| |
|
Nonalcoholic Fatty Liver Disease and the Development of Metabolic Comorbid Conditions in Patients With Human Immunodeficiency Virus Infection
|
| |
| |
Download the PDF here
Nonalcoholic fatty liver disease (NAFLD), defined as fat accumulation in the liver in the absence of excessive alcohol consumption, is an epidemic entity affecting approximately 25% of the world’s population [1].
After application of exclusion criteria (Figure 1), 485 patients were included in this study. At baseline, the prevalence of NAFLD was 38.1%. Severe hepatic steatosis affected 81 (16.7%) patients. Suspected significant liver fibrosis and cirrhosis affected 72 (14.8%) and 12 (2.5%) of the patients, respectively. Overall, 57 (11.9%) and 102 (21%) patients were exposed to didanosine and stavudine, respectively.
In our study, we found that PWH with NAFLD were more likely to acquire T2DM and dyslipidemia, with aHRs of 5.13 (95% CI, 2.14-12.31) and 2.35 (1.34-4.14), respectively. Moreover, suspected significant liver fibrosis was an independent predictor of T2DM development, with an aHR of 2.71 (95% CI, 1.11-6.61). A recent cross-sectional study of factors associated with liver fibrosis and steatosis in patients with HIV monoinfection demonstrated an association of T2DM with liver fibrosis (odds ratio, 3.78; 95% CI, 1.48-9.68) [36]. Our finding suggests a possible link between the progression of NAFLD-associated liver disease and insulin resistance in the context of HIV infection. Current use of PIs was an independent predictor of dyslipidemia development. This class of ART is associated with a less favorable lipid profile, in particular elevated triglyceride and total cholesterol levels, especially when boosted with ritonavir [37]. We also observed a trend for NAFLD to predict development of hypertension. Conversely, we did not observe any link between NAFLD and incident CKD. In the general population, a cohort study of 41 430 adults in Korea reported an aHR for the development of CKD of 1.22 (95% CI, 1.04-1.43) for sonographically diagnosed NAFLD. We speculate that the pathogenesis of CKD in PWH may be more complex, and HIV-related factors may overcome the effect of NAFLD in this setting [3].
The increasing burden of NAFLD in the HIV-infected population, combined with our findings of its impact on metabolic comorbid conditions, indicate that PWH with NAFLD may be at particularly high risk of cardiovascular-associated disease and death. Indeed, cardiovascular events are the leading cause of death in patients with NAFLD, and PWH have higher cardiovascular risk than the general population [39-41]. This is thought to reflect the effects of systemic inflammation and endothelial dysfunction, as well as the disproportionate presence of the established risk factors of insulin resistance and dyslipidemia. In addition, the effect of NAFLD on long-term outcomes in the HIV-infected population is still not completely understood. Importantly, in HIV-uninfected patients with NAFLD, the severity of liver fibrosis affects not only liver-related, but also all-cause mortality rates [39]. As such, guidelines recommend cardiovascular risk stratification for all patients with NAFLD, particularly those with liver fibrosis [42, 43]. Management of NAFLD in HIV infection may include treating modifiable metabolic risk factors, diet and exercise, and nutritional supplements as well as optimizing HIV-related factors and ART [44-46].
After adjustments, NAFLD and preexisting hypertension were independent predictors of T2DM development (Table 4). Incident hypertension was independently predicted by older age. NAFLD and current use of PIs as ART regimen were independent predictors of development of dyslipidemia. Incident CKD was independently predicted by older age and preexisting T2DM. We also conducted a multivariable analysis including suspected significant liver fibrosis as exposure. After adjustment for age (per 10 years; aHR, 1.14; 95% CI, .78-1.67), male sex (0.84; .35-1.97), and black ethnicity (2.15; .93-4.97), the independent predictors of T2DM development were suspected significant liver fibrosis (2.71; 1.11-6.61; P = .03) and preexisting hypertension (4.22; 1.81-9.80; P = .001). There was no effect of suspected significant liver fibrosis on development of other metabolic comorbid conditions (results not shown).
