharmacological medication was used to treat pruritus, regardless of dosage, route of medication, or the duration of follow-up. Internal as well as external application were considered eligible for inclusion in the review. We did not focus on pharmacological interventions targeting the treatment of underlying diseases but rather on pharmacological interventions for treating pruritus as an accompanying symptom of advanced diseases. Complementary medical interventions and non-pharmaceutical treatments such as photodynamic therapy or TENS were excluded and will be evaluated in a separate review, which is being conducted by our palliative care research group.
Types of outcome measures
Primary outcomes included subjective measurements of pruritus on validated and reliable scales, such as:
- uni-dimensional scales (e.g. visual analogue scales (VAS), numeric rating scales (NRS), categorical scales); or
- patient-reported pruritus according to non-validated pruritus scores (e.g.1 to 3 or 1 to 4), which was substituted by estimations by nursing or medical staff if self-assessment was not possible.
Secondary outcomes included:
- quality of life,
- patient satisfaction,
- adverse effects.
Search methods for identification of studies
There were no language restrictions for either the searching strategies or study inclusion.
We identified studies by searching the following databases:
- MEDLINE via Ovid (including MEDLINE In-process and other non-indexed citations) (1950 to August 2012);
- The Cochrane Library via Wiley (including Cochrane Database of Systematic Reviews (CDSR), Cochrane Central Register of Controlled Trials (CENTRAL), Database of Abstracts of Reviews of Effects (DARE)) (August 2012);
- EMBASE via Ovid (including EMBASE Alert) (1980 to August 2012);
- BIOSIS previews via Ovid and via Web of Knowledge, respectively (1969 to August 2012);
- CINAHL via EBSCOhost (1982 to August 2012);
- PsycINFO via EBSCOhost (1806 to August 2012);
- Cochrane Pain, Palliative and Supportive Care Trials Register (August 2012).
We also searched the following registers of ongoing trials:
- Current Controlled Trials at http://www.controlled-trials.com (August 2012);
- WHO International Clinical Trials Registry Platform (ICTRP) at http://www.who.int/trialsearch (August 2012);
- National Institute of Health Randomized Trials Records at http://clinicaltrials.gov (August 2012).
A search strategy was developed for use in MEDLINE and revised appropriately for other databases, after consultation with the Cochrane Pain, Palliative Care and Supportive Care Cochrane Review Group. The subject search used a combination of controlled vocabulary and free text terms based on the search strategy for MEDLINE via Ovid (please see Appendix 1 for the MEDLINE search strategy in Ovid and Appendices 2 to 5 for the other search strategies used for this review [Appendix 2, Appendix 3, Appendix 4, Appendix 5, Appendix 6]).
The Cochrane highly sensitive search strategy (CHSSS) for identifying randomised trials in MEDLINE, a sensitivity maximizing version as referenced in Chapter 188.8.131.52 and detailed in box 6.4.c of The Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (Higgins 2011), was used to identify randomised controlled trials. A similar study design filter was used for other databases, as appropriate.
We included all suitable studies which were published until January 2012. Suitable studies published from February to August 2012 have not been included in this review. They are cited in the chapter 'Studies awaiting classification'.
Searching other resources
Reference lists from retrieved publications and key websites were screened for additional publications and for further relevant studies. We also checked the reference lists of relevant review articles on the subject (Gooding 2010; Khurana 2006; Phan 2010; Tandon 2007; Weisshaar 2003). In addition, we handsearched the reference lists of the following key textbooks: Oxford Textbook of Palliative Medicine (Doyle 2003; Hanks 2009), Oxford Textbook of Palliative Nursing (Ferrell 2010), Textbook of Palliative Medicine (Bruera 2009),Management of Advanced Disease (Sykes 2004), Palliative Care Formulary 3 (Twycross 2007, Oxford Handbook of Palliative Care (Watson 2009), Palliative Medicine: A case based manual (MacDonald 2005), Pruritus in Advanced Disease (Zylicz 2004), Pruritus (Zylicz 2009).
In an effort to identify further published, unpublished, and ongoing trials, we searched the bibliographies of relevant trials and reviews, contacted experts in the field, and searched available proceedings of conferences such as the meetings of the American Society of Clinical Oncology (ASCO), the American Society of Haematology (ASH), the European Haematology Association (EHA), and the European Association of Palliative Care (EAPC).
Data collection and analysis
Selection of studies
All titles and abstracts of studies identified by the search strategies were screened for relevance by two review authors (CX, JM). Disagreement was resolved by consensus and after discussion with a third review author (GB). If it was not possible to accept or reject a study with certainty, we obtained the full text of the study for further evaluation. Three independent review authors assessed the full text of all potentially relevant studies in accordance with the above inclusion criteria (CX, JM, DG). Any differences in opinion at this stage were resolved by consensus and discussion with a fourth review author (GB). We kept a record of all excluded studies and the reasons for exclusion.
Data extraction and management
Data from the selected studies were extracted independently by at least two review authors (CX, JM or DG) using a standardised coding form. Differences in data extraction were discussed and the input of a third review author was sought as necessary. The data extraction form, specifically designed for the review, included the following.
Study ID and publication details
- Study aim
- Study design (randomised, not randomised, controlled, prospective etc.)
- Primary and secondary outcomes
- Type of control group
- Number of participants in each group
Quality of the study
- Randomisation procedure
- Concealment of treatment allocation
- Details of blinding
- Per protocol analysis or intention-to-treat analysis
- Number of withdrawals described
- Management of missing data
- Follow-up data
- Details of analysis
- Status or course of disease
- Type and stage of treatment
- Type of pruritus
- Drug characteristics
- Duration of therapy
- Pharmacological regimen of drug treatment with the drug of interest (dose, frequency of application)
- Description of placebo
- Description of alternative treatment
- Description of additional non-pharmacological techniques if additionally used during similar regimens
- Primary outcome, including the measurement of pruritus (mean, standard deviation) and the change in level of pruritus
- Secondary outcomes, including the measurement of quality of life, patient satisfaction, measurement of depression, adverse effects of treatments
- Patient narrative comments etc.
Authors of studies were contacted, if possible, to provide unpublished data if required for analysis.
Assessment of risk of bias in included studies
The risk of bias assessment for RCTs was performed as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) (Review Manager 2011). Three independent review authors (CX, JM, DG) assessed the quality of included studies using the Cochrane risk of bias tool (Higgins 2011) (Review Manager 2011).
Random sequence generation:
- low risk, every participant had an equal chance to be selected for either treatment and the investigator was unable to predict which treatment the participant would be assigned to;
- unclear risk, no information given;
- high risk, for example, randomisation by date of birth or date of admission.
- low risk, methods to conceal allocation included central randomisation, serially numbered, opaque, sealed envelopes, or other descriptions with convincing concealment;
- unclear risk, authors did not adequately report the method of concealment;
- high risk, investigators enrolling participants could possibly foresee assignments because of the use of high risk methods to conceal allocation, such as an open random allocation schedule (e.g. a list of random numbers), assignment envelopes without appropriate safeguards (e.g. if envelopes were unsealed, nonopaque, or not sequentially numbered), alternation or rotation, date of birth.
Blinding of participants and personnel:
- low risk, blinding of participants and providers stated and unlikely that the blinding could have been broken;
- unclear risk, blinding not adequate, but the outcome measurement is not likely to have been influenced by lack of blinding;
- high risk, no blinding or incomplete blinding, and the outcome or outcome measurement is likely to have been influenced by lack of blinding.
Blinding of outcome assessors:
- low risk, blinding of providers and outcome assessor stated and unlikely that the blinding could have been broken;
- unclear risk, blinding of outcome assessment not adequate, but the outcome measurement is not likely to have been influenced by lack of blinding;
- high risk, no blinding or incomplete blinding of outcome assessment, and the outcome or outcome measurement is likely to have been influenced by lack of blinding.
Incomplete outcome data:
- low risk, no missing outcome data, or reasons for missing outcome data are unlikely to be related to true outcome;
- unclear risk, insufficient information to permit judgement;
- high risk, reasons for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups, or 'as-treated' analysis done with substantial departure of the intervention received from that assigned at randomisation.
Selective outcome reporting:
- low risk, reports of the study free of selective outcome reporting;
- unclear risk, insufficient information to permit judgement;
- high risk, reports of the study suggest selective outcome reporting.
Other sources of bias:
- low risk of bias, the trial appears to be free of other components that could put it at risk of bias;
- uncertain risk of bias, the trial may or may not be free of other components that could put it at risk of bias;
- high risk of bias, there are other factors in the trial that could put it at risk of bias, e.g. for-profit involvement, authors have conducted trials on the same topic, etc.
The review authors were blinded to each others' assessments. Any disagreements were resolved by discussion. No study was automatically excluded as a result of a rating of 'unclear' or 'high risk' or based on a low quality score. Trials assessed as having 'low risk of bias' in all of the specified individual domains were considered to be 'trials with low risk of bias'. Studies having 'uncertain' risk of bias in one or more of the specified individual domains were considered as 'trials with unclear risk of bias'. Finally, studies assessed as having 'high' risk of bias in one or more of the specified individual domains were considered 'trials with high risk of bias'.
Measures of treatment effect
Measurement of itch
Pruritus measurement is problematic because of its subjective nature and poor localization. In addition, itch has multidimensional aspects (for example severity, duration, frequency, spatial distribution, and quality). Although several authors have suggested that VAS are subjective and represent an inadequate and unreliable method of assessing pruritus (Jones 1999), more sophisticated and objective methods pose several practical difficulties. Since the main goal of pruritus treatment is to improve patients´ well-being and quality of life, this can only be measured subjectively as pruritus is primarily based on the subjective perception of the patient. Therefore, for the direct evaluation of itch, we have to rely on the patients’ own ratings of their subjective symptoms and on the assumption that the participant is able to relate their experiences accurately. In the studies included in this review different scales were used for the scoring of pruritus (see Table 1).
Categorical scales are commonly employed research scales and consist of discrete divisions of the frequency of the measured dimension. For example, pruritus: never, rarely, occasionally, frequently, always (an example is the pruritus score by Duo 1987 or Mettang 2002).
Continuous scales like the VAS or numeric rating scale (NRS) are a line of defined length (in most studies 100 mm) with descriptive anchors at the extremes, for example 'no pruritus' and 'pruritus as bad as it can be imagined'. Since the VAS is validated (Reich 2008), simple, accurate, and supposedly the most sensitive approach to measuring pruritus intensity, it is probably the most commonly used scale in pruritus research (Wallengren 2010; Weisshaar 2003).
Another approach to measurement of pruritus is scratching behaviour measurement. In contrast to pruritus, scratching activity can be quantified. Scratching activity may be evaluated by report of scratching behaviour, for example with hand-activated counters to record scratching (Melin 1986). Since this method is not suitable for recording nocturnal scratching, other methods may be used to observe nocturnal scratching; for example nocturnal bed movement measured by a vibration transducer on one of the legs of the bed, and limb or forearm activity measured by movement-sensitive meters. Nocturnal scratching may also be observed by infrared videotaping or by direct observation during the night.
One instrument for recording daily and nocturnal scratching is the pruritometer, which processes the signals of a piezoelectric vibration sensor fixed on the middle finger of the patient's dominant hand and sent to a counter worn by the patient like a wristwatch (Wallengren 2010).
Scores and measurement of scratching may also be determined through questionnaires asking for more information regarding the pruritus, for example 'The Worcester Itch Index' or 'The Eppendorf Itch Questionnaire' (Weisshaar 2003).
In this systematic review, treatment effect was evaluated as measured by estimations of nursing or medical staff if self-assessment was not possible.
Since no gold standard concerning treatment or improvement of pruritus exists, we considered a reduction of pruritus symptoms by 30% as moderate and a reduction by 50% as substantial, assuming that there were no other specifications given in the studies. This is consistent with the recommendations of Turk et al, as introduced in the IMMPACT recommendations (Turk 2008).
Quality of life, patient satisfaction, depression, and adverse effects were recorded as secondary outcomes.
For measuring quality of life, the following scales or methods were used.
Quality of life scores:
- Short Form 36 and Liver Disease Symptom Index 2.0 (Kuiper 2010);
Patient satisfaction was measured by a seven-point scale, where 0 meant indifferent, a value of -3 meant extremely poor, and a value +3 meant excellent (Zylicz 2003).
Depression was surveyed by using:
- the Hamilton depression rating scale (includes items intrinsic to medical conditions (i.e. fatigue, sleep) and concern about health) (Bergasa 2006);
- the Structured Clinical Interview Questionnaire (SCID) for DSM IV, Axis I Disorders (a measure for the diagnosis of depression and anxiety syndromes) (Bergasa 2006); and
- the 30-item Inventory of Depressive Symptomatology-Self-Report (IDS-SR
30) (Mayo 2007).
Adverse effects were reported for every study included in the review.
Unit of analysis issues
The data of the RCTs were evaluated. Identified studies had to evaluate and report the effect of a pharmacological treatment on pruritus compared with a placebo, no treatment, or an alternative treatment in individuals. Cross-over studies were also included, considering specific challenges such as possible carry-over effects.
Dealing with missing data
We did not impute missing outcome data. Data were analysed on an endpoint basis, including only participants for whom final data were obtained. It was not assumed that participants who dropped out after randomisation had a negative outcome.
Assessment of heterogeneity
We investigated heterogeneity using visual inspection of the forest plots as well as the I
Assessment of reporting biases
There were insufficient studies in each of the meta-analyses to assess reporting bias. We had planned funnel plots corresponding to meta-analysis of the primary outcome to assess the potential for small study effects such as publication bias.
RevMan 5.1 (Review Manager 2011) was used for data entry, statistical analysis, and creation of graphs. Each drug class was analysed separately and compared with its respective control group or alternative intervention. Most outcomes in this review were presented as continuous variables. Continuous outcomes, including the mean change in pruritus score between treatment and placebo, were reported either as mean difference (MD) or standardised mean difference (SMD) with 95% confidence interval (CI) depending on whether results were reported on the same or different scales. We anticipated that some individual studies would have used final scores and others would use change scores and even analysis of covariance in their statistical analyses of the results. In this case, we combined these different types of analysis as MD. The fixed-effect model was used in all meta-analyses.
We decided not to pool the results in cases of significant clinical heterogeneity. We calculated 95% CI for each effect size estimate.
The following treatment comparisons were made:
- naltrexone versus placebo;
- nalfurafine versus placebo;
- ondansetron versus placebo;
- gabapentin versus placebo;
- rifampin versus placebo;
- capsaicin versus placebo;
- patient satisfaction (secondary outcome).
Studies with parallel groups and cross-over design were included in the review. Data from cross-over trials were handled according to the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions section 16.4 (Higgins 2011). If all necessary data were provided in the publications of cross-over trials, we included the results of a paired analysis in the meta-analyses. If the required data were not provided, we included only data from the first period of the cross-over trial (if available) and thus treated this trial as a parallel group trial.
Subgroup analysis and investigation of heterogeneity
The following subgroup analyses were determined a priori and were performed if possible:
- chronic kidney disease (CKD)-associated (chronic kidney disease patient (CKDP)) or uraemic pruritus (UP) versus cholestatic pruritus (CP);
- parallel group versus cross-over study design.
