CMV shedding and the risk of end-organ diseasez

ACTG 181 was natural history study which was nested within ACTG081, a RCT of 3 PCP prophylactic drugs. Patients were monitored every 4 weeks for evidence of CMV shedding in the urine or end-organ disease, and every 12 weeks for evidence of CMV in their blood. A Cox model was fit for the time until first shedding. Time was measured from the date a first sample was taken, until the date of 1st indication of positive shedding, or most recent follow-up (censored) if the subject did not shed. CD4 count was used as a time-varying covariate. A Cox model was also fit for time until CMV diagnosis. In this analysis, both CD4 count and shedding status (whether a patient ever shed) were used as time-varying covariates. Results: 181 evaluable patients, 69 (39%) shed CMV in the urine and 29 (16%) shed CMV in the blood during the study. The one year actuarial rate of shedding in the urine and blood was 32% and 9% respectively. Urine shedding was intermittent (at least one negative after a positive result) in 68% of patients while blood shedding was intermittent in only 43% of cases. 33 (18%) patients had a diagnosis of end-organ disease while on the study. In models that adjusted for CD4 count, either shedding in the blood or urine were independent predictors for risk of CMV disease (blood shedding RR=9.2, p = .0001; urine shedding RR=3.4, p = .01). The 3 and 6 month rate of CMV diagnosis following CMV shedding (in the blood or urine) was 5% and 16% respectively. Conclusions: Shedding and intermittent shedding of CMV is common in HIV and is associated with an increased risk of near-term end-organ disease.