Is the CD4 Cell Percentage a Better Marker of Immunosuppression than the Absolute CD4 Cell Count in HIV-Infected Patients with Cirrhosis?

McGovern et al. [1] recently showed that cirrhotic persons without HIV infection had low CD4 cell counts and that, despite this, the majority had normal CD4 cell percentages, probably secondary to portal hypertension and splenic sequestration of CD4 cells. In an editorial, Gandhi [2] hypothesized that these characteristics might also be found in HIV-infected cirrhotic patients and that, as a consequence, the absolute CD4 cell count may be less accurate than the CD4 cell percentage for assessing the degree of immunosuppression. Ghandi [2] also suggested that, if this hypothesis was true, many cirrhotic individuals could start receiving antiretroviral therapy (ART) when their absolute CD4 cell count is lower than the count currently recommended by treatment guidelines [3, 4], because their CD4 cell percentage would be normal. The basis for this potential change in recommendations should be carefully investigated, because there are clinical arguments against delaying therapy initiation for HIV-infected persons with cirrhosis; such patients are more susceptible than noncirrhotic HIV-infected persons to coinfection as a result of changes in gut motility, mucosal defenses, and microflora that allow for translocation of enteric bacteria into mesenteric lymph nodes and the bloodstream [5].

Ghandi [2] noted the lack of studies that address the relationships between the CD4 cell count and percentage and the incidence of AIDS-defining illnesses. In this study, we evaluated the Italian Cohort of Antiretroviral-Naive Patients (ICoNA) to compare the value of the absolute CD4 cell count with that of the CD4 cell percentage for predicting the risk of developing an AIDS-defining illness in persons with cirrhosis.

Patients and methods. To assess the hypothesis that the CD4 cell percentage is more reliable than the absolute CD4 cell count as a marker of the degree of immunosuppression in HIV-infected persons with cirrhosis, we estimated the incidence of AIDS-defining illnesses among patients in ICoNA with cirrhosis or other liver disease and those without liver disease. All patients were ART naive at enrollment. A patient was classified as having cirrhosis if a histological or clinical diagnosis of cirrhosis was reported for that individual in the ICoNA database.

Three study groups were identified at enrollment on the basis of the following criteria: current hepatitis C virus (HCV) and hepatitis B virus (HBV) serological status, results of clinical evaluation, and alanine aminotransferase (ALT) concentration. Group 1 comprised patients without liver disease (defined as HCV and HBV negative, with an ALT level of <40 IU/L), group 2 comprised patients coinfected with HCV (defined as detection of antibodies against HCV) and/or HBV (defined as detection of HBV surface antigens) who did not have a clinical or histological diagnosis of cirrhosis, and group 3 comprised patients with histological and/or clinical evidence of cirrhosis. Because we do not systematically screen individuals in ICoNA for cirrhosis, it is possible that some patients with compensated cirrhosis might have been included in group 2.

The most recent CD4 cell counts and percentages were included in the analysis. Incidences of AIDS-defining illness are reported irrespective of whether patients were currently receiving ART, although these findings were similar to estimates of the rate of AIDS-defining illness while patients were treatment naive.

Results. Our study population included 6126 subjects; 1838 (30%) were women, and 5679 (92.7%) were white. All patients had data on CD4 cell counts, and 5917 patients had data on CD4 cell percentages. Rates of risk factors for HIV infection were as follows: injection drug use, 38% (2328 subjects); heterosexual sex, 36.1% (2211); and male-male sex, 19.8% (1213). At enrollment, the median CD4 cell count was 427 cells/mm³ (range, 1–1361 cells/mm³), the median CD4 cell percentage was 23% (range, 1%–81%), the median viral load was 4.3 log₁₀ copies/mL (range, 1.3–6.8 log₁₀ copies/mL), 38% of patients were coinfected with HCV, and 5% were coinfected with HBV. Four percent of patients had an AIDS-defining illness at enrollment.

