HIV Medicine (2004), 5, 334–343

r

2004 British HIV Association

ORIGINAL RESEARCH

Lopinavir/ritonavir treatment in HIV antiretroviralexperienced patients: evaluation of risk factors for liver enzyme elevation P Meraviglia,1 M Schiavini,1 A Castagna,2 P Vigano`,3 T Bini,4 S Landonio,5 A Danise,2 MC Moioli,6 E Angeli,1 M Bongiovanni,4 H Hasson,2 P Duca7 and A Cargnel1 1 2nd Department of Infectious Diseases, Sacco Hospital, Milan, 2Institute of Infectious Diseases and Tropical Medicine, University of Milan, Milan, 3Department of Infectious Diseases, Civile Hospital Legnano, 4Department of Clinical Infectious Diseases, Sacco Hospital, Milan, 51st Department of Infectious Diseases, Sacco Hospital, Milan, 6Department of Infectious Diseases, Niguarda Hospital, Milan, and 7Medical Statistics Unit, Preclinical Sciences Department, Milan University, Milan, Italy Objectives

To evaluate the risk factors for lopinavir/ritonavir (LPV/r)-related liver enzyme elevation (LEE) in HIV antiretroviral-experienced patients. Methods

An open prospective observational study was carried out to analyse the incidence and time of LEE development during LPV/r treatment, and to determine whether LEE development was correlated with epidemiological, clinical and biochemical data, immune and virological profiles, concomitant hepatic diseases, antiretroviral therapy, or histological and ultrasonography liver examination results. A diagnosis of LEE was considered when LEE symptoms occurred after LPV/r introduction and was confirmed by a second control within 2 weeks. Results

A total of 782 HIV-positive outpatients have been enrolled in six different Infectious Diseases Departments in Northern Italy since August 2000. Of these patients, 71 (9.1%) developed LEE within 115
85 days (mean
standard deviation); 13 of these subjects discontinued LPV/r and four were hospitalized. Of the patients with LEE, 74.6% and 25.4% had grade 2 and 3 toxicity, respectively. No correlation between LEE and sex, baseline CD4 cell count, viral load, HIV stage, triglyceride values, histological and ultrasonography liver examination results, nevirapine use, or increase in CD4 cell count was observed. Higher baseline alanine aminotransferase (ALT) and g-glutamyltransferase (GGT) values (P o 0.0001 and P 5 0.004, respectively), younger age (P 5 0.008), previous hepatitis B virus (HBV) infection (P 5 0.012), efavirenz use (P 5 0.04), and hepatitis C virus (HCV) and/or HBV coinfection (P o 0.0001, relative risk 4.78) were significantly related to LEE. No correlations between LEE and the same risk factors as investigated in the whole study population were found in subgroups of patients with HCV and/or HBV infection. Conclusions

HCV and HBV testing and measurement of baseline ALT values are essential for screening subjects at risk of LEE before starting LPV/r. Strict monitoring of clinical and biochemical parameters should be performed in these patients. Keywords: liver enzyme elevation, lopinavir/ritonavir, viral hepatitis coinfection Received: accepted 11 March 2004

Introduction Highly active antiretroviral therapy (HAART) is very effective in reducing morbidity and mortality in HIVpositive subjects.

Correspondence: Dr Paola Meraviglia, 2a Divisione Malattie Infettive, Ospedale ‘L. Sacco’, Via G.B. Grassi 74, 20157 Milano, Italy. Fax: 1 02 38200909; e-mail: [email protected]

334

LEE onset in LPV/r treatment 335

The benefits of HAART may be limited by the development of side effects, in particular liver toxicity. The incidence of, and risk factors for, this adverse event caused by antiretroviral drugs have been reported in various studies [1–10]. Several authors [7–11] have observed an increased risk of liver enzyme elevation (LEE) in HIV-positive patients presenting hepatitis B virus (HBV) and/or hepatitis C virus (HCV) coinfection compared to HIV-infected patients without hepatic involvement. Various pathogenetic mechanisms have been proposed to explain the increased risk of hepatotoxicity related to antiretroviral therapy in coinfected patients. The first mechanism is based on the intrinsic toxic effect of the drug and is therefore dose-related. All protease inhibitors (PIs) are metabolized in the liver by the cytochrome P450 3A4 enzyme system, and drug metabolism impairment may induce LEE through a cumulative dose in patients with chronic liver disease. The second proposed mechanism is related to the recovery of cell-mediated immunity induced by HAART, leading to immune-mediated HBV- and HCVspecific liver cell damage and transaminase elevation [12,13]. Lopinavir (LPV) is metabolized by the CYP3A isoenzyme, and ritonavir is a potent CYP3A inhibitor which increases the plasma levels of LPV. In the literature, it has been reported that patients with liver failure have presented higher LPV/r plasma concentrations than patients with normal liver function [14], and a study of HAART-naive HBV- and/or HCV-coinfected patients showed that these subjects had a significantly increased risk of transaminase elevation [15,16] during LPV/r treatment. Moreover, because of its high efficacy and tolerability, LPV/r is frequently used as part of an antiretroviral regiment in pluri-experienced patients. The primary aim of this prospective study was to evaluate the incidence of, and associated risk factors for, LEE in antiretroviral pluri-experienced patients receiving LPV/r.

