Severe Hepatotoxicity During Combination Antiretroviral Treatment: Incidence, Liver Histology, and Outcome
The authors conclude immune reconstitution may play a role. Liver failure occurred only in patients with severe CD4 depletion (<200 CD4s) before starting HAART and who had hepatitis.
Comments from Jules Levin: This study in Italy found about 1% of patients out of about 700 starting HAART developed liver failure and all who developed liver failure had CD4s counts before starting HIV therapy of <200. The authors feel that immune reconstitution due to CD4 increase may be the reason for liver failure. Although low CD4 count was not associated with severe hepatotoxicty, it was associated with liver failure. 300 of the 700 patients in this study had CD4 counts <200 before starting HAART, so only a few patients with chronic hepatitis and CD4+ cell counts <200 cells/mm3 at beginning of ART developed liver failure. 14 of the 26 patients who developed severe hepatotoxicity had <200 CD4s. All 26 patients had hepatitis: 25 with HCV and the I without HCV had HDV & HBV. Patients developing severe hepatoxicity were significantly more likely to have hepatitis than patients not developing severe hepatoxicity: HCV (96% vs 67%), HBV HbsAg-positive (19% vs 7%), HDV-antibody (19% vs 5%). But only 7 patients developed liver failure-all 7 had <200 CD4s. As well, for all the 26 patients with severe hepatotoxicity there was a direct correlation observed between increase in ALT and the increment of CD4 over baseline at ALT peak; in other words, higher CD4 count increases after starting ART were associated with high ALT levels, suggesting immune reconstitution plays a role in ALT peaks and perhaps hepatoxocity. The authors conclude their findings indirectly underline the need for early treatment of chronic HCV in patients with HIV coinfection before immune depletion occurs, and for patients who start ART at low CD4 counts, HCV treatment before or during ART may be effective in preventing or ameliorating severe hepatoxocity, and this needs further prospective study in clinical trials.
Seven hundred fifty-five patients underwent 915 treatment episodes for a duration of 611 person-years...The incidence of SH was not significantly different by treatment regimen (4 per 100 person-years in patients treated with two NRTIs plus one PI, 6 per 100 person-years in those treated with two NRTIs, and none in those treated with two NRTIs plus one NNRTI). No difference was found when comparing the incidence of SH according to prescribed drugs..One hundred five cases of relevant hepatoxticity were observed...Twenty-six of 105 cases of relevant hepatoxicity satisfied the definition criteria for severe hepatoxicity with an incidence of 4.2 per 100
person-years of treatment; 7 of 26 patients also developed LF during or no more than 1 month after ALT flare (1.1 per 100 person-years)..RISK FACTORS: Patients who developed SH during combination ART differed from those who did not by the following factors: they were more often male; had intravenous drug use as risk factor for HIV acquisition; were younger; were more often coinfected with HCV, HBV, and HDV; and had higher baseline ALT and bilirubin values and longer PT. No difference in follow-up time was found between patients with or without these characteristics..Patients who developed severe hepatoxicity were more likely to have HCV: 96% of patients developing severe hepatotoxicity had HCV antibody vs 67% of patients who did not develop severe hepatotoxicity (p<0.01).as well as HDV (19% vs 5%).
.....In the 26 patients with SH, a direct correlation was observed between increase of ALT and the increment of CD4 over baseline at ALT peak...
...Sixteen patients agreed to undergo liver biopsy less than 3 months after the occurrence of SH (see discussion below).. Seven of 26 (27%) patients showed LF 3 to 25 days after the ALT peak occurred, and all died..Among 17 patients who underwent rechallenge, 7 (41%) showed a relapse of SH 2 to 4 months after treatment reinitiation. The remaining 10 patients did not show SH relapse during a retreatment period of 20 to 32 months..Among the 7 patients who had SH relapse after rechallenge, 2 refused additional therapy and 5 (all with chronic hepatitis C on liver histology) were treated with interferon alpha-2a at 6 MU thrice weekly for 12 months...without severe hepatoxocity reoccurring during 30 months follow-up...The 13 patients who had rechallenge with ART after the initial SH underwent a second liver biopsy 18 to 32 months after the first liver biopsy was performed. All these patients showed better necroinflammatory scores and stable or better fibrosis scores according to the Ishak classification..