At enrollment, group 1 had 3470 patients, group 2 had 2502 patients, and group 3 had 154 patients (14 of whom had a histological diagnosis of cirrhosis). During the follow-up period, 85 additional patients received a diagnosis of cirrhosis (16 diagnoses were made on the basis of histological findings). Eleven (1.3%) of 839 patients in group 1, 8 (1.5%) of 529 in group 2, and 1 (2.4%) of 42 in group 3 had a CD4 cell count of <200 cells/mm³ and a CD4 cell percentage of >28% at enrollment (P = .59, by the Fisher exact test).

A total of 402 new cases of AIDS-defining illness were reported in the ICoNA database during 25,693 person-years of follow-up, calculated as the interval between the date of ICoNA enrollment and the date of the first new AIDS-defining illness (such as an opportunistic infection) or the final clinical follow-up examination, whichever occurred first. The overall incidence of AIDS-defining illness was 1.6 cases per 100 person-years of follow-up (95% CI, 1.4–1.7 cases per 100 person-years of follow-up). In each study group, the estimated rate of AIDS-defining illness among subjects with a low CD4 cell count (i.e., ⩽200 cells/mm³) differed from that among subjects with a high CD4 cell count (i.e., >500 cells/mm³); the greatest difference was detected in group 1, in which the P value for interaction was .001 (table 1). Similarly, in each group the risk of developing an AIDS-defining illness for patients with a low CD4 cell percentage (i.e., <14%) differed from that for patients with a normal CD4 cell percentage (i.e., >28%); however, in contrast to findings from the analysis of the CD4 cell count, subjects in group 3 (i.e., those without liver diseases) had the greatest difference in risk, with a P value for interaction of .03 (table 1).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Rates of new AIDS-defining illness among subjects in the Italian Cohort of Antiretroviral-Naive Patients, according to CD4 cell percentage and absolute CD4 cell count.

CD4 cell counts and CD4 cell percentages were also analyzed as continuous variables. For group 1, higher CD4 cell counts (relative risk [RR], 0.58; 95% CI, 0.39–0.85; P = .006 per 100 cells/mm³ increase) but not higher CD4 cell percentages (RR, 0.61; 95% CI, 0.30–1.24; P = .17 per 10% increase) were significantly associated with a lower risk of AIDS-defining illness; however, the statistical power of these analyses was low. A low CD4 cell count was significantly associated with a higher risk of AIDS-defining illness in group 2 (RR, 0.75; 95% CI, 0.67–0.83; P = .001 per 100 cells/mm³ increase) and group 3 (RR, 0.52; 95% CI, 0.42–0.63; P = .001 per 100 cells/mm³ increase). A low CD4 cell percentage was also significantly associated with a higher risk of AIDS-defining illness in both groups, although the CD4 cell percentage seemed to be slightly less predictive than the absolute CD4 cell count in group 3 (RR, 0.61; 95% CI, 0.43–0.87; P = .006 per 10% increase), compared with the predictive value of this variable in group 2 (RR, 0.54; 95% CI, 0.43–0.67; P = .001 per 10% increase). After adjusting for current degree of immunosuppression, persons in group 2 (adjusted RR, 1.22; 95% CI, 0.98–1.52; P = .07) and persons in group 3 (adjusted RR, 1.81; 95% CI, 1.16–2.82; P = .009) had an increased risk of AIDS-defining illness, compared with persons in group 1.

Discussion. These results from a population representative of patients treated in clinical practice do not provide evidence to support the hypothesis of Gandhi [2], according to which CD4 cell percentage should be used to guide therapy for HIV-infected patients with cirrhosis. Indeed, we found that, for each study group, the CD4 cell count was a better predictor of the risk of developing an AIDS-defining illness than the CD4 cell percentage. Consistent with the findings of McGovern et al. [1], we found that the proportion of patients with a normal CD4 cell percentage was slightly higher in group 1 than in group 2 and group 3. However, this difference was small and not statistically significant.

Our results might be limited by some factors, such as the accuracy of the clinical diagnosis of hepatic disease and the possible lack of power of the analyses because of the small number of subjects for whom cirrhosis was diagnosed. Furthermore, HCV RNA was available for analysis from only a minority of patients, and therefore, it is likely that we incorrectly classified aviremic HCV antibody—positive patients as having HIV-HCV coinfection. Although use of ART could be a confounder in our analysis, results were similar when we restricted the analysis to patients who were ART naive (data not shown).

