By Carol A. Kemper, MD, FACP

Clinical Associate Professor of Medicine, Stanford University, Division of Infectious Diseases, Santa Clara Valley Medical Center

SOURCE: Haessler S, Lindenauer PK, Zilberberg MD, et al. Blood cultures versus respiratory cultures: 2 different views of pneumonia. Clin Infect Dis 2020;71:1604-1612.

To examine whether current guidelines for the treatment of pneumonia remain appropriate, researchers conducted a large multicenter study of adults with pneumonia admitted to 177 U.S. hospitals between 2010 and 2015. Patients admitted with a principal diagnosis of pneumonia, respiratory failure, acute respiratory distress syndrome, or sepsis with a secondary diagnosis of pneumonia, and who also had blood and/or respiratory cultures obtained on admission were included in the analysis. A total of 138,561 hospitalizations met criteria, of which 68% were considered community-acquired pneumonia (CAP) and 32% were deemed healthcare-associated pneumonia (HCAP).

Blood cultures were obtained on admission in 99% of hospitalizations, and respiratory cultures were obtained in 18%. Positive cultures were infrequent. Only 9.3% of all admissions tested positive, including 4.6% with a positive respiratory culture alone, 4.3% with a positive blood culture alone, and 0.3% with both positive respiratory and blood cultures. In those able to produce a sputum specimen, respiratory cultures were positive in 28%, and patients with HCAP were more likely than those with CAP to produce a positive sputum culture (33% vs. 25.4%; P < 0.001). Of all blood cultures obtained, only 4.7% were positive.

Among those with positive blood cultures alone, Streptococcus pneumonia (33%) and Staphylococcus aureus (22%) were the most common organisms isolated, followed by Escherichia coli (11.8%), Klebsiella spp. (4.6%), Pseudomonas aeruginosa (3.5%), group B strep (2.7%), Haemophilus influenzae (2%), and Proteus mirabilis (1.6%). More than one-third of S. aureus bacteremias were methicillin-resistant (36%). In contrast, in those with only positive respiratory cultures, S. aureus (33.6%) and Pseudomonas aeruginosa (17%) were the most common isolates. In those with both positive blood and respiratory cultures, S. aureus was more common (44.5%; 41% were methicillin-resistant), followed by S. pneumoniae (32%) and Pseudomonas aeruginosa (7.7%).

The prevalence of resistance to recommended first-line CAP antibiotics (i.e., ceftriaxone plus azithromycin or a respiratory quinolone) was assessed by organism and by culture site. A total of 209 patients were excluded because their organisms lacked clear Clinical & Laboratory Standards Institute (CLSI) breakpoints. Overall, 42% of admissions with a positive culture grew an organism resistant to first-line therapy for CAP, including 27% of those with positive blood cultures. Gram-negative organisms isolated in either blood or respiratory cultures were more likely to be resistant to CAP therapy than gram-positives (51.8% vs. 35.4%). Patients with only positive respiratory cultures were twice as likely to yield organisms resistant to CAP therapy, but their outcomes were better, suggesting some organisms represented colonizers rather than true pathogens.

Although two-thirds of patients in this study were considered to have CAP and one-third HCAP, empirical antibiotic therapy administered at the time of admission did not necessarily reflect these designations. For those with only positive respiratory, only positive blood, or both positive respiratory and blood cultures, anti-methicillin-resistant S. aureus (MRSA) antibiotics were administered to 42%, 48%, and 66%, respectively (P < 0.001). Similarly, HCAP-guideline antibiotics were administered in 11.8%, 15.7%, and 27%, respectively, and four or more antibiotics were administered in 17.5%, 21%, and 33%, respectively. This suggests providers were cognizant of the severity of disease at presentation and the risk of MRSA and multidrug-resistant organisms (MDRO) in some patients.

Despite these efforts, patients with both positive blood and sputum cultures generally exhibited more acute and chronic illness, with significantly higher case fatality rates (25%) than those patients with only positive blood (12%) or respiratory cultures (11%). Also, they recorded significantly longer lengths of stay.


Predicting the bacterial etiology of pneumonia on presentation to the hospital, when empirical antibiotic therapy must be chosen, is challenging. The choice depends on many factors, including acuity of the presentation, chronicity of underlying disease, recent residence in long-term care, and the anticipated flora. Not mentioned in this article is the benefit of “flagging” those patients with recognized MDROs from prior cultures in an electronic system, as well as the use of nares MRSA polymerase chain reaction to identify those patients at risk for MRSA pneumonia. These data suggest CAP therapy may no longer be relevant for many patients with CAP, and the required use of the current CAP bundle with limited antibacterial therapy choices should be re-assessed.