Subgroup analysis was conducted as recommended in the Cochrane Handbook for Systematic Reviews of Interventions section 9.6 (Higgins 2011).
Sensitivity analysis was performed to assess whether the quality of the chosen trials influenced the results of the meta-analysis or whether analysis by the fixed-effect or the random-effects model changed the results.
Due to the small numbers of studies for a single comparison, sensitivity analyses based on quality criteria were not conducted.
Description of studies
Please see the Characteristics of included studies table for full information on the included studies.
Results of the search
The database searches and handsearches of conference proceedings, journals, and reference lists resulted in 883 unique citations. Two review authors (EM, SB) searched the databases in June 2010, and one review author updated the search in November 2011 and January 2012. Overall, the literature search yielded a total of 771 citations. Furthermore, we identified one additional study by handsearching the reference lists of the original studies, textbooks, and websites. Personal contact with several investigators did not yield any additional studies or unpublished data. We excluded 144 duplicates. The titles and abstracts of 628 studies were evaluated and relevant studies were selected if they met the inclusion criteria. At this stage, 546 studies were excluded and sorted into the following exclusion groups:
- reviews (27 reports),
- different design (non-RCT) (116 reports),
- different patient population (178 studies),
- pruritus not as primary outcome (97 studies),
- pharmacokinetic study (54 studies),
- non-pharmacological interventions (32 studies),
- others (e.g. letters) (42).
All articles in languages other than English had abstracts available in English. We excluded studies on the basis of the information contained in the abstracts if eligibility could be assessed at this juncture. One study (written in Turkish) required translation. We identified 82 studies as being potentially eligible and subsequently obtained a full copy of each paper for more detailed evaluation. Following an examination of the full text and, in some cases, further contact with the authors of the studies, we found 38 papers (reporting on 40 studies) that met the eligibility criteria. Forty-four studies were excluded for the reasons listed below. Furthermore, it was not possible to ascertain the eligibility of two papers (one abstract and one letter) despite extensive efforts to obtain these reports.
Reasons for exclusion:
- not an RCT (32 studies),
- study intervention targeted underlying disease (one study),
- inclusion criteria concerning palliative care patients not met (seven studies),
- non-pharmacological interventions (one study),
- doubly published data (one study),
- only abstract available (one study),
- insufficient data given (one study).
In total, we identified 40 RCTs reported in 38 articles for inclusion in the review. See Figure 1 for the study flow diagram.
|Figure 1. Study flow diagram.|
Of the studies included, 18 were cross-over studies. The remaining 22 studies had a parallel group design. Two studies were already pooled in a meta-analysis. The studies were conducted in 24 different countries within Europe, North America, and Asia. Twelve studies were multicentre trials.
The 40 identified studies yielded a broad spectrum of different drugs for the treatment of pruritus associated with different underlying diseases. Data from 1286 participants were analysed. The participants suffered from chronic kidney disease (CKD)-associated pruritus (947 participants), (cholestatic) pruritus (CP) caused by hepatobiliary diseases (276 participants), pruritus associated with malignancies (23 participants), and pruritus as a symptom associated with HIV (40 participants). The drugs assessed were antiphlogistic substances, psychotropic drugs, antagonistic drugs, anaesthetics, adsorbent substances and topical treatments. All studies measured pruritus response and adverse effects. A few studies also measured quality of life (three studies), depression (two studies), and patient satisfaction (one study).
In this systematic review we reported the study results first organised by type of pruritus and then by type of intervention. The majority of included studies researched the effect of different interventions on pruritus in patients suffering from advanced diseases and associated with CKD-related or uraemic pruritus (24 studies) or cholestatic or hepatogenic pruritus (14 studies). Two studies explored pharmacological interventions in palliative care patients treated in palliative care wards and in patients with HIV-infection, respectively. A total of 30 different interventions were explored. Among these, 12 treatments focused on uraemic pruritus, nine interventions focused on the treatment of CP, four medications centred on both treatment of CKD-related and CP, four treatments focused on HIV-associated pruritus, and one medication treated palliative care patients with pruritus of different origin. For the drugs researched and the total numbers assigned to the drugs see Table 2.
Palliative care patients with pruritus of different origin
Selective serotonin reuptake inhibitor (SSRI) paroxetine
Zylicz 2003 (see the Characteristics of included studies table): a prospective, double-blind, randomised, controlled, multicentre within patient comparison of the serotonin uptake inhibitor paroxetine and placebo was performed in two hospices. The patient population of 26 patients with solid tumours (17/26), haematological malignancies (4/26), and various non-malignant or idiopathic conditions (5/26) was heterogeneous and representative of palliative care patients. After a run-in period, patients were randomly assigned to treatment with 20 mg paroxetine or placebo. Due to the advanced nature of their disease, the trial had to be short and the cross-over took place after seven days. The intensity of the pruritus was measured subjectively with a numerical analogue scale (10-point NAS). The primary endpoint of the trial was the mean pruritus score, measured for seven days after randomisation and after cross-over. The secondary endpoint was individual global response to the treatment. The response was defined as at least a 50% reduction in the intensity of the pruritus in the last three days of the treatment period versus baseline. Adverse effects, patient satisfaction, and preferences were also recorded. Two patients discontinued treatment because of adverse effects of paroxetine. Twenty-four patients treated with paroxetine had lower pruritus intensity scores over the seven treatment periods when compared to placebo (NAS 5.2 ± 0.32 versus NAS 6.0 ± 0.32, 7 days means). Vomiting, nausea, and sleepiness were observed as side effects among the patients participating in the study; two participants withdrew because of adverse effects.
Patients with advanced diseases suffering from chronic kidney disease-related or cholestatic pruritus
Opioid antagonist naltrexone
Peer 1996 (see the Characteristics of included studies table): in a randomised, double-blind, placebo-controlled cross-over trial, 50 mg naltrexone was administered per day by mouth to 15 haemodialysis patients with severe, resistant pruritus. The naltrexone or placebo periods lasted seven days, each with a seven-day washout between the two periods. Pruritus was assessed by the participants on a VAS from 0 (no pruritus) to 10 (maximum), and mean daily scores were calculated. The administration of naltrexone was associated with heartburn in two patients and upper abdominal discomfort in three.
Pauli-Magnus 2000 (see the Characteristics of included studies table): a placebo-controlled, double-blind cross-over study of uraemic patients with persistent, treatment-resistant pruritus was performed in 23 patients. Participants were started either with a four-week naltrexone sequence (50 mg/d) or matched placebo. This was followed by a seven-day washout, and participants continued with a four-week sequence of the alternate medication. Pruritus intensity was scored daily by a VAS and weekly by a detailed score assessing scratching activity, distribution of pruritus, and frequency of pruritus-related sleep disturbance. The study was completed by 16 of 23 participants. Nine of 23 participants complained of gastrointestinal disturbances during the naltrexone period, in contrast to only one of 23 participants during the placebo period.
Legroux-Crespel 2004 (see the Characteristics of included studies table): in a comparative study researching naltrexone versus loratadine, 52 participants with uraemic pruritus were treated for two weeks with naltrexone (50 mg/day; 26 participants) or loratadine (10 mg/day; 26 participants), after a washout of 48 hours. Pruritus intensity was scored by a VAS. There was no significant difference in the mean VAS scores after treatment, but naltrexone caused a dramatic decrease in VAS scores in seven participants (> 3 units: marked improvement on VAS scale). Adverse events (mainly nausea and sleep disturbances) were observed in 10/26 participants.
Two double-blind RCTs researched the antipruritic effect of the opioid antagonist naltrexone in participants suffering from cholestatic pruritus and described a significant effect of naltrexone on cholestatic pruritus (Terg 2002; Wolfhagen 1997)
Wolfhagen 1997 (see the Characteristics of included studies table): in a double-blind randomised placebo-controlled study,16 participants with pruritus associated with chronic cholestasis were randomised to receive 50 mg/day of naltrexone (eight participants) or placebo (eight participants) daily for four weeks. Pruritus, quality of sleep, fatigue (all by VAS), side effects, and liver function were also assessed. In four naltrexone-treated participants, side effects (transient in three cases) consistent with opiate-withdrawal syndrome were noted.
Terg 2002 (see the Characteristics of included studies table): in a randomised, double-blind, placebo-controlled cross-over study, 20 participants with pruritus and cholestasis were included and randomised to receive 50 mg/day of naltrexone or placebo for two weeks. Subsequently, there was a one-week washout period, and participants were crossed over to the other therapy for two additional weeks. Pruritus was assessed daily with a VAS from 0 to 10. The most frequent side effects were dizziness (10 participants), nausea (eight participants), vomiting (six participants), headache and abdominal cramps (five participants).
κ-Receptor agonist nalfurafine hydrochloride
Three included RCTs investigated the effect of the kappa-receptor agonist nalfurafine hydrochloride on severe itch in haemodialysis participants. Two of these studies were combined in a meta-analysis.
Wikström 2005a and Wikström 2005b (see the Characteristics of included studies table): a meta-analysis of two RCTs showed that treatment with nalfurafine could reduce itch in participants on haemodialysis, although the effect was quite small. Two multicentre, randomised, double-blind, placebo-controlled studies enrolled 144 participants with uraemic pruritus to postdialysis intravenous treatment with either nalfurafine or placebo for two to four weeks. The first study (study one) was a parallel-group design with treatment lasting for four weeks. Seventy-nine participants were randomly assigned in this study, and 74 completed the four weeks of treatment. After the run-in period, participants were randomly assigned to receive nalfurafine 5 g (N = 26) or placebo (N = 25) three times weekly by intravenous infusion, immediately after completion of each haemodialysis (HD) session during the four weeks. A follow-up visit was performed two weeks after administration of the final dose of study medication.
The second study (study 2) was a cross-over design in which participants were randomly chosen to receive an intravenous infusion of either nalfurafine 5 g or placebo three times weekly for two weeks. At the completion of the first treatment period, participants underwent a three-week washout period followed by another one-week run-in period. Participants were then crossed over to the other study medication for an additional two weeks of therapy. Thirty-four participants were randomly assigned to this study, and 31 completed the four weeks of treatment. The severity of pruritus and the treatment effectiveness were evaluated by the participants and by the investigators on a VAS scale. Statistically significant reductions in worst itching, itching intensity, and sleep disturbances were noted. Improvements in itching and excoriations were noted for the nalfurafine-treated participants. Nalfurafine showed similar types and incidences of drug-related adverse events as the placebo.
Kumagai 2010 (see the Characteristics of included studies table): in a phase III, randomised, double-blind, placebo-controlled study the efficacy and safety of nalfurafine hydrochloride were prospectively investigated by randomly (1:1:1) administering 5 μg (N = 114 participants) or 2.5 μg (N = 112 participants) of the drug or a placebo (N = 111 participants) orally for 14 days, using a double-blind design, in 337 haemodialysis participants with itch. Pruritus was assessed by the participants using a VAS scale. The most common adverse reaction in the study was insomnia (sleep disturbance), found in 24 of the 226 nalfurafine participants. Moderate to severe adverse effects reported in the 2.5 μg group were insomnia (moderate in one participant), sudden hearing loss (moderate in one participant), hypertension (moderate in one participant), vomiting (moderate in one participant), nausea (moderate in one participant), and increased eosinophils (moderate in one participant). Moderate headache was observed in one participant in the placebo group. All other adverse reactions were mild. The study showed that orally taken nalfurafine hydrochloride effectively reduced itch.
3 antagonist ondansetron
Three RCTs researched the effect of the emetic agent ondansetron, a peripherally and centrally acting selective serotonin 5-HT
Ashmore 2000 (see the Characteristics of included studies table): in a prospective, randomised, double-blind, placebo-controlled cross-over study, 16 haemodialysis participants with persistent pruritus were randomly assigned to treatment with ondansetron (8 mg) or placebo three times daily for two weeks. The study period consisted of two washout periods of seven days and two treatment periods of 14 days. Participants scored the intensity of their pruritus daily on a 0 to 10 VAS. The second outcome was the daily use of antihistamines as escape medications. Ondansetron showed no advantage over placebo in either relief of pruritus or reduction of antihistamine use. No adverse effects were reported.
Murphy 2003 (see the Characteristics of included studies table): in a double-blind, randomised, placebo-controlled trial, 24 participants on haemodialysis were enrolled and blindly allocated on a random basis to the ondansetron–placebo sequence (14 participants) or to the placebo–ondansetron sequence (10 participants). During the treatment, participants received either 8 mg of ondansetron three times a day or a placebo tablet three times a day for two weeks. Participants were asked to record the severity of their pruritus on a VAS twice a day. At the end of the study, participants were asked blindly which treatment they had preferred. The washout period between the cross-over treatment periods was seven days. One participant left the study because of constipation while on ondansetron.
Özaykan 2001 (see the Characteristics of included studies table): an open, randomised, comparative trial investigated the antipruritic effects of ondansetron and cyproheptadine in 20 haemodialysis participants. Ten participants were given 8 mg/d ondansetron, and the other 10 participants were given 8 mg/d cyproheptadine orally, for 30 days. Pruritus was assessed by a pruritus score developed by Duo and Mettang (Duo 1987; Mettang 1990). In both groups, no side effects, adverse reactions, or other complications were observed. The study was published in Turkish.
O'Donohue 2005 (see the Characteristics of included studies table): in a double-blind, placebo-controlled study, a total of 19 participants with resistant pruritus were randomised to receive either ondansetron 8 mg or placebo as a single intravenous bolus, followed by oral ondansetron 8 mg or placebo twice daily for two days. The response was assessed by both a subjective evaluation of the severity of pruritus using a VAS and an objective measurement of scratching activity using a piezo-electric vibration transducer. The adverse effects reported included moderate increases from baseline in serum alkaline phosphatase and bilirubin levels in one participant, and constipation in four participants, in the ondansetron group. Conversely, nausea in three participants and headache in two participants were reported only in the placebo group. The results of the study showed that ondansetron did not benefit this group of pruritic participants during short-term treatment.
Selective serotonin reuptake inhibitor (SSRI) sertraline
Mayo 2007 (see the Characteristics of included studies table): initially, 21 participants with chronic pruritus due to liver disease underwent an open-label, dose escalation to determine the dose with optimal efficacy and tolerability. After a washout period, 12 of the participants entered a randomised, double-blind, placebo-controlled cross-over trial. Participants were treated for six weeks, then had a washout period of four weeks, and crossed over to the other therapy for six weeks. Participants quantified their pruritus using a 0 to 10 VAS, and pruritus was assessed for distribution, timing, degree of disability, and physical evidence of scratching. Dermatologic examinations were conducted at each visit. As secondary outcomes, scratching lesions were noted and graded (as worsened, stable, improved, or resolved) and the effect of sertraline on depressive symptoms was evaluated by the self-report version of the 30-item Inventory of Depressive Symptomatology. In addition, sertraline and des-methylsertaline levels obtained at each visit were measured by mass spectroscopy. At the end of the double-blind treatment, participants were asked which treatment (first or second) they thought was the sertraline, why they thought so, and which treatment they preferred. Side effects included dizziness (1), loose stools (1), and a positive effect of increased mood stability.
Two RCTs examined the effect of the antiepileptic gabapentin in participants with uraemic pruritus (Gunal 2004; Naini 2007). One study researched gabapentin in participants with cholestatic pruritus (Bergasa 2006).
