Colchicine for pericarditis

  • Protocol
  • Intervention

Authors


Abstract

This is the protocol for a review and there is no abstract. The objectives are as follows:

To review all randomised controlled trials and determine the best evidence of the clinical effects of colchicine alone or combined, versus any other treatment to prevent first or further recurrences of pericarditis, in people with acute or recurrent pericarditis.

Background

Description of the condition

Definitions

Pericarditis is the inflammation of the pericardium, the membranous sac surrounding the heart. Acute pericarditis is idiopathic without an obvious aetiology in 80% to 90% of cases but has a presumed viral origin (Dudzinski 2012). Other causes include tuberculosis, or bacterial and neoplastic diseases that are more common in low-income countries (Zayas 1995).

Recurrent pericarditis is both the most common and most troublesome complication of acute pericarditis and is mostly idiopathic. This is because the exact cause of the recurrence of pericarditis is not known, but appears to be autoimmune as indicated by the presence of autoantibodies and response to steroids (Cantarini 2013). There are two types of recurrent pericarditis, intermittent or incessant. In the incessant type, discontinuation of non-steroidal anti-inflammatory therapy (NSAIDs) usually causes a relapse in less than six weeks. In the intermittent type, people have varying symptom-free intervals, usually longer than six weeks, without therapy (Soler-Soler 2004).

Incidence

The actual incidence of acute pericarditis in unknown, but it is estimated to be 28 cases per 100,000 population/year (Imazio 2008a). It is responsible for 4% of all causes of chest pain (Launbjerg 1996) and 0.1% of all hospitalisations (Pölzl 2011). Recurrent pericarditis can occur in up to 20% to 30% of people who have experienced acute pericarditis (Fowler 1990; Adler 1998); this figure increases to 50% after the first recurrence (Imazio 2005a). The rate of recurrence varies and can be a single episode in some people; however, other people can experience more frequent episodes over many years. Almost 45% of people experience two episodes, 40% have between three and five episodes, and 10% have more than five episodes (Soler-Soler 2004; Shabetai 2005).

Presentation and diagnosis

The manifestation of acute pericarditis is a pleuritic chest pain with a sign or symptom marking the activity of the disease, such as fever, a pericardial rub, electrocardiography (ECG) changes (a widespread ST-segment elevation or PR-segment depression), pericardial effusion and raised inflammatory markers (erythrocyte sedimentation rate or C-reactive protein (CRP)) (Spodick 2003; Troughton 2004). Acute pericarditis is diagnosed if at least two of these manifestations are met (Dudzinski 2012).

Recurrent pericarditis is a repeat episode of acute pericarditis and can have similar symptoms, although it tends to be milder (Soler-Soler 2004; Adler 2011). There are no uniform diagnostic criteria for recurrent pericarditis (Imazio 2007); however, observational studies identified pleuritic chest pain, increased CRP and ECG changes as the minimum criteria for diagnosing a recurrent episode of acute pericarditis (Brucato 2006; Khandaker 2010).

Prognosis

The first relapse usually occurs within 18 months after the initial pericarditis episode (Imazio 2005; Imazio 2005a). However, people can have many relapses that manifest as severe chest pain lasting from several hours to several days. These painful and disabling episodes impair quality of life and cause a severe clinical problem (Soler-Soler 2004).

Acute and recurrent pericarditis can be complicated by life-threatening consequences, such as pericardial effusion, tamponade or constriction, which may increase mortality (Soler-Soler 2004; Dudzinski 2012). These complications can occur in up to 3.5% of people in recurrent pericarditis and even more frequently in people with acute pericarditis (Imazio 2007). However, in the long term, complications are rare and the prognosis of recurrent pericarditis is good (Brucato 2006).

Description of the intervention

Treatment aims to manage the acute episode of pericarditis and to then prevent subsequent recurrences. For a long time, high-dose steroids were the mainstay of treatment. Yet, high-dose steroids caused numerous serious adverse effects (Shabetai 2005), and their prolonged use has actually worsened the prognosis by increasing the recurrence rate of pericarditis and lengthening the course of the disease (Artom 2005; Imazio 2005; Imazio 2008). Therefore, identifying interventions with a safer adverse effect profiles was essential in order avoid worsening the natural course of recurrent pericarditis in other ways.

