The FEAST trial of fluid bolus in African children with severe infection
2012; Elsevier BV; Volume: 379; Issue: 9816 Linguagem: Inglês
10.1016/s0140-6736(12)60260-8
ISSN1474-547X
AutoresKathryn Maitland, Abdel Babiker, Sarah Kiguli, Elizabeth Molyneux,
Tópico(s)Global Maternal and Child Health
ResumoWe acknowledge that the assimilation of the FEAST trial results1Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar is challenging, since they question decades of practice in resource-rich countries and our understanding of the pathophysiology of shock in severe infection. This is precisely why carefully done randomised clinical trials are so important, since they generate robust data to guide and change clinical practice. We now welcome discussion of the possible explanations and interpretations of our findings.Trevor Duke's Comment (Nov 12, p 1685)2Duke T What the African fluid-bolus trial means.Lancet. 2011; 378: 1685-1687Summary Full Text Full Text PDF PubMed Scopus (32) Google Scholar contains two main criticisms of FEAST: (1) that the inclusion criteria were broader than those used in most shock classification systems, prompting him to question any extrapolation of the results to children with "shock"; and (2) that final diagnoses were not reported, although the characteristics of the trial population suggested that many of the children had disorders (severe anaemia, pneumonia, meningitis, or encephalopathy) that can be adversely affected by bolus administration. Finally, he voices concern over the protocol amendment to increase the volume of fluid boluses.First, fluid boluses were associated with increased mortality, irrespective of the clinical criteria used to define shock, including the stringent definition used by WHO (presence of cold hands or feet with a capillary refill time of ≥4 s and a weak, fast pulse). Of the 50 children in this category who received fluid boluses, 24 (48%) died, compared with three of 15 (20%) in the no-bolus group—an absolute risk difference of 28% (95% CI 3·4–52·5; appendix table 51Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar and subsequent correspondence3Maitland K Akech SO Russell EC for the FEAST Trial GroupMortality after fluid bolus in African children with sepsis.N Engl J Med. 2011; 365: 1351-1353Google Scholar). Table 5 also shows increased mortality in children with moderate hypotension (192 children) and those with any one of the WHO clinical criteria for shock who received bolus therapy (1890 children).Second, final clinical diagnoses are shown in appendix table 2.1Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar Duke's concern about children with oxygen saturation of less than 90%, "many of whom probably had pneumonia", is countered by the greater mortality associated with boluses in children without hypoxaemia than in those with hypoxaemia: risk ratio 1·91 versus 1·09 (p=0·04 for heterogeneity).Finally, Duke expresses concern that the protocol amendment to increase the volume of boluses was made at a time when it was already clear that boluses were harmful. This is incorrect; the independent data monitoring committee (IDMC) would not have permitted the trial to continue if there had been safety concerns at any of the interim analyses. The amendment was proposed by the study team (who were blind to all interim results) in June, 2010, because they were concerned that the trial might fail to reach a conclusive result, since bolus volumes were much lower than standard practice in resource-rich countries. The amendment was endorsed by the trial steering committee. The published data include additional children recruited after the penultimate review of the IDMC but before the implementation of the amendment from August, 2010, after ethics committee approval. The investigators unanimously support recommendations made by the IDMC in the conduct of the FEAST trial.We declare that we have no conflicts of interest. We acknowledge that the assimilation of the FEAST trial results1Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar is challenging, since they question decades of practice in resource-rich countries and our understanding of the pathophysiology of shock in severe infection. This is precisely why carefully done randomised clinical trials are so important, since they generate robust data to guide and change clinical practice. We now welcome discussion of the possible explanations and interpretations of our findings. Trevor Duke's Comment (Nov 12, p 1685)2Duke T What the African fluid-bolus trial means.Lancet. 2011; 378: 1685-1687Summary Full Text Full Text PDF PubMed Scopus (32) Google Scholar contains two main criticisms of FEAST: (1) that the inclusion criteria were broader than those used in most shock classification systems, prompting him to question any extrapolation of the results to children with "shock"; and (2) that final diagnoses were not reported, although the characteristics of the trial population suggested that many of the children had disorders (severe anaemia, pneumonia, meningitis, or encephalopathy) that can be adversely affected by bolus administration. Finally, he voices concern over the protocol amendment to increase the volume of fluid boluses. First, fluid boluses were associated with increased mortality, irrespective of the clinical criteria used to define shock, including the stringent definition used by WHO (presence of cold hands or feet with a capillary refill time of ≥4 s and a weak, fast pulse). Of the 50 children in this category who received fluid boluses, 24 (48%) died, compared with three of 15 (20%) in the no-bolus group—an absolute risk difference of 28% (95% CI 3·4–52·5; appendix table 51Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar and subsequent correspondence3Maitland K Akech SO Russell EC for the FEAST Trial GroupMortality after fluid bolus in African children with sepsis.N Engl J Med. 2011; 365: 1351-1353Google Scholar). Table 5 also shows increased mortality in children with moderate hypotension (192 children) and those with any one of the WHO clinical criteria for shock who received bolus therapy (1890 children). Second, final clinical diagnoses are shown in appendix table 2.1Maitland K Kiguli S Opoka RO et al.Mortality after fluid bolus in African children with severe infection.N Engl J Med. 2011; 364: 2483-2495Crossref PubMed Scopus (1057) Google Scholar Duke's concern about children with oxygen saturation of less than 90%, "many of whom probably had pneumonia", is countered by the greater mortality associated with boluses in children without hypoxaemia than in those with hypoxaemia: risk ratio 1·91 versus 1·09 (p=0·04 for heterogeneity). Finally, Duke expresses concern that the protocol amendment to increase the volume of boluses was made at a time when it was already clear that boluses were harmful. This is incorrect; the independent data monitoring committee (IDMC) would not have permitted the trial to continue if there had been safety concerns at any of the interim analyses. The amendment was proposed by the study team (who were blind to all interim results) in June, 2010, because they were concerned that the trial might fail to reach a conclusive result, since bolus volumes were much lower than standard practice in resource-rich countries. The amendment was endorsed by the trial steering committee. The published data include additional children recruited after the penultimate review of the IDMC but before the implementation of the amendment from August, 2010, after ethics committee approval. The investigators unanimously support recommendations made by the IDMC in the conduct of the FEAST trial. We declare that we have no conflicts of interest. The FEAST trial of fluid bolus in African children with severe infection – Authors' replyI think the FEAST Trial Group might have misunderstood some of my comments. I did not say that the results could not be extrapolated to children with "shock" in their setting. I agree with Kathryn Maitland and colleagues that they can be. What I said was: "The main message—fluid boluses are dangerous in children with malaria, and in other common febrile illnesses in which ADH secretion is likely to be high—is important to disseminate clearly" and "This trial does not inform about the management of other shock conditions: dengue shock syndrome in Asia, or hypovolaemic shock from diarrhoea and vomiting. Full-Text PDF
Referência(s)