Methods From August 2000 to August 2002, HIV-positive antiretroviral-experienced outpatients receiving LPV/r as part of an antiretroviral regimen, and referred to six Infectious Diseases Departments in Northern Italy (the 1st, 2nd and Clinical Infectious Diseases Departments, Sacco Hospital, Milan; the Department of Infectious Diseases, Niguarda Hospital, Milan; the Department of Clinical Infectious Diseases, Vita and Salute University, Milan, and the Infectious Diseases Unit, Legnano Hospital, Legnano), were evaluated for the development of LEE and prospectively

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

followed up for at least 12 months or until LPV/r discontinuation if side effects appeared. The standard dose of LPV/r was 400/100 mg twice a day (bid) independent of body weight, or 533/133 mg bid if nevirapine or efavirenz was used as part of an antiretroviral combination therapy. Patients with baseline ALT values 4250 UI/L and subjects with current high alcohol intake levels (450 g/ day) were excluded from the study. No patients used statins for hypercholesterolaemia. Patients were recommended to include omega-3 fatty acids or fibrates in their diet for the management of LPV/rassociated lipid elevations.

Patient evaluations Evaluations of epidemiological and clinical features (sex, age, risk factors for HIV infection, and HIV stage according to the 1992 revised Centers for Disease Control and Prevention HIV classification), biochemical parameters [aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (AP), g-glutamyltransferase (GGT), total cholesterol, triglyceride and glucose concentrations], and immune and virological profiles [CD4 cell count and viral load (branched-DNA Bayer (Bayer Diagnostic, Milan, Italy) lower limit 50 HIV-1 RNA copies/mL)] were performed at baseline, after 1 month and then every 3 months. At baseline, serological parameters for HBV and HCV infection status were measured: hepatitis B surface antigen (HBsAg) and antibody (HbsAb), core antibodies (HBcAb) and antibodies against hepatitis C virus (anti-HCV). HBV and HCV chronic infection was defined by the persistent positivity of HBsAg and HCV serum antibody. Percutaneous liver biopsy using Menghini’s needle (Securcut, Milan, Italy) was performed in 70 coinfected patients within 1 year before LPV/r treatment. The Knodell score system, known as the histological activity index, modified by Ishak (1995 revised 2000), was used to evaluate liver biopsy specimens [17,18]. Liver ultrasonography B-mode with convex probe (3.5 MHz) was performed within 1 year before LPV/r treatment in HCV- and/or HBV-positive patients. The image patterns of the liver parenchyma were classified into four groups: ‘negative’, if no structural alterations were present; ‘fatty liver’, defined as diffuse increased liver echogenicity (bright liver echo-pattern) and possibly distal sound extinction; ‘fibrosis’, defined as widespread inhomogeneous liver texture and slight irregularity of the liver surface; ‘cirrhosis’, if three or more of the following criteria were present: lobulated boundaries, gross irregularity of the liver surface, coarse echo-pattern, distorted hepatic

336 P Meraviglia et al.

veins, caudal liver edge deformation, left and caudal lobe enlargement and a progressive decrease in right liver lobe size. Left liver lobe hypertrophy was assessed as the ratio of right lobe thickness to left lobe thickness. Caudal liver hypertrophy was calculated from the caudal transverse diameter lobe (C) and the right transverse diameter lobe (RL) ratio. A C/RL ratio  0.65 was taken to be significant for a diagnosis of cirrhosis.

analysed using the nonparametric Mann–Whitney test and unpaired t-test. Differences in proportions were tested using Fisher’s exact test or the w2 test. Logistic regression analysis was used to assess risk factors associated with the development of hepatotoxicity. Variables considered in the multivariate analysis included age, HCV and/or HBV positivity, baseline ALT value and the concomitant use of efavirenz.

Clinical end points The objective was to evaluate the appearance of LEE, scored according to an increase in ALT value of 2.6–5 times the upper normal limit (toxicity grade 2), 5.1–10 times the upper normal limit (toxicity grade 3) and 410 times the upper normal limit (toxicity grade 4). The upper limit of normal values was 40 UI/L for AST and ALT. Furthermore, to avoid bias caused by patients with high baseline serum ALT and AST levels (i.e. subjects with chronic hepatitis), LEE was defined as an absolute increase of ALT and AST of more than 100 IU/L from the baseline value. Patients who experienced LEE during follow up as a consequence of another condition, with a definitive diagnosis (i.e. acute viral hepatitis, right cardiac failure, or conditions caused by other hepatotoxic drugs), were not included in the analysis. The decision to discontinue LPV/r therapy was based on clinical and biochemical evaluations for each patient and on the remaining therapeutic possibilities.