Episodes of pericarditis are currently treated with aspirin or other NSAIDs and with steroids for refractory cases (Maisch 2004; Soler-Soler 2004). Colchicine has been used for the prevention of recurrences (Brucato 2006a).

Colchicine has anti-inflammatory actions and antiproliferative effects (Robert 2009). It inhibits many of the functions of neutrophils, such as the adhesion to endothelium and the release of a chemotactic factor from neutrophil lysosomes (Nuki 2008).

Colchicine is considered a safe treatment (Imazio 2007); however, in high doses, it has many toxic effects and, in addition, it has a narrow therapeutic window (Robert 2009). The maximum therapeutic dose is 4 mg/24 hours, while a fatal dose can be as low as 7 mg/24 hours with a higher fatality rate if it the dose exceeds 0.5 mg/kg (Niel 2006; Finkelstein 2010). Overdose is associated with gastrointestinal, hepatic, renal, neuromuscular and cerebral toxicity; bone marrow damage; and high mortality (Nuki 2008; Finkelstein 2010). Colchicine is excreted mainly by the liver after 20 to 40 hours (Niel 2006) and can accumulate in people with advanced liver disease (Rudi 1994).

The recommended dose of colchicine used in gout and in recurrent pericarditis is 1 mg/day by oral administration (Adolph 1990; Adler 1998). Analgesia with colchicine is evident within 12 to 14 hours of oral administration (Imazio 2009). The most common adverse effects of the therapeutic dose are nausea, vomiting, diarrhoea and abdominal pain (Niel 2006).

How the intervention might work

Colchicine is used in treating several inflammatory diseases such as gout and familial Mediterranean fever (FMF) (Famaey 1988; Niel 2006). Considering the possible autoimmune inflammatory pathophysiology of recurrent pericarditis (Caforio 2010; Cantarini 2013), and its response to immunosuppressive and anti-inflammatory treatment (Marcolongo 1995), it is deemed acceptable and logical to determine the effects of colchicine in the management of recurrent pericarditis.

Why it is important to do this review

People with recurrent pericarditis can have a number of relapses over many years causing severe chest pain. These episodes of pain limit both the functionality of patients and their quality of life causing both a social and psychological burden for the patients and an economical burden on the hospitals taking care of them (Soler-Soler 2004). The high incidence of recurrent pericarditis in almost one-third of patients with acute pericarditis increases this burden. Therefore, there is a need to find and examine therapies that decrease the number of recurrences.

Observational studies have shown that colchicine might be effective in treating recurrent pericarditis (Rodríguez de la Serna 1987; Guindo 1990; Millaire 1994; Soler-Soler 2004; Imazio 2005; Brucato 2006a). However, randomised controlled trials have only recently studied the effect of colchicine on pericarditis. Therefore, there is a need to systematically assess and critically appraise these trials in order to obtain a more definite clinical answer for both patients and clinicians dealing with recurrent pericarditis.

A similar review has been published in the BMJ Heart journal (Imazio 2012). The main differences in our review are that we will not include postcardiac injury syndromes due to the different aetiology and pathophysiology from acute or recurrent pericarditis. In addition, trials of acute pericarditis will be analysed separately from trials of recurrent pericarditis. Any differences between our review and the Heart journal review will be mentioned explicitly in the full review in the section 'Agreements and disagreements with other studies or reviews'.

Objectives

To review all randomised controlled trials and determine the best evidence of the clinical effects of colchicine alone or combined, versus any other treatment to prevent first or further recurrences of pericarditis, in people with acute or recurrent pericarditis.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials with any length of follow-up and no limitations on language or publication status. We will exclude quasi-randomised studies.

Types of participants

  1. People with acute idiopathic pericarditis treated to prevent recurrences.

  2. People with recurrent idiopathic pericarditis who have had a documented episode of acute pericarditis defined by any clinically valid diagnostic criteria such as described in the 'Description of the condition' section and who have evidence of recurrent pericarditis will be included in the analysis.