...striking was the observation that LF
occurred only in patients with severe CD4 depletion at baseline...why did liver failure occur in patients with low CD4 counts..more plausible is the interpretation that extensive liver damage could have been related to the shock caused by immune reconstitution against hepatitis viruses in subjects with severe immune depletion before starting ART..This interpretation is also supported by the fact that relapse of SH after a 4-month
wash-out period was observed only in 41% of patients retreated with combination ART..
....Another important consideration is that in patients with SH relapse after rechallenge with ARV, interferon treatment was able to prevent a third relapse of SH after ART despite the persistence of HCV active replication in most patients. Even though not complete, inhibition of HCV replication by interferon could reduce the burden of HCV antigens in the liver cells, and this, in turn, could reduce the intensity of immune-mediated liver cell necrosis after immune reconstitution induced by ART. Controlled prospective trials are necessary to prove the efficacy of interferon with or without ribavirin in combination to prevent SH in HIV patients with HCV coinfection who are put on ART..
...When the CD4+ T-cell count is 200 to 350 cells/mm3, the risk of SH resulting in LF may be low according to our data; thus, ART could be started quite safely. Only a few patients with chronic hepatitis and CD4+ cell counts <200 cells/mm3 at beginning of ART developed liver failure, so our data do not support an earlier start of ART with higher CD4 counts to prevent liver failure in patients with chronic hepatitis C. Nevertheless, they indirectly underline the need for early treatment of chronic hepatitis C in patients with HIV coinfection before severe immune depletion occurs. Finally, for patients who start ART at low CD4+ T-cell counts, antihepatitis pre- or comedication could be an effective preventive or curative measure for SH,
and this merits prospective evaluation in controlled clinical trials.
Combination antiretroviral treatment (ART), decreasing HIV RNA, and increasing CD4 cell counts have slowed disease progression, decreased mortality, and improved the quality of life for many persons with HIV (1). Issues concerning severe adverse events are becoming increasingly evident, however, limiting benefits in a significant proportion of patients (2).
Typical hepatic drug toxicity is exhibited by all classes of antiretrovirals; it is shown by a rise in transaminase levels and occasionally by signs of drug hypersensitivity or steatohepatitis. Last, but most important, liver toxicity is exacerbated by viral hepatitis, which is common in HIV-seropositive patients, especially those with a history of exposure to blood or blood products. Incidence of severe hepatic cytolysis defined as grade III to IV elevation of liver enzymes has recently been reported as 5 to 10 per 100 person-years during ART. Occurrence of liver disease and death has been described anecdotally in patients undergoing combination antiretroviral regimens with (10-13) and without (14) protease inhibitors (PIs). The occurrence of hepatic decompensation was reported in 2% of 187 patients undergoing treatment with two nucleoside analogues (NRTIs) and a PI and in 14% of 51 patients with chronic hepatitis C treated with these regimens.
Thus far, liver histology, etiology, and outcome of severe hepatotoxicity (SH) during ART have yet to be defined as well as risk factors associated with death due to decompensated liver disease in patients with SH.
Preliminary observations have suggested that among the factors involved (18), immune reconstitution may or may not play a role in the occurrence of liver damage. Moreover, the impact of ART on liver histology after resolution of acute hepatotoxicity has not yet been explored.
In the current study, all patients who initiated any combination of ART during an 18-month period have been closely followed with clinical, biochemical, and histologic assessment to define prospectively the incidence and factors associated with the occurrence of SH. Moreover, patients have been followed with the aim of assessing clinical, biochemical, and histologic outcome from SH.
The objective of this study was to assess incidence, risk factors, histology, and outcome of severe hepatotoxicity (SH) during antiretroviral treatment (ART). Seven hundred fifty-five HIV-seropositive patients consecutively prescribed new ART were selected in the HIV/AIDS Clinic of the Institute of Infectious Diseases of the University of Brescia, Italy from January 1997 to June 1998. Liver function were assessed at baseline, after 1 month, and every 4 months thereafter. Liver biopsy was recommended in case of SH (i.e., increase in liver enzymes >10 times the upper limit of normal or 5 times baseline if markedly abnormal).