Nonalcoholic Fatty Liver Disease and the Development of Metabolic Comorbid Conditions in Patients With Human Immunodeficiency Virus Infection
JID April 2020 - Thomas Krahn,1 Myriam Martel,1 Ruth Sapir-Pichhadze,2 Nadine Kronfli,3 Julian Falutz,3 Giovanni Guaraldi,4 Bertrand Lebouche,3,5 Marina B. Klein,3
Philip Wong,1 Marc Deschenes,1 Peter Ghali,1 and Giada Sebastiani1,3,
1Division of Gastroenterology and Hepatology, Department of Medicine, McGill University Health Centre, Montreal, Quebec, Canada, 2Division of Nephrology, Department of Medicine, McGill University Health Centre, Montreal, Quebec, Canada, 3Chronic Viral Illness Service, Department of Medicine, McGill University Health Centre, Montreal, Quebec, Canada, 4University of Modena and Reggio Emilia, Modena, Italy, 5Department of Family Medicine, McGill University Health Centre, Montreal, Quebec, Canada
New Drugs for NASH & HIV
Abstract
Background
Cardiovascular and liver disease are main causes of death in people with human immunodeficiency virus (HIV) (PWH). In HIV-uninfected patients, nonalcoholic fatty liver disease (NAFLD) is associated with incident metabolic complications. We investigated the effect of NAFLD on development of metabolic comorbid conditions in PWH.
Methods
We included PWH undergoing a screening program for NAFLD using transient elastography. NAFLD was defined as a controlled attenuation parameter ≥248 dB/m with exclusion of other liver diseases. Incident diabetes, hypertension, dyslipidemia, and chronic kidney disease were investigated using survival analysis and Cox proportional hazards.
Results
The study included 485 HIV-monoinfected patients.
During a median follow-up of 40.1 months (interquartile range, 26.5-50.7 months), patients with NAFLD had higher incidences of diabetes (4.74 [95% confidence interval, 3.09-7.27] vs 0.87 [.42-1.83] per 100 person-years) and dyslipidemia (8.16 [5.42-12.27] vs 3.99 [2.67-5.95] per 100 person-years) than those without NAFLD.
With multivariable analysis, NAFLD was an independent predictor of diabetes (adjusted hazard ratio, 5.13; 95% confidence interval, 2.14-12.31) and dyslipidemia (2.35; 1.34-4.14) development.
Conclusions
HIV-monoinfected patients with NAFLD are at higher risk of incident diabetes and dyslipidemia. Early referral strategies and timely management of metabolic risk may improve outcomes.
Discussion
The current study, based on a cohort of HIV-infected patients consecutively screened for liver disease, shows that NAFLD predicts the development of important metabolic comorbid conditions, including T2DM and dyslipidemia. Furthermore, NAFLD severity, represented by suspected significant liver fibrosis, predicts development of T2DM. Our findings mirror what has already been reported in the general population and suggest that fatty liver is a central barometer of metabolic health in PWH as well [28]. To the best of our knowledge, ours is the first cohort study linking NAFLD with the development of important metabolic comorbid conditions.
NAFLD is increasingly recognized as the most frequent liver disease in people aging with HIV [13, 29-32]. One meta-analysis situates the prevalence of NAFLD in those with HIV monoinfection at 35% [6], higher than the 25% global prevalence reported for the general population [1]. In the same study, the pooled prevalence of significant liver fibrosis was 22%. One reason for this excess may be that PWH present with particularly high frequency of metabolic conditions. A higher incidence of T2DM in the HIV-infected population has been well described, though the role of the liver in this process has not been investigated. A longitudinal study with a median follow-up of 4 years reported a cumulative incidence of T2DM of 10% in PWH, compared to 3% in uninfected controls [8]. A 2018 meta-analysis reported a pooled incidence rate for T2DM of 13.7 (95% CI, 13-20) per 1000 PY of follow-up [33]. Frequent hypertension and a higher cardiovascular risk have also been consistently reported [10, 11]. Several virological and ART-related factors have been implicated in the pathophysiological mechanism of T2DM and hypertension in HIV infection, including chronic inflammation, immune reconstitution, and lipodystrophy [34]. Dyslipidemia is also common, owing to both HIV chronic infection and lifelong use of ART, particularly PIs [9]. CKD is a frequent complication of HIV infection, occurring in 3.5%-48.5% of the patients, owing to HIV infection itself and as a consequence of ART, such as tenofovir disoproxil fumarate [12, 27].
Our study provides novel longitudinal data on the increased risk of T2DM, hypertension and dyslipidemia in PWH with NAFLD. This is consistent with findings in the general population with NAFLD. A large meta-analysis, including 19 observational studies and 296 439 individuals followed up for ≥1 year, reported a 2-fold increased risk of incident T2DM in patients with NAFLD [4]. A community study including 3869 NAFLD subjects and 15 209 controls followed up for a median of 7 years found that a subject with NAFLD and without T2DM, hypertension or dyslipidemia was >2 times more likely (relative risk, 2.62; 95% CI, 2.31-2.96) to develop ≥1 of these comorbid conditions than an age- and sex-matched control [35].