Results Baseline evaluation A total of 782 HIV-positive patients were included in the study. Their baseline characteristics are presented in Table 1. The mean values for biochemical parameters examined pretreatment and during follow-up are reported in Table 2. At baseline, 513 patients had normal ALT values and 269 (34.4%) patients had ALT values above the upper normal limit. Eighty-three of 269 patients were negative for HCV and/or HBV infection. Baseline AST and ALT levels were significantly higher in HCV- or HBV-coinfected patients than in uninfected patients (AST: mean 64
45 versus 35
21, median 48 versus 31, respectively; ALT: mean 63
43 versus 37
28, median 50 vs. 28, respectively; Po0.0001 for both). Table 1 Baseline characteristics in the study population n (%)

Statistical analysis The software packages Graph Pad Prism (version 3.02) for Windows, Graph Pad Instat (version 3.05) for Windows 95 (Graph Pad Software, San Diego, CA; www.graphpad.com) and Stata 7.0 (Stata Corporation, Lakeway Drive College Station, TX; www.stata.com) were used. To evaluate changes in viral load, CD4 cell count, and AST, ALT, AP, GGT, triglyceride and glucose concentrations, the Wilcoxon matched-pairs signed-rank test was used, while for total cholesterol a paired t-test was used. Risk factors investigated included demographic variables (age, gender and HIV risk group), clinical stage of disease, baseline CD4 cell count, viral load, AST and ALT concentrations, HCV and/or HBV serostatus and concomitant antiretroviral therapy. The Kaplan–Meier plot (log rank test) was used to identify differences in time of appearance of LEE between HCV- and/or HBV-positive patients and HCV/HBV-negative patients, and between patients with elevated baseline ALT values and those with normal baseline ALT values. Differences in the biological parameters studied between patients with and without LEE were

Sex Male Female Age (years) Mean
standard deviation Range Risk group for HIV infection Homosexual Heterosexual Injecting drug use in the past Other CDC class A B C Hepatitis markers Negative Anti-HCV positive HbsAg positive Anti-HCV and HbsAg positive Previous HBV infection Protease inhibitor associated with LPV/r Nelfinavir Indinavir Saquinavir Amprenavir

r

590 (75.4) 192 (24.6) 40.7
7.9 19–79 187 239 331 25

(23.9) (30.6) (42.3) (3.2)

250 (31.9) 224 (20.7) 308 (39.4) 426 304 25 27 341 76 5 18 21 32

(55.5) (38.9) (3.2) (3.4) (46.7) (9.7) (0.6) (2.3) (2.7) (4.1)

2004 British HIV Association HIV Medicine (2004) 5, 334–343

LEE onset in LPV/r treatment 337

Table 2 Variation of different parameters during follow up Mean
SD Variable

Baseline

3 months

6 months

9 months

12 months

CD4 count (cells/mL) P Viral load (log10 copies/mL) P AP (UI/L) P GGT (UI/L) P AST (IU/L) P ALT (IU/L) P Total cholesterol (mg/dL) P Triglycerides (mg/dL) P Glucose (mg/dL) P

238
203

319
212 o0.0001* 2.84
1.2 o0.0001* 214
123 0.857* 109
144 0.729* 49
47 0.043* 54
63 0.006* 198
59 0.0007w 312
225 o0.0001* 97
43 0.130*

343
215 o0.0001* 2.88
1.3 o0.0001* 218
159 0.634* 108
160 0.143* 49
44 0.090* 56
95 0.005* 202
103 o0.0001w 331
327 o0.0001* 106
118 0.053*

343
223 o0.0001* 2.96
1.3 o0.0001* 218
107 0.060* 104
160 0.345* 50
67 0.108* 55
101 0.048* 197
68 0.001* 303
246 o0.0001* 98
58 0.072*

376
235 o0.0001* 2.87
1.3 o0.0001* 212
109 0.061* 94
114 0.521* 48
47 0.221* 50
53 0.059* 209
98 o0.0001w 324
316 o0.0001* 92
28 0.841*

4.24
1.1 211
128 102
125 48
37 52
49 178
58 237
214 92
31

*Wilcoxon matched-pairs signed-rank test; wpaired t-test.

Normal range of laboratory examinations: CD4 count, 600–1500 cells/mL; viral load, o50 copies/mL; AST and ALT, 10–40 IU/L; total cholesterol, o190 mg/dL; triglycerides, o180 mg/dL; glucose, 70–110 mg/dL; AP, 50–300 UI/L in males and 50–260 in females; GGT, 10–50 UI/L in males and 7–32 in females. SD, standard deviation; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AP, alkaline phosphatase; GGT, gammaglutamyltransferase.

Histological data for the 70 pretreatment liver biopsies available for HCV- and/or HBV-coinfected patients indicated nine cases of minimal chronic hepatitis, 33 cases of mild chronic hepatitis and 28 cases of moderate chronic hepatitis; fibrosis grade 0 was found in two patients, grade 1 in 14 patients, grade 2 in 14 patients, grade 3 in 23 patients, grade 4 in four patients, grade 5 in 12 patients and grade 6 in one patient. An ultrasonography abdominal examination performed in HCV- and/or HBV-coinfected patients showed a pattern of steatosis in 25% of patients, fibrosis in 36%, morphological signs of cirrhosis in 6% and no alteration in 33%. The most common HIV therapies associated with LPV/r were: 18% didanosine 1 stavudine and zidovudine 1 lamivudine, 13% lamivudine 1 stavudine, 5% didanosine 1 lamivudine, 4% zidovudine 1 lamivudine 1 abacavir, and 3% zidovudine 1 didanosine. Efavirenz and nevirapine were used, respectively, in 11% and 7% of the patients. Seventy-six patients (9.7%) were treated with LPV/r associated with another protease inhibitor (Table 1).

