Ancestry and Risk of Hepatic Fibrosis and Inflammation in Patients With HCV Infection

Worldwide, ~184 million people have chronic hepatitis C virus (HCV) infection.¹ Persistent racial disparities in outcomes are observed among HCV-infected patients. Hispanic patients with chronic HCV are more likely than non-Hispanic white (NHW) patients to develop advanced hepatic fibrosis and inflammation.^2,3 Conversely, black patients with HCV infection are at lowest risk. The factors that contribute to this racial disparity are multifactorial, including lifestyle, genetics, and medical care. Limited data in other diseases suggest that genetic ancestry determined using ancestry-informative markers (AIMs) may help explain racial and ethnic differences in disease risk or severity.⁴ AIMs are sets of single-nucleotide polymorphisms (SNPs) that determine a person’s ancestral continent of origin and the genetic ancestry proportions assigned to each individual serves as a proxy for his or her genetic ancestral background. We examined the risk of hepatic fibrosis and inflammation in HCV-infected patients according to both genetic ancestry and self-reported race/ethnicity.

Methods

This was a cross-sectional study among NHW, black, and Hispanic male patients with chronic HCV infection from a single Veterans Affairs (VA) hospital.⁵ Participants completed a study survey, underwent a blood draw and FibroSURE, and had anthropometric measurements taken. Patients with FibroSURE-determined fibrosis stage F3/F4–F4 were defined as advanced fibrosis cases. Those with F0–F3 were defined as mild fibrosis controls. We defined as advanced inflammatory activity cases those patients with inflammatory grades A2/A3–A3, and mild activity controls as those patients with inflammatory grades A0–A2.

We genotyped DNA from blood using the Illumina HumanOmni 2.5–8 microarrays (Illumina, San Diego, CA) and performed SNP clustering and quality-control steps. We identified ~10,000 autosomal informative unlinked markers of continental ancestry information in admixed populations,^6,7 of which we mapped 6436 to the HumanOmni 2.5–8 microarrays. After excluding the monomorphic SNPs, SNPs with call-rate <95% or with minor allele frequency <10%, there were 6406 AIMs left in the panel. Of these, 6186 common AIMs were available for the subjects from HapMap Project phase 3 dataset (n = 1115) from 4 different ancestral groups (239 European ancestry [EUA], 317 Asian ancestry [ASA], 488 African ancestry [AFA], and 71 Mexican ancestry [MXA] subjects) and 11 distinct populations. The final genome-wide AIM set included 2398 common AIMs from both our HCV-infected patients and the HapMap phase 3 reference subjects.

We performed principal component analysis and assigned patients to the closest HapMap population.⁸ Odds ratio and 95% confidence interval for associations among self-reported race and ethnicity, AIMs-defined genetic ancestry, and risks of advanced hepatic fibrosis and advanced hepatic inflammation were estimated using logistic regression. Analyses were performed using R version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria).

Results

We included 283 self-reported NHW, 154 black, and 22 Hispanic HCV-infected patients. The mean age was 57 years (range, 39–69). There were 270 (58.8%) patients with mild fibrosis and 189 (41.2%) with advanced fibrosis, and 248 (54.0%) with mild inflammation and 211 (46.0%) with advanced inflammation.

In the PCA analysis, 33.5% of total variation occurred between NHW or Hispanic and black populations compared with 0.7% between Hispanic and NHW or black populations. Patients were classified into 3 genetic ancestry groups (EUA, MXA, and AFA) and 4 populations (Utah residents with Northern and Western European ancestry from the Centre d’Etude du Polymorphisme Humain collection, Toscans in Italy, Mexican ancestry in Los Angeles, California [MEX], and African ancestry in the Southwest United States). While there was near perfect agreement between self-reported race and ethnicity and genetic ancestry or population for blacks and Hispanics, of 283 self-reported NHWs, 266 were AIMs classified as EUA ancestry (172 Utah residents with Northern and Western European ancestry from the Centre d’Etude du Polymorphisme Humain collection and 94 Toscans in Italy), 13 as MXA/MEX, and 4 as AFA or African ancestry in the Southwest United States.

