© 2003 American Thoracic Society
CorticosteroidsThe "Terminator" of All Untreatable Serious Pulmonary IllnessVanderbilt University School of Medicine Nashville, Tennessee The SARS epidemic has opened yet another steroid treatment controversy. This respiratory syndrome strikes without warning, results in a high morbidity, often leads to death, and its pathophysiology is not completely understood. When the medical community is faced with such situations, it turns to those therapies that are familiar, even if not clearly effective. About the same time as the SARS-CoV was identified, standard treatment protocols began to emerge, many of which included the use of corticosteroids alone or in combination with other therapies, while other reports warned that corticosteroids should be avoided (1–3). There has been criticism over the use of corticosteroids alone and in combination with other drugs because the necessary prospective randomized placebo-controlled clinical trials have not been done. Given the circumstances, however, it is completely understandable that information regarding efficacy has been hard to come by (4, 5). In this issue of the Journal (pp. 1449–1456), Ho and coworkers (6) provide one of the earliest reports of attempts to treat SARS pharmacologically. It is a tribute to these investigators that an interventional effort could be mounted on such short notice, especially given the uncertainty that surely must have permeated the hospital atmosphere and the resulting chaos such mortal fear engenders. While diagnosis of SARS is based on presence of fever, cough, and respiratory symptoms, it also clearly causes acute lung injury/acute respiratory distress syndrome (ALI/ARDS) in some patients (7). The latter complication is the most feared and appears to be the most lethal aspect of SARS. For more than a quarter of a century, the critical care community has researched, reported, and debated whether to administer corticosteroids in the treatment of ALI/ARDS, often regardless of cause, and to date that issue remains unresolved (8–9). There are a few acute diseases of the lung parenchyma in which steroids are likely effective in well selected cases, such as bronchiolitis obliterans (10), tuberculosis (11), and Pneumocystis carinii pneumonia (12). But beneficial effects in other conditions associated with ALI/ARDS, especially severe sepsis, and in this case, viral pneumonia, are far less clear (8, 9, 13). Proving a benefit is becoming increasingly imperative given the complications of high-dose corticosteroids, including opportunistic infections and potentially prolonged severe steroid myopathy (2, 14). The report of Ho and coworkers (6) describes early high-dose (pulse) versus lower-dose (nonpulse) corticosteroids in SARS. These authors cite a study of lower-dose, late steroid treatment in ARDS by Meduri and coworkers as part of the justification for use of steroids in SARS (9). Early high-dose steroid treatment of ALI/ARDS was studied several years ago and failed to improve outcome (8). Ironically, though the latter trial of early high-dose corticosteroids showed neither effects on mortality, lung compliance, oxygenation, nor chest radiograph score, Ho and coworkers found early high-dose (pulse therapy) to be more effective than nonpulse intervention. The rapid changes in thinking about the best care for SARS patients is reflected in the lack of standardization of timing, dose, and duration of treatment with steroids in this report. The entry/exclusion criteria selected a patient population with a fairly low mortality (6%) and low ICU admission rate (8%), such that the study did not have statistical power to address effects on these major endpoints that define true clinical benefit. Despite the severe limitations inherent in a nonrandomized, retrospective study, this is an important report. Enough patients were exposed to steroids to allow one to conclude corticosteroid administration in these large doses, regardless of pulse or nonpulse issues, does not magically cure SARS, nor is there substantial short-term toxicity—at least with regard to pulse versus nonpulse administration. Lesser effects, positive or negative, cannot be ruled out. For example, Ho and coworkers showed pulse steroids given early in SARS may affect chest radiograph score (a very difficult endpoint to quantify) and requirement for oxygen (though there were imbalances already existent at baseline that may have confounded this analysis) more than nonpulse treatment. This study also provides very valuable information for the design of future studies, such as field testing inclusion/exclusion criteria and study endpoints, as well as injecting some increased confidence into the process of estimating the effect size that steroid treatment might provide to SARS patients, thus allowing more rigorous power calculations to determine prospective sample size. Ho and coworkers are wise to be circumspect in the conclusions they draw—namely, that this report helps guide SARS therapy but only until data from randomized trials become available (6). Twenty-five years ago (1978) Craddock wrote about the "hypercortisolism" of severe acute illness suggesting that the immunologic response to self-antigens exposed by disease or trauma may be suppressed by corticosteroids to offset the likelihood of autoimmune attack (15). This kind of thinking has provided support for the notion of corticosteroid supplementation as treatment for acute, critical illness, and additional support can be drawn from the report of Annane and coworkers (16). No doubt, with the next outbreak of SARS, studies of its pathophysiology will continue, as will well controlled clinical trials comparing treatments and combinations of treatments, including corticosteroids. Until those results are available, the medical community will have very limited information with which to determine whether steroids are an appropriate treatment for SARS, but will have some important decisions to make if SARS returns. Once SARS pathophysiology becomes better understood and we accumulate the results of clinical trials that are consistent and reproducible, we may have the answer to the corticosteroid question. If not, we may still be debating the use of steroids in SARS for another 25 years. FOOTNOTES Conflict of Interest Statement: G.R.B. has no declared conflict of interest. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
