The Etiology of CAP Remains Elusive – A Call to Change Antibiotic Prescribing Patterns

CAP slider 2

Once upon a time – not so long ago – we were taught the following:

  • Streptococcus pneumoniae was the cause of 90% of CAP cases. That statement was later modified to mean S. pneumoniae was the underlying pathogen in 90% of bacterial cases of CAP, i.e., those with positive cultures from sputum or blood;
  • S. pneumoniae was the “captain of the men of death” and that pneumonia caused by the organism was associated with a high mortality.

It was the time when students, residents and fellows had to check a sputum Gram stain of every CAP patient to see whether an encapsulated diplococcus was present.  Nobody quite understood why this had to be done at 3 o’clock in the morning when the patient ended up on penicillin anyways but it was taught and defended as good practice. Occasionally you were rewarded with an unequivocal specimen giving away the diagnosis.

Now we learn from a large study of CAP patients requiring hospitalization that S. pneumoniae is a rare pathogen being confronted in US hospitals[1].  Despite state-of-the-art diagnostics which included not only traditional Gram stain and cultures but also serology, urinary antigen testing and PCR for a long list of respiratory viruses, S. pneumoniae was positively identified in only 5% (sic!) of cases.  

Only 115 of 2259 hospitalized patients (5%) had a bona fide pneumococcal infection
JAIN et al.  NEJM 373; 5: 415

This is a surprisingly low percentage for which the authors invoke the ‘herd immunity’ effect of successful childhood immunization with polyvalent vaccine.  Prior antibiotic use, a figure not provided in the article, may have resulted in culture-negativity but should not have affected urinary antigen detection.

Perhaps less surprisingly, a bacterial etiology could be proven in only 14% of cases of which only 11% were purely bacterial in nature. Despite exhaustive testing, no pathogen (viral, bacterial, fungal or chlamydial) could be identified in 62% of CAP patients.

It seems that we need to revise our approach to CAP.  We need to think inflammation, not infection; viral, not bacterial, in the overwhelming majority of cases.

Is there really a need to treat every CAP patient with antibiotics when the overall mortality is as low as 2% even in hospitalized patients?

 Would watchful watching be a better approach than the current standard which is combination antibiotic therapy in accordance with 2007 ATS/IDSA CAP treatment guidelines [2] ?  While not mentioned in the publication, we can safely assume that in the cohort of 2259 patients every single one received systemic antibiotics for the duration of hospitalization.

Are we giving antibiotics to CAP patients, because we don’t trust negative cultures and feel they can exclude a bacterial etiology?  No, it is more likely that US doctors practice ‘defensive medicine knowing that the lawyers will have a heyday if they withhold ‘life-saving’ drugs, or don’t follow ‘state-of-the-art’ practices.

Before we invoke ‘antibiotic stewardship’ and talk about ‘inappropriate use’ of antibiotics, we need to acknowledge the ‘elefant in the room’: Even in 2015, our ability to diagnose the different etiologies of what presents as ‘pneumonitis’ (i.e., new infiltrate on CXR plus fever and leukocytosis) is abysmally poor.  This is quite embarrassing to admit.

Maybe it is time for a radical break with traditional approaches.  Instead of working on new guidelines, we need a fundamental understanding of the disease and improved diagnostics.  What we are doing therapeutically just does not make sense anymore.

Recently, the need for macrolide therapy in CAP has been challenged in a widely quoted study[3].  How about the approach proffered by FDA a few years ago to demand placebo-controlled trials in CAP (or CABP)?  It did not get much traction at the time but a few more studies confirming the Jain and Postma findings may trigger a sea change.

Somehow one would believe that we should not wait for the FDA to tell us that “the science has changed”.

[1] S Jain. N Engl J Med 2015; 373: 415
[2] Mandell L Clin Infect Dis. 2007;44: S27
[3] D Postma. N Engl J Med 2015; 372:1312

Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published. Required fields are marked *