We will include participants regardless of the number of pericarditis recurrences, gender, age or ethnicity. We will exclude acute or recurrent pericarditis that has bacterial or neoplastic causes, as in these cases the known cause has to be treated and managed. Pericarditis as a result of postcardiac injury such as postmyocardial infarction pericarditis (Dressler's syndrome), postpericardiotomy syndrome and post-traumatic pericarditis will not be included in this review.

Types of interventions

  1. Colchicine: in any dose, duration, intensity or means of administration and alongside any additional therapy, on the condition that the additional therapy was also used at the same or similar dose in the control group. Toxic doses of colchicine (> 4 mg/24 hour) will be excluded.

  2. Controls: any inactive control intervention (e.g. placebo or no treatment) or any active control intervention (e.g. aspirin, NSAIDs or steroids).

Types of outcome measures

Study eligibility will be considered regardless of the outcomes investigated or presented.

Primary outcomes
  1. Time to first recurrence expressed using hazard ratios (HRs).

  2. Adverse effects: general and specific during treatment and on withdrawal of treatment.

Secondary outcomes

We will consider the following secondary outcomes where the information is available and data are reported.

  1. Rate of first recurrences of pericarditis as expressed by the risk ratio (RR) at the following periods: short term (six months), medium term (12 months) and long term (splitting > 12 month into categories e.g. 18 months, 24 months).

  2. Relief of symptoms during the pericarditis episode.

Search methods for identification of studies

Electronic searches

We will search the following databases:

  • Cochrane Central Register of Controlled Trials (CENTRAL);

  • MEDLINE (Ovid);

  • EMBASE (Ovid);

  • Web of Science Conference Proceedings Citation Index- Science (CPCI-S).

We will adapt a preliminary search strategy for MEDLINE (Ovid) for the other databases (Appendix 1) and will apply no language or date restrictions.

Searching other resources

We will handsearch two databases of ongoing trials:

We will search the references of all identified studies for additional studies. We will contact the first author of each included study for information about trials that have not been published.

Data collection and analysis

Selection of studies

Two review authors (SA, MQ) will independently review the titles and abstracts identified from the searches. We will obtain full-text publications if necessary and two independent review authors will determine eligibility. A third review author (JBC) will resolve any disagreements and uncertainties. If it is not possible to resolve any disagreements, we will try to contact the authors of the study for further information and clarification.

Data extraction and management

Two review authors (SA, GJI) will independently assess methodological quality and extract data from the studies fulfilling the inclusion criteria using an agreed data extraction form. We will include essential items mentioned in the Cochrane Handbook for Systematic Reviews of Interventions table 7.3a (Higgins 2011a) regarding methods, participants, intervention, outcomes and results. In addition, we will record data relevant to the review on the data extraction forms. We will pilot the data extraction form using one of the studies and make changes as necessary.

We will resolve any differences in the forms by discussion and, if necessary, in consultation with a third review author (JBC). We will attempt to acquire missing data by contacting the study authors.

'Summary of findings' table

We will use the GRADE approach, adopted by The Cochrane Collaboration, to interpret findings (Schünemann 2011), and we will use the GRADE profiler (GRADEpro) program to import data from Review Manager 5 (RevMan 2011), to create the 'Summary of findings' tables. The GRADE system involves an assessment of the risk of bias for each individual outcome. In GRADEpro, the risk of bias for each outcome is separately rated as high, moderate, low and very low quality. We will start the rate of the outcomes of all randomised trials as high and downgrade them depending on: limitations in the design of the selected studies, high risk of bias, indirectness of evidence, unexplained heterogeneity, imprecisions of results and high probability of publication bias.

The outcome-specific ratings will be produced in tables by GRADEpro and will give information about the overall risk of bias of each included study. We will select all primary outcomes for inclusion in the 'Summary of findings' table.

Assessment of risk of bias in included studies

Two review authors (SA, GJI) will independently assess the risk of bias in each study using the 'Risk of bias' tool of The Cochrane Collaboration. We will resolve any disagreements about the bias assessment by discussion and, if necessary, in consultation with a third review author. We will attempt to contact the study authors to obtain missing data if insufficient information of randomisation and other aspects of the trials are provided.

