Sex Differences, Cocaine Use, and Liver Fibrosis Among African Americans in the Miami Adult Studies on HIV Cohort

Abstract

Background: HIV infection disproportionally affects African Americans. Liver disease is a major cause of non-HIV morbidity and mortality in this population. Substance abuse accelerates HIV disease and may facilitate progression of liver disease. This study investigated the relationship between sex differences and cocaine use with liver injury, characterized as hepatic fibrosis.

Materials and Methods: A cross-sectional study was conducted on 544 African Americans [369 people living with HIV (PLWH) and 175 HIV seronegative] from the Miami Adult Studies on HIV (MASH) cohort. Cocaine use was determined with a validated self-reported questionnaire and confirmed with urine screen. Fasting blood was used to estimate liver fibrosis using the noninvasive fibrosis-4 (FIB-4) index.

Results: Men living with HIV had 1.79 times higher odds for liver fibrosis than women living with HIV (p = 0.038). African American women had higher CD4 count (p = 0.001) and lower HIV viral load (p = 0.011) compared to African American men. Fewer women (PLWH and HIV seronegative) smoked cigarettes (p = 0.002), and fewer had hazardous or harmful alcohol use (p < 0.001) than men. Women also had higher body mass index (kg/m²) (p < 0.001) compared to men. No significant association was noted among HIV seronegative participants for liver fibrosis by sex differences or cocaine use. Among African Americans living with HIV, cocaine users were 1.68 times more likely to have liver fibrosis than cocaine nonusers (p = 0.044).

Conclusions: Sex differences and cocaine use appear to affect liver disease among African Americans living with HIV pointing to the importance of identifying at-risk individuals to improve outcomes of liver disease.

Introduction

Sex differences have been noted for several diseases and disease outcomes for a variety of pathogens, mainly due to genetic and physiological factors, as well as behavioral lifestyle patterns.^1–3 Men are reported to consume more alcohol and use more illicit drugs compared to women,³ while women have been reported to have lower susceptibility to viral infection and heightened recognition and response to viruses due to stronger immune responses than men.⁴ Moreover, men are 6.5 times more likely to die from chronic liver disease and cirrhosis than women.⁵ Several studies have shown that among people living with HIV (PLWH), HIV RNA levels of women are significantly lower,^6–8 and the CD4 cell count is significantly higher.^9,10 In contrast, women experience more adverse reactions to antiretroviral therapy (ART)¹¹ and have a higher risk of progressing to AIDS.¹²

African American men and women are disproportionally affected by HIV compared with any other ethnic group in the United States.¹³ Florida is one of the states in the United States particularly burdened with the HIV epidemic. In 2017, Florida had the highest number (49,547) of African Americans living with HIV, the highest number of new diagnoses of HIV infection (16,690), and the highest number of deaths associated with HIV (6,795) in the United States.¹⁴ HIV infection has turned into a manageable condition for individuals with access to ART increasing their life span.¹⁵ As PLWH live longer and productive lives, they face a new set of adverse health outcomes. Liver disease is a significant cause of death in PLWH.^16,17 Liver injury, regardless of underlying etiology, is associated with hepatic fibrosis, a wound healing process characterized by deposition of the extracellular matrix and scar formation,¹⁸ leading to increased morbidity and mortality.¹⁹ HIV infection is implicated in the development of hepatic fibrosis even in the absence of hepatitis B virus (HBV) and hepatitis C virus (HCV), alcohol, and ART.²⁰

There are important distinctions between cocaine use behaviors and its effects by sex differences. Epidemiological data from the United States showed that the prevalence of cocaine use among women, but not men, increased from 2011 to 2015.²¹ While men and women seem to be equally susceptible to the pharmacological effects of cocaine,²² women appear to be more sensitive to the euphoric effects of cocaine,²³ which may explain why women progress from first use to dependence faster than men.²⁴ Cocaine-dependent women are also more likely to suffer from psychiatric, medical, and social problems than cocaine-dependent men.²⁵ Among African American women, those who are heavy drinkers are more likely than moderate drinkers to use crack cocaine.²⁶