Gunal 2004 (see the Characteristics of included studies table): in a double-blind, placebo-controlled, cross-over study, the effectiveness of gabapentin against renal itch was assessed in 25 adult participants on haemodialysis who were asked to record the severity of their pruritus on a VAS daily. The participants were randomly assigned to receive gabapentin for four weeks followed by placebo for four weeks, or vice versa. Gabapentin (300 mg) or placebo was administered thrice weekly, at the end of the haemodialysis sessions. Mild to moderate somnolence, dizziness, and fatigue were the most common side effects of gabapentin. They commonly occurred after the first dose of the drug and subsided within seven days from the initiation of the treatment.
Naini 2007 (see the Characteristics of included studies table): in a double-blind, placebo-controlled trial to evaluate the efficacy of gabapentin in controlling uraemic itch, 34 adult participants on maintenance haemodialysis were enrolled and assigned to receive four weeks of treatment with either gabapentin (400 mg) or placebo administered twice weekly after haemodialysis sessions. Pruritus scores were measured using a VAS. Mild to moderate somnolence, dizziness, and nausea were the most reported side effects of gabapentin and usually subsided within five to 10 days.
Bergasa 2006 (see the Characteristics of included studies table): in a double-blind, randomised, placebo-controlled trial, the effect of gabapentin on the perception of pruritus and its behavioural manifestation, scratching, in cholestasis was studied in 16 women with chronic liver disease and chronic pruritus. Participants were randomised to gabapentin or placebo, starting with a divided dose of 300 mg gabapentin orally per day and increasing to a maximum of 2400 mg per day until relief from pruritus. Hourly scratching activity (HSA) was continuously recorded for up to 48 hours at baseline and during treatment for at least four weeks in an inpatient setting. The perception of pruritus was assessed by interviews and by a VAS. Gabapentin caused several adverse effects such as fatigue, dizziness, worsening symptoms of carpal tunnel syndrome, vomiting, and dizziness. On increasing the doses, a fluctuating rise in serum creatinine could be observed.
Semiantibiotic rifampin or rifampicin
Ghent 1988 (see the Characteristics of included studies table): in a double-blind, randomised cross-over trial, nine participants with primary biliary cirrhosis receiving 300 mg to 450 mg/day of rifampin and placebo sequentially, in random order, were studied. Each treatment was administered for 14 days, with a 14-day washout between treatment. Endpoints included participant preference, changes in daily VAS pruritus score, and amount of cholestyramine ingested. Adverse effects were identified as a secondary outcome.
Bachs 1989 (see the Characteristics of included studies table): in a randomised, crossover, clinical trial, the antipruritic effects of rifampicin (10 mg/kg) and phenobarbitone (3 mg/kg) were assessed in 22 participants with primary biliary cirrhosis. Each agent was administered for 14 days, with a 30-day washout period between treatments. The severity of pruritus was assessed by a four-item pruritus score. Secondary outcomes were adverse effects, total bile acids, and different laboratory values. One participant did not complete rifampicin therapy because of developing haemolytic anaemia and renal failure.
Podesta 1991a (see the Characteristics of included studies table): in a randomised, double-blind, placebo-controlled cross-over study, 14 participants with cholestatic pruritus were studied for three weeks after a 15-day washout period. During the first and third week participants received 600 mg of rifampin or placebo. No treatment was administered during the second week (washout period and cross-over). Pruritus was subjectively scored on a scale from 0 to I00. Secondary outcomes were adverse effects. Subsequently, the persistence of antipruritic effect and safety of rifampin was evaluated in an open study over an eight-month period.
Pour-Reza-Gholi 2007 (see the Characteristics of included studies table): an RCT with a cross-over design was performed on 24 participants who were assigned to two groups and received either placebo or oral doxepin, 10 mg, twice a day for one week. After a one-week washout period, the two groups were treated conversely. The subjective outcome was determined by asking the participants to describe their pruritus as completely improved, relatively improved, or remained unchanged or worsened. Adverse effects were recorded as a secondary outcome. Twelve participants (50%) complained of drowsiness that was alleviated after two days on average. One participant refused to continue the treatment due to its sedative effect.
Silverberg 1977 (see the Characteristics of included studies table): 10 participants with uraemic pruritus were examined in a double-blind, randomised, placebo-controlled trial during four weeks of treatment with either 5 g cholestyramine twice daily (five participants) or placebo (five participants). Severity of pruritus was rated several times every day on a three-point pruritus scale, and a daily pruritus score was calculated by taking the mean of all daily scores. Secondary outcomes were adverse effects and several different laboratory values. Data on all participants could be evaluated.
Duncan 1984 (see the Characteristics of included studies table): in a single-blind, randomised, controlled, cross-over trial, the antipruritic activity of cholestyramine (4 g twice daily), chlorpheniramine (4 mg up to thrice daily), and placebo (lactose 200 mg up to thrice daily) were compared to terfenadine (60 mg up to thrice daily) in eight participants with cholestatic pruritus over a treatment period of two weeks for each drug. Daily pruritus scores were reported in diaries. Secondary outcomes were adverse effects (experienced in all treatment groups), psychometric testing, and electroencephalography.
Kuiper 2010 (see the Characteristics of included studies table): in a randomised, double-blind, investigator-initiated, multicentre trial, participants with cholestatic pruritus, both treatment-naive and previously treated, received 1875 mg of colesevelam or an identical placebo twice daily for three weeks. The effect on pruritus was assessed with daily VAS, quality of life scores, and evaluations of cutaneous scratch lesions. The predefined primary endpoint was the proportion of participants with at least a 40% reduction in pruritus VAS scores; 38 participants were included, and 35 were evaluable. As secondary outcomes, fatigue and quality of sleep were evaluated with comparable VAS scores. In addition, laboratory investigations including measurements of the total serum bile acid, total bilirubin, albumin, alkaline phosphatase, aspartate aminotransferase, and alanine aminotransferase levels were performed and adverse effects were recorded.
Silva 1994 (see the Characteristics of included studies table): in a randomised, double-blind, placebo controlled cross-over trial, 29 participants with uraemic pruritus were treated with 100 mg/day of thalidomide or placebo for seven days. After a washout period of one week, drugs were crossed over for another treatment period of one week. Participants scored their pruritus from zero to three times a day. Participants were labelled as 'responders' or 'non-responders'. Response was defined as reduction of at least 50% in the pruritus scoring. Eighteen participants completed the study. Blood samples were collected for complete blood counts and determination of plasma levels of different parameters to identify factors that could be implicated in the absence of a response to the drug.
Leukotriene receptor antagonist montelukast
Nasrollahi 2007 (see the Characteristics of included studies table): the study was conducted as a randomised, single-blind, placebo-controlled cross-over multicentre trial on participants with refractory uraemic pruritus. The participants were divided into two groups to first receive montelukast 10 mg daily for 20 days and then placebo, or vice versa. The washout period was 14 days. Assessment of pruritus was conducted using the detailed pruritus score introduced by Duo (Duo 1987). The scores for sleep disturbances and intensity of pruritus were added and the final score at the beginning and at the end of the study were calculated. No adverse effects were observed.
Turner 1994a (see the Characteristics of included studies table): the initial trial was a double-blind, randomised, placebo-controlled study investigating the effect of low dose flumecinol (600 mg once weekly) in 50 participants with cholestatic pruritus for three weeks (referred to as 'low dose' (LD) flumecinol). After a seven-day baseline period, participants were randomised to flumecinol 600 mg or the identical placebo once weekly for three weeks. Pruritus was assessed by scoring a 100 mm VAS (0 = no itch, to 100 = severe, continuous, day and night intolerable itch) and also a 'quality of life' VAS (VAS 0 = able to cope with normal activities, to 100 = completely incapacitated) daily, reflecting the preceding 24 hours of symptoms, for seven days as secondary outcome. Participants took the medication on days 0, 7, and 14 and continued to score the VAS diary daily until day 21. For each participant, mean VAS scores (itching and quality of life) for consecutive seven-day periods of observation were computed as summary measures. The change between the mean VAS for the baseline period and the mean VAS for the last seven days (days 14 to 20) was calculated, and the calculated VAS score changes for those participants receiving flumecinol and those receiving placebo were compared. As an additional outcome, liver function tests (bilirubin, albumin, alkaline phosphatase, aspartate transaminase, serum biochemical) and haematological parameters were conducted at study commencement and on completion to monitor for treatment effects.
Turner 1994b (see the Characteristics of included studies table): subsequent to the low dose trial, another double-blind, randomised, placebo-controlled study was performed to investigate the effect of high dose flumecinol. At least one month after completing the initial low dose trial, 19 participants with cholestatic pruritus were enrolled (two of the 19 high dose participants had not participated in the low dose trial). Again, the participants completed a seven-day baseline VAS assessment of pruritus and quality of life and then were randomised to take either the treatment with 300 mg of flumecinol or identical placebo daily, for 21 days. Participants kept a VAS diary, daily for the three weeks while taking the medication. Alanine transaminase (during the high dose trial), serum biochemical, and haematological parameters were determined at study commencement and on completion to monitor for treatment effects.
De Marchi 1992 (see the Characteristics of included studies table): in a 10-week double-blind, randomised, placebo-controlled cross-over study, participants with severe uraemic pruritus were enrolled to investigate the effects of recombinant human erythropoietin on pruritus and plasma histamine levels; 20 participants with uremia (10 with severe pruritus and 10 without) were randomised to receive either erythropoietin intravenously (36 U/kg of body weight three times weekly) or placebo for five weeks and then crossed over. The severity of pruritus was screened weekly by a scoring system proposed by Duo 1987 and modified by Mettang 2002 with minor modifications. The scoring system included severity, frequency, and distribution of pruritus; the frequency of scratching; and the number of times the participant was awakened. Secondary outcomes included plasma histamine concentrations, measured before and at the end of each five-week treatment period by radioimmunoassay. Adverse effects were not mentioned.
Mast cell stabilizer cromolyn sodium
Vessal 2010 (see the Characteristics of included studies table): in a double-blind, randomised, placebo-controlled study, 62 haemodialysis (HD) participants with pruritus were randomly assigned to receive cromolyn sodium or placebo (135 mg, three times daily) for eight weeks. Participants were asked to record the severity of their pruritus during each dialysis session on a VAS, during the eight weeks of treatment and four weeks following discontinuation of treatment. Serum tryptase levels were determined at baseline, after eight weeks of treatment, and four weeks after discontinuation of treatment. All adverse effects were gastrointestinal (GI) and consisted of nausea and diarrhoea.
Activated oral charcoal
Pederson 1980 (see the Characteristics of included studies table): in a randomised, double-blind, placebo-controlled, cross-over trial, 11 participants with uraemic pruritus were treated with oral charcoal 6 g daily or placebo and then vice versa in two consecutive eight-week treatment periods. A washout period between the treatments was not mentioned. The intensity of itching was indexed at eight-week intervals using a self-administered questionnaire as suggested by Lowrie 1975 and Ingham 1969, comprising six statements rating the intensity of pruritus. The statements were arranged in nine paired response alternatives, which allowed the severity of itching to be ranked on a continuum of 1 (no itching) to 10 (severe constant itching). As secondary outcomes, skin lesions were examined and laboratory tests on different blood parameters were obtained.
Borgeat 1993 (see the Characteristics of included studies table): in a prospective, randomised, double-blind, cross-over, placebo-controlled study, 10 participants with cholestatic pruritus received two doses of propofol (1 ml = 15 f) and two doses of placebo (1.5 ml of intralipid) during a four-day study period. Pruritus was assessed by a verbal rating scale from 0 to 10. Treatment success was defined as a decrease of pruritus of at least four points on the verbal rating score. Secondary outcomes were sedation, pain during injection, dizziness, mood changes, hallucinations, and haemodynamic parameters. In addition, preference of treatment was analysed. As adverse effects, discomfort on injection and slight dizziness were observed.
Villamil 2005 (see the Characteristics of included studies table): in a double-blind, placebo-controlled trial, 18 participants with cholestatic pruritus were randomised (2:1) to receive 100 mg lidocaine (5 ml saline) intravenously over five minutes or placebo (5 ml saline). Electrocardiographic monitoring was done during infusion and vital signs were recorded every 15 minutes for the first hour after infusion. The occurrence of adverse events was assessed during this period. Severity of pruritus was recorded on a VAS at baseline and every 12 hours. A similar VAS was used to record the severity of fatigue. As additional secondary outcomes, serum liver tests were done and monoethylglycinexylidide (MEGX) concentrations, a metabolite of lidocaine, were measured.
Breneman 1992b (see the Characteristics of included studies table): in a randomised, double-blind, vehicle-controlled trial conducted to evaluate the efficacy and safety of capsaicin 0.025% cream in the treatment of localized areas of pruritus in participants with uraemic pruritus, seven participants were treated with either capsaicin 0.025% cream or the vehicle for six weeks. Each participant was provided with two sets of tubes, one contained capsaicin 0.025% cream and the other contained the vehicle. Therapies were assigned on a random basis to either arm in a double-blinded fashion. The tubes were identical except for the designations of right and left. Participants were instructed to apply the cream four times daily and to apply medication from one set of tubes only, and specifically to one arm, and medication from the other set of tubes likewise to the other arm for six weeks. The degree of relief of itching was measured with a four-point scale (1 = no relief; 2 = minimal relief; 3 = marked relief; 4 = itching resolved). The degree of cutaneous burning or stinging sensations associated with cream application was assessed during each week for the area being treated using a comparable scale (1 = no burning or stinging; 2 = mild transient burning or stinging; 3 = moderate transient burning or stinging; 4 = marked or prolonged burning or stinging). The appearance of the skin and erythema of the treated area were assessed with a similar scale (1 = no dryness; 2 = mild dryness; 3 = moderate dryness; 4 = marked dryness with significant scaling). For analysis, five participants were considered evaluable. Four participants completed the entire trial, and one participant withdrew after five weeks because of a worsening medical status. No significant burning or stinging sensations or erythema were reported by any participants.
Tarng 1996 (see the Characteristics of included studies table): to assess the efficacy and safety of capsaicin 0.025% cream in the treatment of uraemic pruritus and to further explore the underlying pathomechanism in a double-blind, placebo-controlled, cross-over, uni-centre study, 19 haemodialysis participants with uraemic pruritus were treated with capsaicin 0.025% cream or placebo during two four-week treatment periods, with a two-week washout phase between treatments and a follow-up of eight weeks without treatment. Treatment was applied to a selected area four times daily. Degree of itching was assessed by a four-item rating score (severe, moderate, mild, no itching) upon entry to the study and at one-week intervals thereafter. As secondary outcomes, the degrees of cutaneous burning or stinging sensations, dryness of skin, and erythema over the treated area were also evaluated at one-week intervals by a four-point scale (marked, moderate, mild, none). Blood samples were taken at the beginning and end of each treatment period for measurement of serum albumin, calcium, inorganic phosphorus, alkaline phosphatase, and intact parathyroid hormone (PTH).
