The committee found that the cumulative data supporting an effect on hospitalization with malaria parasites, anemia and all-cause hospitalization were more modest and less consistent across the trials than the effect on episodes of clinical malaria (Table 18). The committee found the efficacy estimates for these additional outcomes to be encouraging but less robust than the cumulative data for efficacy against clinical malaria. Accordingly, the committee remained cautious in drawing conclusions concerning the effect of IPTi-SP in preventing these other outcomes. The analyses from randomization through 5 months after the last dose of IPTi-SP leave open the possibility that studies with much larger sample sizes might have demonstrated a statistically more convincing protective effect.
Depending on the specific outcome event measured, the committee found mixed evidence regarding the existence of a rebound. In no case was the rebound sufficiently large to negate the overall benefit of IPTi-SP. Based on its review of all the data and the analyses presented, the committee concluded that the extent of rebound is small and that the benefits of IPTi-SP outweigh this negative effect.
TABLE 18 Overall Pooled Estimates of Efficacy Against Clinical Malaria and Other Relevant Outcomes During Two Periods of Follow-Up
|
|
Period of Observation |
|
|
Outcome |
During the 5 months Starting 35 days After the Last Dose of SP or Placebo (Rebound Period) |
From Randomization Up to 5 Months After the Last Dose of SP or Placebo |
|
Clinical malaria |
|
|
|
% PE |
0 |
21 |
|
95 CI (%) |
(−10, 9) |
(11, 29) |
|
p-value |
0.99 |
<0.001 |
|
Hospitalizations of children with malaria parasites |
|
|
|
% PE |
−20 |
21a |
|
95 CI (%) |
(−60, 10) |
(−2, 38)a |
|
p-value |
0.23 |
0.066a |
|
All-cause hospitalizations |
|
|
|
% PE |
−11 |
18a |
|
95 CI (%) |
(−30, 6) |
(9, 27)a |
|
p-value |
0.22 |
0.001a |
|
Subjects with Anemia |
|
|
|
% PE |
2 |
10a |
|
95 CI (%) |
(−8, 11) |
(4, 17)a |
|
p-value |
0.74 |
0.003a |
|
a The committee used Stata to calculate the overall efficacy and its 95 percent confidence limit. SOURCE: Data from the Report of the Statistical Working Group (IPTi Consortium, 2007b). |
||
Recommendation: In view of the importance of the unpublished analyses by the SWG in showing a net benefit for IPTI-SP, and whereas the committee had no information about how the SWG or the individual study teams ensured quality control of the individual study data and hence the uniformly defined outcomes, the committee recommends that the SWG obtain an independent technical audit of the accuracy of the study-level data and analyses included in the pooled analysis. If this