Cocaine contributed to the highest number of overdose deaths among African Americans from 2000 to 2015 in the United States.²⁷ The recent South Florida Medical Examiners' report shows that the number of cocaine-related deaths has more than doubled since 2003 in Miami-Dade County, Florida, the location of the Miami Adult Studies on HIV (MASH) cohort.²⁸ Cocaine use has been shown to lower access to HIV treatment, reduce ART adherence,^29,30 and enhance HIV disease progression.³¹ In addition, we have shown that cocaine use [odds ratio (OR) = 5.10, 95% confidence interval (CI) 1.60–16.10, p = 0.006] and liver fibrosis (OR = 1.53, 95% CI 1.20–1.90, p < 0.001) were independent predictors of mortality in PLWH in the MASH cohort.³² Cocaine produces cytokine alterations³³ and promotes fibrotic changes in hepatic cells and in the livers of animals.^34–36 While numerous animal models and in vitro experiments have shown the adverse effects of cocaine use on the course and pathogenesis of HIV infection and liver disease,^33,37–40 studies examining the role of cocaine in liver disease progression in humans and the differences between men and women are lacking. Therefore, the objective of the study was to investigate the relationship of sex differences and cocaine use on hepatic fibrosis in African American adults.

Materials and Methods

Study population

This observational study analyzed data from the MASH cohort. A subset of 544 African Americans (369 PLWH and 175 HIV seronegative) from a larger cohort of >1,500 participants in the MASH cohort were included in this study. The study was approved by the Institutional Review Board at Florida International University. Inclusion criteria were 40 years of age or older, non-Hispanic blacks, and HIV status documented in the medical chart. The exclusion criteria were HBV or HCV infection documented in the medical chart, pregnancy, and excessive use (>twice/week) of other drugs besides cocaine determined by questionnaire and confirmed with urine drug screen.

Assessment of substance abuse and ART adherence

Cocaine and other illicit drugs, alcohol, and tobacco questionnaires were administered by trained interviewers. The drug use questionnaire was previously validated for accuracy in 98 participants, against blood metabolites and urine drug profiles, p < 0.0001.⁴¹ Urine drug screen was performed at the same visit. Urine was collected under observation at the study visit. The American Bio Medica Rapid Drugs (Rd, Kinderhook, NY) Kit was used to confirm self-report for cocaine and other illicit drugs. The Alcohol Use Disorders Identification Test (AUDIT) questionnaire is validated to detect hazardous or harmful alcohol use (AUDIT score ≥8).⁴² Adherence with ART and type of ART combinations were obtained from pharmacy records, and the AIDS Clinical Trials Group (ACTG) validated adherence questionnaires to ascertain 12-day recalls. Adherence with ART was defined as never missing any type of ART combinations in the past week. Records of medications were also obtained from medical charts for all participants with their written consent and were followed via a standardized toxicity grade scale used by the ACTG.

Anthropometries

Weight and height were obtained in light clothing and no shoes, utilizing a standard scale calibrated before each measurement, to calculate body mass index (BMI) using the standard formula dividing weight in kilograms by the square of height in meters (kg/m²).

Laboratory assessments

All blood samples were drawn fasting and sent for routine laboratory procedures to Quest Diagnostics Laboratory [chemistry, metabolic panels, and complete blood count (CBC)]. HIV disease progression parameters (CD4 cell count and HIV viral load) were abstracted from the medical charts with the participants' written consent. Metabolic Panel and CBCs included liver enzymes and platelet count (PLT).

Staging liver disease

Liver fibrosis was estimated using the fibrosis-4 (FIB-4) index. FIB-4 was calculated as follows: [age (years) × aspartate aminotransferase (AST) (U/L)]/[PLT (10⁹ cells/L) × alanine aminotransferase^1/2 (U/L)].⁴³ Stages of liver fibrosis for this study were defined as FIB-4 < 1.45 corresponding to no fibrosis and ≥1.45 to fibrosis.^43,44 The negative predictive value using a cutoff for FIB-4 < 1.45 to exclude individuals with advanced liver fibrosis was reported to be 90% with 70% sensitivity.⁴³ Sterling et al.,⁴³ reported that 87% of individuals would be properly classified with FIB-4 values <1.45 (absence of advanced fibrosis) or >3.25 (advanced fibrosis). However, only a small percentage of participants in this study had FIB-4 > 3.25 as follows: two cocaine users living with HIV, seven cocaine nonusers living with HIV, and two HIV seronegative cocaine users, and none of the HIV seronegative cocaine nonusers had FIB-4 > 3.25. Thus, due to this small number of the participants with FIB-4 > 3.25, we merged the FIB-4 index categories into FIB-4 < 1.45 to classify the groups as “no fibrosis” and ≥1.45 as “fibrosis.”

