Bacterial pneumonia

August 2005     View PDF     En español
Reprinted from www.aidsmap.com, Great Britain

Key research into bacterial pneumonia

Grau studied 142 cases of invasive pneumococcal disease caused by Streptococcus pneumonia in HIV-infected people between 1986 and 2002, comparing the pre and post-HAART eras (1986-1996 and 1997-2002). Overall incidence fell from 24 to 8 cases per 1000 patients. In the later period, bacterial pneumonia was associated with advanced disease and severe co-morbidities such as cirrhosis, and was associated with poorer outcomes. In the HAART era, a quarter of patients died within 30 days compared to 8% in the pre-HAART era.

Sullivan assessed the incidence of and risk factors for bacterial pneumonia in 1,898 HIV-infected patients with CD4 cell counts below 200 who attended the Johns Hopkins HIV Clinic between 1993 and 1998. 352 cases of bacterial pneumonia were reported during 2,310 patient-years of follow-up. Incidence of bacterial pneumonia was 22.7 episodes/100 person-years in early 1993, 12.3 episodes/100 patient-years in early 1996, and 9.1 episodes/100 patient-years in late 1997 (p<0.05). The use of protease inhibitor-containing regimens was associated with a decreased risk of bacterial pneumonia. Lower CD4 cell count, injecting drug use and prior PCP were associated with a greater risk of bacterial pneumonia.

Greenberg reported that in a review of 2983 patient records in New York, bacterial pneumonia was more common among injecting drug users than gay men. Non-Haitian black people were more likely to develop Salmonella sepsis that white people. There are no differences in bacterial infections between men and women.

French conducted a randomised, placebo-controlled trial of the 23-valent pneumococcal polysaccharide vaccine. 1323 HIV-infected adults were randomised, with final analysis of 1158 individuals. There was no significant difference in the number of first events of invasive pneumococcal disease and all pneumonia between the two groups. Cases of pneumonia were significantly more common in the vaccine group. Forty new cases (56.9 per 1000 person years) of all-cause pneumonia occurred in the vaccine group compared to 21 cases (30 per 1000 person years) in the placebo group (0.02). Adjusted first event analysis found this difference highly statistically significant (p<0.008) while the all event, non-adjusted figure was also significant (p<0.03).

Bekele found that 9% of nearly 600 HIV admis-sions between 1995-1998 were due to bacterial pneumonia. Median CD4 count was 38 versus 66 cells/mm3 for all other admission groups. Length of hospital stay was longer, and intensive care admission and case-fatality rates were higher. The most common bacteria was Pseudomonas aeruginosa (32 admissions), followed by Strepto-coccus pneumoniae (22 admissions), Staphylo-coccus aureus (16 admissions), and Haemophilus influenzae (11 admissions). Thirty-three of the pneumonias were bacteraemic, which was more common in pneumococcal than in pseudomonal pneumonia groups.

Navin reviewed 211 HIV-infected cases of pneumonia between 1994-1996, comparing them with age and CD4-matched individuals hospitalised for other reasons. Bacterial pneumonia was associated with breathing chemical irritants such as insect spray, petrol or paint fumes in the previous month or hospitalisation for pneumonia within the last 6 months on multivariate analysis.

Chaisson (1997) reported that in a group of 2,888 patients prospectively followed between 1989 and 1996, 14% of drug users developed bacterial pneumonia versus 11% of non-drug users. 14% of blacks developed bacterial pneumonia versus 9% of whites. People with lower CD4 counts were at greater risk. People receiving co-trimoxazole did not appear to be at reduced risk, and a history of PCP increased the relative risk of developing bacterial pneumonia, suggesting that the bacteria might have acquired resistant to co-trimoxazole. There was no decrease in bacterial pneumonia during 1996, even though other opportunistic infections were becoming less common due to the use of protease inhibitors.

Garcia-Leoni conducted a retrospective (13-month) and prospective (14-month) study of 106 adult hospitalized patients with pneumococcal pneumonia, 22% of whom were HIV-positive. The estimated attack rate was 5.9 per 1000 for HIV-infected patients and 0.31 per 1000 for HIV-seronegative patients. Pneumococcal pneumonia was the first manifestation of HIV infection in 48% of cases. Seventy-two percent of patients younger than 40 years of age with pneumococcal pneumonia were HIV-infected.

Steinhoff compared H. influenzæ type b poly-saccharide (PRP) vaccine to the polysaccharide-mutant diphtheria toxoid conjugate vaccine (PRP-CRM). In asymptomatic and early sympto-matic HIV-positive men (and in HIV-negative men), the PRP-CRM vaccine caused a threefold greater antibody response than the PRP vaccine. However, in men with AIDS, the PRP vaccine caused a greater response.

Glaser reported that AZT improves response to pneumococcal vaccine among people with AIDS or ARC.

Schuster reported on 16 people with Pseudomonas Aeruginosa pneumonia. Most had AIDS with a mean CD4 count of 27. Traditional risk factors were often missing. Cavitary infiltrates were present on admission chest radiograph in 50% of cases. In-hospital mortality was 19%, and an additional 25% developed chronic or recurrent disease.

Hirschtick reported that the risk of bacterial pneumonia in HIV-positive people increases sharply when CD4 counts fall below 200, particularly among smokers. Injecting drug users were twice as like to develop bacterial pneumonia than gay men with the same CD4 count. Taking co-trimoxazole for PCP prophylaxis appears to reduce the risk of bacterial pneumonia by about a third.

Bacterial infections of the lungs

LAST      NEXT