Cho 1997 (see the Characteristics of included studies table): a randomised, double-blind, placebo-controlled, cross-over uni-centre study of capsaicin 0.025% cream with two, four-week treatment periods and a 14 day washout period was conducted with 22 participants with uraemic pruritus to evaluate the role of parathyroid hormone (PTH) and substance P in uraemic pruritus and to elucidate the underlying mechanisms. For this purpose, in the first phase of the study the correlation between the intensity of itching and serum levels of intact PTH was tested. For the second phase, participants were further stratified into two subgroups with low intact PTH (≤ 35 pg/ml) and high intact PTH (> 35 pg/ml). Subsequently, the double-blind cross-over trial with topical capsaicin was conducted in the two subgroups. Capsaicin or placebo creams were applied four times daily to a pre-selected area of skin throughout each treatment period. Self-assessment of pruritus was made upon entry into the study and at one-week intervals thereafter. As an additional outcome, the intensity of cutaneous burning or stinging sensations, dryness of skin, and erythema in the treated area related to cream application were also evaluated at one-week intervals. Each factor was given a score from 1 to 4 (1 = none, 4 = severe). In addition, blood samples were taken at the beginning and end of each treatment period for measurement of serum calcium, phosphate, and intact PTH levels (measured by an immunoradiometric assay). As adverse effects, tolerable local burning or stinging sensations (or both) in 11 participants and cutaneous erythema in five participants were reported.
Makhlough 2010 (see the Characteristics of included studies table): the randomised, double-blinded, cross-over clinical trial was performed on 34 participants receiving haemodialysis with uraemic pruritus to research the effect of topical capsaicin compared with placebo in a single centre clinical setting. The participants were divided into two groups, one group received capsaicin 0.03% and the other placebo, for four weeks. Treatment was stopped for two weeks during the washout period and was then continued following a cross over. Pruritus was assessed by a modified score proposed by Duo (Duo 1987), according to the severity of pruritus, distribution of pruritus, and sleep disorder (total scoring of pruritus = (severity of pruritus × distribution of pruritus) + sleep disorder scoring). All of the participants received clinical examinations and were questioned regarding adverse effects. One participant in the verum group reported severe skin burning.
Duque 2005 (see the Characteristics of included studies table): in a randomised, double-blind, vehicle-controlled, multicentre study, 22 participants with end-stage renal disease and uraemic pruritus were studied during a treatment period of four weeks to assess the efficacy of tacrolimus ointment 0.1%. Tubes (30g) were used and the number of tubes used during the study varied between one and eight per participant, with an average of four per participant. Medication was applied only on pruritic areas three times weekly by one of the investigators and by the participants at home twice a day for four weeks. The severity of pruritus was assessed by a VAS for four different temporal states: at the time the participant was examined; at the time of the worst pruritus; at the time when the pruritus was best; and comparing the itch after a mosquito bite. This measurement consisted of a 10 cm line labelled from one extreme 0 (no itch) to the other extreme 10 (very strong itch). The participants were asked to mark the intensity of itch at the time of interview, at week four, and at week six. In addition, the VAS score was measured every visit only for the intensity of itch in the last 24 hours. Clinical evaluation was performed for excoriations, scaliness, lichenification, and overall severity at baseline and three times a week for four weeks. Follow-up took place two weeks after completion of the treatment. This was assessed by a three-point Likert scale (0 (no skin signs) to 3 (severe excoriations or severe scaliness or severe lichenification, or a combination). Side effects were warm sensations in eight participants in the tacrolimus group and a significant burning sensation in one participant, which subsided with continuous applications.
Young 2009 (see the Characteristics of included studies table): in a randomised, double-blind, controlled comparative trial, the efficacy of a commercially available anti-itch lotion containing pramoxine hydrochloride was investigated versus control lotion in the treatment of uraemic pruritus in adult haemodialysis participants. The study was conducted in 28 individuals treated in a community haemodialysis centre during a one-week treatment period. Fourteen participants were randomised to receive the 1% pramoxine hydrochloride lotion and 14 participants received the bland emollient. Daily itching severity was assessed by VAS score. Erythema, xerosis, and lichenification were also assessed using a three-point Likert scale with 0 indicating no symptoms and 3 representing severe symptoms. Additonal measures were the individual pruritus history and assessment questionnaire, the Investigator Global Assessment (IGA) of response to treatment, the skin hydration measurements using the MoistureMeter pico, and adverse events as reported by the participants and investigator. Itch reduction was defined as 1-(mean VAS at end of study)/(mean VAS at baseline)*100. One participant had to be excluded from the analysis. No adverse effects were reported.
Patients with pruritus because of HIV infection
Hydroxyzine hydrochloride, pentoyxifylline, triamcinolone, and indomethacin
Smith 1997 (see the 'Characteristics of included studies' table): a prospective, randomised, controlled trial with four groups investigated the antipruritic effect of hydroxyzine hydrochloride with or without doxepin hydrochloride, pentoxifylline, triamcinolone, and indomethacin in patients with advancing HIV disease. Altogether, 40 participants (10 participants in each treatment group) were included in the study. Duration of the treatment was four to six weeks. All medications showed some antipruritic effect, while patients treated with indomethacin had the greatest improvement in pruritus among the groups. Side effects were multiple (stomach ache, drowsiness, sleepiness, dry eyes and mouth, headache, indigestion, nausea) except in participants on topical steroid therapy. Three participants discontinued their medication because of side effects.
In the final screening we excluded 44 of 82 studies because they did not meet the inclusion criteria (32 not an RCT; seven did not meet the inclusion criteria concerning palliative care patients, one study intervention targeted the treatment of underlying disease, one showed doubly published data, and one did not focus on a pharmacological intervention). Two studies could not be included because only the abstract was available and data were reported inadequately. For both of these studies, we were not able to contact the authors despite multiple attempts. Details regarding reasons for exclusions are provided in the Characteristics of excluded studies table.
Risk of bias in included studies
All included studies were RCTs or a meta-analysis of RCTs. We included 13 studies that used appropriate methods of randomisation (for example drawing lots, flip of coins, computer-generated table of random numbers) and 27 trials that reported using randomisation but failed to state the method of randomisation. Ten trials described allocation concealment whereas the remainder of the trials (30 studies) did not mention the use of allocation concealment. Thirty-four trials were described as double-blind, but only eight studies described blinding of the outcome assessment. Two of the included studies were single-blind (Duncan 1984; Nasrollahi 2007); two studies were open trials (Bachs 1989; Özaykan 2001); and in another two studies the use of blinding was unclear (Legroux-Crespel 2004; Smith 1997). To our knowledge, no analyses of the efficacy of blinding was carried out in the included studies.
With regard to incomplete outcome data, the quality of the included studies was quite heterogenous. The dropout rate was low for most trials. In 13 studies there were no withdrawals. However, five trials reported dropout rates of more than 25%, with a maximum dropout rate of 57% (Breneman 1992b). This high dropout rate may be due to the small number of participants included in the trials (only seven participants were included in the trial with the highest dropout rate). On the other hand, it may also be related to the advanced stage of disease of the participants included in the studies. Most trials reported reasons for dropout. Detailled information on dropouts and reasons for dropout is included in the Characteristics of included studies.
Risk of bias concerning selective reporting was unclear in 16 of 40 studies. Since we did not have access to the protocols of the included studies, there was not enough information to assess selective reporting bias in detail. Nevertheless, three studies showed a high risk of selective reporting (Breneman 1992b; Gunal 2004; Legroux-Crespel 2004). Please see the Risk of bias in included studies table for more details. Other potential sources of bias may have related to the small sample sizes in some of the included studies (sample size < 20 in 16 studies), missing washout periods leading to carry-over effects in cross-over studies (Cho 1997; Tarng 1996), and inadequate study design (Smith 1997; Tarng 1996).
As a result, only four studies (Kuiper 2010; O'Donohue 2005; Turner 1994a; Vessal 2010) were at low risk of bias for objective outcomes and satisfied all criteria used to assess the risk of bias. The remaining studies had an unclear risk of bias or a high risk of bias (seven studies). The main reason for being given a high risk of bias rating was a lack of blinding, missing data, or inadequate reporting of data. Risk of bias is detailed for each study in the risk of bias tables included with the Characteristics of included studies. In addition, an overall assessment of risk of bias can be found in Figure 2. The overall risk of bias is summarised graphically in Figure 3.
|Figure 2. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.|
|Figure 3. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
Effects of interventions
We researched pruritus as the primary outcome. An overall meta-analysis was not possible because of the diversity of the different kinds of pruritus and interventions included in this review. Therefore, several different treatment comparisons were considered (comparisons 1 to 6). If data could not be summarised in meta-analyses, results were summarised and reported descriptively according to the different pharmacological interventions and the underlying kind of pruritus. The description of results focuses on pruritus scores. Additional outcomes of the studies are given in the Characteristics of included studies tables.
Palliative care participants with pruritus of different origin: selective serotonin reuptake inhibitor (SSRI) paroxetine
One study researched the effect of the selective serotonin reuptake inhibitor paroxetine (SSRI) on pruritus in palliative care patients (Zylicz 2003). In this randomised, controlled, cross-over study, paroxetine showed an antipruritic effect in palliative care participants with opioid-induced, paraneoplastic, or haematologic pruritus. Secondary outcomes were participant satisfaction and preference. Twenty-four of 26 participants treated with paroxetine had lower pruritus intensity scores over the seven treatment periods (mean ± SE: 5.2 ± 0.32) when compared to participants receiving the placebo (6.0 ± 0.32). The mean difference (MD) between placebo and paroxetine was 0.78 (95% CI 0.37 to 1.19). Nine of 24 participants (37.5%) fulfilled the response criteria. The onset of antipruritic action was typically observed after two or three days, irrespective of the order of treatment. Two patients discontinued treatment because of severe adverse effects (nausea and vomiting), presumably because there was no opportunity to titrate the dose to the effect of the medication in this trial. The results of the study indicated that paroxetine may be effective in the treatment of intractable, severe non-dermatological pruritus associated with advanced cancer and other systemic diseases. This was the only study that particularly researched palliative care participants treated in palliative care units or palliative care settings.
Participants with advanced diseases suffering from chronic kidney disease (CKD)-related pruritus or cholestatic pruritus
Five trials including 126 participants examined the antipruritic effect of naltrexone in participants suffering from pruritus. Three studies included 90 participants with CKD-related or uraemic pruritus (Legroux-Crespel 2004; Pauli-Magnus 2000; Peer 1996), and two studies (Terg 2002; Wolfhagen 1997) evaluated naltrexone in patients with cholestatic pruritus (36 participants).
Data from two trials could not be included in the meta-analysis. In one study the necessary information was not provided in the publication (Peer 1996). Another study was excluded from the meta-analysis due to poor methodological quality and lack of data (Legroux-Crespel 2004) (see Risk of bias in included studies).
Naltrexone in uraemic pruritus
Results of the RCTs regarding its effects in uraemic pruritus were contradictory (Pauli-Magnus 2000; Peer 1996). The results of an RCT by Peer 1996, involving 15 participants, showed that administration of the oral mu-receptor antagonist naltrexone was associated with a significant decrease in pruritus perception. The median pruritus scores at the end of the naltrexone treatment were 2.1 (interquartile range 1.5 to 2.15) for the naltrexone-placebo sequence and 1.0 (0.4 to 1.15) for the placebo-naltrexone sequence. The respective values before naltrexone was given were 9.9 (9.85 to 9.95) and 9.9 (9.3 to 10.0). The results of this study suggested short-term efficacy with few side effects for the amelioration of uraemic pruritus with naltrexone (Mettang 2010; Peer 1996). The study by Pauli-Magnus 2000, involving 23 participants, showed no statistically significant difference between the naltrexone and the placebo treatment periods. During the naltrexone period, pruritus decreased by 29.2% (95% CI 18.7 to 39.6) on the VAS and by 17.6% (95% CI 4.2 to 31.1) on the detailed score (pruritus score proposed by Duo 1987). The difference between the naltrexone and the placebo treatment periods was not statistically significant (Pauli-Magnus 2000). The third study, researching the effect of naltrexone in participants suffering from uraemic pruritus, was a comparative study comparing naltrexone versus loratadine in 52 participants. In this study, seven of 52 participants showed a dramatic improvement when using naltrexone (> 3 units, marked improvement), whereas the mean VAS score was identical to the alternative medication the H
Naltrexone in cholestatic pruritus
Two studies researched the antipruritic effect of the opioid antagonist naltrexone in participants suffering from cholestatic pruritus and reported a significant effect on cholestatic pruritus (Terg 2002; Wolfhagen 1997).
Wolfhagen 1997: 16 participants with cholestatic pruritus were researched for four weeks. Pruritus mean changes in VAS with respect to baseline were significantly different, in favour of the naltrexone group, for daytime itching (-54% versus 8%; P < 0.001) and night-time itching (-44% versus 75; P = 0.003). For five of the eight participants in the naltrexone group, the total pruritus score was reduced by half or more after four weeks of treatment.
Terg 2002: the VAS showed greater and more significant changes with naltrexone than with placebo (P = 0.0003). At the end of naltrexone treatment, pruritus decreased significantly compared to basal (mean daytime pruritus VAS decreased from 6.29 ± 2.28 to 3.55 ± 2.39; P = 0.0003; and night-time itching improved from 5.89 ± 2.49 to 3.55 ± 2.42; P = 0.001). These outcomes with naltrexone were seen in both groups separately. In nine of 20 participants (45%) receiving naltrexone, pruritus decreased 50% compared to the basal value, including five cases where the pruritus disappeared completely.
Three studies were included in the meta-analysis of pruritus on VAS comparing naltrexone with placebo or standard medication (Pauli-Magnus 2000; Terg 2002; Wolfhagen 1997) ( Analysis 1.1). After pooling the studies to examine the effect of naltrexone, the results showed a statistically significant effect, with a MD of -2.1 (95% CI 2.91 to -1.37, P < 0.0001). Substantial statistical heterogeneity existed (I
Subgroup analysis by nature of pruritus of the two studies on cholestatic pruritus (CP) (Terg 2002; Wolfhagen 1997) (N = 36) and one study on chronic kidney disease patients (CKDP) (N = 23) (Pauli-Magnus 2000) resulted in a significant effect for the CP studies (MD -2.50; 95% CI -3.41 to -1.59). In the analysis of one study on CKDP (Pauli-Magnus 2000), there was no significant difference between the study groups with naltrexone and placebo (MD -1.23; 95% CI -2.67 to 0.21). The test for subgroup differences was not significant (P = 0.14) ( Analysis 1.2).
Subgroup analysis by study design ( Analysis 1.3) for the studies on CP resulted in an MD of -3.77 (95% CI -5.28 to -2.25) for the study with a parallel group design (Wolfhagen 1997) and MD of -1.79 (95% CI -2.9 to -0.66) for the cross-over study (Terg 2002). The difference between the two subgroups was statistically significant (P = 0.004); however, both studies showed a statistically significant treatment effect. For the sensitivity analysis see Analysis 1.4.
k-Receptor antagonist nalfurafine
In a meta-analysis of two RCTs researching 144 participants suffering from CKD-related pruritus, Wikström and colleagues showed that treatment with the κ-receptor antagonist nalfurafine could reduce itch in participants on haemodialysis, although the effect was quite small (Wikström 2005a; Wikström 2005b). Statistically significant reductions in worst itching (P = 0.0212), itching intensity (P = 0.0410), and sleep disturbances (P = 0.0003) were noted in the nalfurafine group when compared with placebo. Analysing data from the first two weeks of both studies, patients who received nalfurafine showed a significant (9.53 mm greater) reduction in worst itching VAS score (weighted mean difference (WMD)) from run-in to week two than those given placebo (95% CI 1.42 to 17.64 mm; P = 0.0212). Significantly more nalfurafine-treated patients responded (achieved 50% reduction in worst itching VAS score) within two weeks of run-in than placebo-treated patients (36% versus 14%; P = 0.0226). The number of days with tolerable itching and the number of nights undisturbed by itching increased significantly more in the nalfurafine groups than in the placebo groups during the first two weeks of treatment (+2.2 days versus +1.4 days; P = 0.0410 and +2.5 nights versus +0.9 nights; P = 0.0003). Improvements in itching (P = 0.0025) and excoriations (P = 0.0060) were noted for the nalfurafine-treated participants.
