Relationship between hepatic enzymes and insulin resistance syndrome markers in HIV with lipodystrophy
To the Editor:
Some adverse effects associated with the highly active antiretroviral therapy (HAART) in human immunodeficiency virus (HIV) patients are the metabolic complications (lipodystrophy syndrome) such as dyslipidemia, abnormal fat distribution, and insulin resistance.1, 2 Several patients develop hepatotoxicity with noticeably elevated liver enzymes3; but other patients present normal or mildly elevated levels of liver enzymes, a fact that may obscure relevant information.4
We evaluated 131 HIV patients: 53 HIV naive of treatment (NAIVE group), 35 HIV under HAART without lipodystrophy (HAART LPD− group), and 43 HIV under HAART with lipodystrophy (HAART LPD+ group). We measured liver enzymes, and lipid and metabolic profile. Lipodystrophy was defined by visible changes in body fat distribution and/or substantial changes in lipid profile (particularly triglycerides). None of the patients had suffered ischemic heart disease or diabetes mellitus, nor were they alcohol abusers. Hepatitis C and B virus status, age, sex, body mass index (kg/m2), alcohol intake status, acquired immunodeficiency syndrome status, and drug addicts percentage were similar among the groups (P > .05).
Homeostasis model assessment (HOMA) was used to assess insulin resistance. Serum levels of uric acid, alkaline phosphatase, alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglycerides (TG), low-density lipoprotein cholesterol, apolipoprotein-B and lipoprotein(a) were measured by standard techniques. Plasma levels of plasminogen activator inhibitor-1 (PAI-1), whose elevated values are a core feature of the insulin resistance syndrome,5 were evaluated by enzyme-linked immunosorbent assay.
The results are shown in Table 1. Within HAART LPD+ group, the linear correlation revealed that both AST and ALT correlated (P < .05) with HOMA (Rho = 0.313 and Rho = 0.375, respectively), PAI-1 (Rho = 0.357 and Rho = 0.366, respectively), and TG (Rho = 0.338 and Rho = 0.324, respectively). In a stepwise multivariate analysis model including AST or ALT as dependent variables, only PAI-1 (β = 0.727, P < .001, and β = 0.482, P = .001, respectively) remained significant. Conversely, in neither NAIVE nor HAART LPD− groups did liver enzymes exhibit any correlation with the studied parameters.
| NAIVE Group n = 53 | HAART LPD− Group n = 35 | HAART LPD+ Group n = 43 | P | |
|---|---|---|---|---|
| Age (years) | 35.0 (29.0–40.5) | 35.0 (30.0–42.7) | 39.0 (33.0–41.0) | .121* |
| HCV+ (%) | 28.3 | 17.9 | 27.9 | .550† |
| BMI (kg/m2) | 24.4 (21.9–27.5) | 24.4 (22.5–26.4) | 23.3 (20.7–25.3) | .172* |
| APO-B (mg/dL) | 82.6 (67.3–93.5) | 107.0 (77.6–130.7) | 117.9 (95.4–133.4) | <.001* |
| Uric acid (mg/dL) | 5.0 (4.2–5.9) | 4.5 (4.0–5.3) | 5.6 (4.7–7.0) | .006* |
| ALKP (U/L) | 89 (78–130) | 93 (80–115) | 94 (74–120) | .875* |
| HOMA (mU.mmol/L2) | 2.1 (1.2–2.8) | 1.8 (1.4–2.4) | 2.4 (1.3–3.8) | .411* |
| LDL-chol (mg/dL) | 90 (76–122) | 147 (89–180) | 148 (118–175) | <.001* |
| Lp(a) (mg/dL) | 19.4 (9.6–39.3) | 34.1 (14.7–61.0) | 35.5 (12.3–82.0) | .073* |
| PAI-1 (ng/mL) | 48 (25–68) | 46 (36–79) | 40 (28–69) | .800* |
| TG (mg/dL) | 121 (85–164) | 147 (115–220) | 210 (129–344) | <.001* |
| AST (U/L) | 29 (25–44) | 29 (22–35) | 32 (24–41) | .379* |
| ALT (U/L) | 34 (27–51) | 30 (21–39) | 42 (30–58) | .008* |
- NOTE. Data are shown as medians with quartiles (25% and 75%) or corresponding percentages.
- Abbreviations: HCV, hepatitis C virus; BMI, body mass index; APO-B, apolipoprotein-B; ALKP, alkaline phosphatase; LDL-chol, low-density lipoprotein cholesterol; Lp(a), lipoprotein(a).
- * Kruskal-Wallis test was used to compare the 3 groups.
- † Chi-square test was used to compare the 3 groups.
Our results showed that HAART LPD+ group had significantly higher levels of ALT in comparison with any of the 2 remaining groups. The remarkable point was that these levels of ALT were within the normal range or slightly elevated and were in close relationship with 3 insulin resistance markers: HOMA, TG, and PAI-1. Recent studies have pointed out hyperinsulinemia and insulin resistance as pathogenic factors in nonalcoholic fatty liver diseases.6 Similarly, plasma PAI-1 concentrations increase in direct proportion to liver fat contents in lipodystrophic HIV patients.7 The lack of radiographic or histological evidence in our study turns the relationship between liver enzymes and fatty liver only into a hypothesis. Although AST levels did not show any substantial differences among the groups, its levels kept the same association as ALT levels with insulin resistance makers in HAART LPD+ group.
The range of “normal” values of hepatic enzymes was determined long time ago. The ALT and AST measure often fails to identify patients with minimal to mild necroinflammatory activity. When metabolic risk factors are taken into account, the range of hepatic enzymes, mainly ALT, is considerably reduced.8 Therefore, the “normal” range used in our study may have underestimated some degrees of liver diseases.
As a whole, in the present study, HAART LPD+ patients had a slight elevation of hepatic enzymes (ALT levels) compared with NAIVE or HAART LPD− patients. Such elevation was in close relationship with insulin resistance markers and essentially with PAI-1. These findings could represent indirect evidence of the presence of fatty liver.
References
Gabriela De Larrañaga*, Silvia Perés*, Jorge Benetucci , * Hemostasis and Thrombosis Laboratory, Hospital of Infectious Diseases “F. J. Muñiz”, Buenos Aires, Argentina, Infectious Diseases Department and Retrovirus Laboratory of FUNDAI, Hospital of Infectious Diseases “F. J. Muñiz”, Buenos Aires, Argentina.
