The relationship between cytomegalovirus retrieved by bronchoalveolar lavage and mortality in patients with HIV.

STUDY OBJECTIVE

To evaluate mortality over 6 months of patients with HIV with cytomegalovirus (CMV) cultured from bronchoalveolar lavage (BAL) compared with those without CMV and to assess the significance of CMV cytologic study, CD4+ counts, and coexistent Pneumocystis carinii pneumonia.

DESIGN

Retrospective evaluation of HIV-infected patients undergoing bronchoscopy with BAL. The 40 most recent HIV-positive patients undergoing bronchoscopy with BAL were included for each of three categories: CMV by cytologic study; CMV by culture only; and CMV absent. Patients for whom survival status at 6 months was unknown were excluded from analysis.

SETTING

University hospital, tertiary care center.

PATIENTS

Group 1 consisted of 36 patients with positive CMV culture and cytologic study and group 2 consisted of 38 patients with only a positive culture for CMV. Group 3 consisted of 40 patients with no evidence of CMV by BAL.

RESULTS

On comparison of the groups, there was no difference in 3-week survival (from date of bronchoscopy). There was a statistically significant increase in mortality in group 1 patients compared with group 3 patients at both 3 and 6 months. Between groups 2 and 3, there was a difference in mortality that approached but did not reach significance at 3 months but did at 6 months. The mortality in group 1 at 3 months = 28%, at 6 months = 47%, whereas mortality in group 2 at 3 months = 26% and at 6 months = 45%. Group 3 had a 3-month mortality of 10% and a 6-month mortality of 15%. While those patients with positive CMV cytologic study had lower mean CD4+ counts, within the group, CD4+ counts were no different between the 3-month survivors and nonsurvivors (survivors, CD4/mm3 median = 38 [0 to 141]; and nonsurvivors, CD4/mm3 median = 16 [3 to 224]). Coinfection with P carinii did not increase mortality at 3 months.

CONCLUSIONS

The CMV retrieved by BAL in HIV-infected patients was associated with significantly greater 3- and 6-month mortality. The CMV cytologic study did not predict a higher mortality and the difference in mortality between patients with and without CMV in BAL fluid was not directly attributed to lower CD4+ counts or P carinii coinfection.