Follow-up The mean time of follow-up was 373
197 days (median 349 days, range 2–783 days); 597 patients completed 12 months of follow-up, and of these 144 completed 24 months of follow-up; 185 (24%) patients discontinued

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

LPV/r therapy, and their mean time of follow-up was 235
189 days (median 198 days; range 2–740 days). Causes of LPV/r discontinuation were: therapeutic failure (57 patients), low compliance (40 patients), side effects (88 patients: four showed metabolic toxicity, seven allergic reactions, 18 neurological toxicity, 42 gastrointestinal toxicity, and 13 hepatic toxicity). Four patients died from HIV-related infections. During treatment, the mean change in CD4 cell count was 1 80 cells/mL at month 3, 1 105 cells/mL at month 6, 1 106 cells/mL at month 9 and 1 138 cells/mL at month 12; the mean decrease in HIV viral load was  1.38 log10 at month 3,  1.35 log10 at months 6 and 9 and  1.28 log10 at month 12. A statistically significant increase in total cholesterol and triglyceride values was observed (Po0.0001) during follow-up, while no significant change in AP, GGT or glucose levels was noted (Table 2). A statistically significant increase in ALT values was found at month three (P 5 0.006), month six (P 5 0.005) and month nine (P 5 0.04) (Table 2): the comparison between HCV- and/or HBV-positive subjects and HCV/ HBV-negative subjects showed that the increase was mainly present in the coinfected patients (Fig. 1a). LEE occurred in 71 patients (9.1%), of whom 52 were male and 19 female, with a mean age of 38 years. This side effect was found in 16.3% of HCV- and/or HBV-positive subjects and in 3.3% of HCV/HCB-negative subjects. LEE

338 P Meraviglia et al.

toxicity was mainly grade 2 (74.6% of patients); grade 3 was present in 19.7% of patients and grade 4 in 5.7%; the mean ALT value was 288
289 UI/L (range 125–1838 UI/ L, median 208 UI/L). The time ALT remained elevated was 115
85 days (median 90 days, range 27–540 days), and there was no statistical difference between HCV- and/or HBV-coinfected patients and HCV/HBV-negative patients (mean 121 days versus 95 days, respectively; P 5 0.21 by the Mann–Whitney test). None of the HBV-coinfected patients developing LEE discontinued lamivudine or showed emergence of lamivudine resistance. Thirteen (18.3%) of 71 patients discontinued LPV/r (four uninfected and nine coinfected); among these, 11 subjects continued the concomitant antiretroviral therapy with normalization of ALT. Clinical data regarding hospitalized patients and subjects with grade 3 or 4 toxicity are shown in Table 3. Four patients were hospitalized (mean time 33 days) for symptomatic hepatitis, but no other concomitant disease or opportunistic infection was diagnosed; only one patient was taking a high dose of benzodiazepine and HIV therapy. Three out of four patients showed a large increase in total bilirubin value (mean value 18 mg/dL) and recovered with conservative therapy after LPV/r discontinuation. The patient undergoing therapy with benzodiazepine died of liver failure, but an autopsy was not carried out.

Risk factors for hepatotoxicity

Fig. 1 (a) ALT values during follow up in all patients (}), in patients with negative hepatitis markers ( & ), in patients with positive hepatitis markers (~) and in patients who developed LEE (). In patients with positive hepatitis markers, the mean baseline ALT value was 63
43 (95% CI 58.951–67.897) and at 12 months it was 72
57 (Po0.01; Wilcoxon matched-pairs signed-rank test; 95% CI 62.361–81.551). In patients uninfected with HCV/HBV, the mean baseline ALT value was 37
28 (95% CI 33.960–39.280) and at 12 months it was 32
41 (Po0.002; Wilcoxon matched-pairs signedrank test; 95% CI 28.276–38.683). (b) Per cent probability of developing LEE in subjects with and without viral hepatic markers. The Kaplan–Meier evaluation revealed a significant difference between the two groups (Po0.0001). After 6 months, survival rates of 83.4% and 96.9% were found in HCV- and/or HBV-positive patients and HCV/HBV-negative patients, respectively. (c) Per cent probability of developing LEE in subjects with and without normal baseline ALT values. The Kaplan–Meyer evaluation revealed a significant difference between the two groups (Po0.0001). After 6 months, survival rates of 84.2% and 94.1% were found in patients with elevated baseline ALT and patients with normal baseline ALT, respectively.