The meta-analysis combined data from a study with parallel group design (Wikström 2005a) and a cross-over study (Wikström 2005b). With the intention to select comparable data from the two studies, only the data from the nalfurafine 5 µg and the placebo groups were used in these analyses because it allowed for pooling of groups with the same dosage. Using only the data from the first two weeks of treatment in ine study and the first two weeks of treatment in the second study was considered appropriate because of concerns about carry-over effects of the drug and regression to the mean. Due to missing data in the constituent studies, subgroup analysis by study design was not possible in the meta-analysis of studies researching nalfurafine.
The results of the studies reported by Wikström were confirmed by a randomised, double-blind, placebo-controlled, parallel group prospective study undertaken by Kumagai et al. In this study, 337 haemodialysis participants with itch that was resistant to currently available treatments were treated (Kumagai 2010). Based on the hypothesis that the activation of ĸ-receptors expressed by dermal cells and lymphocytes might lead to the suppression of pruritus, the study by Kumagai tested whether κ-receptor antagonists (nalfurafine) were able to reduce CKD-associated pruritus (Mettang 2010). A total of 337 participants were studied, with 114 participants receiving nalfurafine 5 μg, 112 participants receiving 2.5 μg nalfurafine, and 111 participants receiving placebo. Pooled data for all morning and evening VAS values showed that in the 5 μg nalfurafine group, decreases in the VAS values (mean 16; 95% CI 13 to 18 during the first seven days of the treatment period and mean 22; 95% CI 18 to 26 during the latter seven days of the treatment period) were significantly greater than those in the placebo group (mean 8; 95% CI 6 to 11 and mean 13; 95% CI 10 to 16, respectively). In the 2.5 μg nalfurafine group, decreases in the VAS values (mean 16; 95% CI 13 to 19 during the first seven days of the treatment period and mean 23; 95% CI 19 to 27 during the latter seven days of the treatment period) were also significantly larger than those in the placebo group. On day seven of the treatment period, VAS changes in the 5 μg group (P < 0.0001) and the 2.5 μg group (P = 0.0101) were significantly greater than the VAS change in the placebo group (t-test, one-sided 2.5% significant level). The VAS changes in the two nalfurafine groups were greater during the latter seven days of the treatment period than during the first seven days of the treatment period. The study, based on VAS evaluations, showed a significant effect for reduction of itch (SMD -0.53; 95% CI -0.76 to -0.29).
Data from the meta-analysis by Wikström and the study by Kumagai et al could be pooled in a meta-analysis, including 481 participants ( Analysis 2.1). Since the studies used different scales for measuring pruritus, we used the standardised mean difference (SMD) as the outcome measure. Meta-analysis of the studies researching ĸ-receptor-antagonist nalfurafine versus placebo resulted in a significant effect in favour of nalfurafine (SMD -0.46; 95% CI -0.65 to -0.28; I
3 antagonist ondansetron
Four studies including 79 participants examined the effect of ondansetron for treatment of pruritus. Three studies compared ondansetron to placebo, and one trial used cyproheptadine as a standard medication in comparison with ondansetron. One study involved 19 participants with cholestatic pruritus (O'Donohue 2005) and three studies, including 60 participants, researched participants with CKD-related pruritus (Ashmore 2000; Murphy 2003; Özaykan 2001). In all studies, a dosage of 8 mg ondansetron was given each day.
For the CKD-related pruritus studies, conflicting results were found. Whereas Murphy et al reported a significant reduction of the VAS score for both treatment with ondansetron and for placebo, Ashmore et al and Özakan et al reported no significant change of VAS scores during active or placebo treatment. For cholestatic pruritus, the study conducted by O'Donohue et al also did not find significant treatment effects.
Ashmore 2000: in this prospective, placebo-controlled, cross-over study researching 16 participants, the median daily pruritus score measured via a VAS did not change significantly during active or placebo treatment (pre-ondansetron 5.3; interquartile range (IQR) 3.4 to 6.3; during ondansetron 3.9; IQR 2.7 to 5.0; P = not significant; pre-placebo 3.7; IQR 3.0 to 4.6; during placebo 3.6; IQR 2.4 to 4.8; P = not significant).
Murphy 2003: this study examined 24 participants with CKD-related pruritus in a placebo-controlled, cross-over study and recorded the severity of pruritus on a VAS. In this study, pruritus decreased by 16% (95% CI 0.5% to 32%) during active treatment and by 25% (95% CI 9% to 41%) during treatment with the placebo. The changes in VAS scores during treatment with ondansetron (P = 0.04) and placebo (P = 0.01) were both significant.
Özaykan 2001: 20 participants with CKD-related pruritus were enrolled in a prospective, parallel group study researching ondansetron versus cyproheptadine, which served as a standard medication. Pruritus was measured by the pruritus scale introduced by Duo (Duo 1987). No significant statistical difference was found among the groups in itching scores at the end of the first and second weeks of the therapy. At the end of the third and fourth weeks, participants receiving ondansetron had statistically lower itching scores (Z = -2.80; P < 0.01).
O'Donohue 2005: this study investigated the effect of ondansetron in 19 participants with cholestatic pruritus in a placebo-controlled, parallel group study. Mean pruritus score using a VAS and scratching activity were reduced on the first treatment day compared with baseline in both the ondansetron and placebo groups (P < 0.05). However, there were no significant differences in mean pruritus perception or scratching activity between the two groups in this study.
In principle, all studies evaluating ondansetron were suitable for meta-analysis. However, one study used different measurements and scales to measure the intensity of pruritus (Özaykan 2001). Since data from the other two studies researching CKD-related pruritus did not allow for the calculation of MDs and SMDs, only data from two studies on CKD-related pruritus could be pooled (Ashmore 2000; Murphy 2003). As information on the standard error of the CKD-related pruritus studies was not available, no reasonable conclusion was possible based on the results of a single study with 19 participants (O'Donohue 2005).
Meta-analysis of the studies using VAS combined data from two studies researching CKD-related pruritus (Ashmore 2000; Murphy 2003) and data from the study researching cholestatic pruritus (Ashmore 2000; Murphy 2003; O'Donohue 2005). Again, the standard errors were not available in the CKD-related pruritus studies, and therefore no reasonable conclusion could be drawn ( Analysis 3.1).
In a third analysis, data on CKD-related pruritus from the third study and data from the study researching the effect of ondansetron in patients with cholestatic pruritus (O'Donohue 2005; Özaykan 2001) were combined ( Analysis 3.1). For the Özaykan study, only data from week four could be used ( Analysis 3.2). The pooled estimate from the two studies did not find support for an effect of ondansetron (SMD -0.15; 95% CI - 0.15 to 1.16, P = 0.13), and a subgroup analysis showed no significant effects for treatment with ondansetron, neither for patients with CKD-related pruritus (SMD 1.01; 95% CI 0.07 to 1.96) (Özaykan 2001) ( Analysis 3.3) nor for patients with cholestatic pruritus (SMD 0.05; 95% CI - 0.85 to 0.96) (O'Donohue 2005) ( Analysis 3.4). Sensitivity analysis concerning different scales ( Analysis 3.5) and different statistical models ( Analysis 3.6; Analysis 3.7) did not show a noticeable effect.
Selective serotonin reuptake inhibitor (SSRI) sertraline
In a randomised, double-blind, placebo-controlled trial undertaken by Mayo et al (Mayo 2007), 12 participants suffering from cholestatic pruritus were enrolled. Participants taking sertraline improved by a mean of 1.86 points on the 0 to 10 VAS , whereas pruritus in participants taking placebo worsened by 0.38 points. The difference between the two (net beneficial effect of 2.24 points) was described as statistically significant (P = 0.009).
Two RCTs examined the effect of the antiepileptic gabapentin in a total of 59 participants with uraemic pruritus (Gunal 2004; Naini 2007), and one study researched gabapentin in 16 participants with cholestatic pruritus (Bergasa 2006).
In a double-blind, placebo-controlled, cross-over study, Gunal et al studied the effectiveness of gabapentin against renal itch in 25 patients. The treatment period was four weeks. The mean pruritus score (VAS) of the cohort before the study was 8.4 ± 0.94. After placebo intake, it decreased to 7.6 ± 2.6 (P = 0.098). After gabapentin administration, the mean score decreased significantly to 1.2 ± 1.8 (P = 0.0001) (Gunal 2004).
An RCT in 34 participants with uraemic pruritus confirmed the positive effect of gabapentin. Participants were also treated for four weeks and the mean decrease in pruritus score (VAS) in the gabapentin and placebo groups was 6.7 ± 2.6 and 1.5 ± 1.8, respectively (P < 0.001) (Naini 2007).
Data from these two studies researching participants with uraemic pruritus could be pooled (Gunal 2004; Naini 2007). Data from the study researching the effect of gabapentin on cholestatic pruritus could not be included because we could not calculate MD or SMD (Bergasa 2006).
The pooled estimate of the two trials (Gunal 2004; Naini 2007) indicated some positive effect (MD -5.20; 95% CI -6.70 to -3.70) ( Analysis 4.1); however, the results must be interpreted with caution because the standard error (SE) of one included study could not be calculated (Gunal 2004). Data from the study by Naini et al showed an estimated MD of -5.20 (95% CI -6.70 to -3.70) ( Analysis 4.2).
The study researching the effect of gabapentin on cholestatic pruritus (Bergasa 2006) did not find a significant therapeutic advantage over the placebo. On the contrary, data suggested a strong placebo effect. The data given in the study did not allow an estimate of the effect.
Semiantibiotic rifampin or rifampicin
Three studies including 45 participants researched the treatment of cholestatic pruritus with semiantibiotic rifampicin either compared to placebo (Ghent 1988; Podesta 1991a) or to phenobarbitone, used as a standard treatment (Bachs 1989).
Ghent et al conducted a double-blind, controlled, cross-over clinical trial of rifampin, including nine participants, during a treatment period of 14 days. VAS pruritus scores showed no significant placebo response or any effect from the order of the treatment. Rather, a significant reduction in pruritus in response to rifampin (P < 0.002) was shown. Analysis confirmed this estimate and resulted in an SMD of -1.46 (95% CI -2.78 to -0.13).
The randomised, controlled cross-over trial conducted by Podesta et al (Podesta 1991a) included 14 participants with cholestatic pruritus that were treated with rifampin and placebo for seven days, or vice versa. A positive effect was shown for rifampin. Since the figures provided in the publication suggested a carry-over effect, we analysed the data and found an interdependence effect (P = 0.0063) (periodical effect P = 0.0002; therapeutical effect < 0.0001). Therefore, in the meta-analysis we only included the results of the first treatment period and treated the study as a parallel group trial. The data showed a significant effect of rifampin (SMD -1.96; 95% CI -2.75 to -1.20).
Bachs et al (Bachs 1989) assessed the antipruritic effects of rifampicin and phenobarbitone in 22 patients with cholestatic pruritus in a cross-over RCT. In contrast to the above study, no interaction between therapy effect and treatment period or carry-over effect could be found. Pruritus improved in 19 patients taking rifampicin compared to eight taking phenobarbitone. Improvement was significantly greater with rifampicin than with phenobarbitone (P < 0.001). This study could be included in the meta-analysis showing a significant effect of rifampicin (SMD -1.81; 95% CI -2.90 to -0.72).
The pooled estimate of the three studies researching rifampin and rifampicin indicated that rifampin may improve pruritus in patients suffering from cholestatic pruritus (SMD -1.96; 95% CI -2.72 to -1.20; I
In an RCT by Pour-Reza-Gholi 2007 with a cross-over design, performed on 24 participants with CKD-related pruritus, complete improvement was described in 58.3% of patients treated with doxepin. This was significantly higher than improvements with placebo (P < 0.001). Overall, doxepin was described as effective in 87.5% of the patients (P < 0.001). Since no raw data were presented, further statistical analysis was not possible (Pour-Reza-Gholi 2007).
Two RCTs researched the effect of cholestyramine on participants with uraemic or cholestatic pruritus, respectively. One study was a parallel group trial including 10 participants with uraemic pruritus (Silverberg 1977). The other trial used a cross-over design and researched eight participants with cholestatic pruritus (Duncan 1984), comparing the effect of cholestyramine, terfenadine, chlorpheniramine, and placebo. Both studies reported some positive effect of cholestyramine and of terfenadine, respectively. However, the very small samples limited the informative value of the results.
The randomised, double-blind, investigator-initiated, multicentre trial by Kuiper et al (Kuiper 2010) aimed to assess the efficacy of colesevelam versus placebo in cholestatic pruritus. No difference in pruritus score was found between participants treated with colesevelam and participants receiving placebo (P = 1.0 for the VAS day score and P = 0.74 for the VAS evening score; predefined primary endpoint = proportion of patients with at least a 40% reduction in pruritus VAS scores).
Data from 18 of 29 participants could be analysed in a double-blind, randomised cross-over trial on thalidomide against uraemic pruritus. Silva et al found a similar proportion of participants in phase one and two responding to thalidomide (responders in phase 1: 55% and responders in phase 2: 57%). The reduction on the three-item pruritus score was 78% ± 6 in Phase 1 and 81% ± 10 in Phase 2, respectively (P < 0.05). The authors found a positive effect of thalidomide in 67% of the study population, with a mean reduction of pruritus scoring of approximately 80% (Silva 1994).
Leukotriene receptor antagonist montelukast
The study researching the effect of the leukotriene receptor antagonist montelukast was conducted as a multicentre, randomised, single-blind, placebo-controlled cross-over study including 16 participants with CKD-related pruritus (Nasrollahi 2007). Results at the end of the treatment with montelukast showed a reduction of pruritus by 35% (95% CI 9.5% to 62.5%) as compared to a reduction of 7% (95% CI 0.5% to 15.9%) with placebo (P = 0.002). The mean change in pruritus score was 16.1 (95% CI 9.5 to 22.5) with montelukast and 7.1 (95% CI 0.5 to 13.7) with placebo. Assessment of pruritus was conducted using the Detailed Pruritus Score introduced by Duo (Duo 1987).
Two randomised, double-blind, placebo-controlled studies examined the effects of low and high doses of flumecinol in participants with cholestatic pruritus during a treatment period of three weeks. In the low dose study (Turner 1994a), 50 patients were enrolled and scored with the VAS daily for a seven-day baseline and for a further 21 days. Subjectively, pruritus improved in 13 of 24 participants taking flumecinol and 10 of 26 taking placebo (x' = 1.24, P = 0.27). The median difference in the reduction of VAS pruritus score between the baseline week (mean score for each individual used) and the last week was 8.0 (95% CI -2.1 to 20.81) and favoured flumecinol over the placebo. In the subsequent study with high dose flumecinol (Turner 1994b), 19 patients were included. The median difference in reduction of the VAS pruritus score was 19.8 mm (95% CI 3.3 to 40.7 mm), in favour of flumecinol over the placebo.
