Clindamycin Plus Quinine for Treating Falciparum Malaria
Clindamycin Plus Quinine for Treating Falciparum Malaria
Studies were considered for inclusion in the review if they were randomized controlled trials designed to compare the efficacy of clindamycin plus quinine with another anti-malarial drug (used alone or in combination) in participants with symptomatic, microscopically-confirmed uncomplicated falciparum malaria. Studies that had enrolled participants with signs of severe malaria were excluded. The review's primary outcome was prospectively defined as parasitological treatment failure evaluated 28 days after starting treatment. Secondary outcomes included day 14 parasitological failure, gametocyte carriage, parasite and fever clearance time, mean haemoglobin and adverse events.
Using a combination of the terms, "malaria", "quinine", "clindamycin" or" dalacin", a search of the following electronic databases was made: Cochrane Infectious Diseases Group specialized register; Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library 2011, Issue 3; MEDLINE (1966 to October 2011); EMBASE (1988 to October 2011); and LILACS (1982 to October 2011). The third, fourth and fifth MIM Pan-African Malaria Conference proceedings were also searched for relevant abstracts. Individual researchers working in the field, organizations including WHO and the East African Network for Monitoring Antimalarial Treatment (EANMAT), and pharmaceutical companies including Pfizer, Sanofi-Aventis, Novartis, and Rhône-Poulenc Rorer were contacted for unpublished and ongoing trials. The reference lists of retrieved studies were also checked for additional studies. An attempt was made to retrieve all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).
Abstracts obtained by the search strategy were screened for potentially relevant trials. Full text articles of the selected trial reports were retrieved. Each trial report was scrutinized for multiple publications from the same data set. If a trial was published more than once, only one publication was assessed. Based on the inclusion criteria, both authors independently assessed the trials for inclusion in the review. For the included trials, data on trial dates, location, publication status, trial methods, participants, interventions and outcomes was independently extracted by both authors onto data abstraction forms. Any disagreements were resolved by referring to the study report and by discussion. Additional information was sought from the trial authors if the available data was insufficient or missing.
Data was extracted to allow for an intention-to-treat analysis. For dichotomous outcome, the number of participants experiencing the event and the number analysed in each treatment group was recorded. For continuous outcome measures (e.g. fever and parasite clearance times), the arithmetic means and standard deviations for each group was extracted in addition to the numbers analysed in each group. The range was extracted if medians were reported.
The methodological quality for each included trial was independently assessed by both authors. Specifically, the following components of methodological quality were assessed: generation of allocation sequence, allocation concealment, blinding and loss to follow up. Generation of allocation sequence and allocation concealment were classified as adequate, inadequate, or unclear. Blinding was classified as open, single, or double blind. The percentage of randomized participants lost to follow up was calculated, and the percentage of participants available for analysis was categorized as adequate, if it was at least 90%.
Data was analysed using Review Manager 5.1 software. Dichotomous data was combined using risk ratios (RR), while continuous data was combined using the weighted mean difference (WMD). Where arithmetic means were reported for an outcome measure where the scale is naturally bound at zero, the ratio of the mean to the standard deviation was used to check the assumption that the data are normally distributed. When data was suspected to be skewed or inappropriately summarized as means and standard deviations (mean/sd > 2) then it was not combined in a meta-analysis. The results are presented as point estimates together with the 95% confidence intervals. P values less than 0.05 were considered statistically significant differences. A fixed effects model was used in pooling data where there was no evidence of heterogeneity.
Statistical heterogeneity was assessed by visually inspecting the forest plots (for overlapping confidence intervals), applying the chi-square test (p value < 0.10 considered statistically significant), and the I statistic with the value of 50% used to denote moderate levels of heterogeneity. If heterogeneity was detected, and if it was still appropriate to pool the data, then a random-effects model was used. Potential sources of heterogeneity were explored by conducting subgroup analyses using participant age (less than versus greater or equal to 5 years), and the effect of different drug dosing regimes on treatment efficacy. A sensitivity analysis was performed using allocation concealment to evaluate the robustness of our conclusions.