Patients who developed LEE were compared with those who did not for sex, age, CDC stage, HIV risk factor, previous HBV infection, baseline CD4 T-cell count, HIV viral load, triglyceride, cholesterol, GGT, AP, AST and ALT values, HBV and/or HCV coinfection and concomitant antiretroviral therapy (Table 4). Fifty-eight of 71 patients (81.7%) developing LEE presented HCV and/or HBV coinfection: 50 subjects were HCVAb positive, three were HBsAg positive and five had HCV and HBV coinfection. The Kaplan–Meier plot (Fig. 1b, c) shows that there was a significantly higher probability of developing LEE in the group with chronic viral hepatitis and elevated pretreatment ALT values. HCV- and/or HBV-positive patients had a 4.78-fold greater risk of developing LEE compared with HCV- and/or HBV-uninfected patients [Po0.0001; 95% confidence interval (CI) 2.67–8.585]. Other risk factors related to LEE were younger age (P 5 0.008), higher baseline ALT and GGT levels (Po0.0001 and P 5 0.004, respectively), lower baseline total cholesterol value (P 5 0.0096), previous HBV infection [P 5 0.012; relative risk (RR) 1.85; 95% CI 1.139–3.018], the use of intravenous drugs in the past (P 5 0.0006;

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

LEE onset in LPV/r treatment 339

Table 3 Clinical data for hospitalized patients or patients with LEE  grade 3 Number of patients (sex)

Mean CD4 at baseline (cells/lL)

Mean HAART (days)

NNRTI associated

HCV Ab Positive

13 (3F, 10M)

194
122

151
131

2 patients

HCV Ab Negative

5 (1F, 4M)

267
158

72
67

1 patient

Mean ALT value (IU/L)*

Liver disease 2 4 4 3 5

cirrhosisw fibrosisz steatosisz negative negative

521
394

620
683

*Higher ALT value during follow up; wliver biopsy; zultrasonography diagnosis.

P-values for parameters in the two groups were not statistically significant. M, male; F, female; NNRTI, nonnucleoside reverse transcriptase inhibitor.

Table 4 (a) Biochemical, immunovirological and clinical parameters in patients with and without LEE; (b) antiretroviral therapy associated with LPV/r (a)

All Male Female Stage of HIV A B or C Intravenous drug users in past HCV and/or HBV coinfection Previous HBV infection Concomitant therapy with EFV Age (years) CD4 count (cells/mL)§ Viral load (log10 copies/mL)§ ALT (IU/L)§ Triglycerides (mg/dL)§ Gammaglutamyltranferase (UI/L)§ Alkaline phosphatase (UI/L)§ Total cholesterol (mg/dL)§

Patients with LEE

Patients without LEE

n (%) 71 (9.1) 52 (73.2) 19 (26.8)

n (%) 711 (90.9) 538 (75.6) 173 (24.4)

21 (29.6) 50 (70.4) 44 (61.9) 58 (81.7) 39 (61.9) 13 (18.3) Mean
SD 38
4.8 211
164 4.35
1.14 66
41 191
115 130
142 214
127 150
47

229 (32.2) 482 (67.8) 287 (40.4) 298 (41.9) 302 (45.3) 74 (10.4) Mean
SD 41
8 241
207 4.26
1.1 47
38 242
220 99
122 211
127 181
58

P

0.661*

0.693* 0.0006* o0.0001* 0.012* 0.040* 0.008w 0.390w 0.370w o0.0001w 0.161w 0.0047w 0.950w 0.0096z

*Fisher’s exact test; wMann–Whitney test; zunpaired t-test.

§ All these parameters are at baseline. ALT, alanine aminotransferase; SD, standard deviation.

(b)

n (%) Patients with LEE

Patients without LEE

ZDV 1 3TC vs. 3TC 1 d4T ZDV 1 3TC vs. ddI 1 d4T 3TC 1 d4T ZDV 1 3TC DdI 3TC D4T ZDV NVP EFV

17 vs. 8 17 vs. 12 6 (8.4) 18 (25.3) 28 (39.4) 34 (47.8) 29 (40.8) 23 (32.4) 3 (4.2) 13 (18.3)

179 vs. 110 179 vs. 147 112 (15.7) 178 (25) 282 (39.6) 382 (53.7) 315 (44.3) 212 (29.8) 54 (7.5) 74 (10.4)

Protease inhibitors

10 (13.1)

66 (86.9)

*ZDV, zidovudine; 3TC, lamivudine; d4T, stavudine; ddI, didanosine; NVP, nevirapine; EFV, efavirenz.

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

P 0.66 0.84 0.11 1.0 1.0 0.3 0.3 0.6 0.4 0.040 RR 1.79; 95% CI 1.024–3.131 0.20

340 P Meraviglia et al.

RR 2.22; 95% CI 1.405–3.510) and concomitant therapy with efavirenz (P 5 0.04; RR 1.79; 95% CI 1.024–3.131) (Table 4a). No significant association was detected between LEE and sex, CDC stage, baseline viral load, CD4 cell count, AP value, triglyceride value, nevirapine use, PI therapy or nucleoside reverse transcriptase inhibitor (NRTI) therapy (Table 4b). No statistical correlation was found between development of LEE and CD4 cell count during follow up. In the multivariate analysis, HCV and/or HBV coinfection [odds ratio (OR) 4.72; 95% CI 2.47–9.01; Po0.0001], younger age (OR 0.66; 95% CI 0.49–0.88; P 5 0.005), higher baseline ALT value (OR 2.17; 95% CI 1.27–3.72; P 5 0.005) and concomitant use of efavirenz (OR 2.57; 95% CI 1.27–5.17; P 5 0.008) remained predictive for the development of LEE. In the analysis of LEE onset time, patients developing LEE within the first 30 days of LPV/r therapy had no added risk factors towards patients developing LEE after one month of LPV/r treatment.