A 10-week, randomised, controlled cross-over trial researching 10 participants with uraemic pruritus and 10 participants without pruritus treated both groups with erythropoietin and placebo, respectively. Eight of the 10 participants suffering from pruritus showed marked reductions in their mean pruritus score (Duo 1987; Mettang 2002), which decreased from 25 ± 3 to 6 ± 1 during treatment with erythropoietin (De Marchi 1992). Since the mean score increased again during the placebo treatment, the authors concluded a positive effect of erythropoietin.
Oral cromolyn sodium
In a double-blind, placebo-controlled trial researching the effect of oral cromolyn sodium in participants with uraemic pruritus, Vessal et al analysed the data from 40 participants (62 participants originally enrolled) (Vessal 2010). Pruritus was recorded on a VAS. The mean pruritus score was 8.48 ± 2.2 (range 4 to 10, median 10) in the placebo group (19 participants evaluated) and decreased to 5.58 ± 3.8 (range 0 to 10, median 6) after eight weeks of therapy (P = 0.004). The mean pruritus score was 8.68 ± 1.8 (range 4 to 10, median 10) in the cromolyn sodium group (21 participants evaluated) and decreased to 0.9 ± 1.8 (range 0 to 6, median 0) after 8 weeks (P < 0.001). Repeated measures analysis of variance showed a significant difference in pruritus scores between the two groups during the period of study (P < 0.001).
Activated oral charcoal
The study by Pederson et al (Pederson 1980) contained two consecutive eight-week treatment periods involving 11 participants with CKD-related pruritus. This randomised, placebo-controlled, cross-over design showed a significantly greater effect for charcoal over placebo during the first study period (P = 0.01) and a nearly significant effect during the second study period (P = 0.05). Missing data did not enable further statistical analysis.
In a randomised, double-blind, placebo-controlled cross-over trial including 10 participants with cholestatic pruritus by Borgeat et al (Borgeat 1993), treatment with propofol was described as successful in 17 of 20 (85%) doses of propofol, compared to two of 20 (10%) doses in the placebo group (P < 0.01). Data did not allow further statistical analysis.
A randomised, double-blind, placebo-controlled study investigated the efficacy of lidocaine on treatment-resistant pruritus in patients with chronic cholestatic liver disease (Villamil 2005). Lidocaine administration resulted in a significant reduction of pruritus severity when compared with placebo administration (P < 0.05). This difference was evident at day 2 (mean VAS 39.1 ± 23.4 versus 70.8± 8.1; P < 0.05 versus placebo) and day 3 (mean VAS 48.7 ± 23.3 versus 72.0 ± 11.7; P < 0.05). Significant differences with respect to baseline were found only for the treatment group (P < 0.05).
Four RCTs tested the efficacy of the topical agent capsaicin in treating pruritus in CKD-related or uraemic pruritus. Three of the studies did not provide sufficient data or appropriate statistical analyses for their findings to be evaluated (Breneman 1992b; Cho 1997; Tarng 1996). In the fourth study (Makhlough 2010), data from week four of treatment could be analysed.
The study of Breneman et al (Breneman 1992b) compared topical capsaicin and placebo applied four times daily to either the right or left arm over six weeks. The sample was very small (N = 7), and only five participants were evaluable. Findings were presented descriptively for these five participants (two capsaicin-treated participants reported complete resolution of itching), and no statistical analysis was available. Breneman stated an improvement of pruritus for the participants treated with capsaicin.
Cho et al (Cho 1997) also researched the effect of topical capsaicin versus placebo applied four times daily to a pre-selected area of skin in a cross-over study involving 22 participants during a treatment period of four weeks. The authors reported a significant effect of capsaicin in reducing itch; however, no raw data and no between-group comparisons were given. In addition, findings for the individual participants were presented graphically and indicated a substantial carry-over effect of the capsaicin treatment, which would make data from the second treatment phase invalid.
A similar cross-over study with two four-week treatment periods of capsaicin and placebo was carried out by Tarng et al. The study involved 19 participants and also reported a significant effect of capsaicin (Tarng 1996). Despite a 14-day washout period between the treatments, a carry-over effect must be assumed according to the graphical data presented. No data allowed for inter-group statistical comparisons to be made for the first phase of the cross-over.
The fourth cross-over study, undertaken by Makhlough et al (Makhlough 2010), researched 34 patients on haemodialysis with uraemic pruritus. Participants received capsaicin 0.03% and placebo for four weeks each, with a two-week washout period between the treatment phases. There was no significant difference in pruritus scores between the two groups before the treatment. However, the difference became more significant with each week (P < 0.001). Repeated measurement tests showed that decreases in pruritus severity in the capsaicin group were more than those in the placebo group during the treatment period (P < 0.001). Analysis of the data in week four of treatment showed a significant effect of topical capsaicin with a MD of -0.80 (95% CI -1.34 to -0.25) ( Analysis 6.1). Sensitivity analysis showed no different effect ( Analysis 6.2).
A randomised, double-blind, vehicle-controlled study by Duque et al assessed the efficacy of tacrolimus ointment 0.1% for the treatment of pruritus in patients undergoing haemodialysis. The study found a similar effect for both tacrolimus ointment and placebo in regard to the reduction of itch (72% to 81%) (Duque 2005).
In a randomised, double-blind, controlled, comparative trial, Young et al found a 61% decrease in the average reported VAS of participants treated with pramoxin hydrochloride as compared to participants treated with placebo. Since data were presented graphically and no raw data were given, statistical analysis of the data was not possible (Young 2009).
Participants with pruritus because of HIV infection: hydroxyzine hydrochloride, pentoyxifylline, triamcinolone, and indomethacin
A randomised parallel group study on 40 participants examined the four different therapies hydroxyzine hydrochloride, pentoxifylline, triamcinolone, and indomethacin in HIV participants suffering from pruritus (Smith 1997). Results showed that participants placed on indomethacin obtained relief more consistently and more completely. Participants on pentoxifylline had the fewest side effects of all oral therapies. Participants on antihistamines with or without doxepin had the highest incidence of side effects. These participants reported a greater degree of relief than participants on pentoxifylline. All participants on oral therapy had greater relief overall than participants using topical steroids.
Only few of the included studies examined the secondary outcomes quality of life, patient satisfaction, and depression.
Quality of life
Quality of life as a secondary outcome was measured in three studies (Kuiper 2010; Turner 1994a; Turner 1994b). Whereas Kuiper et al used the Short Form 36 and Liver Disease Symptom Index 2.0 to measure quality of life, Turner et al used a VAS.
Regarding quality of life in patients receiving colesevelam for treatment of cholestatic pruritus, Kuiper et al found no statistically significant changes with respect to the domains of physical functioning (P = 0.67), role physical functioning (P = 0.50), bodily pain (P = 1.00), general health (P = 0.48), vitality (P = 0.90), social functioning (P = 0.37), emotional functioning (P = 0.17), and mental health (P = 0.26). Results were based on the Short Form 36 questionnaire in the colesevelam group before and after treatment.
For patients receiving low dose flumenicol (Turner 1994a), the difference in median improvement in quality of life assessment between flumecinol and placebo, measured via the VAS, was 5.0 mm (95% CI 0.4 to 13.0, P = 0.02), in favour of flumecinol. Quality of life was not significantly improved by the higher dose of flumecinol. The difference in median improvement between the two groups was 3.5 mm (95% CI - 5.9 to 24.9 mm) (Turner 1994b).
Patient satisfaction with the treatment regimen was studied in only one of 40 included studies (Zylicz 2003). Using a non-validated seven-point scale, where 0 meant indifferent, -3 meant extremely poor, and 3 meant excellent, patients treated with paroxetine had, on average, higher satisfaction scores (mean 0.41, SE = 0.36) when compared to patients who received placebo (mean - 0.66, SE = 0.36), regardless of the order in which they were received. The mean difference was -1.08 (95% CI -0.19 to -96) between paroxetine and placebo and was significant (P = 0.027) ( Analysis 7.1; Analysis 7.2).
Two studies examined depression as a secondary outcome (Bergasa 2006; Mayo 2007). Depression was measured using the Hamilton depressing rating scale, the Structured Clinical Interview Questionnaire (SCID) (Bergasa 2006) and the 30-item Inventory of Depressive Symptomatology-Self-Report (IDS-SR
evaluation were available for 13 of the 16 patients (Bergasa 2006).
Mayo et al found that all four participants with moderate or severe depression improved with sertraline (12 participants included). Participants with mild depressive symptoms, however, did not reliably improve their IDS-SR
Due to the absence of satisfactory data, the feasibility of pooling the secondary outcomes data was limited. For detailed results please see the 'Secondary outcomes' table ( Table 3).
Data on adverse effects were collected in all but three included studies (Ashmore 2000; De Marchi 1992; Pederson 1980). Seven studies did not observe any adverse effects (Ghent 1988; Özaykan 2001; Podesta 1991a; Silva 1994; Turner 1994a; Turner 1994b; Young 2009).
Ten studies described adverse effects in the intervention group leading to withdrawal (Kumagai 2010; Legroux-Crespel 2004; Murphy 2003; Nasrollahi 2007; Pauli-Magnus 2000; Pour-Reza-Gholi 2007; Terg 2002; Wikström 2005a; Wikström 2005b; Zylicz 2003). In contrast, only four studies found adverse effects in the placebo groups or the standard medication group (Kumagai 2010; Legroux-Crespel 2004; Pauli-Magnus 2000; Wikström 2005a).
Most adverse effects observed were mild or moderate. Two interventions also showed multiple major adverse effects (naltrexone and nalfurafine).
Because of the diversity of different interventions and the absence of satisfactory data, meta-analysis for adverse effects was not possible. In many cases, the number of adverse events exceeded the number of patients, making the results of statistical analyses incoherent. On the one hand, one patient may experience several adverse events simultaneously. On the other hand, especially in the cross-over studies, the data does not allow us to identify which patient in which group of the study experienced the adverse event. Therefore, it was not possible to allocate the adverse events described in the studies to a special patient and to adequately calculate the number of adverse events in relation to the factual number of patients included. For this reason, the adverse events and the number of withdrawals for each intervention have been summarised in a synoptical table. For detailed information regarding adverse events, please see Table 4, 'Adverse effects according to the different studies', and Table 5, 'Adverse effects according to different interventions'.
Summary of main results
We identified 40 studies assessing the effects of 30 different interventions against pruritus of multiple origin in patients with advanced diseases. The majority of palliative care patients included in the studies were treated on normal wards due to the fact that many hospitals still do not have a palliative care ward. Thus, only 26 participants included were treated in palliative care settings like palliative care wards (Zylicz 2003). All other participants have been identified as palliative care patients but have not been treated in palliative care settings. Due to the diversity of interventions and the different origins of pruritus studied, no final and conclusive conclusions may be drawn regarding the treatment of pruritus in palliative care patients at large. Nevertheless, this systematic review demonstrates that pruritus in palliative care is a complex phenomenon. Until now, there is only one RCT that has researched pruritus as a symptom in palliative care patients (Zylicz 2003). All other studies focused on different kinds of pruritus based on the underlying disease of the patients. The kind of pruritus most frequently researched was chronic kidney disease (CKD)-related pruritus (947 participants). Consequently, the CKD patient group comprises more than 80% of the evidence. Overall, a total of 38 studies researching 12 treatments for CKD-related (that is uraemic) pruritus, nine interventions for cholestatic pruritus (CP) and four treatments focusing on both uraemic and cholestatic pruritus could be included in this review. Furthermore, one study researching HIV-related pruritus could also be included. The main results of this review were reported according to the structure resulting from these findings.
Palliative care patients - pruritus of different origin
The study involving palliative care patients with a heterogenous, representative palliative care patient population researched the effect of paroxetine on severe non-dermatological pruritus. Patients with solid tumours, haematological malignancies, and various non-malignant or idiopathic conditions were included in the study. Several patients experienced severe adverse effects, mainly nausea and vomiting, presumably because there was no opportunity to titrate the dose to the effect of the medication in this trial. However, the secondary endpoint of treatment, which was a 50% reduction in itch intensity, was robust and possibly compensated for other shortcomings in the study. The outcome of the study indicates that paroxetine is effective in the treatment of severe pruritus of non-dermatological origin. Still, the number of patients included in this study was quite small, the treatment period was brief, and there was no follow-up. Furthermore, the dose was not related to the effect of the medication. The study population was dominated by patients with pruritus in the course of advanced neoplastic disease and the conclusions cannot be extrapolated to all patients with severe pruritus. In addition, two participants dropped out because of adverse effects.
CKD-associated and cholestatic pruritus
Four interventions aimed at the treatment of both CKD-related pruritus and cholestatic pruritus.
Naltrexone has been shown to be effective against pruritus in patients with advanced diseases in four studies. Two studies focused on uraemic pruritus (Legroux-Crespel 2004; Peer 1996) and two on cholestatic pruritus (Wolfhagen 1997; Terg 2002). Thereby, Legroux-Crespel et al found that naltrexone was effective only in a subset of patients and indicated that the differences in efficacy and tolerance between patients might be due to metabolism. The results of the study by Pauli-Magnus et al on patients with uraemic pruritus were contrary to these findings and show an advantage of naltrexone compared to the placebo (Pauli-Magnus 2000). A pooled estimation for the studies researching naltrexone in uraemic pruritus was not possible. Thus, the results of the studies pertaining to uraemic pruritus are conflicting. For at least the two studies by Pauli-Magnus and Peer, differences cannot be explained by variations in participant compliance, naltrexone dose, study design, or methodological quality. However, the trial of Legroux-Crespel et al should be interpreted with caution because of missing data, high dropout rates, and high risk of bias. Concerning cholestatic pruritus, naltrexone was shown to be effective and pooled estimation resulted in a significant effect in favour of naltrexone. Still, the study population included was quite small (16 and 20 participants respectively), and naltrexone has the highest incidence of adverse effects.
Ondansetron was also applied for either CKD-associated pruritus or cholestatic pruritus. Three of four included studies reported no advantage over placebo, either for use in cholestatic pruritus (O'Donohue 2005) or for uraemic pruritus (Ashmore 2000; Murphy 2003). One study researching ondansetron for treatment of uraemic pruritus found a benefit for patients treated with ondansetron compared to patients treated with the standard medication cyproheptadine (Özaykan 2001). The quality of the studies researching uraemic pruritus was weak, and the data did not allow pooling. Interestingly, in all studies comparing ondansetron with placebo, the placebo group showed an improvement in pruritus, which was even significant in one study (Murphy 2003). The authors hypothesize the possibility of a Type II error. Overall, ondansetron is not recommended for treatment of uraemic or cholestatic pruritus.
