A recent article in the Archives of Internal Medicine by Dr. D. Shefet et al who performed a systematic review and an editorial review of that article with comments in Courtlandt Forum might make you think there is no Legionairre's disease and therefore no need to empirically include antibiotic coverage for that possibility in patients hospitalized with community acquired pneumonia (CAP).The Archives article and the comments in Courtlandt are similar to a meta-analysis published in the BMJ in 2005 by Mills et. al. from New Zealand and comments in the Cleveland Clinic Journal that I wrote about last year.That MA focused on less than severely ill patients and found beta-lactams did as well as antibiotics that are effective against the atypicals except in the subgroup with Legionella infection.The authors concluded that beta-lactam should remain (that is or was the standard recommendation in UK) the antibiotics of choice in mild or moderately severe cases of CAP which is contrary to the guidelines of both the Infectious Disease Society of America (IDSA) and the American Thoracic Society.My reading of the Shefet article is that they did show an advantage to appropriate coverage in cases of Legionella although the conclusion does not give that impression nor does the review in Courtlandt. Further, the regimens which were compared were a beta-lactam and either a fluroquinolone or a macrolide given as single drugs.While the IDSA guidelines include the option of a fluroquinolone alone the alternative recommendation and one that is commonly used is the combination of a macrolide plus either of two specific cephalosporins ( ceftriaxone or cefotaxime).The Archives systematic review had no data on that issue. The results section said "these regimens[those with coverage for atypical pathogens] showed a significant advantage in clinical success for Legionella pneumophilia." The summary said " "no benefit of survival or clinical efficacy".
The section in the Courtlandt Review is called "evidence-based medicine"and purports to highlight "important new studies applicable to primary care".The section editor,Dr. Daniel Tobin, concluded that even though the systematic review may be underpowered to show a difference that "there does not appear to be a benefit to survival or clinical efficacy" to atypical coverage not quoting the results section comments regarding the clinical success advantage. He stops short of recommending not to give atypical coverage but this seems to be implied by his approach or at least could easily be interpreted in that way unless the original article is studied and the issue considered in light of "all of the evidence".
I consider this analysis typical of what I believe is an increasingly common "application" of what passes for EBM, namely emphasizing a meta-analysis as if it should be determinative and ignoring the context of the totality of evidence and other concerns regarding a particular issue.In this regard here is some of the other evidence that is necessary to consider in antibiotic management of CAP: 1)L.pneuomophilia is a recognized cause of serious and at times fatal cases of CAP 2)fluroquinolones and macrolides are effective treatment for L.pneumophilia infection.3)there is no test nor set of clinical findings that can exclude with reasonable certainty that L.pneumophilia is present at the time of clinical presentation of CAP.4)In seriously ill CAP patients antibiotics need to be given promptly and be correct (ie cover the usual suspects). From this set of evidentiary material the conclusion to cover atypicals (basically to cover L.pneumophilia as mycoplasma and chamydia are generally less potentially serious) follows. That- in part- is the evidence likely considered by the panels responsible for the current U.S. guidelines. In addition, macrolides and fluroquinolones are not particularly toxic or difficult to use medications and the practice of giving broader coverage initially and then focused therapy if and when a specific organism is identified continues to be sound practice.A systematic review for which the statistical power is suspect is not the type of trump card evidence that should overturn current practice.
Evidence based medicine should be based on the totality of the evidence and then blended with clinical judgment and in those regards there is more to factor in than simply randomized trials and systematic reviews of coarse grained outcome data.Systematic reviews should not be dismissed out of hand but neither should their results be determinative. They should be considered and given appropriate weight in the analytic process,