HCV- and/or HBV-coinfected patients In the univariate analysis, age, sex, baseline CD4 cell count, ALT, triglyceride and total cholesterol values, concomitant

antiretroviral therapy and previous HBV infection were not related to the development of LEE (Table 5). During follow up, no statistically significant difference in CD4 cell count increase was found between coinfected patients who did and did not develop LEE. Analysis of the liver biopsy specimens did not show a statistically significant association between the histological data for the 15 patients who developed LEE and those for the 55 subjects who did not develop LEE (P 5 0.9; w2 test for independence). There were also no correlations between LEE onset and ultrasonography diagnosis (P 5 0.58; Fisher’s exact test).

Discussion Antiretroviral drugs, including PIs, have frequently been associated with hepatotoxicity, as demonstrated in several studies [1,2,5–10,19–21]. In a recent work, den Brinker et al. reported that, under HAART, patients with chronic hepatitis B and/or C developed LEE in 37% of cases compared to 12% of patients without such coinfection [4]. In the literature, 8–35% of HIV-positive subjects have been reported to be HCV-coinfected [22–24]; the higher frequency (38.9%) revealed by our data is attributable to a higher rate of intravenous drug use.

Table 5 Comparison of biochemical, immunovirological and clinical parameters between patients developing LEE and those not developing LEE in the subgroup of HBV- or HCV-coinfected patients

All Male Female Stage of HIVz A B or C CD4 count o200cells/mLz Concomitant therapy with EFV Previous HBV infection Gradingz Minimal chronic hepatitis Mild chronic hepatitis Moderate chronic hepatitis Stagingz F0 F 1 Ultrasonographyz Normal Not normal Age (years)z CD4 count (cells/mL)z Viral load (log10 copies/mL)z ALT (IU/L)z

Patients with LEE

Patients without LEE

n (%) 58 40 (68.9) 18 (31.1)

n (%) 298 226 (75.8) 72 (24.1)

16 42 31 10 32

210 88 152 28 90

(27.5) (72.5) (53) (17.2) (64)

(70.4) (29.6) (51) (9.3) (64.5)

P

0.321*

0.874* 0.775* 0.100* 1*

2 7 6

7 26 22

1*

0 15

2 53

0.412*

22 36 Mean value
SD 38
5 213
162 4.28
1.1 71
43

97 201 Mean value
SD 39
5 224
182 4.27
1.1 63
43

0.441*

0.228w 0.838w 0.925w 0.051w

*Fisher’s exact test; wMann–Whitney test. z

All these parameters are at baseline.

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

LEE onset in LPV/r treatment 341

We observed LEE onset in 9.1% of treated subjects (71 patients); toxicity was mainly grade 2, with an incidence of 16.3% in coinfected patients versus 3.3% in uninfected patients. On analysing the risk factors, we found that HCVand/or HBV-positive subjects had a relative risk of 4.78. Similarly, Arribas et al. described, in antiretroviralnaive patients, a relative risk of 5.2 for development of a grade 3–4 hepatotoxicity based on baseline hepatitis status [15]. The rate of LPV/r discontinuation for LEE in our study was low (1.66% of subjects), with an incidence of 2.53% in coinfected subjects versus 0.93% in negative subjects. Among the patients who stopped LPV/r because of side effects, the incidence of hepatotoxicity was 14.7%; this value was lower than that found for LPV/r interrruption for other reasons. In HCV- and/or HBV-positive patients, no correlation between LEE and the risk factors analysed in the whole study population was found, supporting the importance of coinfection itself in LEE development. Cell-mediated immunity has a central role in the pathogenesis of HBV liver disease. In fact, HIV immune suppression could increase HBV replication in HBV carriers [4]. The immunological improvement produced by HAART therapy potentially increases the specific immunological response against HBV, causing greater liver damage. Recent reports have shown that the HBV genome is frequently detectable in an episomal form in HBsAgnegative and core antibody-positive patients (antiHBc), suggesting that a large number of these patients are latent HBV carriers, with viral replication in their hepatocytes [25]. The prevalence of latent HBV carriers among our patients is relevant; in our study, we found a significant correlation between LEE onset and previous HBV infection, with a relative risk of 1.85 in antiHBc-positive patients. It is possible that in these subjects there was latent liver damage, with greater hepatocyte susceptibility to drug toxicity. In contrast, the part played by HCV infection in increasing the risk of HAART-related hepatotoxicity is not completely clear and requires further evaluation. A higher incidence of HCV-RNA positivity, a higher HCVRNA viral load, greater liver fibrosis and faster progression to cirrhosis have been widely reported in HIV/HCVcoinfected patients compared with patients infected with HCV alone [12,13,26–28]. Liver damage pathogenesis in chronic HCV/HIV coinfection is still controversial. One hypothesis proposes that the immune recovery produced by PIs results in an increase in specific HCV CD4 and CD8 cell counts, inducing cytolysis in infected hepatocytes and the reactivation of HCV replication [2,12]. This has not been confirmed by other authors [29].