Methods
Search Strategy and Selection Criteria
Studies were considered for inclusion in the review if they were randomized controlled trials designed to compare the efficacy of clindamycin plus quinine with another anti-malarial drug (used alone or in combination) in participants with symptomatic, microscopically-confirmed uncomplicated falciparum malaria. Studies that had enrolled participants with signs of severe malaria were excluded. The review's primary outcome was prospectively defined as parasitological treatment failure evaluated 28 days after starting treatment. Secondary outcomes included day 14 parasitological failure, gametocyte carriage, parasite and fever clearance time, mean haemoglobin and adverse events.
Using a combination of the terms, "malaria", "quinine", "clindamycin" or" dalacin", a search of the following electronic databases was made: Cochrane Infectious Diseases Group specialized register; Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library 2011, Issue 3; MEDLINE (1966 to October 2011); EMBASE (1988 to October 2011); and LILACS (1982 to October 2011). The third, fourth and fifth MIM Pan-African Malaria Conference proceedings were also searched for relevant abstracts. Individual researchers working in the field, organizations including WHO and the East African Network for Monitoring Antimalarial Treatment (EANMAT), and pharmaceutical companies including Pfizer, Sanofi-Aventis, Novartis, and Rhône-Poulenc Rorer were contacted for unpublished and ongoing trials. The reference lists of retrieved studies were also checked for additional studies. An attempt was made to retrieve all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).
Data Extraction and Management
Abstracts obtained by the search strategy were screened for potentially relevant trials. Full text articles of the selected trial reports were retrieved. Each trial report was scrutinized for multiple publications from the same data set. If a trial was published more than once, only one publication was assessed. Based on the inclusion criteria, both authors independently assessed the trials for inclusion in the review. For the included trials, data on trial dates, location, publication status, trial methods, participants, interventions and outcomes was independently extracted by both authors onto data abstraction forms. Any disagreements were resolved by referring to the study report and by discussion. Additional information was sought from the trial authors if the available data was insufficient or missing.
Data was extracted to allow for an intention-to-treat analysis. For dichotomous outcome, the number of participants experiencing the event and the number analysed in each treatment group was recorded. For continuous outcome measures (e.g. fever and parasite clearance times), the arithmetic means and standard deviations for each group was extracted in addition to the numbers analysed in each group. The range was extracted if medians were reported.
Assessment of the Methodological Quality of Included Studies
The methodological quality for each included trial was independently assessed by both authors. Specifically, the following components of methodological quality were assessed: generation of allocation sequence, allocation concealment, blinding and loss to follow up. Generation of allocation sequence and allocation concealment were classified as adequate, inadequate, or unclear. Blinding was classified as open, single, or double blind. The percentage of randomized participants lost to follow up was calculated, and the percentage of participants available for analysis was categorized as adequate, if it was at least 90%.
Data Analysis
Data was analysed using Review Manager 5.1 software. Dichotomous data was combined using risk ratios (RR), while continuous data was combined using the weighted mean difference (WMD). Where arithmetic means were reported for an outcome measure where the scale is naturally bound at zero, the ratio of the mean to the standard deviation was used to check the assumption that the data are normally distributed. When data was suspected to be skewed or inappropriately summarized as means and standard deviations (mean/sd > 2) then it was not combined in a meta-analysis. The results are presented as point estimates together with the 95% confidence intervals. P values less than 0.05 were considered statistically significant differences. A fixed effects model was used in pooling data where there was no evidence of heterogeneity.
Statistical heterogeneity was assessed by visually inspecting the forest plots (for overlapping confidence intervals), applying the chi-square test (p value < 0.10 considered statistically significant), and the I statistic with the value of 50% used to denote moderate levels of heterogeneity. If heterogeneity was detected, and if it was still appropriate to pool the data, then a random-effects model was used. Potential sources of heterogeneity were explored by conducting subgroup analyses using participant age (less than versus greater or equal to 5 years), and the effect of different drug dosing regimes on treatment efficacy. A sensitivity analysis was performed using allocation concealment to evaluate the robustness of our conclusions.