French researchers report a new strategy that they believe might help decrease the antibiotic exposure of patients in the intensive care unit (ICU) without increasing patient mortality. However, the study had numerous design flaws that undermine the findings.
Reporting in the Lancet, the research team expanded an existing hypothesis that the hormone procalcitonin is a fairly specific biomarker for severe bacterial infection in patients with suspected sepsis. The studies of patients admitted to emergency departments or to the hospital with lower respiratory tract infections have corroborated the utility of this biomarker in reducing antibiotic use. However, no studies have shown whether this could be useful within the ICU.
This randomized, open-label study enrolled patients in eight ICUs who were: (1) expected to be in the ICU for at least 3 days, (2) had a suspected bacterial infection, and (3) were at least 18 years of age. They were randomized to a procalcitonin group (n=311) or to a control group (n=309). The physicians treating patients in the procalcitonin group were instructed to start or stop antibiotics per a treatment algorithm using a preset range of calcitonin levels; the patients in the control group received antibiotics according to current guidelines. The study’s primary endpoints were death from any cause at 28 and 60 days and differences in antibiotic use.
At 30 days, mortality in the procalcitonin group was 21.2% versus 20.4% in the control group. The absolute difference was 0.8%, which was below the preset 10% difference required to establish noninferiority. At 60 days, the corresponding figures were 30% and 26.1% (absolute difference, 3.8%). The patients in the procalcitonin group achieved 14.3 days without antibiotics compared with 11.6 days in the procalcitonin group. The absolute difference of 2.7 days was significant at P<0.0001. However, despite having been less exposed to antibiotics, the procalcitonin group showed no difference in the rate of emerging multidrug-resistant bacteria compared with the control group.
The limitations to the study include the open design and a low number of surgical patients, which may limit the applicability of the findings. As well, 53% of patients in the procalcitonin group did not received therapy guided by the algorithm either due to physician choice or because the patient left the ICU. However, the authors state that the results would remain significant even if these patients were excluded.
While this approach to curtailing unnecessary antibiotic use in critically ill patients is interesting, the design flaws limit its use in any widespread infectious disease program at this time. It is still not clear whether the ideal long-term strategy will involve the use of a serum marker such as procalcitonin or a locally applied, clinically validated practice protocol.