In logistic models, the magnitude and direction of associations with advanced hepatic fibrosis and advanced hepatic inflammation were the same regardless of whether patients were defined by self-report or genetic ancestry (Table 1). Upon mutual adjustment, although the confidence intervals were wide, the point estimates for self-report race and ethnicity went to the null whereas the estimates for race and ethnicity defined by genetic ancestry or population remained unchanged.

Table 1.

Associations Among Self-Reported Race/Ethnicity, Ancestry, and Population With Advanced Hepatic Fibrosis or Inflammation Risk in Patients With Hepatitis C Virus Infection (N = 459)

	Advanced Fibrosis (F3/F4-F4, Cirrhosis vs F0-F3)			Advanced Inflammation (A0-A2 vs A3)
Variable	OR (95% CI)a	OR (95% CI)b	OR (95% CI)c	OR (95% CI)a	OR (95% CI)b	OR (95% CI)c
Race (self-reported)
Black	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)
White	0.73 (0.49–1.09)	0.80 (0.53–1.20)	1.14 (0.20–6.33)	1.12 (0.76–1.66)	1.18 (0.78–1.77)	0.93 (0.17–5.05)
Hispanic	0.81 (0.33–2.00)	0.71 (0.27–1.84)	1.06 (0.15–7.39)	1.05 (0.43–2.59)	0.83 (0.33–2.11)	1.04 (0.15–7.15)
Genetic ancestry
AFA	1 (Ref)	1 (Ref)	—	1 (Ref)	1 (Ref)	—
EUA	0.72 (0.49–1.08)	0.79 (0.52–1.19)	—	1.13 (0.76–1.68)	1.21 (0.80–1.82)	—
MXA	0.78 (0.37–1.64)	0.71 (0.32–1.54)	—	0.94 (0.45–1.97)	0.78 (0.36–1.68)	—
Genetic population
ASW	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)	1 (Ref)
CEU	0.69 (0.44–1.07)	0.77 (0.49–1.21)	0.68 (0.12–3.80)	1.12 (0.73–1.73)	1.22 (0.78–1.91)	1.31 (0.24–7.18)
TSI	0.79 (0.47–1.33)	0.83 (0.49–1.41)	0.73 (0.13–4.18)	1.15 (0.69–1.93)	1.20 (0.71–2.02)	1.29 (0.23–7.18)
MEX	0.78 (0.37–1.64)	0.71 (0.32–1.54)	0.65 (0.11–3.94)	0.94 (0.45–1.97)	0.78 (0.36–1.68)	0.78 (0.13–4.62)

NOTE. AFA, African ancestry; ASW, African ancestry in the Southwest United States; CEU, Utah residents with Northern and Western European ancestry from the Centre d’Etude du Polymorphisme Humain collection; CI, confidence interval; EUA, European ancestry; MEX, Mexican ancestry in Los Angeles, California; MXA, Mexican ancestry; OR, odds ratio; TSI, Toscans in Italy.

^aUnadjusted model.

^bEach variable examined in a model also adjusting for age group, body mass index, chronic alcohol abuse, diabetes, and log 10 viral load.

^cEach variable examined in a model also adjusting for age group, body mass index, chronic alcohol abuse, diabetes, and log 10 viral load and mutually adjusted for self-reported race and ethnicity and ancestry-informative marker–defined population.

Discussion

Race is a complex term and the way race and ethnicity is defined can have implications for how we interpret study findings.⁴ Here, we found that self-reported race and ethnicity was a good proxy for genetic ancestry in self-reported Hispanic and black patients, but there was poorer agreement among NHWs. Genetic ancestry inferred from AIMs may be a better predictor for advanced fibrosis in HCV-infected patients compared with self-report. By using genetic ancestry instead of self-report, misclassification may be avoided and we may more accurately characterize the real magnitude and source of observed racial disparity in risk. However, our findings are only suggestive due to the small sample size and potential lack of power. It is important to note the relative lack of association between race and degree of fibrosis observed here. This may be related to some common features of a VA-based cohort including period of acquisition, male sex, high background alcohol intake, and similar HCV genotypes. Further larger studies using genetic ancestry verification of race and ethnicity in non-VA HCV-infected patients are needed, and may have important implications for ongoing research examining disparities in these patients.