E-mail Abstract Author Session Search Abstracts Program


Session 115 Poster Presentations
Determinants of Disease Progression and Mortality
Session Day and Time: Tuesday 1:30 - 3:30 pm
Room: Hall B


909
Modelling the Variation of CD4 Cell Count Depending on Virologic Suppression in HIV-infected Patients Receiving Antiretroviral Therapy
I. Kousignian*1, S. Abgrall1, X. Duval2, V. Joly2, D. Descamps2, S. Matheron2, D. Costagliola1, The Clinical Epidemiology Group from the French Hospital Database on HIV3
1INSERM EMI 0214, Paris, France; 2Hosp Bichat-Claude Bernard, Paris, France; and 3France

Background: To study the durability of the virologic suppression with HAART, and to evaluate the influence of different levels of virologic rebounds on the dynamic of CD4 lymphocytes by use of a Markov process.
Methods: Between January 1997 and December 1999, 3,736 HIV-1 infected patients (pts), from the French Hospital HIV Database, who became undetectable (plasma viral load [VL] below 500 cp/ml) after having undergone HAART, were included. Four (4) levels of virologic suppression were defined according to the VL values during the first year after initial undetectability on HAART: 1) "Durable Virologic Suppression" (all VL < 500 cp/ml); 2) "Low Rebound Level" (at least 2 specimens of VL between 500 and 5,000 cp/ml); 3) "Intermediate Level" (at least 2 specimens of VL between 500 and 10,000 cp/ml with at least 1 between 5,000 and 10,000 cp/ml); and 4) "High Rebound Level" (at least 2 specimens of VL > 500 cp/ml with at least 1 > 10,000 cp/ml). A continuous time homogeneous Markov process with reversible states characterizing evolution of CD4 count was developed. For each specimen, pts were classified into 1 of 5 reversible states defined as follows: CD4 cells count between: 1) 0 and 199/mm3; 2) 200 and 349/mm3; 3) 350 and 499/mm3; 4) 500 and 649/mm3; and 5) CD4 count above 650/mm3. We used the estimates of transition intensities between states to evaluate the mean temporal evolution of the continuous predictive variation of the CD4 count in the 4 groups of virologic suppression.
Results: There was a continuous increase in CD4 count over time in every group. However, the lower the virologic rebound, the higher increase in CD4 count (p < 0.0001). At 2 months (mos), the increase of CD4 count levels were about 22, 18, and 19/mm3 in the groups 1, 2, and 3, respectively, vs 5/mm3 for group 4. The CD4 cell count increment between the 2nd and the 6th mos were 26, 20, 11, and 2/mm3 for the group 1, 2, 3, and 4, respectively. The increment decreased after the 6th mo, and the rate of this decrease differed significantly from one virologic group to another (p < 0.0001). For instance, between 6 mos and 1 yr 14/mm3 for group 1; about 7/mm3 for group 2 and 3; and 1/mm3 for group 4.
Conclusion: After a few mos, CD4 count ceased to increase only in the "High Rebound level" group. An immunologic reconstitution is thus possible with a low level of viral replication under potent antiretroviral combination therapies.