Spinal cord stimulation for cancer-related pain in adults

  • Review
  • Intervention

Authors

  • Peng Lihua,

    1. The First Affiliated Hospital, Chongqing Medical University, Department of Anaesthesia and Pain Medicine, Chongqing Municipanity, Chongqing Municipanity, China
    Search for more papers by this author
  • Min Su,

    Corresponding author
    1. The First Affiliated Hospital, Chongqing Medical University, Department of Anaesthesia and Pain Medicine, Chongqing Municipanity, Chongqing Municipanity, China
    • Min Su, Department of Anaesthesia and Pain Medicine, The First Affiliated Hospital, Chongqing Medical University, No 1 Youyi Road, Yuanjiangang community, Yuzhong District, Chongqing Municipanity, Chongqing Municipanity, 400016, China. minsu89011068@yahoo.com.cn.

    Search for more papers by this author
  • Zhou Zejun,

    1. The First Affiliated Hospital, Chongqing Medical University, Department of Anaesthesia and Pain Medicine, Chongqing Municipanity, Chongqing Municipanity, China
    Search for more papers by this author
  • Wei Ke,

    1. The First Affiliated Hospital, Chongqing Medical University, Department of Anaesthesia and Pain Medicine, Chongqing Municipanity, Chongqing Municipanity, China
    Search for more papers by this author
  • Michael I Bennett

    1. University of Leeds, Leeds Institute of Health Sciences, Leeds, UK
    Search for more papers by this author
    • Lancaster University, Institute for Health Research, Bowland Tower East Lancaster, LA1 4YT UK.


Abstract

Background

Cancer-related pain places a heavy burden on public health with related high expenditure. Severe pain is associated with a decreased quality of life in patients with cancer. A significant proportion of patients with cancer-related pain are under-treated.There is a need for more effective control of cancer-related pain. Spinal cord stimulation (SCS) may have a role in pain management. The effectiveness and safety of SCS for patients with cancer-related pain is currently unknown.

Objectives

This systematic review evaluated the effectiveness of SCS for cancer-related pain compared with standard care using conventional analgesic medication. We also appraised risk and potential adverse events associated with the use of SCS.

Search methods

We searched the following bibliographic databases in order to identify relevant studies: the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Libary (from inception to 2012, Issue 6); MEDLINE; EMBASE; and CBM (Chinese Biomedical Database) (from inception to July, 2012). We also handsearched relevant journals.

Selection criteria

We planned to include randomised controlled trials (RCTs) that directly compared SCS with other interventions with regards to the effectiveness of pain management. We also planned to include cross-over trials that compared SCS with another treatment. We planned to identify non-randomised controlled trials but these would only be included if no RCTs could be found.

Data collection and analysis

The initial search strategy yielded 430 articles. By scrutinising titles and abstracts, we found 412 articles irrelevant to the analytical purpose of this systematic review due to different scopes of diseases or different methods of intervention (intrathecal infusion system; oral medication) or aims other than pain control (spinal cord function monitoring, bladder function restoration or amelioration of organ metabolism). The remaining 18 trials were reviewed as full manuscripts. No RCTs were identified. Fourteen sporadic case reports and review articles were excluded and four before-and-after case series studies (92 participants) were included. Two review authors independently selected the studies to be included in the review according to the pre-specified eligibility criteria. A checklist for methodological quality of non-randomised controlled trials was used (STROBE checklist) and all review authors discussed and agreed on the inclusion of trials and the results of the quality assessment.

Main results

Four before-and-after case series studies (a total of 92 participants) met our criteria for inclusion. All included trials adopted a visual analogue scale (VAS) to evaluate pain relief. Heterogeneity existed in terms of baseline characteristics, electrode and stimulator parameters, level of implantation and route of implantation; data reporting was different among all trials. In two trials, pain relief was achieved in 76% (48/63) of patients at the end of the follow-up period. In the third trial, pre-procedure VAS was 6 to 9 (mean 7.43 ); the one-month post-implant VAS was 2 to 4 (mean 3.07); the 12-month post-implant VAS was 1 to 3 (mean 2.67). In the fourth trial, the pre-procedure VAS was 6 to 9 (mean 7.07); 1 to 4 (mean 2.67) at one-month; 1 to 4 (mean 1.87) at 12 months. Analgesic use was largely reduced. The main adverse events were infection of sites of implantation, cerebrospinal fluid (CSF) leakage, pain at the sites of electrodes, dislodgement of the electrodes and system failure, however, the incidence in patients with cancer could not be calculated. Since all trials were non-randomised controlled trials, they carried risk of all types of bias.

Authors' conclusions

Current evidence is insufficient to establish the role of SCS in treating refractory cancer-related pain. Future randomised studies should focus on the implantation of SCS in patients with cancer-related pain.

Résumé scientifique

Stimulation de la moelle épinière contre la douleur liée au cancer chez l'adulte

Contexte

La douleur liée au cancer représente un lourd fardeau sur la santé publique avec les dépenses élevées qui y sont associées. La douleur sévère est associée à une qualité de vie dégradée chez les patients atteints de cancer. Une proportion significative de patients souffrant de douleur liée au cancer est sous-traitée. Il est nécessaire d'effectuer un contrôle plus efficace de la douleur liée au cancer. La stimulation de la moelle épinière (SME) peut jouer un rôle dans la prise en charge de la douleur. L'efficacité et l'innocuité de la SME chez les patients souffrant de douleur liée au cancer sont à l'heure actuelle indéterminées.

Objectifs

Cette revue systématique a évalué l'efficacité de la SME contre la douleur liée au cancer comparée aux soins standard utilisant un médicament analgésique conventionnel. Nous avons aussi estimé les risques et les événements indésirables potentiels associés à l'utilisation de la SME.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans les bases de données bibliographiques suivantes afin d'identifier des études pertinentes : le registre Cochrane des essais contrôlés (CENTRAL) dans The Cochrane Libary (de son origine jusqu'à l'année 2012, numéro 6) ; MEDLINE ; EMBASE ; et la base CBM (Chinese Biomedical Database) (de son origine jusqu'à juillet 2012). Nous avons également effectué des recherches manuelles de revues pertinentes.

Critères de sélection

Nous avions prévu d'inclure les essais contrôlés randomisés (ECR) qui ont directement comparé la SME à d'autres interventions en termes d'efficacité de la prise en charge de la douleur. Nous avions aussi prévu d'inclure les essais croisés ayant comparé la SME à un autre traitement. Nous avions prévu d'identifier les essais contrôlés non-randomisés, mais ces derniers ne devaient être inclus que si aucun ECR n'avait pu être trouvé.

