The cytomegalovirus (CMV) DNA load was determined in polymorphonuclear leukocytes (PMNL) and plasma samples from 106 human immunodeficiency virus-infected subjects at risk of developing CMV disease (group 1) and from 27 AIDS patients with documented CMV disease (group 2). For both groups, the number of CMV copies in PMNL was significantly higher than in plasma when results were derived from an equivalent blood volume (P < .001, PMNL vs. plasma). Additionally, group 2 (symptomatic) patients had a greater viral DNA load than group 1 (asymptomatic) subjects (P < .001 for both PMNL and plasma). The sensitivity, specificity, and positive and negative predictive values of qualitative polymerase chain reaction using PMNL (PCR-PMNL) for the presence of CMV disease were 100%, 58%, 38%, and 100%, respectively, compared with 70%, 93%, 74%, and 92% for qualitative PCR-plasma and 93%, 92%, 76%, and 98% for quantitative PCR-PMNL using a cutoff of 16,000 copies/mL. Thus, the best strategy for diagnosing CMV disease in these individuals relies on quantitative assessment of the viral DNA load in PMNL.

The prevalence of the most common CMV UL97 mutations associated with ganciclovir resistance in PMNL of patients with AIDS treated for > or = 3 months (30.8%) appears to be higher than the rate of emergence of ganciclovir-resistant CMV isolates as previously reported using phenotypic assays (about 8%). Moreover, the detection of these mutations is associated with a considerable increase in the CMV DNA load in the blood as well as with progression of CMV retinitis during ganciclovir therapy.

It has been suggested that human cytomegalovirus (HCMV) glycoprotein B (gB) genotypes could be used as a marker for viral virulence in patients with AIDS. The present study was designed to evaluate a possible association between specific gB genotypes, the presence of HCMV retinitis, and the HCMV viral load. Fifty-four blood samples were obtained from 54 HIV- and HCMV-infected patients. Twenty-seven of these patients were asymptomatic for HCMV, whereas the other 27 patients had been diagnosed recently with HCMV retinitis. HCMV gB genotyping was carried out by using restriction enzyme analysis of PCR-amplified PMNL extracts. Determination of the HCMV viral load in the same specimens was carried out using a quantitative-PCR. HCMV gB genotype 2 was found more frequently than other genotypes in PCR-amplified polymorphonuclear leukocytes (PMNL) of patients with AIDS (P < 0.05) but not more frequently in samples from patients with HCMV retinitis. No significant association was found between any HCMV gB genotypes and the viral load in blood. In conclusion, the actual HCMV gB genotyping system using PMNL provides no additional benefit over the viral load in blood for identification of HIV-infected subjects at risk of HCMV disease.

The AMPLICOR cytomegalovirus (CMV) test, a new qualitative assay for the detection of CMV DNA in plasma, was compared to conventional methods and quantitative PCR (Q-PCR) assays by using leukocytes and plasma from 179 blood samples from subjects with AIDS. For the diagnosis of CMV disease, cell-based assays such as a Q-PCR with polymorphonuclear leukocytes (Q-PCR-PMNL) and a pp65 antigenemia assay had the highest sensitivities but suffered from a lack of specificity. The best agreement between the results of the Q-PCR-PMNL assay and those of the AMPLICOR test was found when a threshold diagnostic value of 690 copies per 105 cells was selected for the Q-PCR-PMNL assay. In that context, the AMPLICOR CMV test had a sensitivity of 96.4% and a specificity of 95.3% when results were compared to results of the cell-based PCR assay. This threshold was close to the one described as associated with the best sensitivity and specificity for the diagnosis of CMV disease in a recently published study (4). Blood samples that tested positive by the Q-PCR-PMNL assay but negative by the AMPLICOR CMV test were associated with viral loads (mean, 785 copies, median, 96 copies per 105leukocytes) lower than the viral loads of blood samples that tested positive by both assays (mean, 21,452 copies; median, 9,784 copies per 105 leukocytes) (P = 0.003). The AMPLICOR CMV test gave positive results at least 48 days before the development of symptomatic CMV disease in a longitudinal analysis of a limited subset of patients (n = 6) from whom sequential specimens were available for testing. In conclusion, the AMPLICOR CMV test is a very convenient assay combining rapidity, simplicity, and the possibility of batch testing. A positive result by this test seems particularly important since this implies, in most instances, the presence or the imminence of CMV disease, although a negative test result does not rule out disease.

The expression of a late cytomegalovirus (CMV) transcript (pp150) was sought in peripheral blood leukocytes (PBL) of subjects with AIDS and correlated with the amounts of CMV DNA in PBL and plasma, by means of quantitative polymerase chain reaction (PCR). The detection of the late CMV transcript was associated with a high number of CMV DNA copies in PBL (P=.0015) and with a positive CMV PCR assay in plasma (P<.001). Expression of CMV pp150 mRNA was best predicted by viral DNA thresholds corresponding to 7058 and 30 copies in PBL and plasma, respectively. The detection of CMV pp150 mRNA was associated with the presence of CMV disease in a univariate analysis but not in a multivariate analysis after controlling for the viral DNA load in PBL. Thus, active viral replication as determined by a high CMV DNA load in PBL is reflected by expression of the late CMV transcript in the same cells and by the presence of CMV DNA in plasma.