Statistical analyses

All analyses were performed with the Statistical Package for the Social Sciences (SPSS) version 24 (IBM, Armonk, NY). Data were examined for distributional normality and outliers before any analyses. Descriptive statistics was generated for all variables of interest included in the analysis, by sex (men/women), by groups (HIV⁺/HIV⁻), and by cocaine use (yes/no). Comparisons consisted of chi-square for categorical variables and t-test (or nonparametric Mann–Whitney) for continuous variables. Logistic regression analysis was used to estimate ORs and 95% CIs to assess the associations of sex (men/women) and cocaine use (yes/no) with liver fibrosis (defined as FIB-4 ≥ 1.45) by study groups (HIV⁺ and HIV⁻) in unadjusted and adjusted models. We included covariates in the logistic regression models that could potentially confound the relationship with liver fibrosis. Confounding variables used to build the final predictive models in the logistic regression analysis were BMI, unsuppressed viral load defined as HIV-1 RNA viral load ≥400 copies/mL, and hazardous or harmful alcohol use defined as AUDIT score ≥8. All associations were considered significant at the alpha level of 0.05.

Results

As shown in Table 1, the combined African American women living with HIV and HIV seronegative women (N = 246) had significantly higher BMI (kg/m²) (p < 0.001), lower percentage of smokers (p = 0.002), fewer engaged in hazardous or harmful alcohol use (p < 0.001), had higher CD4 cell count (p = 0.001), lower HIV viral load (p = 0.011), and fewer had liver fibrosis (p < 0.001) than African American men (N = 298).

Table 1.

Characteristics of Combined People Living with HIV and HIV Seronegative African Americans from the Miami Adult Studies on HIV Cohort by Sex

Variables	Women (N = 246)	Men (N = 298)	p
Age (years)	52.0 (48.0–57.0)	53.0 (48.1–57.3)	0.472
PLWH	156 (63.4%)	213 (71.5%)	0.045
BMI (kg/m2)	31.2 (26.8–36.9)	26.8 (24.1–30.7)	<0.001
Smoking	161 (65.4%)	230 (77.2%)	0.002
AUDIT ≥8	45 (18.4%)	100 (33.7%)	<0.001
Cocaine use	92 (38.7%)	135 (46.6%)	0.068
CD4+ cells/μLa	612.0 (346.2–888.2)	402.0 (265.7–632.2)	<0.001
HIV-1 viral load ≥400 copies/mLa	28 (18.7%)	63 (30.6%)	0.011
Receiving ARTa	142 (91.0%)	191 (90.1%)	0.763
Reported ART adherencea	83 (55.9%)	105 (57.6%)	0.686
FIB-4 ≥ 1.45	39 (16.3%)	97 (33.9%)	<0.001

Data presented as n (%) or median and interquartile range.

^aResults only for PLWH.

ART, antiretroviral therapy; AUDIT, Alcohol Use Disorders Identification Test; BMI, body mass index; PLWH, people living with HIV; FIB-4, fibrosis-4.

As shown in Table 2, there were 369 PLWH and 175 HIV seronegative African Americans in this study. PLWH were younger (p < 0.001), majority were men (p = 0.045), had lower BMI (kg/m²) (p < 0.001), used less cocaine (p < 0.001), and had more liver fibrosis (p < 0.001), compared to the HIV seronegative group.

Table 2.