Recueil et analyse des données

La stratégie de recherche initiale a abouti à 430 articles. En examinant minutieusement les titres et les résumés, nous avons trouvé 412 articles inappropriés à l'examen analytique de cette revue systématique en raison des différentes ampleurs des maladies ou des différentes méthodes d'intervention (système de perfusion intrathécale ; traitement oral) ou des buts autres que le contrôle de la douleur (surveillance de la fonction de la moelle épinière, restauration de la fonction vésicale ou amélioration du métabolisme organique). Les autres 18 essais ont été examinés dans leur texte intégral. Aucun ECR n'a été identifié. Quatorze études de cas sporadiques et articles de revue ont été exclus et quatre études de séries de cas avant-après (92 participants) ont été incluses. Deux auteurs de la revue ont, de manière indépendante, sélectionné les études à inclure dans la revue conformément aux critères d'éligibilité pré-spécifiés. Une liste de contrôle de la qualité méthodologique des essais contrôlés non-randomisés a été utilisée (liste de contrôle STROBE) et tous les auteurs de la revue ont discuté et accepté l'inclusion des essais et les résultats de l'évaluation de la qualité.

Résultats principaux

Quatre études de séries de cas avant-après (un total de 92 participants) répondaient à nos critères d'inclusion. Tous les essais inclus ont adopté une échelle visuelle analogique (EVA) pour évaluer le soulagement de la douleur. Il existait une hétérogénéité en termes de caractéristiques initiales, de paramètres d'électrode et de stimulateur, de niveau d'implantation et de voie d'implantation ; en outre, les données étaient rapportées de façon différente entre les essais. Dans deux essais, le soulagement de la douleur a été obtenu chez 76 % (48/63) des patients à la fin de la période de suivi. Dans le troisième essai, l'EVA pré-procédure était de 6 à 9 (moyenne de 7,43) ; l'EVA un mois post-implantation était de 2 à 4 (moyenne de 3,07) ; l'EVA 12 mois post-implantation était de 1 à 3 (moyenne de 2,67). Dans le quatrième essai, l'EVA pré-procédure était de 6 à 9 (moyenne de 7,07) ; de 1 à 4 (moyenne de 2,67) au bout d'un mois ; de 1 à 4 (moyenne de 1,87) au bout de 12 mois. L'utilisation d'analgésiques a été considérablement réduite. Les principaux événements indésirables étaient l'infection au niveau des sites d'implantation, les fuites du liquide céphalo-rachidien (LCR), la douleur au niveau des sites des électrodes, le délogement des électrodes et les pannes de système, toutefois, l'incidence chez les patients atteints de cancer n'a pas pu être calculée. Puisque tous les essais étaient des essais contrôlés non-randomisés, ils comportaient des risques de biais de tous types.

Conclusions des auteurs

Les preuves actuelles sont insuffisantes pour établir le rôle de la SME dans le traitement de la douleur liée au cancer réfractaire. Les futures études randomisées devraient être consacrées à l'implantation de la SME chez les patients souffrant de douleur liée au cancer.

Plain language summary

Spinal cord stimulation for intractable cancer-related pain

Cancer-related pain is an emerging heavy burden on public health. Spinal cord stimulation (SCS) is a minimally invasive and potentially effective tool against chronic pain.This systematic review intended to evaluate the efficacy and effectiveness of SCS for cancer-related pain compared with standard care using conventional analgesic medication. No randomised controlled trials were identified. Four before-and-after case series studies (92 participants) were included in this systematic review. Current evidence is insufficient to establish the role of SCS in treating refractory cancer-related pain in comparison with other analgesic approaches. In addition, the studies reported significant side effects such as local infection.

Résumé simplifié

Stimulation de la moelle épinière contre la douleur liée au cancer réfractaire

La douleur liée au cancer est un lourd fardeau émergent sur la santé publique. La stimulation de la moelle épinière (SME) est une technique mini-invasive et potentiellement efficace contre la douleur chronique. Cette revue systématique est destinée à évaluer l'efficacité et les performances de la SME contre la douleur liée au cancer comparée aux soins standard utilisant un médicament analgésique conventionnel. Aucun essai contrôlé randomisé n'a été identifié. Quatre études de séries de cas avant-après (92 participants) ont été incluses dans cette revue systématique. Les preuves actuelles sont insuffisantes pour établir le rôle de la SME dans le traitement de la douleur liée au cancer réfractaire en comparaison avec d'autres méthodes analgésiques. En outre, les études ont rapporté des effets secondaires significatifs tels qu'une infection locale.

Notes de traduction

Traduit par: French Cochrane Centre 1st March, 2013
Traduction financée par: Pour la France : Ministère de la Santé. Pour le Canada : Instituts de recherche en santé du Canada, ministère de la Santé du Québec, Fonds de recherche de Québec-Santé et Institut national d'excellence en santé et en services sociaux.

Background

Description of the condition

World-wide, cancer-related pain has increasingly become a heavy burden on public health with related high expenditure. It has been estimated that world-wide nearly seven million people suffer moderate-to-severe cancer-related pain each year caused directly by cancer or from cancer treatment. An epidemiological study revealed that some 15% of these patients fail to achieve acceptable pain relief with conventional management (Running 2011; Yakovlev 2008). Severe pain is associated with a decreased quality of life and unwanted life events such as depression, anxiety and even suicide. Conventional treatment is based on the World Health Organization (WHO) guidelines for cancer pain management which consists of a three-step ladder: (1) non-steroidal anti-inflammatory drugs (NSAIDs), aspirin, and acetaminophen for mild-to-moderate levels of cancer pain; (2) weak opioids for mild-to-moderate pain that does not respond to NSAIDs alone; and (3) strong opioids for moderate-to-severe levels of cancer pain (Schug 1990). Adjuvant medications, such as antiepileptics and tricyclic antidepressants, can also be added at any step of the ladder for optimal pain relief with this notion. Reduced pain intensity and standardised protocols have been used globally for improving cancer pain management. When this approach fails (10% of the patients), interventional pain management has been proposed for this group of patients with refractory pain (Miguel 2000).

Among all the diversified procedures of interventional management approaches for pain control, alternative strategies are needed such as (1) neuroaxial analgesia (spinal, epidural); (2) neurostimulation; (3) neurolysis (sympathetic blockades with phenol or alcohol), (4) thermal neurolysis (radiofrequency) (Slavik 2004). The most commonly used forms of neuromodulation are (1) neurostimulation: the electric stimulation of peripheral nerves, the spinal cord (spinal cord stimulation (SCS)), and brain (deep brain stimulation); (2) another common form of neuromodulation: intrathecal drug delivery system by means of programmed infusion pumps. To date, different techniques of neuromodulation are among the more frequently used types of interventional procedures in the treatment of non-cancer pain (Isagulian 2008).

