[Incidence of bacterial pneumonia in HIV-positive patients treated with preventive co-trimoxazole or pentamidine].

BACKGROUND

Besides Pneumocystis carinii bacterial pathogens represent the most common aetiology of pulmonary infections in HIV-positive patients. However, the impact of PCP prophylaxis on the incidence of bacterial pneumonia in HIV-positive patients using pentamidine or co-trimoxazole is still unknown.

PATIENTS AND METHODS

We analysed retrospectively the data of 80 consecutive HIV-positive patients with a CD4-cell count < 300/microliter. The total observation period was 1993 patient months. Type and duration of chemoprophylaxis, frequency of bacterial pneumonia, PCP, extrapulmonary bacterial infections and cerebral toxoplasmosis were documented in a standardised manner. For statistical analysis we used the Kaplan-Meier test for the time to a recurrence of the various infections under both prophylaxis regimens and the Odds ratio for determination of the relative risk.

RESULTS

We followed up 47 patients inhaling 300 mg pentamidine monthly for a total of 1133 months and 33 patients taking 480 mg co-trimoxazole per day p.o. for a total of 860 months. There were no statistically significant differences between the two groups in respect of demographic parameters, stage and therapy of HIV infection and distribution of risk groups. We found seven bacterial pneumonias in the co-trimoxazole group and 13 in the pentamidine group (not significant); the most common causative organisms were S. pneumoniae (n = 4), S. aureus (n = 3) and H. influenzae (n = 3). Furthermore, in the pentamidine group 12 PCP and nine cases of toxoplasma encephalitis were observed, whereas none of these infections occurred in the co-trimoxazole group (p < 0.05). Two of the patients taking co-trimoxazole and 15 of those inhaling pentamidine had extrapulmonary bacterial infections (p < 0.05), the most frequently identified pathogen being S. aureus (n = 7). The two prophylaxis groups did not differ significantly with regard to laboratory data, course and therapy of the bacterial pneumonias.

CONCLUSION

There was no significant influence of chemoprophylaxis on the incidence of bacterial pneumonia in patients with advanced HIV-disease in our study. Since S. pneumoniae represents the most common causative agent, we suggest immunisation with a polyvalent pneumococcal vaccine at an early stage of HIV-infection.