– A substantial failure rate of antibiotic monotherapy was found in patients with community acquired pneumonia (CAP), according to a presentation given at an international conference of the American Thoracic Society.

In a study of 413,801 patient records with confirmed CAP, an average of 25% of patients reported treatment failure, according to James A McKinnell, MD, an infectious disease specialist at LA BioMed and an assistant professor at the University of California, Los Angeles.

Adult outpatient records with a diagnosis of CAP and a prescription for antibiotics were gathered from the period of 2012-2015, with treatment failure defined as a refill or change in the medication prescribed, a visit to the emergency department, or a hospitalization, according to Dr. McKinnell and the other investigators.

When broken down, the failure rates in patients given beta-lactams (25.7%), macrolides (22.9%), tetracycline (22.5%), and fluoroquinolones (20.8%), were all found to increase when patients’ Charlson Comorbidity Index (CCI) score increased (odds ration [OR] = 1.16 [1.13-1.20] for CCI = 1, OR = 1.22 [1.18-1.26], for CCI = 2, OR = 1.44 [1.39-1.49], for CCI greater than or equal to 3).

These medications have been shown to be effective through the usual array of controlled tests. While these trials do confirm overall efficacy, they are not always accurate in predicting how they will affect individual patients, Dr. McKinnell noted.

“I want to know the best drug for my patient, [and] unfortunately randomized clinical trials are not completely generalizable,” Dr. McKinnell said during his presentation. “Pathogen distribution and resistance is different in a clinical trial compared to the patients we see, and there’s a measuring bias, so there’s a lot of limitations when just using clinical trials.”

When analyzing failure endpoints, the investigators found 79%, 73.4%, 80.8%, and 64% of patients switched their antibiotics while taking beta-lactams, macrolides, tetracycline, or fluoroquinolones, respectively. The investigators interpreted this as a sign that patient treatment plans must be better fitted for their personal circumstances.

This is where the idea of “big data” would apply; using large-scale, “real-world” data of current and previous CAP patients could be instrumental to test the benefits and limitations of certain treatment options on patients with certain comorbidities, according to Dr. McKinnell and his fellow investigators.

When breaking down comorbidities among patients, the investigators found that many of the comorbid conditions had a “significant predictor value” of treatment failure, according to Dr. McKinnell.

Investigators were not surprised that hemiplegia or paraplegia, which increased the odds of antibiotic failure by 33%, were independent factors; however, comorbidities such as peptic ulcer disease (OR: 1.15) was less expected, Dr. McKinnell noted.

When looking at the mortality rate of patients 18 years of age and older with treatment failure, 18.1% (10,087) died (P less than .0001), with an even higher mortality rate of 24.3% (3,299) among those at least 65 years of age, he said.

If big data studies could decrease the number of treatment failures, the implications would be significant in decreasing the number of mortalities, the investigators noted.

“Prescribers should be aware of those CAP patients most at risk for poor outcomes and consider these factors to guide a comprehensive treatment plan,” said Dr. McKinnell.

Cempra Pharmaceuticals funded the study. The researchers did not report any conflicts of interest during their presentation.

On Twitter @eaztweets