In our study, we found that PWH with NAFLD were more likely to acquire T2DM and dyslipidemia, with aHRs of 5.13 (95% CI, 2.14-12.31) and 2.35 (1.34-4.14), respectively. Moreover, suspected significant liver fibrosis was an independent predictor of T2DM development, with an aHR of 2.71 (95% CI, 1.11-6.61). A recent cross-sectional study of factors associated with liver fibrosis and steatosis in patients with HIV monoinfection demonstrated an association of T2DM with liver fibrosis (odds ratio, 3.78; 95% CI, 1.48-9.68) [36]. Our finding suggests a possible link between the progression of NAFLD-associated liver disease and insulin resistance in the context of HIV infection. Current use of PIs was an independent predictor of dyslipidemia development. This class of ART is associated with a less favorable lipid profile, in particular elevated triglyceride and total cholesterol levels, especially when boosted with ritonavir [37]. We also observed a trend for NAFLD to predict development of hypertension. Conversely, we did not observe any link between NAFLD and incident CKD. In the general population, a cohort study of 41 430 adults in Korea reported an aHR for the development of CKD of 1.22 (95% CI, 1.04-1.43) for sonographically diagnosed NAFLD. We speculate that the pathogenesis of CKD in PWH may be more complex, and HIV-related factors may overcome the effect of NAFLD in this setting [3].
Liver and cardiovascular disease are the primary non-AIDS-related causes of disease and death among PWH [38]. Although HCV coinfection has driven much of the liver-related mortality in the past, the implementation of effective direct antiviral-agents has inverted this trend. The increasing burden of NAFLD in the HIV-infected population, combined with our findings of its impact on metabolic comorbid conditions, indicate that PWH with NAFLD may be at particularly high risk of cardiovascular-associated disease and death. Indeed, cardiovascular events are the leading cause of death in patients with NAFLD, and PWH have higher cardiovascular risk than the general population [39-41]. This is thought to reflect the effects of systemic inflammation and endothelial dysfunction, as well as the disproportionate presence of the established risk factors of insulin resistance and dyslipidemia. In addition, the effect of NAFLD on long-term outcomes in the HIV-infected population is still not completely understood. Importantly, in HIV-uninfected patients with NAFLD, the severity of liver fibrosis affects not only liver-related, but also all-cause mortality rates [39]. As such, guidelines recommend cardiovascular risk stratification for all patients with NAFLD, particularly those with liver fibrosis [42, 43]. Management of NAFLD in HIV infection may include treating modifiable metabolic risk factors, diet and exercise, and nutritional supplements as well as optimizing HIV-related factors and ART [44-46].
The main strength of our study is the longitudinal design, able to capture dynamics over time and providing novel data regarding the effects of NAFLD on development of metabolic comorbid conditions in PWH. Moreover, we included only consecutive patients in an ongoing screening program for liver disease at a single center.
We also acknowledge several limitations of our study. First, the study was retrospective. Second, we did not have more robust diagnostic tools to diagnose hepatic steatosis, such as liver biopsy or magnetic resonance imaging [14, 32]. However, it would be costly and time consuming to use these approaches in a large cohort. Moreover, transient elastography has not been widely validated against liver biopsy in the context of HIV infection. Although a CAP cutoff of ≥248 dB/m has a good sensitivity for detection of any grade steatosis, higher cutoff values have been proposed and its accuracy for grading steatosis is lower [14]. Third, we could not correlate our findings with actual cardiovascular outcomes. Fourth, given the limited number of outcomes, we were unable to study the influence of specific ART drugs and to include all metabolic comorbid conditions in the multivariable models. Fifth, BMI was missing for 22% of cases, so we could not account for its effect on multivariable models, although we found a tendency to predict incident T2DM, hypertension and CKD in univariable model. BMI might be an easy predictor to classify patients at risk of metabolic complications in clinical practice. Sixth, the observational design of this study does not allow for conclusions regarding causality. Specifically, our findings do not necessarily implicate NAFLD as a cause of diabetes. Finally, shared risk factors could contribute to NAFLD, T2DM, and dyslipidemia, including medication effects, genetics, type of diet, and other lifestyle factors.
In conclusion, our study suggests that NAFLD is an independent predictor for the development of important metabolic comorbid conditions in PWH. This finding configures NAFLD as a barometer for metabolic health in the setting of HIV infection. Metabolic comorbid conditions might perpetuate the NAFLD spectrum and contribute to increased cardiovascular risk in PWH. In the context of HIV infection, patients with NAFLD should be closely monitored for the development of metabolic outcomes and modifiable risks be managed in a timely fashion. Our findings might be considered a further argument to advocate for screening for NAFLD in PWH, as it is the case for patients with diabetes [47]. Indeed, the European AIDS Clinical Society guidelines recommended screening for NAFLD in PWH with metabolic syndrome, and expansion of these criteria to PWH with any metabolic comorbid condition has been proposed [48, 49]. The implications of our findings on long-term cardiovascular outcomes should be investigated in future studies.
