© 2006 American Thoracic Society
The Sound and the FuryTo the Editor:The recent editorial by Drs. Fowler and Annane (1) raises a number of unresolved questions about clinical science. How much evidence should be required to change physicians' practice (assuming they assess and apply the evidence)? When evidence is preliminary but promising, what "threshold" warrants a change in medical practice until stronger evidence is available? When is evidence so overwhelming that a question is answered? Intensive insulin therapy (IIT) joins recent examples of dietary fat, activated protein C and early goal-directed therapy for sepsis, to name a few also cited by Drs. Fowler and Annane. Each offers a different paradigm. After just one study demonstrating favorable outcomes of surgical patients (2), IIT was generalized to include medical patients and embraced to treat patients in many ICUs (and some hospital wards) around the world. Is physiologic plausibility and one study enough? In 1,200 medically ill patients, IIT failed to improve mortality even after those least expected to benefit were excluded from study (3). In addition, as Fowler and Annane cite (1), a German study of septic patients was suspended due to lack of effect and excess toxicity in the IIT group. Will IIT, which is very labor intensive and uncomfortable for patients, be used less in ICUs? Are we willing to reconsider after new evidence refutes initial, hopeful reports? The story for activated protein C differs slightly. One well-constructed study demonstrated mortality benefit of patients with severe sepsis (4). The reproducibility of this observation may never be tested because of the high cost of a second study. Additional reports suggest that toxicity was underestimated in the initial study (5). Should we give it or not? To whom should it be given? Low-fat diets were associated with no reduction in cardiovascular mortality in women (6). Various reasons have been entertained about why the intervention may not have worked. Could it have been that the hypothesis was wrong? The study is unlikely to be reproduced. Should we prescribe low-fat diets or not? The history of goal-directed resuscitation for sepsis is littered with negative studies (e.g., Reference 7). Yet when one study yields the hoped-for result (8), we embrace it and devise arguments about why previous studies were inconsistent. It has been 5 years without a second study to reproduce the initially promising results. It is only common sense that we should commence resuscitations of septic patients in the emergency department. But should we use venous oximetry–guided endpoints or not? We physicians are a hopeful group. We're attracted to interesting new ideas and interventions that may help our patients. History shows that sometimes we cling to ideology at the expense of our patients, overreacting to initial studies or disregarding negative results until a study is published to support attractive therapeutic ideas. Unfortunately, there is no universal litmus test that can be applied to the medical literature to help guide us. As the recent Vioxx and stem cell debacles demonstrate, we are easily misled. Reproducibility, then, becomes the final test of validity. Nonetheless, it is very likely that millions of medically ill patients will continue to receive early goal-directed therapy with or without activated protein C for severe sepsis, IIT for stress hyperglycemia and a low fat diet for good measure. All these hypotheses may stand the test of time. But that possibility is improbable (9).
Bridgeport Hospital and Yale University School of Medicine, Bridgeport, Connecticut FOOTNOTES Conflict of Interest Statement: C.A.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