Although the aetiology of cancer pain is not yet fully understood, altered peripheral nociception (the ability to feel pain) and central sensitisation involving the level of SCS act as one pivotal role in its pathogenesis (Schmidt 2010). Within the cancer microenvironment, cancer and immune cells produce and secrete mediators that activate and sensitise primary afferent nociceptors. In addition, neuropathic mechanisms are also prevalent and cancer pain is often regarded as a mixed-pain mechanism (Ro 2005).

As our understanding of the peripheral and central mechanisms that underlie cancer pain improves, targeted analgesics for the patient with cancer will likely follow, especially in relation to the spinal cord (Boswell 2010; Christo 2008). Thus, when pharmacotherapy for severe and intractable cancer-related pain such as opioids and potent COX-2 inhibitors are ineffective, interventional management approaches have received considerable attention in an attempt to provide pain relief for patients with cancer pain. These offer important additional approaches to the WHO analgesic ladder to control cancer-related pain. Neurostimulation in particular has been recognised in non-cancer pain as having the potential for long-term effectiveness with minimal side effects observed clinically. Currently, the evidence that neurostimulation is effective for the long-term treatment of non-malignant painful conditions such as angina, limb ischaemia, and lower back pain has been established (Kemler 2010; Klomp 2009; North 2008; Taylor 2009). Since chronic cancer pain has some features in common in its pathogenesis with non-malignant pain, systematic reviews, sporadic case series and cohorts of observational studies have reported a marked reduction of pain intensity (Mailis 2004; Ubbink 2005; Yakovlev 2008) using this approach.

SCS can provide long-term relief in managing patients with failed back surgery syndrome and the level of evidence recommendation is Level II-1 or II-2 (Michael 2009). SCS has also been recorded to be effective in reducing the chronic neuropathic pain of complex regional pain syndrome (CRPS) type I (Simpson 2009); this evidence has helped to establish the potential role of SCS in treating patients with cancer-related pain. However, the effectiveness and relative safety of SCS for cancer pain has not been adequately established. Therefore, in this systematic review, we intended to provide scientific evidence as to the efficacy and safety of patients receiving SCS, and to identify which patients are most likely to benefit.

Description of the intervention

SCS consists of placing electrodes in the epidural space on the dorsal surface of the spinal cord. The electrodes can be placed either by using an open procedure in which the dura is exposed (surgical laminotomy), or a closed procedure via epidural needles. The electrodes are connected subcutaneously to an impulse generator that is also inserted under the skin. The impulse generator is programmed using an external device to deliver impulses continuously or in preset patterns throughout a 24-hour period. The technique is reversible and minimally invasive (in contrast to nerve ablation) (Costantini 2005) and appears to result in no adverse effects such as sedation or lethargy, commonly associated with centrally-acting analgesic drugs.

How the intervention might work

The basic scientific background of SCS is based on the gate control theory of Melzack and Wall (Stephen 2005). It has been demonstrated in multiple studies that dorsal horn neuronal activity caused by peripheral noxious stimuli could be inhibited by concomitant stimulation of the dorsal columns. Various other mechanisms, which may play a significant role in the mechanism of action of SCS (Stojanovic 2002), include the suppressive effect of SCS on tactile allodynia (pain produced from a stimulus that would not normally produce pain), increased dorsal horn inhibitory action of gamma-aminobutyric acid (GABA), prevention or abolition of peripheral ischaemia, and effects on human brain activity. Thus, the use of SCS to treat cancer-related pain can be mechanism-based and tailored to the needs of the individual patient. Although opioids remain the mainstay of analgesia for cancer pain (IASP 2008), SCS can be used in addition to, or instead of, conventional approaches. The mixed-pain mechanisms that can result from cancer suggest that multi-modal approaches are likely to result in better outcomes for patients (Herr 2004).

Why it is important to do this review

Despite a few case series and expert recommendations of the potential significance of SCS for cancer-related pain, few cohort studies and fewer randomised controlled trials (RCTs) have been conducted to establish the efficacy of this approach in cancer pain. It is important to identify this in a systematic review so that it can help inform the need for further RCTs in this area and ultimately clinical practice.

Objectives

This systematic review aimed to evaluate the efficacy and effectiveness of SCS for cancer pain compared with standard care using conventional analgesic medication. We also planned to appraise risk and potential adverse events associated with its use.

Methods

Criteria for considering studies for this review

Types of studies

For the purpose of generating high-quality evidence, we planned to include RCTs that directly compared SCS with other interventions for pain management. We also intended to include cross-over trials comparing SCS with another treatment. Non-randomised controlled trials were included as no RCTs were identified .

Types of participants

Adult participants aged between 18 to 80 years old with cancer-related pain who were eligible for the implantation of SCS and treated with this intervention accordingly for cancer-related pain management.

Types of interventions

  1. Participants receiving SCS versus participants receiving conventional medical treatments.

  2. Participants receiving SCS plus conventional medical treatments versus participants receiving conventional medical treatments only.

  3. Participants receiving SCS versus participants receiving physical therapies or complementary therapies.

  4. Participants receiving SCS versus participants receiving other invasive interventions such as surgery or neuro-ablation therapies.

Types of outcome measures

Primary outcomes

Effectiveness of pain management:

  • at least 50% pain reduction of pain (visual analogue score (VAS) as the primary parameter);

  • health-related quality of life;

  • physical and functional abilities;

  • pain-related anxiety and depression.

Secondary outcomes

Rate of procedural complications (bleeding, infection, spinal cord compression etc), incidence of technical failures and withdrawal rate, incidence of treatment-related mortality.

Search methods for identification of studies

Electronic searches

We searched the following bibliographic databases for relevant studies:

  • the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Libary (from inception to 2012, Issue 6);

  • MEDLINE (from inception to July 2012);

  • EMBASE (from inception to July 2012);

  • CBM (Chinese Biomedical Database) (from inception to July 2012).

We used text words combined with MeSH terms: i.e. electric stimulation therapy; clinical trial; comparative study; cancer; randomised controlled trials (RCTs); prospective studies; treatment outcomes; neoplasm; pain (see Appendix 1). Missing information or data in the study reports were sought by direct contact with the principal investigator of the relevant study where possible. We searched for cohort studies (Appendix 2), and ongoing trials by using a similar strategy to that used for searching for published RCTs in the Cochrane Central Register of Controlled Trials (CENTRAL), the trial registry in National Cancer Institute, WHO International Clinical Trials Registry Platform, and clinicaltrials.gov supported by U.S. National Institutes of Health for eligible trials. We searched abstracts of international conferences related to cancer pain management using the term 'spinal cord stimulation'.