Risk of bias for an outcome within a study (across domains)

The specific characteristics assessed will include random sequence generation, allocation concealment, blinding of participants, personnel and outcome assessors, incomplete outcome data, selective reporting and other sources of bias.

Risk of bias for an outcome across studies (e.g. for a meta-analysis)

We will summarise the risk of bias of the outcomes for each domain included in the meta-analyses and incorporate the judgements into the 'Summary of findings' tables.

We will assess the risk of bias in each domain using the following judgements: 'Low risk', 'High risk' or 'Unclear risk' of bias. We will take the criteria for the judgement from the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions table 8.5.d (Higgins 2011b).

Measures of treatment effect

We will follow the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions sections 9.2 and 9.4 (Higgins 2011c) for measuring the effects of different data types.

Dichotomous data

We will use RR with 95% confidence interval (CI) for binary data (i.e. recurrence rates, adverse effects). We will calculate the number needed to treat for an additional beneficial outcome (NNTB) from the absolute risk reduction (ARR) if available.

Continuous data

For continuous data, we will calculate a mean difference (MD) with 95% CI between groups to summarise the effects of the outcome. We will use the standardised mean difference (SMD) with 95% CI if the studies assess the same outcome but measure it in a variety of ways and are of such similarity to allow pooling.

Counts and rates

We will express count data (i.e. recurrence rate of pericarditis) as rate ratios.

Time-to-event

We will use HRs with 95% CIs to express events such as time until the first recurrence of pericarditis. We will use the methods described in Tierney 2007 to calculate approximate HRs.

Unit of analysis issues

Our primary outcome will use per-patient analysis.

Events that may re-occur

As recurrent pericarditis can happen to a person more than once, one outcome will be the number of times these events occurred per patient rather than simply whether each person experienced any event. Therefore, recurrence rate should be expressed as total person-years as opposed to total number of patients. We will try to ensure that the original analysis has also used this and that the authors were aware that the same person may have contributed more than once to the total number of events (Higgins 2011c). We will also analyse time to the first event.

Studies with multiple treatment groups

Where a trial has multiple treatment arms, if relevant, we will included the additional arms in the comparisons. Otherwise, they will not be studied and the additional data will not be used.

We do not anticipate finding cluster trials. Similarly, cross-over trials are not appropriate for unstable conditions such as recurrent pericarditis (Elbourne 2002).

Dealing with missing data

Whenever possible, we will contact the original investigators to request missing data. We will try to make assumptions about the cause of the missing data and if the data were missing at random or because of a specific outcome. Where possible, we will perform an available-data analysis. However, we will endeavour to apply an intention-to-treat (ITT) analysis by assigning the missing data with replacement values based on the best-case and worst-case scenarios as sensitivity analyses. This is to assess how sensitive the results are to reasonable changes in the assumptions that are made.

Finally, we will address the potential impact of the missing outcomes on the findings of the included studies in the assessment of risk of bias and we will describe its impact on the results of the review in the discussion section.

We will follow the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions about missing data (Higgins 2011, chapter 16).

Assessment of heterogeneity

Clinical heterogeneity

Clinical heterogeneity might be due to differences in the population (i.e. age or ethnicity), differences in interventions (i.e. different doses, duration, intensity or delivery method) or differences in the way outcomes are measured (i.e. such as different criteria for pericarditis recurrences). All studies with outlying situations will be discussed fully. We plan to do a subgroup analysis of the different causes of clinical heterogeneity.

Methodological heterogeneity

We will investigate all included trials for unpredicted outlying methods. When such methodological issues arise, we will discuss these with a statistician.

Statistical heterogeneity

We will investigate statistical heterogeneity by visual inspection and carrying out both a Chi2 test and an I2 test. The Chi2 test with a small P value provides evidence of heterogeneity. However, because of the low power of the Chi2 test, we will use a P value of 0.1 to determine statistical significance. We will use the I2 statistic to quantify the statistical inconsistency and assess the impact of heterogeneity on the meta-analysis. We will interpret an I2 greater than 50% to demonstrate high heterogeneity.