Twenty-six cases of severe hepatotoxicity were observed with an incidence of 4.2% person-years. Liver failure (LF) was rarely seen (1.1 per 100 person-years). Liver damage was invariably observed in patients with chronic viral hepatitis. Liver histology showed exacerbation of viral hepatitis in all 16 patients for whom a liver biopsy was available at the time of SH. A direct correlation was found between alanine aminotransferase increase and increase in CD4+ T-cell count in patients with SH (r = 0.53, p < .001). Death occurred during follow-up in 7 of 26 (27%) patients, all of whom showed liver failure and baseline CD4+ count less than 200 cells/mm3 (7/7 patients = 100% vs. 8/19 patients without LF; p < .01). Relapse of SH was observed after ART was recommenced in 7 of 17 (41%) patients. Five of these 7 patients did not show further SH relapse after treatment with interferon.
The authors concluded that this study provides estimates of SH and LF in a large population-based setting where hepatitis C virus coinfection is highly prevalent and provides indications that liver damage may be caused by immune reconstitution and related exacerbation of viral hepatitis. A strict follow-up for hepatotoxicity is mandatory when ART is initiated in patients with <200 CD4+ T cells/mm3. Antihepatitis pre- or comedication could be an effective preventive or curative measure.
JAIDS Journal of Acquired Immune Deficiency Syndromes 2003; 32(3):259-267
*Massimo Puoti; *Carlo Torti; *Diego Ripamonti; *Francesco Castelli; *Serena Zaltron; *Barbara Zanini; *Angiola Spinetti; *Valeria Putzolu; *Salvatore Casari; *Lina Tomasoni; *Eugenia Quiros-Roldan; Maurizio Favret; āLuisa Berchich; Piergiovanni Grigolato; Francesco Callea; *Giampiero Carosi; HIV-HCV Co-Infection Study Group
According to a standardized toxicity grade scale [modified from the scale used by the AIDS Clinical Trial Group (13)], two classes of outcomes have been defined: relevant hepatotoxicity (RH) and SH. RH was defined as any
increase of liver enzymes of at least 5 times the ULN or 2.5 times baseline if markedly abnormal (i.e., over twice the ULN); in these patients, ART was generally not discontinued by the treating physician. SH was defined as an
increase of liver enzymes of at least 10 times the ULN or 5 times baseline if markedly abnormal (i.e., over twice the ULN), followed by hepatic enzyme improvement after ART discontinuation. Liver failure (LF) was defined as
the presence of noninflammatory ascites plus all the following: PT less than 75% of the control after at least 1 week of vitamin K administration, total serum bilirubin greater than 3.5 mg/dL, serum albumin less than 2.5 g/dL in the absence of wasting syndrome, and ultrasound-observed space-occupying lesions in the liver. In the presence of SH, all patients underwent the following: medical examination; alcohol and nonmedicinal drug consumption
assessment; liver ultrasound examination and computed tomography scan when indicated; biochemical and hematologic profile, including arterial blood lactic acid and bicarbonate measurements, hepatitis C virus (HCV)
RNA, hepatitis B virus (HBV) DNA in HBsAg+, hepatitis D virus (HDV) antibody (Ab), and IgM; and a complete workup intended to exclude any concurrent opportunistic infections or neoplasms in the liver (including cytomegalovirus [CMV] DNA detection and quantification by PCR technique; blood cultures for bacteria, fungi, and mycobacteria; detection of cryptococcal soluble antigens; serology for Bartonella sp.; and stool examinations for ova and parasites). Consenting patients without major contraindications underwent liver biopsy no more than 3 months after the occurrence of SH. Patients showing SH were followed up for at least 36 months. Combination
ART was restarted after 4 months of wash-out in all consenting patients who had normalization or values equal to baseline or less than 5 times the ULN elevation of liver function enzymes. Drug regimens were not changed unless
genotype resistance testing showed HIV resistance to antiretrovirals. Patients with chronic hepatitis showing SH relapse were treated with recombinant interferon [alpha]-2a (Roferon; Roche, Basel, Switzerland) at a dose of 6 million units (MU) three times weekly for 12 months; a third cycle of combination ART was provided during interferon treatment.