Searching other resources

We handsearched major international journals such as Pain, The Clinical Journal of Pain, European Journal of Pain, and conference articles of internationally renowned associations of pain such as the International Association for the Study of Pain (IASP), ESA, APS for preliminary reports of high-quality studies on a week-to-week basis. We checked reference lists of up-dated articles of importance. We also searched international conference proceedings and seminars for potential studies. There were no language restrictions.

Data collection and analysis

Selection of studies

Two review authors (Lihua Peng and Wei Ke) independently selected the studies to be included in the review according to the pre-specified eligibility criteria. Disagreements were resolved by discussion. If this did not resolve the disagreement, we consulted a third review author (Mike Bennett). We based decisions for inclusion or exclusion on the whole content of the studies if available. We recorded data from included studies on a specially developed data extraction form.

Data extraction and management

We extracted and reported data in specially designed data extraction forms. One review author (Peng Lihua) extracted data and these were checked by a second review author (Min Su). Data entry into the database Review Manager (RevMan, Review Manager 2011) was also double-checked. We resolved disagreements concerning data extraction by reaching a consensus based on the inclusion criteria. Where we could not resolve disagreements, we consulted a third review author (MB).

We recorded the following data for each study:

  • details of methodology including whether the study was randomised, whether the methods of sequence generation, allocation sequence concealment, and blinding were reported;

  • details of the participants including age, gender, and diagnosis before SCS;

  • details of the experimental and control interventions including the intervention type, name, dosage and schedules.

Assessment of risk of bias in included studies

We planned to assess the risk of bias in included studies using the following aspects: sequence generation, allocation sequence concealment, blinding.

Sequence generation: adequate (random numbers generated by a computer or similar) or inadequate (other methods or not described).

Allocation concealment: adequate (sealed envelopes or similar) or inadequate (open table of random numbers or similar or not described).

We also planned to assess the quality of included studies using different aspects: adequate allocation concealment, scientific methods of randomisation and balanced enrolment of participants between different interventional arms, follow-up of adequate time and inclusion of the intention-to-treat (ITT) principle during data analysis. As SCS is a minimally invasive yet prominent intervention against cancer pain, we assessed blinding of participants to genuine stimulation or sham stimulation along with conventional therapy. We indexed these as either adequate (independent pain physicians or investigators who assess the subjective outcome such as Visual Analogue Scale (VAS) score or quality of life (QoL)) or inadequate (not performed or similar). However, as we only included non-randomised controlled trials, we used STROBE (Strengthening the Reporting of Observational Studies in Epidemiology), a 22-item check list (see Table 1), to assess the overall quality of the studies.

Table 1. STROBE checklist
StructureItemRecommendation
Title and abstract1Indicate the study’s design with a commonly used term in the title or the abstract; provide in the abstract an informative and balanced summary of what was done
and what was found.
Introduction  
Background/rationale2Explain the scientific background and rationale for the investigation being reported.
Objectives3State specific objectives, including any prespecified hypotheses.
Methods  
Study design4Present key elements of study design early in the paper.
Setting5Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection.
Participants6Give the eligibility criteria, and the sources and methods of selection of participants.
Variables7Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable.
Data sources/
measurement
8For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe comparability of assessment methods if there
is more than one group.
Bias9Describe any efforts to address potential sources of bias.
Study size10Explain how the study size was arrived at.
Quantitative variables11Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why.
Statistical methods12

(a) Describe all statistical methods, including those used to control for confounding.
(b) Describe any methods used to examine subgroups and interactions.
(c) Explain how missing data were addressed.

(d) If applicable, describe analytical methods taking account of sampling strategy.

Results  
Participants13(a) Report numbers of individuals at each stage of study—e.g. numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and
analysed.
(b) Give reasons for non-participation at each stage.
Participants
(c) Consider use of a flow diagram.
Descriptive
data
14(a) Give characteristics of study participants (e.g. demographic, clinical, social) and information on exposures and potential confounders.
(b) Indicate number of participants with missing data for each variable of interest.
Outcome data15Report numbers of outcome events or summary measures.
Main results16If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period.
Other analyses17Report other analyses done—e.g. analyses of subgroups and interactions, and sensitivity analyses.
Discussion  
Key results18Summarise key results with reference to study objectives.
Limitations19Discuss limitations of the study, taking into account sources of potential bias or imprecision.
Discuss both direction and magnitude of any potential bias.
Interpretation20Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence.
Generalisability21Discuss the generalisability (external validity) of the study results.
Other information  
Funding22Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based

Measures of treatment effect

We planned to perform all analyses according to the ITT principle including all randomised participants. For dichotomous outcomes such as rate of adverse events, we planned to record percentages with 95% confidence intervals (CI). We intended to calculate the NNTB (number-needed-to-treat-to-benefit) from the risk ratio (RR) or risk difference (RD) for RCTs. For continuous outcomes such as VAS scores, questionnaires or scores measuring quality of life, we used use medians and standard errors (SEs) or interquartile ranges with CIs to summarise the value in each group. We used mean difference (MD) for continuous outcomes. If different scales had been used to measure continuous data, we would have used standardised mean differences (SMD).

Unit of analysis issues

We planned to assess whether groups of individuals were randomised together to the same intervention, whether individuals undertook more than one intervention and whether multiple investigators observed the same outcome.

Dealing with missing data

We planned to contact the original investigators to request missing data whenever possible in person, by mail or by phone. If we had been unable to obtain missing data, we would have imputed the missing data using mean values. We planned to perform sensitivity analyses to assess how sensitive results were to reasonable changes in the assumptions that were made, and we would have addressed the potential impact of missing data on the findings of the review in the 'Discussion' section. We collected and reported dropout rates in the 'Risk of bias’ table. We used available case analysis for extracted data. If the total dropout rate had exceeded 20%, we planned to use available case analysis and perform subsequent sensitivity analysis to test the effects of missing data from dropout patients. If the dropout rate was less than 20%, we planned to perform ITT analysis.

Assessment of heterogeneity

We used the Chi2 test to assess statistical heterogeneity. If significant heterogeneity was found, we re-checked that the data were correct and explored the reason for the heterogeneity.

Assessment of reporting biases

We performed comprehensive searches for studies that met the eligibility criteria, including unpublished studies and trial registries if possible, as authors with financial aid from pharmaceutical companies or authors of studies with negative outcomes tend to selectively report incomplete outcomes. We extracted all important outcomes of clinical relevance to attempt to eliminate this type of bias as far as possible.