Characteristics of People Living with HIV and HIV Seronegative African Americans from the Miami Adult Studies on HIV Cohort

Variables	PLWH (N = 369)	HIV seronegative (N = 175)	p
Age (years)	51.0 (47.0–56.0)	55.0 (52.0–60.0)	<0.001
Sex: men	213 (57.7%)	85 (48.6%)	0.045
BMI (kg/m2)	27.9 (24.5–32.0)	30.5 (25.9–35.7)	<0.001
Smoking	266 (72.1%)	125 (71.4%)	0.873
AUDIT ≥8	103 (28.1%)	42 (24.0%)	0.317
Cocaine use	131 (36.1%)	96 (58.2%)	<0.001
CD4+ cells/μL	468.0 (301.0–738.0)	—	—
HIV viral load ≥400 copies/mL	91 (25.6%)	—	—
Receiving ART	333 (90.5%)	—	—
Reported ART adherence	188 (56.3%)	—	—
FIB-4 ≥ 1.45	109 (31.0%)	27 (15.5%)	<0.001

Data presented as n (%) or median and interquartile range.

Table 3 shows cocaine users had significantly lower BMI (kg/m²) (p = 0.014), engaged more in hazardous or harmful alcohol use (p = 0.005), and had higher rates of unsuppressed viral load (≥400 copies/mL) (p = 0.019) compared to cocaine nonusers. A relatively small sample of participants used other drugs than cocaine as follows: methamphetamine users living with HIV, n = 10 (1.8%), HIV seronegative methamphetamine users, n = 2 (0.3%), heroin users living with HIV, n = 24 (4.4%), and HIV seronegative heroin users, n = 11 (2.0%). Due to these small sample sizes, we did not include methamphetamine and heroin in the statistical analyses.

Table 3.

Characteristics of Combined People Living with HIV and HIV Seronegative African Americans from the Miami Adult Studies on HIV Cohort by Cocaine Use

Variables	Cocaine nonuser (N = 317)	Cocaine user (N = 227)	p
Age (years)	51.0 (46.8–55.2)	51.0 (47.0–56.0)	0.256
BMI (kg/m2)	28.1 (24.9–33.0)	27.4 (23.9–31.0)	0.014
Smoking	198 (65.8%)	179 (78.9%)	0.001
AUDIT ≥8	66 (21.9%)	74 (32.9%)	0.005
CD4+ cells/μLa	478.0 (309.0–753.0)	454.0 (292.7–685.2)	0.948
HIV-1 viral load ≥400 copies/mLa	47 (21.0%)	41 (32.3%)	0.019
Receiving ARTa	209 (90.1%)	121 (93.1%)	0.336
Reported ART adherencea	124 (59.3%)	62 (50.8%)	0.132
FIB-4 ≥ 1.45	67 (23.3%)	64 (28.7%)	0.163

Data presented as n (%) or median and interquartile range.

^aResults only for PLWH.

Unadjusted ORs indicate that men living with HIV were 1.91 times more likely to have liver fibrosis (OR = 1.91, 95% CI 1.20–3.06, p = 0.006) than women living with HIV. In addition, cocaine users were 2.96 times more likely to have liver fibrosis than cocaine nonusers (OR = 2.96, 95% CI 1.80–4.87, p < 0.001) among PLWH. No significant association was noted among HIV seronegative participants for liver fibrosis by sex differences or cocaine use.

Table 4 shows adjusted logistic regression model to determine the association of sex differences and cocaine use, on the likelihood of liver fibrosis in PLWH adjusting for BMI, hazardous or harmful alcohol use, and unsuppressed viral load. The logistic regression model was statistically significant, χ²(5) = 48.72, p < 0.001. The model explained 19.1% (Nagelkerke R²) of the variance in liver fibrosis and correctly classified 71.8% of cases. Men living with HIV had 1.79 times higher odds of exhibiting liver fibrosis than women living with HIV (p = 0.038). Among African Americans living with HIV, cocaine users were 1.68 times more likely to have liver fibrosis compared to cocaine nonusers (p = 0.044). PLWH with higher BMI (OR = 0.91, 95% CI 0.86–0.96, p < 0.001) had lower odds for higher liver fibrosis, while unsuppressed viral load (OR = 2.61, 95% CI 1.49–4.57, p = 0.001) increased the likelihood of exhibiting liver fibrosis. The interaction effect of cocaine use * unsuppressed viral load was not associated with liver fibrosis in adjusted logistic regression model for PLWH. Table 5 shows that there were no significant differences in the likelihood of liver fibrosis among the HIV seronegative participants by sex differences and cocaine use in the adjusted logistic regression model.

