Community-acquired necrotizing pneumonia due to S. aureus- secreting PLV toxin is a highly lethal infection, affecting a young and healthy population group . The hallmarks are an influenza-like prodrome, leukopenia, rapid progression to septic shock, and respiratory distress, with multilobar necrosis and haemoptysis [5, 6, 14].
In this series, the mortality rate was 41%, which is lower than most of previously published rates [3, 5, 14, 15]. On multivariable analysis, influenza-like prodrome predicted fatal outcome. The true proportion of influenza infection is difficult to assess, because influenza testing is not routinely performed and rapid test sensitivity is only 50-70% . Influenza impedes phagocytic killing and damages the trachea-bronchial epithelium with subsequent impairment of secretion clearance and facilitated bacterial adhesion [42–44]. The influenza-like prodrome also may be caused by other respiratory viruses or by S. aureus itself.
We found a significant reduction of mortality for patients with skin and soft-tissue infection on admission. This result is new in the context of PVL-associated S. aureus necrotizing pneumonia, although one study mentioned a nonsignificant trend toward lower mortality for patients with a history of furuncles  and a recent retrospective study found a significant protective effect of a history of PVL-associated infections . Similar results also have been published in studies on S. aureus carriers. Approximately 20-30% of healthy persons are persistently colonized with S. aureus
[44, 46]. When hospitalized, these carriers have an increased risk of developing severe S. aureus infection caused by their colonizing strain , but mortality of S. aureus bacteremia is much lower in carriers than in noncarriers . The likely explanation for this protective effect is the stimulation of an immune response, which lowers the intensity of a subsequent invasive infection . A PVL vaccine has already been successfully tested on mice models and may find a human application in the near future .
Another issue raised by this study is the high rate of inadequate initial antibiotic treatment regimens. None of the 32 published cases received an antibiotic with an antitoxin effect as part of their first-line treatment, and all but three patients received beta-lactams. The use of beta-lactams as first-line treatment is controversial, because drug levels below the minimum inhibitory concentration may increase toxin release and stimulate toxin production [37, 38, 51]. The former effect is due to release of intracellular toxin secondary to cell wall lysis. In vivo, low drug levels in target tissues are a consequence of extensive tissue necrosis, leading to poor antibiotic diffusion and a sepsis-related increase of the volume of distribution. However, the stimulatory effect on toxin release is reversed when beta-lactams are given in association with clindamycin, linezolid, or rifampicin . The high rate of inadequate initial antibiotic treatment may be explained by the low prevalence of necrotizing S. aureus pneumonia, the low specificity of initial clinical signs and symptoms, and the adherence to treatment guidelines for community-acquired pneumonias. However, even after overt clinical suspicion or microbiological confirmation of PVL-secreting S. aureus, only 36% (5/14) of second-line antibiotics were adequate. Not surprisingly, the rate of adequacy was higher among more recent case reports. Since 2007, the Infectious Diseases Society of America (IDSA) recommends adding vancomycin or linezolid in case of severe pneumonia due to CA-MRSA . In the United Kingdom, the Health Protection Agency (HPA) recommends a combination of clindamycin, linezolid, and rifampicin but explicitly dissuades from the use of beta-lactams . Based on the discussed in vitro findings for beta-lactams, a recent recommendation by Gillet et al. suggests a third-generation cephalosporin with vancomycin and clindamycin or linezolid as first-line antibiotherapy. In the case of MSSA, vancomycin can be replaced by oxacillin .
Intravenous immunoglobulin may be an important adjunct to antibiotherapy. As illustrated in the above and in other case reports, it has been used successfully on a sporadic basis [30, 36, 56]. It was studied in vitro and was shown to neutralize PVL-induced pore formation and cytopathic effect . The HPA recommends intravenous immunoglobulin at a dose of 2 g/kg to be repeated after 48 hours if there is persistence of septic shock or failure to respond .
The significance of our study results is limited by its reliance on a relatively small number of case reports. Also, many variables, such as thrombocytopenia or kidney function, were reported infrequently and thus could not be included in the statistical analysis. We suspect that only the most severe cases of CAP caused by PVL-secreting MSSA are reported and that many cases are not detected, making it difficult to describe the full spectrum of clinical illness and to form meaningful conclusions based on the case reports. To improve our knowledge of the epidemiology, diagnosis, and treatment, there is a need to establish an international database.