Data synthesis

For the outcome of pain relief, we used random-effects to provide a descriptive analysis of extracted data and no statistical pooling was made. (Analysis 1.1; Analysis 2.1).

Subgroup analysis and investigation of heterogeneity

We planned to analyse the association between different kinds of stimulation apparatus and intervention effects.

As sites of cancer-related pain (lower extremities, trunk or other sites) or different systems of implantation systems may impact on the efficacy of SCS, if possible, we would have considered the above factors as parameters when performing subgroup analyses.

Sensitivity analysis

If we had identified and included RCTs, we would have performed sensitivity analysis comparing studies that had or had not reported: allocation concealment, adequate blinding, or studies without full methodological detail (e.g. published as abstracts only). For included RCTS, we planned to include all studies at first, then eliminate those studies with moderate or poor quality or those only with abstracts one at a time to see if it altered the results. Finally, we planned to perform the analysis with data from studies of good methodological quality; thus, the sensitivity analysis would have been performed in a multiple-step way. Variation among included studies might cause the issue of heterogeneity. First, we planned to use the Chi2 test to test the statistical significance of heterogeneity.

As only non-randomised trials were included, we analysed each trial in a descriptive way.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

Results of the search

The initial search strategy yielded 430 articles. By scrutinising titles and abstracts, we excluded 412 articles due to different scopes of diseases or different methods of intervention (intrathecal infusion system; oral medication) or aims other than pain control (spinal cord function monitoring, bladder function restoration or amelioration of organ metabolism).The remaining 18 trials were reviewed as full manuscripts. No randomised controlled trials were identified; 14 sporadic case reports were excluded and four uncontrolled longitudinal studied were included. See the flowchart (Figure 1).

Figure 1.

Study flow diagram.

Included studies

Comprehensive searching yielded 412 articles but there were no randomised controlled trials (RCTs) that met the inclusion criteria. One seemingly well-conducted RCT that reported on the use of transcutaneous spinal electro-analgesia was excluded because it did not meet the definition of spinal cord stimulation (SCS). After consultation with group editors and a group discussion, we modified our inclusion criteria to include non-randomised controlled trials for this review. By reading abstracts we identified eighteen non-randomised controlled trials, and after scrutinising these eighteen potentially relevant articles, thirteen case reports of individual patients plus one review were also excluded. Four case series studies (Meglio 1989; Shimoji 1993; Yakovlev 2010; Yakovlev 2011) were included in the review. We carried out quality assessment according to the STROBE statement which aims at assessing methodological quality of non-randomised controlled trials..

Four studies (92 participants) as case series met our criteria for inclusion; please see the 'Characteristics of included studies' table. Meglio 1989 retrospectively reported on 11 patients with cancer from a total of 109 patients who were diagnosed with six categories of diseases that were eligible for the implantation of SCS. The rest included patients with vasculopathic pain; lower back pain; paraplegic pain; deafferentation pain and post-herpetic pain. A visual analogue scale (VAS) was used to assess the analgesic effect of the procedure. Patients with a 50% reduction of pain intensity were considered to be responders.

Shimoji 1993 reported a large survey of clinical outcomes using percutaneous, low-frequency SCS to alleviate pain caused by several types of diseases including cancer, post-herpetic neuralgia, spinal trauma, phantom limb pain etc. Visual analogue scales were used for the evaluation of pain. Percentage of pain relief, long-term efficacy and adverse events were also reported.

One author published two consecutive articles on treating patients with cancer-related pain with SCS. The first article (Yakovlev 2010) reported on 14 patients who received spinal cord stimulator placement after surgical or radiological intervention against lung cancer. Significant pain relief was calculated as at least a 50% reduction of the VAS score.The follow-up duration was 12 months and the safety of the procedure was investigated. Yakovlev 2011 retrospectively analysed 15 patients with lower back pain after surgical resection or radiation therapy because of metastatic disease of adjacent organs.These two trials reported the percentage of opioid use before and after SCS implantation, pre-procedure, one month post-implant and 12 months post-implant using VAS.

Excluded studies

Most the papers recognised by our searches were individual case report on spinal cord stimulation for cancer-related pain or experimental studies which did not contain clinical data. After obtaining full texts of potential eligible trials, 14 additional articles were excluded. Seven individual case reports of using spinal cord stimulation to treat cancer-related pain were ruled out because of limited clinical data obtained from the articles (Cata 2004; Eisenberg 2002; Hamid 2007; Lee 2009; Nouri 2011; Ting 2007; Tsubota 2009). Two further case reports were excluded; one included two patients (Yakovlev 2008), the other (Yakovlev 2009) was an individual case report. Three retrospective case series (Clavo 2004; Clavo 2009; Robaina 2007) focused on the effect of SCS on metabolism and blood flow of cerebral cells, no information on pain control was provided. One review (Lee 2006) discussed the indications and outcomes of SCS was also excluded. We also excluded one article (Rainov 2007) that reported hardware failure of SCS in benign pain. In general, the scarcity of literature suggests a lack of high-quality clinical trials. Reasons for exclusion are listed in the Characteristics of excluded studies.

Risk of bias in included studies

Since all included trials were non-randomised trials, we used the STROBE 22-item checklist (Vandenbroucke 2007) to evaluate the quality of observational studies. The CONSORT statement (Moher 2001) aimed at evaluating allocation, blinding, incomplete outcome data and reporting bias could not be used for non-randomised trials.

We used the STROBE checklist to assess the overall quality of each study; this checklist is specially designed for observational studies (see Table 1). All 22 items were rated as yes, no or unclear; 'yes' means that the study was conducted and reported in accordance with the checklist; 'no' means that the study was not conducted as required by the checklist; 'unclear' means no information related to each item could be drawn from the article. Two review authors (PLH and WK) independently rated each article and disagreement was resolved by group discussion. Methodological quality was generally poor and lacked the components of "prospective " trial design. Of all the 22 items, five to 10 items were considered fulfilled for all included trials (see Table 2). One of the common issues was the lack of statistical methods to examine or control possible confounding factors. For all included trials, the enrolment of participants lacked preset eligibility criteria, the reporting of primary outcome as pain relief generally lacked subgroup analysis or intervention interaction (analgesic use and implantation of spinal cord stimulation) and all trials lacked a rational explanation of how the sampling sizes were decided, therefore, we concluded that all of the included trials were at high risk of bias.