Etiology of Alanine Aminotransferase Flares and Biopsy
Anti-HCV and HCV RNA reactivity were detected in all but 1 of 26 patients with SH; HBsAg was detected in 5 patients, with none of them showing serum HBV DNA reactivity; and anti-HDV IgM reactivity was detected in 5
patients. The patient without HCV-RNA reactivity showed HBsAg and HDV-Ab IgM reactivities. These data were comparable with those obtained at baseline in all patients except 1, who showed HCV RNA but not anti-HCV
reactivity at baseline; this patient showed less than 50 CD4 cells/mm3 at baseline. In this patient, liver histology showed HCV-related cirrhosis. None of these 26 patients showed evidence of opportunistic infections and/or
In the 26 patients with SH, a direct correlation was observed between increase of ALT and the increment of CD4 over baseline at ALT peak (Spearman rank correlation coefficient: r = 0.53, p < .001).
One hundred five cases of RH were observed (17 per 100 person-years of treatment). Treatment was stopped in only 9 of 105 patients. Multivariate analysis revealed that only two variables were independently associated with the occurrence of RH: anti-HCV reactivity (OR = 4.0, 95% CI: 2.16-7.69; p < .001) and baseline ALT levels (OR = 1.04, 95% CI: 1.01-1.07 for each 10 U/mL; p < .001). No patient with RH continuing ART showed SH and/or LF at the following visits.
Sixteen patients agreed to undergo liver biopsy less than 3 months after the occurrence of SH. The highest scores regarding necroinflammatory scores according to Ishak classification (piecemeal necrosis, confluent necrosis, focal necrosis, portal inflammation) were observed for portal inflammation and piecemeal necrosis, usually in association with immune reaction. Piecemeal necrosis involved most of the portal areas in 15 of 16 cases. Zone 3 necrosis was absent (i.e., confluent necrosis [CN] scores were less than 2) in 14 of 16 patients. The patient who showed anti-HCV seroconversion showed histologic evidence of chronic hepatitis C with cirrhosis; thus, seroconversion was interpreted as restoration of anti-HCV reactivity previously lost because of immune depletion and not as acute hepatitis C. Mild bile duct damage was present in 5 of 16 biopsies, lymphoid
follicles in 3 of 16 biopsies, mild steatosis in 8 of 16 biopsies, and moderate steatosis in 2 of 16 biopsies; mild eosinophil granulocyte infiltration was present in 7 biopsies, and it was moderate in 2 of 16 biopsies; and Mallory
inclusion bodies were present in 2 biopsies.
7 of 16 patients undergoing biopsy had stage 2; 1 patient had stage 0; 1 patient had stage 1; 3 patients had stage 3; 1 patient had stage 4; and 1 patient had stage 6.
Distribution of fibrosis scores in the 16 liver biopsies according to Ishak et
al. (23). 0, no fibrosis; 1, fibrous expansion of some portal areas with or
without short fibrous septa; 2, fibrous expansion of most portal areas with
or without short fibrous septa; 3, fibrous expansion of most portal areas
with occasional portal to portal (P-P) bridging; 4, fibrous expansion of
portal areas with marked bridging P-P as well portal to central (P-C)
bridging; 5, marked bridging (P-P and/or P-C) with occasional nodules
(incomplete cirrhosis); 6, cirrhosis probable or definite.
In 2 patients, a former liver biopsy performed 6 to 18 months before starting ART was available. Compared with the former biopsies pre-ART, the post-ART biopsies showed worsening of necroinflammatory activity score (from
5 to 9 and from 3 to 8, respectively).
Outcome of Severe Hepatotoxicity
The outcome of the 26 patients who suffered from SH is described below in more detail. None of the patients without SH developed LF or died as a result of liver disease in the subsequent follow-up.