Nonalcoholic fatty liver disease (NAFLD), defined as fat accumulation in the liver in the absence of excessive alcohol consumption, is an epidemic entity affecting approximately 25% of the world’s population [1]. In the general population, NAFLD is not only associated with liver-related disease and death related to cirrhosis and hepatocellular carcinoma, but it has also been linked with increased mortality rates associated with extrahepatic diseases, particularly cardiovascular disease [2]. Moreover, NAFLD is strongly associated with the metabolic syndrome and has been linked with increased risks of diabetes, hypertension, dyslipidemia, and chronic kidney disease (CKD) [3-5].
NAFLD is expected to increasingly comprise the burden of liver disease in people with HIV (PWH) owing to the widespread availability and effectiveness of direct-acting antivirals for the treatment of chronic hepatitis C. The prevalence of NAFLD in HIV monoinfection is currently estimated to be ≥35% [6]. The prolonged exposure to HIV infection and antiretroviral therapy (ART), as well as increased intestinal dysbiosis and bacterial translocation, are thought to contribute to the development of NAFLD in PWH, in addition to risk factors shared with the general population, including obesity, insulin resistance, and dyslipidemia [7]. Importantly, the classic metabolic risk factors for NAFLD are more frequent in PWH. In HIV-infected men, a 4 times higher prevalence of type 2 diabetes mellitus (T2DM) has been described, compared with that in HIV-uninfected men [8]. Dyslipidemia is also common. owing to both HIV chronic infection and lifelong use of ART [9]. Hypertension and a higher cardiovascular risk have also been consistently reported [10, 11]. HIV-infected patients are also at higher risk for CKD [12].
Given that HIV infection and NAFLD independently increase the risk of metabolic comorbid conditions, the aim of the study was to investigate the effect of NAFLD on development of T2DM, hypertension, dyslipidemia, and CKD in a cohort of PWH, by means of transient elastography with associated measurement of controlled attenuation parameter (CAP) [13]. This diagnostic tool for NAFLD and associated liver fibrosis has been validated against liver biopsy results in both HIV-uninfected and HIV-infected patients [14-16].
RESULTS
After application of exclusion criteria (Figure 1), 485 patients were included in this study. At baseline, the prevalence of NAFLD was 38.1%. Severe hepatic steatosis affected 81 (16.7%) patients. Suspected significant liver fibrosis and cirrhosis affected 72 (14.8%) and 12 (2.5%) of the patients, respectively. Overall, 57 (11.9%) and 102 (21%) patients were exposed to didanosine and stavudine, respectively. Table 1 reports the main demographic, clinical, biochemical and immunovirological characteristics of the study population by NAFLD status at baseline. Patients with NAFLD were older, more likely to be white, and had higher BMIs. Moreover, they had longer time since HIV diagnosis. Finally, they had higher prevalence of hypertension, lower high-density lipoprotein cholesterol, levels, higher triglyceride, glucose, alanine aminotransferase, aspartate aminotransferase, and albumin levels, and higher liver stiffness measurement. At baseline, patients with NAFLD had higher prevalences of hypertension and dyslipidemia (Table 1), and those with suspected significant liver fibrosis had higher prevalence of all metabolic comorbid conditions (Figure 2).
Incidence and Predictors of Metabolic Comorbid Conditions
Patients were followed up for a median of 40.1 months (interquartile range, 26.5-50.7 months). There were 68, 86, 190, and 84 patients with T2DM, hypertension, dyslipidemia, and CKD, respectively, who were excluded from the longitudinal analysis for having the outcome at baseline. Overall, incidence rates for T2DM, hypertension, dyslipidemia, and CKD were 2.2 (95% CI, 1.6-3.3), 4.2 (3.2-5.5), 5.3 (4.0-7.1), and 2.7 (1.9-3.8) per 100 PY, respectively. Tables 2 and 3 shows incidence rates for metabolic comorbid conditions by NAFLD and suspected significant liver fibrosis status. Patients with NAFLD had a higher incidence rate of T2DM and dyslipidemia than those without NAFLD (Figures 3A and 3C). There was also a tendency for a higher incidence rate of hypertension (Figure 3B), whereas there was no difference for CKD (Figure 3D).
| |
| |
| |
|
|
|