Table 2. Result of STOBE Checklist
  1. Y:Yes; N:No; U:Unlear

Item No.Meglio 1989Shimoji 1993Yakovlev 2010Yakovlev 2011
1YYYY
2NYYY
3NYYY
4NNNN
5NNYY
6NNNN
7NNNN
8NNNY
9NNNN
10NNNN
11YYYY
12NNNN
13NNNN
14NNNN
15YYYY
16NNNN
17NNNN
18YYYY
19NNNN
20YYYY
21YYYY
22NNYY

Other potential sources of bias

None known.

Effects of interventions

Heterogeneity existed among all included trials and statistical pooling was not carried out.

Pain Relief

All included trials adopted visual analogue scales (VAS) to evaluate pain relief. In the earliest article, Meglio (Meglio 1989) reported on 11 patients with cancer pain; three patients reported satisfactory analgesia (at least a 50% reduction of pain intensity) and received permanent implantation; the mean level of reduction pain was 75% in these three patients. One patient lost the therapeutic effect after one month of implantation, the two remaining patients were reported to have experienced a 50% reduction of pain until death at 2.5 and five months after implantation (baseline and post procedure VAS scores were not provided).

Shimoji (Shimoji 1993) retrospectively analysed a total of 454 patients receiving implantation of SCS for various conditions; subjective pain relief (at least 50% reduction of original VAS) was observed in 45 out of 52 patients with cancer-related pain. When the authors used a 2 x 2 Chi2 test to examine the relationship of background diseases with pain relief, the number of patients who rated pain relief at more than 50% was significantly larger in patients with carcinoma/sarcoma than the overall effect (253/454), yet, the study did not provide accurate scores of VAS in this group of patients and electrodes were withdrawn at the terminal stage in 49 cases of cancer-related pain. Analgesia use during SCS is also a parameter of clinical efficacy. In 454 patients, medication was stopped in 52 patients (11%), reduced analgesic use was observed in 263 patients (58%); 323 patients reported partial to complete pain relief (over 30% of pain reduction).

Yakovlev reported two consecutive before-and-after case series. The first study (Yakovlev 2010) enrolled 14 patients with intractable cancer-related chest pain. All patients received permanent implantation of an electrode at T3-T4-T5 level. Pain duration before implant was nine to 23 months (median duration was 16 months). The rate of opioid use before implantation was 100% (14/14) and 29% (4/14) after the implant with a decreased dose. Mean values of pre procedure VAS was 7.43 (Standard Deviation 0.94); one month post implant the VAS was 3.07 (Standard Deviation 0.62); 12 months post implant VAS was 2.07 (Standard Deviation 0.83).

The second study (Yakovlev 2011) reported on 15 patients with intractable cancer-related lower back pain receiving SCS; all patients had leads inserted at T11-12 or T12-L1 level. Pain duration before implant was 14 to 26 months (median duration was 19 months ). Rate of opioid use before implant was 100% (14/14) and 47% (7/15) after implant with a decreased dose. Mean values of pre procedure VAS was 7.07 (SD 1.03); one month post implant VAS was 2.07 (SD 0.9); 12 months post-implant VAS was 1.87(SD 0.83). Since no comparison could be made against other interventional groups, before-and-after comparisons of this outcome were reported and analysed in narrative forms (Analysis 1.1; Analysis 2.1).

Adverse events

Adverse events were reported in patients in two earlier studies (Meglio 1989; Shimoji 1993) with all diseases eligible for SCS. Meglio 1989 revealed three cases of infection of implantation, four cases of aseptic meningitis, two cases of rejection of the electrode leads, three cases of cerebrospinal fluid (CSF) leakage, three cases of subcutaneous haematoma, two cases of pain at the electrode sites, three cases of dislodgement of the electrodes, and four cases of system failure. Other minor side effects included five cases of headache, five cases of asthenia, five cases of dizziness and six cases of muscle twitches/contractions in a total of 109 patients (including 11 patients with cancer-related pain). Shimoji 1993 reported six cases of CSF leakage, 27 cases of infection of implantation,19 cases of pain at the electrode sites, 22 cases of dislodgement of the electrodes and eight cases of electrode dysfunction in a total of 454 patients (including 52 patients with cancer-related pain). In two other recent studies (Yakovlev 2010; Yakovlev 2011), no complications of SCS implantation were reported.

Discussion

Summary of main results

In the four before-and-after case series studies included in this systematic review, clinical efficacy was reported as modest (Meglio 1989) to excellent (Shimoji 1993; Yakovlev 2010; Yakovlev 2011). Over 80% of patients reported at least a 50% reduction of pain intensity, more than 50% of patients reported decreased use of opioid medications. Major complications were infection of sites of implantation, CSF leakage, pain at the sites of electrodes, dislodgement of the electrodes and system failure although the incidence was considerably low. The follow-up period varied from one week to more than one year. However, all these studies were at high risk of bias.

Overall completeness and applicability of evidence

In this systematic review, the lack of randomised controlled trials (RCTs) related to this topic left the question of effectiveness unanswered. Four cases series including 92 patients were included for descriptive analysis. These four studies varied greatly in clinical setting, patient characteristics, electrode and stimulator parameters, level of implantation, route of implantation (subarachnoid cavity or epidural cavity) and methods of electrode implantation (laminectomy or percutaneous insertion). Meglio 1989 did not mention the types of cancer and sites of pain and three out 11 patients with cancer reported excellent pain relief after implantation. Shimoji 1993 reported outcomes of 52 patients with cancer-related pain in a cohort of 454 patients. Sites of pain included head and face, neck and upper extremities, trunk and lower extremities.Types of cancer and pre-procedure VAS scores were not provided. Adverse events were reported in patients, not only with cancer-related pain, but also with chronic pain of non-cancer origin. In the two later studies the author clarified types of cancer and sites of pain. One of these (Yakovlev 2010) described 14 patients with lung cancer and intractable chest wall pain. In this study, pain relief at one-year of follow-up was excellent without complication. In another study (Yakovlev 2011), 15 patients with cancer-related lower back pain from metastasis related to colon, anal cancer, and angiosarcoma of the sacrum were described. All patients reported significant pain relief (over 50% of a reduction) that was maintained for at least one year. A major limitation of the evidence base is that all included studies lacked preset eligibility of participants and comparison with control.

