|
Effet de la dapsone sur la survie des patients infectes par le VIH: une meta-analyse des essais termines [Effect of dapsone on survival in HIV -infected patients: a meta-analysis of finished trials] |
Saillour-Glenisson F, Chene G, Salmi L R, Hafner R, Salamon R |
|
|
Authors' objectives To study the effect of dapsone on survival in HIV-infected patients.
Searching Computerised searches of MEDLINE, AIDS TRIALS and AIDS DRUGS databases were performed from 1983 to January 1996. The clinical trials registries of collaborative groups studying HIV infection in France and the USA were also searched. Additional studies were identified by examining the abstract books of international conferences on AIDS and infectious diseases, consulting experts in HIV infection, and reviewing reference lists of retrieved articles. Papers published in English, French or Spanish were sought.
Study selection Study designs of evaluations included in the reviewRandomised clinical trials (RCTs) with at least one arm evaluating dapsone, and with any length of follow-up, were included. Only trials incorporating an intention to treat analysis were included in the meta-analysis. The mean length of follow-up ranged from 10 to 39 months.
Specific interventions included in the reviewDapsone given as primary or secondary prophylaxis of Pneumocystis carinii pneumonia (PCP), used alone or in combination with other drugs. Cotrimoxazole and aerosolised pentamidine were the agents most frequently compared to dapsone. The cumulative weekly dosage of dapsone ranged from 100 to 700 mg. Dapsone was studied in combination with pyrimethamine in 10 trials and with iron protoxalate in 1 study. One trial studied the efficacy of dapsone specifically for secondary prophylaxis of PCP. The others included only primary prophylaxis or both primary and secondary prophylaxis.
Participants included in the reviewStudies recruiting adults with HIV infection were included. The main inclusion criteria for most of the individual trials was CD4+ cell count less than 200/mm3.
Outcomes assessed in the reviewThe outcome was survival.
How were decisions on the relevance of primary studies made?The authors do not state how the papers were selected for the review, or how many of the reviewers performed the selection.
Assessment of study quality The trials were assessed using the scoring system proposed by Chalmers et al. (see Other Publications of Related Interest no.1). The score was expressed as a proportion of fulfilled criteria, and ranged from 0 (lowest quality) to 100% (highest quality). The actual criteria used were not listed in the paper. The authors do not state how the papers were assessed for validity, or how many of the reviewers performed the validity assessment.
Data extraction The authors of trials were contacted and asked to complete a standardised questionnaire which covered the following: year of publication; dosage of dapsone; number and type of comparative groups; and for each arm - number of patients, duration of follow-up in person-years, mortality, PCP and toxoplasmosis end points, adverse effects occurring during follow-up, and baseline characteristics (proportion of patients with AIDS, mean CD4+ cell count, mean haemoglobin level, and proportion of patients receiving anti-retroviral therapy). The trialists were also asked to provide files of individual patient data wherever possible. Each individual data file was checked to obtain accurate and complete data from all patients. Aggregated data were derived from the individual patient data where available. If individual data were not available, aggregated data were extracted from the questionnaire or from the published paper. Two reviewers used a standardised form to independently record the aggregated data. Data were extracted in an unblinded fashion.
Methods of synthesis How were the studies combined?An analysis of aggregated data was undertaken.
Using the maximum likelihood method (see Other Publications of Related Interest no.2), summary odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to estimate the overall effect of dapsone on survival. Estimates were generated using both random-effects and fixed-effect models.
Studies were combined by analysis of individual data.
Survival probability was estimated using the Kaplan-Meier product-limit method (see Other Publications of Related Interest no.3). A proportional hazards model was used to estimate the effect of dapsone on survival by a hazard ratio (HR) stratified on studies (see Other Publications of Related Interest no.3).
How were differences between studies investigated?Differences between the studies were investigated by analysis of aggregated data.
Heterogeneity between studies was tested and explored using a Galbraith plot (see Other Publications of Related Interest nos.2 and 4). Possible sources of heterogeneity were investigated using a meta-regression (see Other Publications of Related Interest no.5). In the meta-regression, the natural log of the study OR was regressed on study characteristics (other treatments such as pentamidine, cotrimoxazole or fansidar, full article or abstract available, weekly cumulative dosage of dapsone, association of dapsone with iron protoxalate or pyrimethamine, the calendar year of initiation, mean length of follow-up) or patient characteristics (mean CD4+ cell count at baseline, proportion of patients with anti-retroviral treatment at baseline, mean age at baseline, and proportion of patients with a diagnosis of AIDS at baseline).
Differences between studies were also investigated by analysis of individual data.
The same variables evaluated in the meta-regression described above were tested as possible sources of heterogeneity, with the addition of type of prophylaxis (primary versus secondary), which could be studied only with individual observations.
Results of the review Overall, 17 RCTs (n=4,343) were identified as being eligible for inclusion. The meta-analysis of aggregated data included 16 trials with a total of 4,267 patients and 1,148 deaths (501 in dapsone group and 647 in control group). Individual data were available for 10 trials with a total of 3,115 patients and 978 deaths (410 in the dapsone group and 568 in the control group).
Validity assessment.
The median quality score for studies published as articles in peer-reviewed journals was 51% (range: 11 - 75). When the lowest scoring trial (published as a letter to the editor) was excluded, the median score remained similar and the range became 43 to 75%.
Meta-analysis of aggregated data.
