Clinical Citations: Predicting antibiotic resistance in invasive pneumococcal infections

September 1, 2005
Volume 5, Issue 9

How do you assess the likelihood of antibiotic resistance in patients presenting with suspected pneumococcal infection? Vanderkooi and associates addressed this in a study of 3339 patients with invasive pneumococcal disease. They found that antibiotic use in the 3 months before the onset of infection was the most important risk factor for antibiotic resistance. The single most important risk factor for resistance to a particular antibiotic was previous use of antibiotics from the same class.

How do you assess the likelihood of antibiotic resistance in patients presenting with suspected pneumococcal infection? Vanderkooi and associates addressed this in a study of 3339 patients with invasive pneumococcal disease. They found that antibiotic use in the 3 months before the onset of infection was the most important risk factor for antibiotic resistance. The single most important risk factor for resistance to a particular antibiotic was previous use of antibiotics from the same class.

Over the 8-year study period, the prevalence of penicillin resistance among pneumococcal isolates increased from 0.91% to 6.23%, ceftriaxone resistance increased from 0% to 1.78%, erythromycin resistance increased from 4.6% to 13.1%, and levofloxacin resistance increased from 0.3% to 1.2%. There was no change in the prevalence of resistance to trimethoprim-sulfamethoxazole (TMP-SMX).

Patients who had previously received macrolides, TMP-SMX, or fluoroquinolones were at least 4 times as likely to have pneumococcal isolates that were resistant to agents from the same antimicrobial class as were patients who had not received such agents. The macrolides were not homogeneous with respect to the development of resistance: compared with erythromycin and clarithromycin, azithromycin preferentially selected for resistance.

Age and underlying illness did not appear to be significant predictors of resistance. The risk factors for resistance that were identified by multivariate analysis are shown in the Table.

The authors concluded that knowledge of previous antibiotic use is essential in determining appropriate therapy for suspected pneumococcal infection.

In an accompanying editorial, Pantosti and Moro caution against generalizing these findings to other areas. The study was conducted in Toronto, where the prevalence of antibiotic resistance among patients with invasive pneumococcal infections is moderate. The prevalence is higher in parts of the United States and is quite variable in Europe.

Pantosti and Moro emphasize the importance of the observation that the macrolides differ with respect to resistance. They also say that the emergence and spread of fluoroquinolone-resistant pneumococci in nursing homes should be closely studied.