Table 4.

Adjusted Logistic Regression Model Predicting the Likelihood of Liver Fibrosis (FIB-4 ≥ 1.45) in People Living with HIV

Variables	OR	95% CI		p
		Lower	Upper
Sex: men	1.79	1.03	3.10	0.038
Cocaine use (yes)	1.68	1.01	2.79	0.044
BMI (kg/m2)	0.91	0.86	0.96	<0.001
HIV-1 viral load ≥400 copies/mL	2.61	1.49	4.57	0.001
AUDIT score ≥8	0.90	0.52	1.55	0.696

CI, confidence interval; OR, odds ratio.

Table 5.

Adjusted Logistic Regression Model Predicting the Likelihood of Liver Fibrosis (FIB-4 ≥ 1.45) in HIV Seronegative Participants

Variables	OR	95% CI		p
		Lower	Upper
Sex: men	1.90	0.75	4.85	0.176
Cocaine use (yes)	0.81	0.30	2.17	0.683
BMI (kg/m2)	1.03	0.97	1.10	0.343
AUDIT score ≥8	1.47	0.52	4.19	0.467

Discussion

Men living with HIV had 1.79 times higher odds for liver fibrosis than women living with HIV. Differences in hepatic fibrosis between men and women may represent a complex interplay between the host and its response to pro-inflammatory extrinsic factors, including markers of HIV disease progression and lifestyle factors, which increase oxidative stress and result in increased hepatocyte apoptosis, collagen deposition, and hepatic fibrosis.^37,45–48 Hepatic oxidative stress, predominantly produced in the mitochondria, induces apoptosis and fibrosis.^49–51 Apoptosis, or programmed cell death of hepatocytes, contributes to fibrogenesis and correlates with the progression of liver fibrosis.⁵² The results of our study are congruent with findings from other investigations where women had a slower progression of liver fibrosis compared to men.^53,54

Our results on higher CD4 cell count and lower HIV viral load in African American women agree with previous literature, which shows that biological and hormonal dissimilarities are probably responsible for sex differences in these variables.^7,9,55 Women exhibit higher cytotoxic T cell activity and upregulated expression of antiviral and pro-inflammatory genes due to estrogen-response elements in their promoters.⁵⁵ In other studies, women's HIV RNA levels were lower compared to those of men^7,9; however, there were no differences between men and women in the risk of HIV disease progression.⁷ The MASH cohort participant-reported adherence to ART, although relatively low (55.9% among women and 57.6% among men), did not show differences between sexes. However, women reported less harmful and hazardous alcohol use, less cigarette smoking, and slightly less use of cocaine than men, most likely accounting for better HIV and liver disease outcomes than those observed among men. These behaviors are consistent with those reported by other studies for lower rates of alcohol consumption, smoking, and cocaine use³ by women, as well as higher cocaine overdose-related deaths among men.²⁷ Unsuppressed viral load (≥400 copies/mL) was significantly associated with higher odds of liver fibrosis in this study. This finding is consistent with a previous report showing that PLWH with unsuppressed viral load had higher odds for liver injury compared to those with suppressed viral load.⁵⁶ We explored the interaction effect of cocaine use*unsuppressed viral load with liver fibrosis and found that it was not associated with liver fibrosis in adjusted logistic regression model for PLWH. Therefore, in our population, the effect of cocaine use on liver fibrosis appears to be independent of the unsuppressed viral load.

In the MASH cohort, sex differences in BMI were noted. African American women had significantly higher BMI than men. This sex difference is consistent with a recent report by the CDC showing higher prevalence rates of obesity in African American women (54.8%) compared to men (36.9%).⁵⁷ In addition, we found that a lower BMI in African American men was associated with higher odds of having liver fibrosis. Paradoxically, a higher BMI has been reported to be a risk factor for liver fibrosis.^58,59 However, in our analyses, lower BMI in men was accompanied by higher HIV viral load, lower CD4 cell count, and more cocaine use that may account for the higher liver fibrosis among African American men. This is in agreement with Quach et al.,⁶⁰ who reported that cocaine use was associated with lower BMI, food insecurity, and microbial translocation in HIV positive individuals.