A total of three studies concentrated on gabapentin for the treatment of uraemic pruritus (Gunal 2004; Naini 2007) and cholestatic pruritus (Bergasa 2006). For the therapy of uraemic pruritus, gabapentin has been proven to be a safe and effective intervention, showing a significant improvement of pruritus ( Analysis 4.1). These results may also support the neuropathic hypothesis of uraemic pruritus. The adverse effects observed were only minor. Although the quality of both studies is limited because of small sample sizes and unclear risk of bias, gabapentin may be an option for treatment of CKD-associated pruritus in patients with advanced diseases.
No significant therapeutic effect was shown concerning gabapentin as an intervention in cholestatic pruritus (Bergasa 2006). The findings of the study even indicate that gabapentin appears to worsen pruritus. On the other hand, a strong placebo effect was observed. In this context, the authors discuss whether gabapentin possibly interferes with the placebo effect (association of a placebo intervention with dopamine release). Even though the small number of patients was a limitation of this study, several scientific questions emerged from this trial, including the role of dopamine neurotransmission in the treatment of pruritus and the neurophysiologic basis for the placebo effect on the pruritus of cholestasis.
The fourth intervention researching uraemic (Silverberg 1977) and cholestatic pruritus (Duncan 1984) was the bile acid sequestrant cholestyramine. The two studies showed some positive effect for treatment of both uraemic and cholestatic pruritus. However, both studies suffered from very small sample sizes (eight to 10 patients included), and they were of low methodological quality. Thus, the validity of the results is limited.
Chronic kidney disease-associated or uraemic pruritus
Nalfurafine for uraemic pruritus was examined in three studies including a total of 422 participants (Kumagai 2010; Wikström 2005a; Wikström 2005b). The meta-analysis of two studies by Wikström et al indicates that two weeks of treatment with nalfurafine could indeed be beneficial in uraemic pruritus. Nevertheless, the improvement in pruritus seen with nalfurafine was rather modest when compared to the placebo, which actually had an impressive benefit. In addition, statistical benefit did not persist when treatment was continued for four weeks in a subgroup of patients. However, the trial by Kumagai et al resulted in a significant effect in favour of nalfurafine. This multicentre study was of good quality and showed a low risk of bias. Analysis of the pooled data showed a significant effect in favour of nalfurafine ( Analysis 2.1).
Additional systemic interventions like erythropoietin (De Marchi 1992), montelukast (Nasrollahi 2007), thalidomide (Silva 1994), activated oral charcoal (Pederson 1980), and doxepin (Pour-Reza-Gholi 2007) were also investigated. All of these interventions showed some positive effect on uraemic pruritus. However, the sample sizes of the studies were small (16 to 29 participants), and in four of the six studies methodological quality was limited (De Marchi 1992; Nasrollahi 2007; Pederson 1980; Silva 1994). Additional trials confirming the results have not been found. A trial by Villamil et al investigating oral cromolyn sodium found a significant reduction in the severity of pruritus in patients with CKD-associated pruritus (Villamil 2005). The overall quality of the study was high and the adverse effects observed were minor. However, no other study has been found confirming these results. There is some evidence that oral cromolyn sodium may be an alternative therapy for patients suffering from pruritus of cholestatic origin.
Six studies investigated the effect of topical agents on uraemic pruritus. Whereas tacrolimus ointment was not more effective than the vehicle in relieving uraemic pruritus (Duque 2005), pramoxine lotion reduced pruritus to a greater degree than the control lotion (Young 2009). Adverse effects for tacrolimus were minor, and for pramoxine hydrochloride no adverse effects were observed. However, both studies had a high risk of bias. Thus, evidence for the use of these topical applications is low.
A total of four studies compared the efficacy of topical capsaicin and placebo creams in treating CKD-associated pruritus (Breneman 1992b; Cho 1997; Makhlough 2010; Tarng 1996). Capsaicin was shown to markedly improve pruritus. A common side effect of topical capsaicin was a transient burning sensation and local erythema with initial application. These may be classified as minor adverse effects. However, three of the studies demonstrated methodological flaws such as inappropriate designs leading to carry-over effects, inadequate depiction of data, and failure to provide appropriate statistical analyses, which resulted in a high risk of bias (Breneman 1992b; Cho 1997; Tarng 1996). The recent study by Makhlough et al (Makhlough 2010) also reported a significant effect of capsaicin over placebo. However, similar to the other studies, this study was at high risk of bias because of possible attrition and reporting bias and also showed methodological shortcomings. Overall, these studies provide no convincing evidence for the effectiveness of topical capsaicin in CKD-related pruritus.
Six interventions targeted cholestatic pruritus in patients with advanced disease. The effect of rifampin or rifampicin was explored in three studies (Bachs 1989; Ghent 1988; Podesta 1991a). Despite the lack of clarity regarding concealment of allocation and the absence of blinding in one of these trials (Bachs 1989), the overall quality of the trials was adequate. All trials clearly stated the primary endpoints, that is relief from pruritus and rifampin-induced side effects. The quantitative results indicate that rifampin is effective for treatment of pruritus associated with chronic cholestasis when compared with a placebo or phenobarbitone ( Analysis 5.2). The incidence of adverse effects was low in patients treated with rifampin for up to 14 days.
Flumecinol was proven to ameliorate pruritus in patients suffering from cholestatic pruritus in two studies by Turner et al using different dosages (Turner 1994a; Turner 1994b). The studies showed that flumecinol at a dose of 300 mg daily for three weeks (in contrast to 600 mg once weekly for three weeks) significantly improved cholestatic pruritus compared to the placebo. Both studies were of high quality and the risk of bias was low. No adverse effects were observed. However, the majority of the patients included suffered from primary biliary cirrhosis and transferability of these results to different kinds of pruritus would have to be explored.
Two studies investigated the effect of the anaesthetics propofol (Borgeat 1993) and lidocaine (Villamil 2005). Both trials state an amelioration of pruritus for patients treated with propofol and lidocaine. Side effects reported were minor. Since the study by Borgeat et al was assessed to have a high risk of bias due to possible attrition bias, the evidence for the use of propofol is low (Borgeat 1993). The study by Villamil et al shows an unclear risk of bias because of possible selective reporting. However, the applicability of lidocaine as an alternative therapy for cholestatic pruritus is limited because of the necessity of hospital or inpatient treatment of the patients concerned.
The selective serotonin reuptake inhibitor sertraline has been shown to be effective and well-tolerated in a trial by Mayo et al (Mayo 2007) on patients with cholestatic pruritus. However, the sample size of 12 participants was very small and there was some precariousness in the blinding of outcome assessment. Additional studies confirming or disagreeing with these results are lacking.
One study by Smith 1997 researched pruritus in patients with HIV and explored four different interventions compared to a placebo. Since no information about blinding was given and additional data were missing, the risk of bias of the study was assessed as high. No conclusive recommendations may be deduced from the study results. However, indomethacin was described as the most effective oral drug.
Conclusions for treatment
Thirty different treatments for pruritus in four different patient groups have been integrated in this review. Conclusions for treatment are discussed for the four patient groups.
For patients in palliative care settings who mainly suffered from pruritus related to solid tumours, haematological malignancies, and various non-malignant or idiopathic conditions, the selective serotonin reuptake inhibitor (SSRI) paroxetine has been shown to be effective. Because serotonin might have a role in pruritus secondary to malignant disease, as well as cholestasis, uremia, and opioids, it is reasonable to try the drug. In order to reduce adverse effects, patients should start with small doses, such as 5 to 10 mg nightly, which will cause fewer side effects than higher doses. Effects can usually be observed within 24 to 48 hours. Results of the study should be confirmed by further trials using these lower doses of paroxetine.
Four interventions aimed at treating both CKD-related pruritus and cholestatic pruritus. The opioid antagonist naltrexone offers a therapeutic alternative for patients suffering from CKD-associated or cholestatic pruritus, as μ-opioid receptor agonists are central mediators of pruritus. However, these drugs are often inappropriate in the palliative population because of the risk of loss of analgesia at higher doses of naltrexone. Pooled data suggest that pain is reported by up to 84% of palliative care patients and by over 70% of patients with advanced neoplasms (Higginson 1997; Potter 2003; Walsh 2000). Palliative care patients might be using opioids for the management of pain or dyspnoea, and opioid antagonists could reverse analgesia or lead to withdrawal symptoms.
The serotonin 5-HT
The major proportion of the included participants was suffering from CKD-related pruritus. Thus, the evidence for treatment of CKD-related pruritus is documented best.
For patients suffering from CKD-associated pruritus, gabapentin may be an option for treatment of CKD-associated pruritus in patients with advanced diseases. It has been proven to be a safe and effective intervention, showing a significant improvement of pruritus and only minor adverse effects. For patients suffering from cholestatic pruritus, gabapentin seems to worsen pruritus and is not recommended.
Sequestrans cholestyramine showed some positive effect for the treatment of cholestatic pruritus (Duncan 1984) and uraemic pruritus (Silverberg 1977). Oral charcoal may be beneficial for patients with uraemic pruritus (Pederson 1980). However, because of very small sample sizes, the evidence is limited. These treatments may be useful as adjuvant or alternative therapies.
The κ-opioid receptor agonist nalfurafine appears to be beneficial in CKD-related pruritus. It shows significant amelioration in the three studies included. Since adverse effects are tolerable, nalfurafine may be recommended for treatment of uraemic pruritus. Interestingly, nalfurafine was officially approved for clinical use as an antipruritic for CKD-associated pruritus in Japan in January 2009.
A significant reduction in the severity of pruritus in patients with CKD-associated pruritus was also found for oral cromolyn sodium. Cromolyn sodium may offer an alternative therapy for patients with refractory uraemic pruritus and should be subjected to further randomised placebo-controlled trials.
Thalidomide has been shown to be effective in over 50% of patients with CKD-related pruritus. Due to its potential for neuropathy and its poor availability, thalidomide should not be used unless other treatments fail.
Evidence for the topical treatments is only low. Capsaicin has been investigated in several studies with limited quality and has shown some benefit. Capsaicin is accompanied by initial burning and stinging and may be useful as additional treatment in mild or moderate uraemic pruritus. This may also hold true for pramoxine hydrochloride lotion.
For patients suffering from pruritus associated with HIV infection, no distinct conclusion may be drawn as the evidence was very low. However, indomethacin was described as the most effective oral drugs.
The review of the interventions targeted on cholestatic pruritus in patients with advanced disease resulted in a positive effect in favour of rifampin or rifampicin. Showing a low incidence of adverse effects, rifampin or rifampicin may be recommended for patients with cholestatic pruritus. Nevertheless, patients should be informed that this drug can possibly change the colour of body fluids. Flumecinol was also proven to ameliorate pruritus in patients suffering from cholestatic pruritus and is recommended for patients with primary biliary cirrhosis. However, transferability of these results to different kinds of pruritus must be explored. Anaesthetic lidocaine may be used to ameliorate cholestatic pruritus in some cases but its applicability is limited because of the need of hospitalisation of the patients. Last but not least, the selective serotonin reuptake inhibitor sertraline has been shown to be effective and well-tolerated in a trial on patients with cholestatic pruritus. However, due to the small sample size, the evidence is limited.
Overall completeness and applicability of evidence
This review included 40 RCTs. Only one study researched pruritus in palliative care patients who received palliative care. According to the presumed definition of palliative care patients as "adult patients in any setting, receiving palliative care or suffering an incurable progressive medical condition", we also included studies researching pruritus in patients with advanced and incurable diseases in different settings. Thus, a total of 30 different interventions for treatment of four different groups of patients suffering from pruritus of different origins was explored. Due to the diversity of the interventions, and the small number of studies per intervention, we could not compare the effectiveness of all interventions. As the overall quality of studies varied, missing data in several studies did not allow for including data from all studies in meta-analyses. Results of several studies had to be described as a narrative summary. Due to the diversity and different character of pruritus and the patient groups researched, the applicability of evidence is restricted to the particular patient group targeted by the intervention.
Quality of the evidence
The overall quality of included studies was variable. Most had adequate randomisation and blinding procedures, although a few studies did not. Two of the included studies were single-blind (Duncan 1984; Nasrollahi 2007); two studies were open trials (Bachs 1989; Özaykan 2001); and another two studies had unclear use of blinding (Legroux-Crespel 2004; Smith 1997). Risk of bias concerning selective reporting was unclear in 16 of 40 studies, and three studies showed a high risk of selective reporting (Breneman 1992b; Gunal 2004; Legroux-Crespel 2004). Potential sources of bias were the small sample sizes in some of the studies (sample size less than 20 in 16 studies), missing washout periods leading to carry-over-effects in cross-over studies (Cho 1997; Tarng 1996), and inadequate study design (Smith 1997; Tarng 1996). Due to the large number of studies at unclear and high risk of bias, the findings of this review must be interpreted with caution.
Potential biases in the review process
Defining the population of palliative care patients has been identified as a problem in previous reviews. Therefore, we used the definition 'adult patients in any setting, receiving palliative care or suffering an incurable progressive medical condition', which has previously been used in other Cochrane reviews (Dorman 2010; Perkins 2009). However, it is still difficult to identify the patients or patient groups who are appropriate to be included in a systematic review on palliative care topics as there is still a lack of original studies, particularly RCTs in palliative care. In a study in which 25 Cochrane reviews in palliative care, published up to December 2007, had been reviewed (Wee 2008), the authors concluded that "Cochrane reviews in palliative care ... fail to provide good evidence for clinical practice because the primary studies are few in number, small, clinically heterogeneous, and of poor quality and external validity". Nevertheless, data on palliative care patients may be hidden in studies not described as palliative care topics. Therefore, it is possible that some published or unpublished trials including palliative care patients have not been identified for this review.
In this review, a comprehensive literature search was performed, inclusion and exclusion criteria were specified, and data analysis was conducted where possible. We searched all available databases, checked reference lists of all relevant trials, contacted relevant experts for identification of unpublished trials, searched registers of ongoing trials and available proceedings of conferences, and checked and included publications without restricting language. We performed a systematic literature search up to January 2012 and updated the search in August 2012. One database (BIOSIS Previews 1969 to 2010, including BIOSIS Previews Archive 1926 to 1968, via OvidSP) could only be searched up to 2010 as the licence ended in 2010. For literature from 2011 up to now we searched 'Web of knowledge' instead. In the literature search two studies meeting all inclusion criteria were identified (Ghorbani 2011; Sja'bani 1997), but data could not be retrieved and the authors could not be contacted or did not answer our enquiry, respectively. Inability to include this data could have led to publication bias or could have changed the results concerning a special intervention. Another two studies (Feily 2012; Najafabadi 2012) have been identified as eligible for the review from the update search in August 2012, but have not been included in the review as the evaluation of the data had already been completed. We cited these studies in the section 'Studies awaiting classification'.
In order to avoid duplicate publication bias, it was crucial to identify all redundant and multiple publications, which can lead to overestimation of intervention effects. One study identified in this review was found to be repetitive or overlapped substantially with an included study and therefore was excluded.
In this review, there was no language restriction. As a result, a Turkish language study (Özaykan 2001) was translated before data extraction. We accept possible lack of clarity due to translation.
