To date there have been 46 trials of SDD and three meta-analyses have recently been performed ( SD.D lIr!aJ.i.StS! CPJ.J§bor§t!V§.G.^oup 1993; Heyland et,§L 1994;
KoJieii i1994). These meta-analyses differ not only in the number of trials and patients analysed, but also in the methodology. KoJJ.efi1994.) analysed only published trials including 2270 patients in 21 trials. H§.yland.ll..e.t...,a.l (1994) analysed 3395 patients in 25 published and unpublished randomized trials. The SDDllI^ia.!iSÍS!
Co||aborativelllG^royP.lllí1993} analysed 4142 patients from 22 randomized trials, both published and unpublished; in order to perform an intent-to-treat analysis they collected all patient data through direct contact with the investigators.
All meta-analyses showed a reduction of 60 to 70 per cent in the relative risk of respiratory infections in patients receiving intensive care. The SDD Trialists' Collaborative Group found that SDD significantly reduced respiratory tract infections (odds ratio (OR), 0.37; 95 per cent confidence interval (CI), 0.31-0.43). The effect of SDD on pneumonia rate was independent of both the criteria used for diagnosis of pneumonia and the blinding of the study ( SDDii lT.ria.!.!St.S!llllC.O.!.!aborative
Group 199.3.; Heyland et...§l 1994). Although the infection rate was lower in studies diagnosing pneumonia by the protected brush technique or by bronchoalveolar lavage than in studies using clinical criteria, it was demonstrated that the reduction of the odds ratio for pneumonia was similar in both subgroups of studies.
The effect of SDD on mortality is less clear. In the meta-analysis by the SDD Trialists' Collaborative Group the value of the common odds ratio for total mortality (OR, 0.90; CI, 0.79-1.04) suggested a moderate treatment effect, reaching statistical significance only when the subgroup of trials using both topical and systemic treatment was considered separately (OR, 0.80; CI, 0.67-0.97). This was confirmed by the meta-analysis by Heyland e.tM (1994). In a subgroup analysis the largest absolute difference was found in the comparison of studies using regimens including intravenous antibiotics (risk ratio (RR), 0.80; CI, 0.68-0.95) and those that excluded systemic antibiotics (RR, 0.99; CI, 0.83-1.19).
The subgroup of studies using a systemic component also differs from those without a systemic component with respect to the decontamination regimen used: in the majority of the former both oral and intestinal decontamination with the PTA regimen were used, whereas in the latter many different topical antibiotic regimens were used and were applied only in the gastrointestinal tract without oropharyngeal decontamination. Thus the mortality difference between the two groups may be explained not only by the use of the systemic component but also by the effectiveness of the topical regimen. The discrepancy between the dramatic reduction in infection rate and the modest reduction in mortality clearly demonstrates that the underlying disease is a more important determinant of mortality than acquired infections. SDD might be particularly beneficial in patients with reversible underlying diseases (e.g. surgical patients)
The clinical relevance of the reduction in mortality can be assessed by the number of patients to be treated to prevent one death. Based on the results from trials in which combined topical and systemic treatment was used, 23 patients (range 13-139) would need to be treated with SDD to prevent one death ( SDD.Trialists'... Collaborative ...Group. .1993).
Was this article helpful?