This is one of the first studies showing a significant effect of cocaine use on increased hepatic fibrosis. Among African Americans living with HIV from the MASH cohort, cocaine users were 1.68 times more likely to have liver fibrosis than cocaine nonusers. Our data indicate that cocaine imposes additional complexity to that posed by the effects of HIV infection on liver disease. Cocaine use creates a continued stimulus for liver fibrosis due to its addictive nature⁶¹ and lack of long-term effective cessation treatment programs.^62,63 Cocaine increases HIV viral load,^31,37,64 activates cytokines associated with progression of liver fibrosis,³⁶ increases apoptosis³⁵ and oxidative stress,⁶⁵ and induces morphological changes in the liver.^33,37 The great majority of studies investigating mechanisms of liver disease progression has been conducted in vitro or in animal models. This study shows a distinct association between cocaine use and liver fibrosis in African Americans living with HIV. Our findings are not in agreement with those of a Canadian cohort of HIV/HCV coinfected individuals which showed that cocaine/crack use was not associated with liver disease progression.⁶⁶ However, in the study by Martel-Laferrière et al.,⁶⁶ HIV/HCV coinfected individuals had probably more competing risk for liver fibrosis. In addition, cocaine/crack use was obtained only from self-reports, relying on participants' willingness and truthfulness to disclose substance abuse information, possibly creating classification bias for cocaine use versus nonuse. In our study, cocaine use was determined with a validated self-reported questionnaire and confirmed with a urine screen test making it possible to objectively distinguish between cocaine use and nonuse.

Health disparities among African Americans are affected by biological, social, and behavioral factors that contribute to a higher risk for the development and progression of liver disease.^67–71 Cocaine is a highly addictive drug that rapidly causes dependence.⁷² African Americans are more likely to develop cocaine dependence within 2 years of cocaine use initiation compared to whites.⁷² Furthermore, racial differences are also observed in biomarkers of liver disease among those that consume alcohol.^69,70 African Americans showed elevated AST and gamma-glutamyl-transferase, even after adjustment for age, BMI, HCV/HBV, education, and smoking.^69,70 Although we found that cocaine users consume more alcohol than cocaine nonusers, alcohol intake was not associated with liver fibrosis in our study. Previous studies have shown that African Americans had high rates of liver cirrhosis.⁶⁹ African Americans were more frequently hospitalized for liver cirrhosis than whites,⁶⁸ and those with heavy drinking alcohol patterns had higher risk of mortality than whites.⁷¹

The cross-sectional design of this study is a limitation of this analysis. Another limitation was the use of an indirect marker (the FIB-4 index) to assess liver fibrosis. While the gold standard for determining the degree of liver injury is histological assessment of liver biopsy,⁷³ this method has its own limitations, including sampling variability,⁷⁴ cost, and risk for complications (e.g., bleeding and infection).⁷³ These drawbacks led to the development and validation of noninvasive biomarkers. We have previously found that the FIB-4 index performs better than several other noninvasive methods to predict significant liver fibrosis⁴⁴ and may be widely used for the screening of HIV infected populations.²⁰ The FIB-4 index <1.45 had a negative predictive value of 90% and sensitivity of 70% to exclude advanced fibrosis.⁴³ These findings need to be confirmed in longitudinal studies that follow the development of liver disease in a population that uses illicit drugs.

Conclusions

Findings from this study underline the importance of sex differences and cocaine use as risk factors for liver fibrosis in African Americans living with HIV. It is crucial to identify at-risk individuals, gain insights into the pathogenesis and outcomes of the disease, and to provide the basis for sex-specific antifibrotic interventions. These studies are critical for program development and for improving health and access to care in the African American population. Further studies are needed to determine how to reduce the adverse health consequences and decrease the prevalence of comorbidities such as liver disease among African American men and women living with HIV.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

National Institute on Drug Abuse (Grant No. 5U01DA040381).