There was no evidence of an overall deleterious effect of dapsone on survival (OR 1.11, 95% CI: 0.96, 1.29, p=0.16), or of heterogeneity between studies. The results of the random-effects model analysis were similar to those of the fixed-effect model analyses (OR 1.11, 95% CI: 0.96, 1.29, p=0.15). Including only control group patients taking pentamidine, the OR was 1.08 (95% CI: 0.89, 1.30). Including only control group patients taking cotrimoxazole, the OR was 1.12 (95% CI: 0.91, 1.37). In a meta-regression including data from all studies, the risk of death increased in patients with lower CD4+ cell counts (OR for a decrease of 50 CD4+ cells/mm3=1.50, 95% CI: 1.01, 2.21). However, when one study which did not use a cut-off level of CD4+ cell count as the main study inclusion criteria was excluded from the meta-regression, no association was found between any tested characteristic and survival. Also it was the only study to use dapsone combined with iron protoxalate.
Analysis of individual data files.
The overall OR of the aggregated data in this subsample for the effect of dapsone on survival was 1.10 (95% CI: 0.93, 1.29); this was similar to the results for all studies. The overall survival probability was 88.7% (95% CI: 87.4, 89.8) at 1 year, 67.7% (95% CI: 65.7, 69.7) at 2 years, 48.1% (95% CI: 45.4, 50.7) at 3 years, and 32.5% (95% CI: 27.4, 37.8) at 4 years. Survival did not differ significantly between the dapsone and control groups. Estimation of the overall effect of dapsone on survival, stratified on studies with available individual data, showed no deleterious effect of dapsone on survival (HR 1.12, 95% CI: 0.99, 1.27). Omitting the data from the trial reporting the greatest effect of dapsone on survival in the meta-analysis of all individual files did not change the results significantly. Including only control group patients taking cotrimoxazole, the HR was 1.15 (95% CI: 0.96, 1.37). The interaction between treatment group and CD4+ cell count was not statistically significant (p=0.8). In patients receiving secondary PCP prophylaxis (n=299), the HR was 1.77. In patients receiving primary PCP prophylaxis, the HR was 1.07. The interaction between treatment group and type of prophylaxis was significant (p=0.02) in that a deleterious effect of dapsone as treatment for secondary prophylaxis was noted. However, when one outlying trial was excluded from the analysis of individual data, no significant interaction existed between the type of prophylaxis and use of dapsone. The stratification on dapsone dosage did not show a dose-response relationship between dapsone and the risk of death. The adjustment for all the collected variables that are known to have a potential effect on survival in HIV-infected patients (e.g. CD4 lymphocyte count) did not change the estimation of the effect of dapsone on survival.
Authors' conclusions The overall results of this meta-analysis showed no evidence of a deleterious effect of dapsone on survival in patients receiving PCP prophylaxis. Dapsone can be used safely as primary prophylaxis for PCP. However, no definitive recommendation may be made for the use of dapsone as secondary PCP prophylaxis, and further research is needed to resolve this issue.
CRD commentary Overall, this is a well-conducted systematic review. The selection criteria for primary studies were clearly explained and some details of individual trials were presented in tables. The methods used for data synthesis were appropriate, between-trial heterogeneity was assessed, and sensitivity analyses were performed in order to explore differences between studies. Assessment of the validity of individual trials was carried out, although the criteria used were not listed and the results of the assessment per trial were not presented in the paper. Relevant sources were searched, and attempts to identify unpublished data were made by contacting experts in the field. Whilst details of the process for data extraction were given (i.e. the number of reviewers involved), more information about the processes relating to study selection and validity assessment would have been useful. The authors' conclusions are supported by the data presented.
Implications of the review for practice and research Practice: The authors state that dapsone can be used safely as primary prophylaxis for PCP. No definitive recommendation may be made for the use of dapsone as secondary PCP prophylaxis.
Research: The authors state that no definitive recommendation may be made for the use of dapsone as secondary PCP prophylaxis, and further research is needed to resolve this issue.
Bibliographic details Saillour-Glenisson F, Chene G, Salmi L R, Hafner R, Salamon R. Effet de la dapsone sur la survie des patients infectes par le VIH: une meta-analyse des essais termines [Effect of dapsone on survival in HIV -infected patients: a meta-analysis of finished trials] Revue d'Epidemiologie et de Sante Publique 2000; 48(1): 17-30 Other publications of related interest 1. Chalmers TC, Smith H, Blackburn B, Silverman B, Schroeder B, Reitman D, et al. A method for assessing the quality of a randomized control trial. Control Clin Trials 1981;2:31-49. 2. Hedges LV, Olkin I. Statistical methods for meta-analysis. San Diego (CA): Academic Press; 1985. 3. Collett D. Modelling survival data in medical research. London: Chapman and Hall; 1994. 4. Galbraith RF. A note on graphical presentation of estimated odds ratios from several clinical trials. Stat Med 1988;7:889-94. 5. Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev 1987;9:1-30.
Indexing Status Subject indexing assigned by NLM MeSH AIDS-Related Opportunistic Infections /prevention & Adult; Anti-Infective Agents /therapeutic use; Anti-Inflammatory Agents, Non-Steroidal /therapeutic use; Confidence Intervals; Controlled Clinical Trials as Topic; Dapsone /therapeutic use; Follow-Up Studies; HIV Infections /drug therapy /mortality; Humans; Pneumonia, Pneumocystis /prevention & Randomized Controlled Trials as Topic; Retrospective Studies; Time Factors; control; control AccessionNumber 12000000734 Date bibliographic record published 31/01/2002 Date abstract record published 31/01/2002 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|