Seven of 26 (27%) patients showed LF 3 to 25 days after the ALT peak occurred, and all died. This incidence of LF was 1.1 per 100 person-years of combination ART treatment. In 6 patients, LF was irreversible; thus, these
patients were not rechallenged with ART, and they died as a result of end-stage liver disease 3 to 12 months after the occurrence of SH. In 1 of these patients, LF was associated with lactic acidosis and pancreatitis, and the patient died after 3 days. The seventh patient showed an improvement of liver function 3 months after withdrawal of indinavir, zidovudine, and lamivudine, but the CD4 cell count dropped to less than 50 cells/mm3, and this patient developed relapsing disseminated Mycobacterium avium infection, candida esophagitis, and a new CMV retinitis. After successful treatment of these opportunistic infections, ART with stavudine plus didanosine plus saquinavir was recommenced. After 2 months, the patient showed a relapse of SH, with occurrence of irreversible LF 10 days
after the ALT increase. He died 3 months later. When looking at the clinical characteristics between the 7 patients with liver failure and the remaining 19 with only severe hepatotoxicity, a lower baseline CD4 count was the only significant difference between the two groups, and a baseline CD4 count <200 appeared more often in patients with liver failure. The nadir CD4 count was a bit lower (177 vs 296).
Among 19 patients who did not show LF, 2 (10%) were lost to follow-up. Seventeen patients showed a decrease of ALT to below five times the ULN but without complete normalization; after a 4-month wash-out period, they
were rechallenged with the previous treatment regimens, with the exception of 2 patients who were treated with didanosine instead of lamivudine because of evidence of HIV genotype resistance in previous resistance tests.
Among 17 patients who underwent rechallenge, 7 (41%) showed a relapse of SH 2 to 4 months after treatment reinitiation. The remaining 10 patients did not show SH relapse during a retreatment period of 20 to 32 months.
Among the 7 patients who had SH relapse after rechallenge, 2 refused additional therapy and 5 (all with chronic hepatitis C on liver histology) were treated with interferon [alpha]-2a at 6 MU thrice weekly for 12 months. One of these 5 patients had normalized ALT and cleared serum HCV RNA and did not show an increase in ALT after ART was recommenced in combination with interferon treatment; the other patients did not show complete ALT
normalization despite a 2 log decrease in HCV RNA and a 50% decrease in ALT activity after 1 month of interferon. These 4 patients underwent another rechallenge during interferon treatment with the same combination
ART and did not show SH over a follow-up interval of at least 30 months, including a period of at least 8 months after interferon had been withdrawn.
The 13 patients who had rechallenge with ART after the initial SH underwent a second liver biopsy 18 to 32 months after the first liver biopsy was performed. All these patients showed better necroinflammatory scores and
stable or better fibrosis scores according to the Ishak classification.
Discussion by Authors
Immune reconstitution against hepatitis virus antigens in liver cells may have played a major role in the pathogenesis of SH during ART in the study cohort as evidenced by the association of the following findings: (1) the occurrence of SH mainly in patients coinfected with hepatitis viruses, (2) the positive correlation between CD4 and ALT increases in patients with SH, (3) the liver histologic picture showing exacerbation of chronic hepatitis, and (4) the absence of relapse of SH of most patients rechallenged with the same ART regimens after a wash-out period and/or after interferon treatment. In addition, even the milder clinical outcome of RH was significantly
associated with HCV coinfection and pretreatment hypertransaminasemia on multivariate analysis. These findings confirm the hypothesis formulated in previous studies in contrast to other recent observations.
The inconsistency of our results with respect to the study reported by Martin-Carbonero et al. might be a result of the fact that lower levels of transaminases were assumed as an outcome measure of liver damage (i.e., [ge]1.5 times the ULN) in that study. This less strict end point may have encapsulated other causes of liver damage (e.g., direct drug toxicity, preexisting or concomitant liver damage by alcohol or concomitant medications); thus, the correlation between immune reconstitution and SH may have been confounded. Alternative explanations could be
the low number of patients in that cohort and/or baseline differences in the patients studied (e.g., naive vs. mixed cohort) and/or differences in the immunologic response to ART.
A limitation of the current study is that multivariate analysis was not performed to assess risk factors because of the extremely strict criteria used to define SH (i.e., grade IV hepatic toxicity according to WHO adverse events severity scale) and the small number of events recorded. Notwithstanding, risk factors for SH identified in our cohort by univariate analysis are all linked to preexisting chronic viral hepatitis, and this is consistent with other European studies using multivariate analysis techniques (4, 6-8). In addition, in the same cohort, multivariate analysis of a
milder clinical outcome, RH, revealed an independent association with chronic hepatitis C.