Quality of the evidence

Only non-controlled case series without interventional comparison were available. All studies had small numbers of patients with cancer and were poorly designed to reach a conclusion about the comparative efficacy of spinal cord stimulation (SCS) for cancer-related pain. Participant attrition, selective reporting and performance bias could have been factors influencing all of the included trials. Meglio 1989 was a retrospective analysis of SCS against chronic pain (cancer-related pain included) in a single institution without a power calculation. Shimoji 1993 did not perform a power calculation nor were the baseline characteristics reported. Yakovlev 2010 and Yakovlev 2011 provided baseline VAS scores and rate of opioid use; a before-and-after comparison was made. Lack of randomisation, allocation concealment or blinding introduced considerable risk of bias. Randomised controlled trials are still needed to clarify clinical efficacy of SCS in cancer-related pain. Optimal patient selection, time of implantation and approaches to minimise its side effects should be analysed.

Potential biases in the review process

All included studies did not comply with the CONSORT statement nor did they meet all essential criteria of the STROBE checklist. All included trials were before-and-after case series and no comparison with other interventions could be made. Furthermore, researchers are more likely to report "positive outcomes" in a selected group of patients while leaving "negative outcomes " overlooked. In summary, all trials carried with them a great risk of bias .

Agreements and disagreements with other studies or reviews

Our review focused on the efficacy and safety of SCS against cancer-related pain and no previous published systematic review was found. Spinal cord stimulation has been utilised for control of chronic cancer and non-cancer pain for nearly 40 years (Miles 1974; Sweet 1974), but the efficacy of SCS has only been established for chronic non-cancer pain (Frey 2009; Grabow 2003; Simpson 2009; Taylor 2006), including failed back surgery syndrome, neuropathic pain, complex regional pain syndrome etc. Neuromodulation has been given attention to alleviate cancer pain with encouraging outcomes (Hurlow 2012). Flagg 2012 recommended that cancer-related pain should be treated at an early stage with an algorithm integrating SCS. Although all included articles reported that patients with cancer-related pain may benefit from SCS (Meglio 1989; Shimoji 1993; Yakovlev 2010; Yakovlev 2011), there is no evidence to support or refute the use of SCS in the treatment of pain in patients with cancer. The bulk of the literature identified in this review were individual case reports with greater risk of bias, and which generally reported positive outcomes. Spinal cord stimulation should not be compared with "sham stimulation" for ethical reasons in patients with cancer, however, the safety and efficacy of SCS should be compared with means of pain control (oral medications; intrathecal drug delivery; transcutaneous electrostimulation) in patients with cancer.

Authors' conclusions

Implications for practice

Current evidence is insufficient to establish the role of spinal cord stimulation (SCS) in treating refractory cancer-related pain in comparison with other analgesic approaches. However, evidence from non-randomised controlled trials is generally positive and is consistent with a stronger evidence base in non-cancer pain.

Implications for research

Future research should focus on the implantation of SCS in patients with cancer-related pain at an early stage, and randomised controlled trials are needed to quantify the benefits and harms of this procedure.

Acknowledgements

PaPaS editorial office for protocol editing.

Data and analyses

Download statistical data

Comparison 1. Pain Intensity after SCS implantation
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Pain intensity---Visual Analogue Scale258Mean Difference (IV, Random, 95% CI)4.38 [3.93, 4.83]
Analysis 1.1.

Comparison 1 Pain Intensity after SCS implantation, Outcome 1 Pain intensity---Visual Analogue Scale.

Comparison 2. Pain intensity---1 month after SCS versus 12 months after SCS
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Pain Intensity---Visual Analogue Scale258Mean Difference (IV, Fixed, 95% CI)0.91 [0.50, 1.32]
Analysis 2.1.

Comparison 2 Pain intensity---1 month after SCS versus 12 months after SCS, Outcome 1 Pain Intensity---Visual Analogue Scale.

Appendices

Appendix 1. Search strategy for MEDLINE

1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
3. randomized controlled trials.sh.
4. random allocation.sh.
5. double blind method.sh.
6. single blind method.sh.
7. or/1-6
8. (animal not human).sh.
9. 7 not 8
10. clinical trial.pt.
11. exp clinical trials/
12. (clin$ adj25 trial$).ti,ab,sh.
13. (singl$ or double$ or trebl$ or tripl$) adj25 (blind$ or
mask$).ti,ab,sh.
14. placebo$
15. placebos.ti,ab,sh.
16. random$ ti,ab,sh.
17. research design.sh.
18. or/10-17
19. 18 not 8
20. 19 not 9
21. comparative study.sh.
22. exp evaluation studies/
23. follow up studies.sh.
24. prospective studies.sh.
25. (control$ or prospective$).mp or volunteer$.ti,ab.
26. or/21-25
27. 26 not 8
28. 27 not (9 or 20)
29. 9 or 20 or 28
30. cancer$ or carcino$ or tumour$ or tumor$ or neoplas$ or malig$
31. exp Neoplasms/ 
32. 30 or 31
33. exp Pain/
34. pain* 
35. 33 or 34
36. exp Electric Stimulation Therapy/
37. exp Spinal Cord/
38. spinal cord stimulation$.tw.
39. scs.tw.
40. dorsal column stimulation.tw.
41. 36 or 37 or 38 or 39 or 40
42. 32 and 35
43. 29 and 41 and 42

Appendix 2. Search strategy for cohort study

1. longitudinal studies.pt.
2. follow-Up Studies.pt.
3. Prospective Studies.pt
4. Retrospective Studies.pt
5. cohort study.pt
6. exp cohort study/
7. cohort*
8. or/1-7
9. (animal not human).sh.
10. 8 not 9
11. clinical trial.pt.
12. exp clinical trials/
13. (clin$ adj25 trial$).ti,ab,sh.
14. (singl$ or double$ or trebl$ or tripl$) adj25 (blind$ or
mask$).ti,ab,sh.
15. placebo$
16. placebos.ti,ab,sh.
17. random$ ti,ab,sh.
18. research design.sh.
19. or/10-17
20. 10 not 19
21. exp evaluation studies/
22. follow up studies.sh.
23. prospective studies.sh.
24. (control$ or prospective$).mp or volunteer$.ti,ab.
25. or/21-24
26. 25 not 9
27. 20 or 26
28. cancer$ or carcino$ or tumour$ or tumor$ or neoplas$ or malig$
29. exp Neoplasms/ 
30. 28 or 29
31. exp Pain/
32. pain* 
33. 31 or 32
34. exp Electric Stimulation Therapy/
35. exp Spinal Cord/
36. spinal cord stimulation$.tw.
37. scs.tw.
38. dorsal column stimulation.tw.
39. 34 or 35 or 36 or 37 or 38
40. 30 and 33
41. 27 and 39 and 40

Contributions of authors

Peng Lihua wrote the protocol draft.

Michael Bennett modified the protocol.

Peng Lihua conceived the idea for this review and gave some suggestions on the protocol.