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

In our study, no significant correlation between increased CD4 cell count and the occurrence of LEE was found. Studies on chronic HCV-positive patients found that CD4 T cells in liver tissue are functionally and clonally different from NS4-specific T cells present in peripheral blood, suggesting a specific T-cell compartmentalization at the site of infection [30,31]. It is possible that, in patients developing hepatotoxicity, the immune reconstitution produced by PIs can increase the numbers of specific HCV CD4 T cells in liver tissue, causing cytolysis; this could not occur in peripheral blood. The other risk factors found in our study – higher baseline ALT and GGT values, lower cholesterol values, intravenous drug-use history and concomitant use of efavirenz – may indicate the presence of underlying impaired liver function which could be further impaired by potentially hepatotoxic drugs. A possible consequence is a modification of LPV/r metabolism, with an increase in LPV/r plasma levels and an adverse effect on liver function. However, in patients developing LEE within the first 30 days of LPV/r therapy, direct iatrogenic liver involvement is possible, rather than damage caused by viral hepatitis reactivation or by immune system reconstitution; a liver biopsy was not performed because of the rapid resolution of the event after LPV/r discontinuation. Some authors have recently demonstrated higher LPV plasma levels in patients with chronic hepatitis compared with HIV-positive control subjects [32,33]: therefore, therapeutic drug monitoring, although not available in our study, is recommended for all subjects presenting with liver damage. In conclusion, HCV and/or HBV coinfection is frequently associated with HIV infection. The higher risk of hepatotoxicity during antiretroviral therapy in these patients represents a real problem in the management of HIV infection. This clinical situation, the possibility of both coinfection and hepatotoxicity, must be carefully evaluated and monitored when an HIV-positive patient starts an antiretroviral regimen, especially one including PIs, or when the regimen is modified. In addition, therapeutic drug monitoring or a complete pharmacokinetic evaluation of PIs can be a useful tool in tailoring antiretroviral therapy for individual patients, in adjusting the drug dose and in monitoring patients to prevent drug-related liver damage in this clinical setting.

References 1 Save`s M, Vandentorren S, Daucourt V et al. Severe hepatic cytolisis incidence and risk factors in patients treated by antiretroviral combination. Aquitaine Cohort, France, 1996–98. AIDS 1999; 13: 115.

342 P Meraviglia et al.

2 Sulkowski M, Thomas D, Mehta S, Chaisson R, Moore RD.

from a phase III blinded randomized clinical trial. 5th

Hepatotoxicity associated with nevirapine or efavirenz-

International Congress on Drug Therapy in HIV Infection

containing antiretroviral therapy. Role of hepatitis C and B

Glasgow, October 2000 [Abstract P369].

infections. Hepatology 2002; 35: 182–189. 3 Rancinan C, Neau D, Save`s M et al. Is hepatitis C co-infection

16 Bernstein B, King M, Cernohous P, Moseley J, Sun E. Treatment

associated with survival in HIV-infected patients treated by combination antiretroviral therapy? AIDS 2002; 16: 1357–1362.

safety and efficacy comparison of LPV/r vs. Nelfinavir from a phase III blinded randomized clinical trial. 8th European

of HIV subjects co-infected with hepatitis B or C: week 60

4 den Brinker M, Wit FW, Wertheim-van Dillen PM et al.

Conference on Clinical Aspects and Treatment of HIV Infection.

Hepatitis B and C virus co-infection and the risk for hepatotoxicity of highly active antiretroviral therapy in HIV-1

Athens, October 2001 [Abstract P228]. 17 Ishak K, Baptista A, Bianchi L et al. Histological grading and

infection. AIDS 2000; 14: 2895–2902. 5 Martinez E, Blanco JL, Arnaiz J et al. Hepatotoxicity in HIV-1-

staging of chronic hepatitis. J Hepatol 1995; 22: 696–699. 18 Ishak K. Pathologic features of chronic hepatitis. A review and

infected patients receiving nevirapine-containing antiretroviral therapy. AIDS 2001; 15: 1261–1268. 6 Bonfanti P, Landonio S, Ricci E et al. Risk factors for

update. Am J Clin Pathol 2000; 113: 40–55. 19 Bonfanti P, Valsecchi L, Parazzini F et al. Incidence of adverse reaction in HIV patients treated with protease inhibitors: a

hepatotoxicity in patients treated with highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 27:

cohort study. J Acquir Immune Defic Syndr 2000; 23: 236–244. 20 Gisolf EH, Dreezen C, Danner SA, Well JL, Weverling GJ. Risk

316–318.

factors for hepatotoxicity in HIV-1-infected patients receiving

7 Sukkowski MS, Thomas DL, Chaisson RE, Moore RD.

ritonavir and saquinavir with or without stavudine.

Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of

Prometheus Study Group. Clin Infect Dis 2000; 31: 1234–1239. 21 Monforte A, Burgarini R, Pezzotti P et al. Low frequency of

hepatitis C or B virus infection. J Am Med Assoc 2000; 283: 74–80.

severe hepatotoxicity and association with HCV coinfection in HIV-positive patients treated with HAART. J Acquir Immune

8 Cooper CL, Parbhakar MA, Angel JB. Hepatotoxicity associated with antiretroviral therapy containing dual versus single

Defic Syndr 2001; 28: 114–123. 22 Soriano V, Kirck O, Antunes F et al. The influence of hepatitis C

protease inhibitors in individuals coinfected with hepatitis C

on the prognosis of HIV-infected persons: the EuroSIDA study

virus and human immunodeficiency virus. Clin Infect Dis 2002;

XIIIth International AIDS Conference. Durban, July 2000

34: 1259–1263. 9 Orenstein R, Tsogas N. Looking beyond highly active antiretroviral therapy: drug-related hepatoxicity in patients with human immunodeficiency virus infection.

[Abstract Th.OrB655]. 23 Thomas D. Hepatitis C and human immunodeficiency virus infection. Hepatology 2002; 36: S201–S209. 24 Ockenga J, Tillmann HL, Trautwein C, Stoll M, Manns MP,

Pharmacotherapy 2002; 22: 1468–1478.

Schmidt RE. Hepatitis B and C in HIV infected patients.

10 Bonnet F, Lawson-Ayayi S, Thiebaut R, Ramanampamionjy R,

Prevalence and prognostic value. J Hepatol 1997; 27: 18–24.

Lacoste D, Bernard N. A cohort study of nevirapine tolerance in

25 Marusawa H, Uemoto S, Hijikata M et al. Latent hepatitis B

clinical practice: French Aquitaine Cohort, 1997–99. Clin Infect

virus infection in healthy individuals with antibodies to

Dis 2002; 35: 1231–1237. 11 Vento S, Garofano T, Renzini C, Casali F, Ferraro T, Concia E.

hepatitis B core antigen. Hepatology 2000; 31: 488–495. 26 Benhamou Y, Di Martino V, Bochet M et al. Factors affecting

Enhancement of hepatitis C virus replication and liver damage in HIV-coinfected patients on antiretroviral combination

liver fibrosis in human immunodeficiency virus and hepatitis C virus-coinfected patients: impact of protease inhibitor therapy.

therapy. AIDS 1998; 12: 116–117. 12 John M, Flexman J, French MAH. Hepatitis C virus-associated

Hepatology 2001; 34: 283–287. 27 Nunez M, Rios P, Martin-Carbonero L, Perez-Olmeda M,

hepatitis following treatment of HIV-infected patients with

Gonzales Lahoz J, Soriano V. Role of hepatitis C virus genotype

protease inhibitors: an immune restoration disease? AIDS

in the development of severe transaminase elevation after the

1998; 12: 2289–2293.

introduction of antiretroviral therapy. J Acquir Immune Defic

13 Powderly WG, Landay A, Lederman MM. Recovery of the immune system with antiretroviral therapy: the end of opportunism? J Am Med Assoc 1998; 280: 72–77. 14 Kaletra (Lopinavir/Ritonavir) Product Monograph. Abbott

Syndr 2002; 30: 65–68. 28 Khalili M, Behm BW. Hepatitis C in the setting of HIV coinfection. Microbes Infect 2002; 4: 1247–1251. 29 Martin-Carbonero L, Nunez M, Rios P, Perez-Olmeda M,

Laboratories, 2001: 55–56.

Gonzales-Iahoz J, Soriano V. Liver injury after beginning

15 Arribas J, Barra C, Gonzales-Lahoz J et al. Treatment of

antiretroviral therapy in HIV/hepatitis C virus co-infected

HIV 1 subjects co-infected with hepatitis B or C: safety

patients is not related to immune reconstitution. AIDS 2002;

and efficacy comparison of ABT-378/ritonavir vs. Nelfinavir

16: 1423–1425.

r

2004 British HIV Association HIV Medicine (2004) 5, 334–343

LEE onset in LPV/r treatment 343

30 Chisari FV, Ferrari C. Viral hepatitis. In: Nathanson N, ed. Viral Pathogenesis. Philadelphia, Lippincott-Raven Publishers, 1997: 745–778. 31 Minutello MA, Pileri P, Unutmaz D et al. Compartmentalization

33 Gonzales de Requena D, Nunez M, Gonzales-Lahoz J, Soriano V. Liver toxicity of lopinavir/ritonavir containing

of T-lymphocyte to the site of disease: intrahepatic CD4 1

regiments in HIV infected patients with or without

T-cells specific for the protein NS4 of hepatitis C virus in patients with chronic hepatitis. J Exp Med 1993; 178: 17–25.

hepatitis C virus (HCV) co-infection. 9th European Aids Conference Warsaw, 25–29 October 2003

32 Arribas J, Pulido F, Peng JZ et al. Evaluation of the multipledose pharmacokinetics of lopinavir/ritonavir in HIV and HCV co-infected subjects with mild or moderate hepatic

r

insufficiency. 9th European Aids Conference Warsaw, 25–29 October 2003 [Abstract F2/6].

2004 British HIV Association HIV Medicine (2004) 5, 334–343

[Abstract 9.6/1].