Liver injury after beginning antiretroviral therapy in HIV/hepatitis C virus co-infected patients is not related to immune reconstitution

Transaminase elevations occur more frequently after beginning antiretroviral therapy in HIV-positive patients co-infected with hepatitis C virus (HCV). The mechanism of liver injury in these individuals is unknown, although immune reconstitution phenomena have been postulated. In 42 HIV/HCV co-infected individuals followed after beginning potent antiretroviral therapy, the development of liver injury was not associated with significant changes in serum HCV-RNA levels nor with greater CD4 cell increases. Underlying chronic hepatitis may thus increase the risk of liver toxicity by other mechanisms.

Liver toxicity is a common complication of highly active antiretroviral therapy (HAART), occurring more often in HIV-positive individuals co-infected with hepatitis C virus (HCV) than in those without chronic hepatitis C [^1-3]. It has been postulated that immune reconstitution phenomena might explain this higher incidence of transaminase elevations. In favour of this hypothesis is the observation of more pronounced increases in the CD4 cell count or HCV seroconversions accompanying liver enzyme elevations, as well as histological findings that argued against direct drug toxicity in liver biopsies [⁴]. More controversy exists as to whether changes in HCV replication occur after the initiation of HAART: whereas some authors have found an increase in serum HCV-RNA levels in patients developing liver damage on HAART [^5,6], others have failed to demonstrate this [⁷]. A biological interference between both viruses or the release of HCV RNA from hepatocytes killed as a result of a more efficient cytotoxicity could explain this association.

To assess whether transaminase elevations in patients receiving HAART may be related to changes in HCV viraemia or to a greater degree of immune recovery, the medical charts of all HIV/HCV co-infected individuals who started HIV treatment between June 1997 and January 2000 in a referral institution were reviewed. Transaminases (alanine aminotransferase and aspartate aminotransferase), CD4 cell counts, plasma HIV-RNA and HCV-RNA levels were measured at baseline and every 3 months during 6 months or more of follow-up. Given the retrospective nature of the study, other parameters that might reflect more accurately the CD4 T cell function were not available. Liver damage was defined as an increase in transaminases of at least 1.25 times the upper limit of normal, or baseline if previously elevated. It was graded as severe when the elevation was at least five times the upper limit of normal (or > 3.5 baseline levels if previously elevated). At the time of maximum hepatic damage, HIV-RNA and HCV-RNA levels and CD4 cell counts were recorded. A comparison of these values at that moment and at baseline in patients who developed hepatotoxicity and in those who did not was made. For these analyses, t-tests or non-parametric tests were performed for quantitative variables, and χ² or Fisher's exact test were used for analysing categorical variables. Differences between groups were considered for P values of less than 0.05 (two-sided). Univariate and multivariate logistic regression analyses were performed to examine the potential determinants of hepatic injury. Variables with a P value of less than 0.2 in the univariate analysis were entered into the multivariate analysis. The SPSS software version 8.0 (SPSS Inc., Chicago, IL, USA) was used for this purpose.

Forty-two patients (32 men and 10 women) were recruited into the study. Most (88%) had acquired HIV infection through needle-sharing. Their mean age was 37.6 years. The antiretroviral regimens prescribed to these patients included two nucleoside analogues plus protease inhibitors (PI) (10 patients), non-nucleoside reverse transcriptase inhibitors (26 patients), or both PI and non-nucleoside reverse transcriptase inhibitors (six patients). All patients were hepatitis B surface antigen negative.

A total of 30 (72%) subjects developed transaminase elevations of any grade, which was severe in six (14%). However, treatment had to be interrupted in only three patients. The peak in liver injury appeared after a mean time of 6 months. At this time, as expected, a significant decrease in HIV-RNA titres was noticed (from a mean of 4.3 logs at baseline to 1.7 logs;P < 0.001). Likewise, a significant increase in the CD4 cell count was recognized (from a mean of 277 cells/mm³ at baseline to 422;P < 0.001).

Mean HCV-RNA titres suffered a modest but significant increase with respect to baseline (from 5.71 logs to 5.88 logs;P = 0.01). The development of transaminase elevations was not associated with sex, risk factor, treatment modality, nor with baseline CD4 cell counts, HIV-RNA or HCV-RNA levels (Table 1). Patients with less than 200 CD4 cells/mm³ at baseline did not develop transaminase elevations more frequently than the rest (71% versus 71%), although immune reconstitution syndromes occurred more frequently in them (two developed mycobacterial diseases, one Pneumocystis pneumonia, and another herpes zoster). Liver injury was not associated with a decrease in HIV-RNA levels, nor with an increase in CD4 cell counts, nor with a greater increase in HCV-RNA levels after starting antiretroviral therapy (Table 2).

Table 1

Table 2

The progression of previously quiescent disorders to symptomatic diseases has been reported after the initiation of HAART. These conditions have been described for infections caused by cytomegalovirus, Mycobacterium avium or Pneumocystis carinii, and were associated with pronounced reductions in plasma HIV-RNA levels and increases in CD4 cell counts, suggesting immune recovery mechanisms [⁸]. A similar pathogenesis has been postulated for the higher incidence of liver injury noticed among HCV/HIV co-infected patients who begin HAART [⁴]. An improvement in HCV-specific immunity could produce a greater destruction of infected hepatocytes, and a further release of transaminases and HCV-RNA molecules. Our results argue against this hypothesis: neither the increases in CD4 cell counts nor in HCV-RNA levels were related to the development of hepatotoxicity.

As previously reported [^6,9], antiretroviral treatment was followed by an increase in HCV-RNA levels, but this greater HCV viraemia was not associated with liver damage. A possible antagonistic relationship between both HIV and HCV could explain why the control of HIV may be followed by a slight increase in HCV replication.

Finally, our study confirms the high rate of transaminase elevations in HCV/HIV co-infected patients who start HAART (overall 72%, severe 14%), although treatment had to be discontinued in only 7% of individuals. Although the close monitoring of liver function tests is warranted in these patients, HAART is thus relatively safe [^1-3].

In conclusion, transaminase elevations are frequently seen after the initiation of HAART in HIV/HCV co-infected patients. Our results do not support a role of immune reconstitution mechanisms in the pathogenesis of this phenomenon, because transaminase elevations were not related to changes in HCV-RNA levels nor to a greater increase in CD4 cell counts. Other mechanisms must therefore explain the increased risk of liver toxicity in patients with chronic hepatitis C.

Luz Martín-Carbonero

Marina Núñez

Pilar Ríos

Mayte Pérez-Olmeda

Juan González-Lahoz

Vincent Soriano