Min Su and Wei Ke did the search and extracted the data for the full review.

Wei Ke carried out the analysis for the full review.

Peng Lihua will be responsible for conducting any update of this review.

Declarations of interest

None known.

Sources of support

Internal sources

  • Department of Anesthesia and Pain Medicine, Chongqing Medical University, China.

External sources

  • Cochrane Pain, Palliative & Supportive Care Review Group, UK.

Differences between protocol and review

Where randomised trial evidence is desired but unlikely to be available, eligibility criteria defines that non-randomised trials would only be included where randomised trials are found not to be available and non-randomised trials will be appraised with commonly used checklists for methodological quality (Reeves 2011).

We intended to include randomised controlled trials (RCTs); however, we did not find any such trials. After consultation with the PaPaS group editors and a group discussion, we modified our inclusion criteria to include non-randomised controlled trials for this review.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Meglio 1989

MethodsPart of a retrospective study to analyse 109 patients with chronic pain who underwent spinal cord stimulation, clinical efficacy was analysed in relation to the aetiology of pain.
ParticipantsFrom 1978-1986,109 participants were enrolled, 11 patients with cancer pain; 40 with vasculopathic pain, 19 with lower back pain; 15 with paraplegic pain; 9 with deafferentation pain,10 with post-herpetic pain.
InterventionsPercutaneous placement of the stimulator electrodes or positioned through a small laminectomy after a test period of 5 to 60 days, two kinds of stimulators were used: the first was a radiofrequency system; the second was programmable stimulators, which were programmed with a pulse width of 210 microseconds and a rate of 85 Hz,64 seconds on,1 to 4 minutes off, amplitude was at will to produce comfortable paraesthesia.
OutcomesReduction of visual analogue scale as percentage of analgesia (0% denotes no effect,100% denotes complete pain relief, a reduction of more than 50% of original pain was considered as responder); adverse events.
NotesNone
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskNo information was provided.
Allocation concealment (selection bias)High riskNo information was provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo information was provided.
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo information was provided.
Incomplete outcome data (attrition bias)
All outcomes
High riskNo information was provided.
Selective reporting (reporting bias)High riskNo information was provided.
Other biasHigh riskNo information was provided.

Shimoji 1993

MethodsA survey of clinical results of using percutaneous epidural low-frequency spinal cord stimulation for chronic pain.
ParticipantsBetween 1970-1991, 454 patients with chronic pain received percutaneous epidural low-frequency spinal cord stimulation, 52 with carcinoma/sarcoma; 126 with post-herpetic neuralgia; 189 with causalgia; 12 with spinal trauma; 9 with SMON; 3 with tabes dorsalis; 8 with phantom pain; 14 with TAO/ASO; 9 with thalamic syndrome; 32 with other pain.
InterventionsAll patients received implantation of electrodes at sites of pain which connected to a stimulator that delivered saw-wave pulses (0.5ms in duration and 0.5-50 Hz in frequency). The frequency of stimulation was adjustable by the patient at between 1.6 and 8.0 Hz, the intensity being 0.5-5.0 V. The mode of stimulation was continuous in nine patients with cancer or occasional (3-12/day for 20-30 min) in 445 patients, depending on patients' complaints.
OutcomesDegree of pain relief as visual analogue scale, 50% of reduction was considered as pain relief; adverse events.
NotesNone
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskNo information was provided.
Allocation concealment (selection bias)High riskNo information was provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo information was provided.
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo information was provided.
Incomplete outcome data (attrition bias)
All outcomes
High riskNo information was provided.
Selective reporting (reporting bias)High riskNo information was provided.
Other biasHigh riskNo information was provided.

Yakovlev 2010

MethodsTo retrospectively analyse the pain relief outcome of spinal cord stimulation in patients with cancer-related chest wall pain.
ParticipantsFrom 2005-2008,14 patients diagnosed with lung cancer underwent thoracotomy or lung resection and postoperative radiation therapy, and complained of intractable chronic chest pain.
Interventions14 patients received percutaneous implantation of permanent leads and stimulators at T3,T4,T5 after a successful trial of at least 2 days; stimulators were programmed with a pulse width of 400 to 450 microseconds and a rate of 50-60 Hz,amplitude ranged from 1.5-2.3 volts.
OutcomesRate of opioid use before and after treatment; pre procedure, 1 month post implant and 12 months post implant visual analogue scale; complication.
NotesNone
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskNo information of randomisation was provided.
Allocation concealment (selection bias)High riskNo information of allocation concealment was provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo information of blinding was provided.
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo information of blinding of outcome assessment was provided.
Incomplete outcome data (attrition bias)
All outcomes
High riskNo information of patient dropout was provided.
Selective reporting (reporting bias)High riskNo information was provided.

Yakovlev 2011

  1. a

    ASO: arteriosclerosis obliterans
    SMON: subacute myelo-optico-neuropathy
    TAO: thromboangiitis obliterans

MethodsTo retrospectively analyse the pain relief of spinal cord stimulation for intractable cancer-related lower back pain.
ParticipantsBetween 2005-2009,15 patients underwent surgical resections and radiation therapy because of metastatic disease related to colon, anal cancer, angiosarcoma of the sacrum, complained of intractable chronic low back pain.
Interventions15 patients received percutaneous implantation of permanent leads and stimulators at T11-12,T12/L1 after successful trial at least 2 days,stimulators were programmed with a pulse width of 390 to 480 microseconds and a rate of 40-60 Hz,amplitude ranged from 1.4-5.2 volts.
OutcomesRate of opioid use before and after treatment; pre procedure ,1 month post implant and 12 months post implant visual analogue scale; complications.
NotesNone
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)High riskNo information was provided.
Allocation concealment (selection bias)High riskNo information was provided.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo information was provided.
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo information was provided.
Incomplete outcome data (attrition bias)
All outcomes
High riskNo information was provided.
Selective reporting (reporting bias)High riskNo information was provided.
Other biasHigh riskNo information was provided.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    SCS: spinal cord stimulation

Cata 2004Individual case report
Clavo 2004Outcomes not related to the topic of systemic review
Clavo 2009Outcomes not related to the topic of systemic review
Eisenberg 2002Individual case report
Hamid 2007Individual case report.
Lee 2006Review article of SCS.
Lee 2009Individual case report.
Nouri 2011Individual case report.
Rainov 2007Outcomes not related to the topic of review.
Robaina 2007Individual case report.
Ting 2007Individual case report.
Tsubota 2009Individual case report.
Yakovlev 2008Case report including only 2 patients.
Yakovlev 2009Individual case report.

Ancillary