Are macrolides now obligatory in severe community-acquired pneumonia?

Abstract

In this issue of Intensive Care Medicine, Martin-Loechesand colleagues [1] present an analysis of a cohort of 218patients with community-acquired pneumonia requiringmechanical ventilation enrolled into a larger, observa-tional, multi-center European study. The major finding ofthe study in this issue was that the 42 patients whoreceived a macrolide antibiotic had half the mortality rateof other patients after adjusting for severity of illness atpresentation.That macrolide antibiotics appear to confer a signifi-cant survival advantage in patients with severecommunity-acquired pneumonia is not a new concept,with multiple observational and retrospective studiesdemonstrating substantial mortality benefits [2–7]. How-ever, as has been pointed out in many editorials andreviews, none of these studies are prospective, random-ized controlled trials. In the absence of scientificallyirrefutable evidence, at what point does the weight of datain favor of using macrolides become so overwhelmingthat their use is obligatory?If we look at the potential downsides of making mac-rolides obligatory, then there is an obvious economic costif they are not needed. However, relative to most costs inhealth care, the economic burden is trivial. Unnecessarymacrolide use could perceivably contribute to increasedantibiotic resistance to this class of antibiotics in thecommunity, but the reality is that patients hospitalizedwith community-acquired pneumonia account for a min-ute portion of total antibiotic use, and this is not asustainable argument given the already widespread use ofthis class of agents in the outpatient setting for upper andlower respiratory tract infections. Overuse of macrolides,like any antibiotic, could theoretically lead to selectionfor multi-resistant pathogens. However, the risk of thisseems to be smaller than for broad-spectrum beta-lactams,third-generation cephalosporins and fluroquinolones,which all have well-documented track records of thisadverse side effect. As with all antibiotics, drug reactionscan occur, but macrolides are generally a very safe classof antibiotics. An increased incidence of arrhythmias hasbeen reported with macrolides because of prolongation ofthe Q-T interval, but overall the risk is no greater than thatassociated with fluroquinolones [8].If there is no major downside to adding a macrolide,the next question is whether these are the best agents orwhether other antibiotics or antibiotic combinations havean equivalent or greater beneficial effect. One of theproposed (and I think the least likely) potential explana-tions for the benefit of macrolides is coveringunrecognized ‘atypical’ pathogens (such as Legionellaspp. or Mycoplasma). If this were the mechanism, thenthere should be equivalent benefit from fluroquinolonesand tetracyclines. However, just as previous observationalstudies have shown [9, 10], the current study by Martin-Loeches and colleagues [1] also clearly demonstrates thatfluroquinolones do not give the same apparent protectiveeffect as macrolides. Although much more limited, thereare also some data suggesting that tetracyclines are alsonot as efficacious as macrolides [10].That the benefit of macrolides is almost certainly notdriven by undiagnosed atypical pathogens invalidates the