A limitation of this review is that, with the exception of one study (Kumagai 2010), all included studies had only small to medium sample sizes, with an average of about 33 participants. Small trials have less power, meaning that there is less chance of detecting a small but true effect as being statistically significant. As we included 24 different interventions, only 11 interventions were explored by two or more trials. For 13 interventions there was only one study available. Consequently, the evidence concerning these interventions was limited and the validity was low. The informative value of this review may also be limited by the inhomogeneous number of participants included. About three-fourths of the participants suffered from CKD-related pruritus and only 26 participants were treated in palliative care settings. Another limitation of this review may be that the majority of secondary outcomes investigated in the different studies were not included in our review. This again is due to the diversity of the interventions included. Addressing all important outcomes was not possible within the constraints of this review.
Agreements and disagreements with other studies or reviews
The results of this review are generally consistent with the findings of other recent reviews conducted within the fields of pruritus. Several drugs and interventions have been summarised in different reviews. However, these reviews did not focus on the field of palliative care, nor did they examine different therapeutic effects like pain or depression (Derry 2012; Derry 2013; Moore 2011).
Reviews and systematic reviews on the following interventions were found, and the results have been compared to the results of this review:
Phan et al reviewed the use of systemic μ-opioid receptor antagonists (MORA) in the treatment of various forms of chronic pruritus. The RCTs identified by Phan et al that researched patients with advanced diseases are similar to the RCTs included in this review (Phan 2010).
The effect of topical capsaicin is explored in the systematic review of Gooding et al (Gooding 2010). Since the review by Gooding was conducted in 2010, the study on topical capsaicin by Makhlough et al (Makhlough 2010) was not included in his review; however, the overall results are comparable with our findings.
Implications for practice
At present there is no universally accepted therapy for itch. Furthermore, the varying pathogenesis of pruritus in different disorders means that a universally accepted therapy is difficult to establish. The therapy of pruritus is challenging and requires an individualistic approach. Therefore, identifying the underlying cause of pruritus is still of prime importance in order to develop tailored treatment plans. Especially in palliative care, patients with pruritus may have more than one origin for their pruritus. The fact that itch affects the skin, immune system, and the peripheral and central nervous system means that complex and combinatory pathways are likely to be more effective than a single-line approach. Clinical management of chronic pruritus should include the use of drugs that reduce neuronal sensitization for pain (such as gabapentin) and selective serotonin and antidepressants, either as mono-therapies or in combination. Topical therapies, which are the mainstay of therapy for mild and localized itch, should be used as a supplement. A future goal is a wide range of topical and systemic therapies that target the various receptors and neural pathways that mediate itch of different types and lead to improved management of all kinds of pruritus.
Implications for research
The findings of this review indicate that the number of topical and systemic drugs used for the different subforms of pruritus is increasing. However, optimal therapy is constrained by the fact that the present understanding of crucial itch mediators and receptors in the various subforms of itch is limited. Especially in the field of palliative care, various forms of pruritus occur and sometimes the origin of the pruritus is difficult to determine. The fact that itch affects the skin, immune system, and the peripheral and central nervous system means that complex and combinatory pathways may be more effective than a single-line approach. Therefore, it is difficult to know which medication to apply for specific subtypes of itch, and why. In the future, larger studies would help delineate the efficacy of available and proposed antipruritics. Studies in the field of palliative care are especially lacking, and the evidence for interventions targeting palliative care patients is low. Therefore, well-designed treatment studies, where possible placebo-controlled and randomised, are needed to further verify the effectiveness of many antipruritic agents currently being used.
The opioid antagonist naltrexone has already been shown to be effective in the treatment of CKD-associated and cholestatic pruritus. As already mentioned, these drugs are often inappropriate in the palliative population because of the risk of loss of analgesia at higher doses of naltrexone. Interestingly, methyl naltrexone, a peripherally acting MORA, was recently approved by the Food and Drug Administration for the treatment of opioid-induced constipation in patients with advanced illness (for example cancer, AIDS) receiving palliative care. Notably, methyl naltrexone does not cross the brain barrier, offering the advantage of peripheral action only and thus significantly fewer adverse effects including addiction. Clinical and laboratory studies performed during the development of methyl naltrexone have indicated that peripheral opioid receptors mediate certain effects, including nausea, vomiting, and pruritus. Although subcutaneous methyl naltrexone has not been used as an antipruritic, it should theoretically be of value in the treatment of pruritus. Researching the efficacy of this medication in larger, high quality studies, especially in the field of palliative care, would be of particular interest. In addition, given the possible role of the opioidergic system in pruritus, the role of the μ-opioid receptor antagonist naltrexone in a topical 1% form in the treatment of severe pruritus might be of interest for palliative care patients and should also be researched in larger studies (Bigliardi 2007).
The authors wish to thank the patients who entered the trials and the investigators who conducted them.
We should like to thank Edith Motschall, Information retrieval specialist at the German Cochrane Center and Institute of Medical Biometry and Medical Infomatics, University Medical Center Freiburg, for her great support with development of literature search strategies.
We thank Mascha Bussmann, Anette Lampert and Karin Jors, graduate assistants at the Department of Palliative Care, University Medical Center Freiburg, for their help with data extraction.
We are grateful to Jessica Thomas, former Managing Editor of the Cochrane Pain, Palliative and Supportive Care Review Group and to Yvonne Roy, Assistant Managing Editor of the Cochrane Pain, Palliative and Supportive Care Review Group and we would like to thank them for their help and editorial advice during the preparation of the protocol and the review.
This review was funded by the German Ministry for Education and Research (BMBF), Germany
Grant No. 01KG0819.
The review was conducted independently of funding from any interested party.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Appendix 1. MEDLINE search strategy
mp=title, original title, abstract, name of substance word, subject heading word; ti=title; pt=publication type; ab=abstract; fs=floating subheading; sh=MeSH subject heading
1. exp Pruritus/
2. (prurit* or itch* or scratch*).ti.
3. ((prurit* or itch* or scratch*) adj10 (prevent* or stop* or alleviat* or relief* or reliev*)).mp.
4. 1 or 2 or 3
5. ((advance* or late or last or end or final) adj4 (stage* or phase*)).mp.
6. (palliat* or terminal* or endstage or end-stage or "end stage" or hospice* or (end adj3 life) or (care adj3 dying)).mp.
7. ((advance* or progressi* or terminal*) adj6 (ill* or disease* or condition*)).mp.
8. (terminal* adj6 (care or therap* or treat*)).mp.
9. exp Palliative Care/ or Terminal Care/ or Terminally Ill/ or Hospice Care/ or exp Home Care Services/ or exp Hospitals, Special/ or Attitude to Death/
10. 5 or 6 or 7 or 8 or 9
11. 4 and 10
For identification of randomised controlled trials (humans or human and animals) the subject search above will be combined with the following search strategy
Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity-maximizing version (2008 revision); Ovid format
1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
5. drug therapy.fs.
9. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8
10. exp animals/ not humans.sh.
11. 9 not 10
Appendix 2. Biosis search strategy
1. (prurit* or itch* or scratch*).m_titl.
2. ((prurit* or itch* or scratch*) adj10 (prevent* or stop* or alleviat* or relief* or reliev*)).mp.
3. 1 or 2
4. ((advance* or late or last or end or final) adj4 (stage* or phase*)).mp.
5. (palliat* or terminal* or endstage or end-stage or "end stage" or hospice or (end adj3 life) or (care adj3 dying)).mp.
6. ((advance* or progressi* or terminal*) adj6 (ill* or disease* or condition)).mp.
7. (terminal* adj6 (care or therap* or treat*)).mp.
8. 4 or 5 or 6 or 7
9. 3 and 8
10. (animals not (humans and animals)).sh.
11. 9 not 10
Appendix 3. CINAHL search strategy
1. MH pruritus
2. MJ prurit* or MJ itch* or MJ scratch*
3. prurit* N10 prevent* or prurit* N10 stop* or prurit* N10 alleviat* or prurit* N10 relief* or prurit* N10 reliev*
4. itch* N10 prevent* or itch* N10 stop* or itch* N10 alleviat* or itch* N10 relief* or itch* N10 reliev*
5. scratch* N10 prevent* or scratch* N10 stop* or scratch* N10 alleviat* or scratch* N10 relief* or scratch* N10 reliev*
6. S1 or S2 or S3 or S4 or S5
7. advance* N4 stage* or advance* N4 phase* or late N4 stage* or late N4 phase* or last N4 stage* or last N4 phase* or end N4 stage* or end N4 phase* or final N4 stage* or final N4 phase*
8. palliat* or terminal* or endstage or end- stage or "end stage" or hospice* or end N3 life or care N3 dying
9. advance* N6 ill* or advance* N6 disease* or advance* N6 condition* or progressi* N6 ill* or progressi* N6 disease* or progressi* N6 condition* or terminal* N6 ill* or terminal* N6 disease* or terminal* N6 condition*
10. terminal* N6 care* or terminal* N6 therap* or terminal* N6 treat*
11. MH Palliative Care or MH Terminal Care or MH Terminally Ill or MH Hospice Care or MH Home Care Services or MH Hospitals, Special or MH Attitude to Death
12. S7 or S8 or S9 or S10 or S11
13. S6 and S12
14. animals not (humans and animals)
15. S13 not S14
Appendix 4. EMBASE search strategy via Ovid
1. exp Pruritus/
5. ((pruri$ or itch$ or scratch$) adj5 prevent$).mp.
7. exp Antipruritic Agent/
9. (Antipruritic Agent or Pruritus or anti-pruritic$).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer, device trade name, keyword]
11. exp Drug Therapy/
12. drug therapy combination$.mp.
13. exp Drug Combination/
14. prevention control.mp.
15. (pharmacol$ adj3 (therap$ or treat$ or intervent$)).mp.
16. (pharmaceutic$ adj3 (therap$ or treat$ or intervent$)).mp.
17. (drug$ adj3 prevent$).mp.
18. pharmaceutic aid$.mp.
19. exp "Pharmaceutical Vehicles and Additives"/
21. (endstage or end-stage).mp.
23. (terminal$ adj3 ill$).mp.
24. palliative care.mp.
25. exp Palliative Therapy/
26. (palliativ$ adj3 car$).mp.
27. (palliative adj3 therap$).mp.
28. (palliative adj3 treat$).mp.
29. (terminal adj3 care).mp.
30. (terminal adj3 disease).mp.
31. (terminal adj3 treat$).mp.
32. (end adj3 life adj3 care).mp.
33. exp Home Care/
34. hospice care.mp.
35. exp Psychological Aspect/
36. exp Terminal Disease/
37. exp Terminal Care/
38. exp Terminally ill Patient/
39. exp Medicare/
40. exp Hospice/
41. exp Hospice Care/
42. exp Hospice Patient/
43. exp Patient Care/
44. (hospice adj3 care).mp.
46. exp hospice nursing/
47. exp hospital patient/
48. exp nursing home/
49. exp nursing home patient/
50. exp home for the aged/
51. ("nursing adj2 home$" or "home for the aged" or "old age home$").mp.
53. 6 and 20 and 52
Appendix 5. PsycINFO search strategy
S1 DE "Pruritus"
S2 MJ prurit* or MJ itch* or MJ scratch*
S3 prurit* N10 prevent* or prurit* N10 stop* or prurit* N10 alleviat* or prurit* N10 relief* or prurit* N10 reliev*
S4 itch* N10 prevent* or itch* N10 stop* or itch* N10 alleviat* or itch* N10 relief* or itch* N10 reliev*
S5 scratch* N10 prevent* or scratch* N10 stop* or scratch* N10 alleviat* or scratch* N10 relief* or scratch* N10 reliev*
S6 S1 or S2 or S3 or S4 or S5
S7 advance* N4 stage* or advance* N4 phase* or late N4 stage* or late N4 phase* or last N4 stage* or last N4 phase* or end N4 stage* or end N4 phase* or final N4 stage* or final N4 phase*
S8 palliat* or terminal* or endstage or end- stage or "end stage" or hospice* or end N3 life or care N3 dying
S9 advance* N6 ill* or advance* N6 disease* or advance* N6 condition* or progressi* N6 ill* or progressi* N6 disease* or progressi* N6 condition* or terminal* N6 ill* or terminal* N6 disease* or terminal* N6 condition*
S10 terminal* N6 care* or terminal* N6 therap* or terminal* N6 treat*
S11 MH Palliative Care or MH Terminal Care or MH Terminally Ill or MH Hospice Care or MH Home Care Services or MH Hospitals, Special or MH Attitude to Death
S12 Palliative or Palliative care or Palliative treatment or Terminal care or Terminally ill or Hospice or Hospice care or Home care service or Attitude to death
S13 S7 or S8 or S9 or S10 or S11 or S12
S14 S6 and S13
Appendix 6. Search strategy for The Cochrane Library
#1 MeSH descriptor Pruritus explode all trees
#6 (antipruritic* NEXT therap*):ti,ab,kw
#7 (#1 OR #2 OR #3 OR #4 OR#5 OR #6)
#8 (drug therap*):ti,ab,kw
#9 (pharmacologic* therap*):ti,ab,kw
#10 (pharmacologic* treat*):ti,ab,kw
#11 (pharmaceutic* therap*):ti,ab,kw
#12 (pharmaceutic* treat*):ti,ab,kw
#13 (#9 OR #10 OR #11OR #12)
#14 (advanced disease*):ti,ab,kw
#15 (advanced NEXT disease*):ti,ab,kw
#16 (palliative care):ti,ab,kw
#17 (palliative NEXT care):ti,ab,kw
#18 (hospice care):ti,ab,kw
#19 (hospice NEXT care):ti,ab,kw
#20 (terminal care):ti,ab,kw
#21 (terminal NEXT care):ti,ab,kw
#22 (terminal* ill*):ti,ab,kw
#23 (#14 OR #22)
#24 (#7 AND #13)
'#25 (#'23 AND #24)
Last assessed as up-to-date: 1 April 2013.
Contributions of authors
CX: developed and wrote the protocol, developed the search strategies and the data extraction form, extracted and screened data from the papers and entered data into RevMan.
SB: was engaged to develop the search strategies for several databases.
GS: provided statistical expertise.
JM, DG, GA: discussed the search strategy, the data extraction form and the protocol; screened retrieved papers, were involved in drawing up the discussion of the review.
GB: developed the idea of the review, contributed to the search strategy and the data extraction form, discussed and approved the protocol, was responsible for the write up of the review, is the guarantor of review and will be responsible for the update of the review.
Declarations of interest
Sources of support
- German Cochrane Centre, Freiburg, Germany.
- German Ministry for Education and Research (BMBF), Germany.Grant No. 01KG0819
Differences between protocol and review
In contrast to our preliminary definition in the protocol the present review only includes randomised controlled trials. We did not include controlled trials or observational studies.
In addition, in the background some modification and improvements concerning the Cochrane Review Management Program have been made. First of all, in RevMan 5.1 an updated version of the ‘Risk of bias tool’ has been implemented. We adapted our assessment criteria as described in the protocol according to the new risk-of-bias tool. Hence, the risk of bias assessment was performed as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011; Review Manager 2011).
Medical Subject Headings (MeSH)
*Palliative Care; Anesthetics [therapeutic use]; Anti-Inflammatory Agents, Non-Steroidal [therapeutic use]; Cholestasis [complications]; HIV Infections [complications]; Pruritus [*drug therapy; etiology; prevention & control]; Receptors, Opioid, kappa [agonists]; Renal Insufficiency, Chronic [complications]; Serotonin Uptake Inhibitors [therapeutic use]
MeSH check words