More Antibiotic Resistance Tied to Flu Season

Antibiotic resistance of common bacterial infections — both Gram positive and Gram negative — tracks seasonally with influenza rates, a study showed.

For each 20% increase in positive influenza tests at a center, every 1,000 admissions brought approximately 90 extra patients with macrolide-non-susceptible pneumoniae, 12 more with penicillin-non-susceptible Streptococcus pneumoniae, and seven more with S. pneumoniae not susceptible to extended-spectrum cephalosporins.

Among Gram-negative bacteria, the same 20% rise in flu rates was associated with about 1% more Enterobacterales and Pseudomonas aeruginosa not susceptible to fluoroquinolones and a 4% increase in Acinetobacter baumannii not susceptible to carbapenems, reported Vikas Gupta, PharmD, BCPS, of Becton Dickinson in Franklin Lakes, New Jersey, and colleagues.

“This result is somewhat surprising as Gram-negative bacteria are less commonly involved in influenza co-infections,” the authors wrote in Open Forum Infectious Disease.

Empiric antibiotics are frequently used for viral infections, since bacterial co-infections occur in 11% to 35% of flu patients, potentially resulting in antibiotic misuse.

The links between Gram-negative resistance and influenza rates “suggest that empiric antibiotic use associated with influenza season may have collateral effects on non-respiratory pathogens and add corroborative evidence that potential bacterial pathogens of public health interest may be important targets for infection prevention practices in a post-COVID-19 era,” the group added.

These resistant infections’ poorer outcomes and increased costs could have meaningful impacts on healthcare systems, they speculated. In practical terms, facilities should incorporate these findings into their antibiotic stewardship policies and programs to help guide more appropriate antimicrobial selection based on season and severity of influenza rates in the local community, Gupta’s group added.

The findings also lend support to strengthening flu shot campaigns, which should “help mitigate the inciting viral infection and thereby reduce secondary bacterial co-infections and antibiotic use,” they noted.

For their study, Gupta and colleagues examined antibiotic susceptibility profiles of 8,250,860 pathogens that were “30-day non-duplicate bacterial isolates” and 3,510,459 influenza tests from Becton Dickinson’s Insights Research database, which covers about 13% of all-cause hospital admissions across the U.S. Bacterial isolates were collected from 38,619,461 adults admitted at 257 acute healthcare centers from 2011 to 2019.

This ecological analysis expressed antibiotic resistance as the proportion of isolates deemed not susceptible (either intermediate or resistant), while inpatients’ resistance was expressed per 100 admissions. Analyses were stratified by pathogen type, isolate source (respiratory or non-respiratory) and setting (inpatient or ambulatory).

Facility-level influenza rates, evidenced by positive results for every 100 tests, were determined by PCR and antigen testing. Analysis of “year-quarters,” accounting for seasons, captured fluctuations in flu rates.

Enterobacterales (77%) and P. aeruginosa (9%) were most frequently identified. Methicillin-resistant Staphylococcus aureus (MRSA) accounted for half (50.3%) of S. aureus isolates.

Of all isolates, 1,502,796 were Gram-negative bacteria non-susceptible to fluoroquinolones, 498,012 were MRSA, 154,841 were Gram-negative bacteria non-susceptible to carbapenems, and 44,131 were S. pneumoniae not susceptible to one of the classes of antibiotics.

Among those pathogens isolated during hospital admissions, influenza rates significantly correlated with the proportion of Gram-negative isolates of Enterobacterales and P. aeruginosa not susceptible to fluoroquinolones (β coefficient, 0.041 and 0.039, respectively), as well as A. baumannii not susceptible to carbapenems (β coefficient 0.205).

Of all Gram-positive isolates, the strongest relationship was seen for S. pneumoniae not susceptible to macrolides (β coefficient, 0.464 overall; β coefficient, 0.253 for respiratory isolates, P<0.001 for both).

Also tied to influenza rates were S. pneumoniae not susceptible to penicillin (β coefficient 0.062, P=0.011) and not susceptible to extended-spectrum cephalosporins (β coefficient 0.033, P=0.036).

Multivariate analysis showed respiratory MRSA per 100 admissions was linked to influenza rates; whereas MRSA overall, accounting for non-respiratory cases as well, was not.

For Gram-positive isolates, greater antibiotic resistance rates were observed in respiratory versus non-respiratory sources. Just 5% of Gram-negative isolates involved a respiratory source.

Across all pathogens and antibiotics, antibiotic resistance rates were greater for inpatients versus outpatients.

The first quarter of the year showed the highest number of resistant isolates, although Enterobacterales not susceptible to carbapenem peaked in the fourth quarter.

The authors acknowledged several limitations to the data, including the lack of flu rates recorded beyond 2019, which does not account for trends once SARS-CoV-2 began circulating. Additional measures such as vaccinations and other respiratory viruses were not assessed, nor were pediatric isolates. They cautioned that the relationships observed do not imply causality.