With regard to irreversible LF after SH, an incidence of 1.1 per 100 person-years of combination ART was observed, unlike in other published cohort studies (4, 6-8). For instance, LF occurred in only 0.3% of patients in
the cumulative analysis of 21 AIDS Clinical Trial Group studies (9). This difference is probably related to the higher number of subjects with hepatitis virus coinfection in our cohort. More striking was the observation that LF occurred only in patients with severe CD4 depletion at baseline. No other differences were found between patients showing LF and those without occurrence of LF over the follow-up period. This observation is consistent with the fact that baseline CD4 T-cell counts were below 200 cells/mm3 in all cases of LF without lactic acidosis reported on
ART until now (6-10).
Three hypotheses may explain the occurrence of LF only in patients with severe CD4 depletion. First, severe CD4 depletion may be associated with more advanced chronic hepatitis. This seems to be ruled out in our study by the fact that baseline liver function parameters were comparable between patients with LF and without LF. Second, opportunistic infections or immune restoration against opportunistic pathogens present in the liver may have played a role in the pathogenesis of LF, but only 1 of the 7 patients with LF showed evidence of a concomitant opportunistic infection. Finally, more plausible is the interpretation that extensive liver damage could have been
related to the shock caused by immune reconstitution against hepatitis viruses in subjects with severe immune depletion before starting ART. This interpretation is also supported by the fact that relapse of SH after a 4-month wash-out period was observed only in 41% of patients retreated with combination ART. Consequently, this "stop, wash-out, and go" strategy allowed ART to be continued even after an episode of SH, indicating that direct hepatotoxicity by itself did not have a major role in the pathogenesis of SH. Moreover, histologic follow-up performed in 13 patients did not reveal any worsening of liver histologic findings after up to 32 months of follow-up.
Another important consideration is that in patients with SH relapse after rechallenge with ARV, interferon treatment was able to prevent a third relapse of SH after ART despite the persistence of HCV active replication in most patients. Even though not complete, inhibition of HCV replication by interferon could reduce the burden of HCV antigens in the liver cells, and this, in turn, could reduce the intensity of immune-mediated liver cell necrosis after immune reconstitution induced by ART. Controlled prospective trials are necessary to prove the efficacy of interferon with or without ribavirin in combination to prevent SH in HIV patients with HCV coinfection who are put on ART.
Although this is the first study focusing on the follow-up of patients with clinically significant ARV hepatotoxicity, it has some weaknesses that need to be acknowledged. First, the association between chronic hepatitis and severe hepatotoxicity is not novel; however, we have confirmed it in relation to outcomes that are clearly relevant from a clinical point of view. Second, although histologic examination was available only in 16 patients, this is, to our knowledge, the largest series of biopsies performed in patients with ARV-induced hepatotoxicity. Third, treatment regimens were chosen by the treating physicians and therefore differed by baseline characteristics; however, only a few patients were treated with full-dose ritonavir or an NNRTI, which have been independently associated with hepatic cytolysis on treatment.
In conclusion, this study provides estimates of the incidence of ART-induced SH and LF in a large clinical setting of patients with a high prevalence of chronic hepatitis. Severe hepatotoxicity was related to preexisting chronic viral hepatitis followed by irreversible LF in a few patients, all with severe CD4+ T-cell depletion (i.e., <200 CD4+ T cells/mm3) before starting ART. Besides the fact that all patients with chronic viral hepatitis should be strictly monitored for liver damage after starting ART, this observation strengthens the importance of careful follow-up in patients with chronic hepatitis and a low CD4+ T-cell count. When the CD4+ T-cell count is 200 to 350 cells/mm3, the risk of SH resulting in LF may be low according to our data; thus, ART could be started quite safely. Only a few patients with chronic hepatitis and CD4+ cell counts <200 cells/mm3 at beginning of ART developed liver failure, so our data do not support an earlier start of ART with higher CD4 counts to prevent liver failure in patients with chronic hepatitis C. Nevertheless, they indirectly underline the need for early treatment of chronic hepatitis C in patients with HIV coinfection before severe immune depletion occurs. Finally, for patients who start ART at low
CD4+ T-cell counts, antihepatitis pre- or comedication could be an effective preventive or curative measure for SH, and this merits prospective evaluation in controlled clinical trials.