In conclusion, our data confirm that HIV-infected patients with cirrhosis are at higher risk of developing an AIDS-defining illness than are HIV-infected patients without cirrhosis. These findings suggest that the absolute CD4 cell count should be used to guide therapy decisions for all HIV-infected patients.

Members of the study group. The Italian Cohort of Antiretroviral-Naive Patients study group includes the following individuals (all cities are in Italy, unless otherwise indicated): M. Montroni, G. Scalise, M. C. Braschi, and A. Riva (Ancona); U. Tirelli and G. Di Gennaro (Aviano [PN]); G. Pastore, N. Ladisa, and G. Minafra (Bari); F. Suter and C. Arici (Bergamo); F. Chiodo, V. Colangeli, C. Fiorini, and O. Coronado (Bologna); G. Carosi, G. Cristini, C. Torti, C. Minardi, and D. Bertelli (Brescia); T. Quirino and S. Melzi (Busto Arsizio); P. E. Manconi and P. Piano (Cagliari); L. Cosco and A. Scerbo (Catanzaro); E. Pizzigallo and M. D'Alessandro (Chieti); D. Santoro and L. Pusterla (Como); G. Carnevale and A. Zoncada (Cremona); P. Viganò and M. Mena (Cuggiono); F. Ghinelli and L. Sighinolfi (Ferrara); F. Leoncini, F. Mazzotta, M. Pozzi, and S. Lo Caputo (Firenze); G. Angarano, B. Grisorio, A. Saracino, and S. Ferrara (Foggia); P. Grima and P. Tundo (Galatina [LE]); G. Pagano, G. Cassola, A. Alessandrini, and R. Piscopo (Genova); M. Toti and M. Trezzi (Grosseto); F. Soscia and L. Tacconi (Latina); A. Orani and P. Perini (Lecco); A. Scasso and A. Vincenti (Lucca); F. Chiodera and P. Castelli (Macerata); A. Scalzini and L. Palvarini (Mantova); M. Moroni, A. Lazzarin, G. Rizzardini, L. Caggese, A. d'Arminio Monforte, D. Repetto, A. Galli, S. Merli, C. Pastecchia, and M. C. Moioli (Milano); R. Esposito and C. Mussini (Modena); N. Abrescia, A. Chirianni, C. M. Izzo, M. Piazza, M. De Marco, R. Viglietti, E. Manzillo, and S. Nappa (Napoli); A. Colomba, V. Abbadessa, T. Prestileo, and S. Mancuso (Palermo); C. Ferrari and P. Pizzaferri (Parma); G. Filice, L. Minoli, R. Bruno, and S. Novati (Pavia); F. Baldelli and G. Camanni (Perugia); E. Petrelli and A. Cioppi (Pesaro); F. Alberici and A. Ruggieri (Piacenza); F. Menichetti and C. Martinelli (Pisa); C. De Stefano and A. La Gala (Potenza); G. Ballardini and E. Rizzo (Ravenna); G. Magnani and M. A. Ursitti (Reggio Emilia); M. Arlotti and P. Ortolani (Rimini); R. Cauda, F. Dianzani, G. Ippolito, A. Antinori, G. Antonucci, M. Ciardi, P. Narciso, N. Petrosillo, V. Vullo, A. De Luca, M. Zaccarelli, R. Acinapura, P. De Longis, M. P. Trotta, P. Noto, M. Lichtner, M. R. Capobianchi, F. Carletti, E. Girardi, P. Pezzotti, and G. Rezza (Roma); M. S. Mura and M. Mannazzu (Sassari); P. Caramello, G. Di Perri, G. C. Orofino, and M. Sciandra (Torino); P. A. Grossi and C. Basilico (Varese); A Poggio and G. Bottari (Verbania); E. Raise and F. Ebo (Venezia); G. Pellizzer and D. Buonfrate (Vicenza); F. Resta and K. Loso (Taranto); and A. Cozzi Lepri (London, United Kingdom)

• Received February 7, 2007.
• Accepted April 11, 2007.

• © 2007 Infectious Diseases Society of America