Assessment of reporting biases

We will assess publication bias by the use of funnel plots if there is a sufficient number of trials and reasons for any asymmetry will be considered. Therefore, tests for funnel plot asymmetry will only be used when there are at least 10 studies included in the meta-analysis, because a funnel test with fewer studies will have too low a power to distinguish chance from real asymmetry (Cochrane Handbook for Systematic Reviews of Interventions section 10.4.3.1; Sterne 2011). We will not use tests for funnel plot asymmetry if all studies are of similar sizes (similar standard errors of intervention effect estimates).

Data synthesis

We will pool data in meta-analyses where they are available and it is clinically acceptable to do so. We will use Review Manager software for meta-analyses (RevMan 2011). For the statistical analyses, we will use a fixed-effect model with 95% CI as the main analysis and perform a sensitivity analysis using a random-effects model.

People with acute pericarditis and recurrent pericarditis have a different baseline risk for recurrences, as people who already have experienced recurrent episodes of pericarditis are more susceptible to recurrences (Soler-Soler 2004; Shabetai 2005). Therefore, the two patient groups will be analysed separately for any outcome comparing recurrence rates. However, outcomes of adverse effects and symptom relief will be combined as they are not affected by the risk for recurrences.

However, if we suspect that the clinical differences between the included trials could possibly lead to substantial heterogeneity, then we will consider the random-effects model as the main analysis and will perform a sensitivity analysis with the fixed-effect model for comparison. This will be decided before any statistical analysis is commenced.

We will use the recommendations for data synthesis in chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Subgroup analysis and investigation of heterogeneity

1. Subgroup analyses

Where possible and appropriate, subgroup meta-analyses will be considered for:

  1. dosage of colchicine used: 1 mg/day is most commonly used dose (Adler 1998), and we suggest that this dose has the largest effect;

  2. age of participants (children and adults): our hypothesis is that colchicine prevents recurrences in children (Yazigi 1998);

2. Investigation of heterogeneity

If the heterogeneity is high (I2 statistic > 50%), this will be reported. First, we will check the entered data of all included studies and ensure that they are correct. Second, we will try to identify the reason why any of the included trials may be different from the other trials and why it could cause heterogeneity before we perform the meta-analysis. If heterogeneity is found, then we will perform a sensitivity analysis to exclude the trial that we thought may have caused the heterogeneity.

Sensitivity analysis

Where possible and appropriate, we will conduct sensitivity analyses on the primary outcomes to analyse the effect of including any of the following studies:

  1. studies that were judged to be at high risk of bias across one or more domains of The Cochrane Collaboration's 'Risk of bias' tool;

  2. studies where we assigned values for missing data;

  3. sensitivity analyses to assess the assumptions of fixed-effect versus random-effects models.

Acknowledgements

We are grateful to the Cochrane Heart Group for their help and comments on the protocol.

Appendices

Appendix 1. Preliminary search strategy for MEDLINE (Ovid)

1. Pericarditis/
2. pericard*.tw.
3. or/1-2
4. Colchicine/
5. colchicin*.tw.
6. colchin.tw.
7. colchicum*.tw.
8. colchily.tw.
9. colchimedio.tw.
10. colchiquim.tw.
11. colchisol.tw.
12. colchysat.tw.
13. colcine.tw.
14. colcrys.tw.
15. colgout.tw.
16. goutichine.tw.
17. goutnil.tw.
18. kolkicin.tw.
19. nsc 757.tw.
20. tolchicine.tw.
21. or/4-20
22. 3 and 21

Contributions of authors

S Alabed: conceiving the review, designing the protocol, co-ordinating the protocol, entering data into RevMan and writing the protocol.

JB Cabello: conceiving the review, co-ordinating the protocol and writing the protocol.

GJ Irving: writing the protocol.

M Qintar: writing the protocol.

M Imazio: participating in the design and commenting on the protocol.

Declarations of interest

None known.

Sources of support

Internal sources

  • Department of Continuing Education - University of Oxford, UK.

    Providing access to journals and books

External sources

  • No sources of support supplied

Ancillary