Incidence of Hepatitis C Virus and Other Bloodborne Virus Infections Among Nonurban Residents of Fairfield and New Haven Counties, Connecticut Who Inject Drugs

Abstract

Opioid overdose and substance abuse treatment entry data suggest that injection drug use is increasing in nonurban locations. We sought to explore the prevalence and incidence of viral infections among people who inject drugs (PWID) residing in Fairfield and New Haven counties but outside of the six largest cities. A longitudinal cohort of PWID was assembled and incidence of HIV-1, hepatitis B virus, and hepatitis C virus infections was determined by annual antibody screening. Data on participants’ socioeconomic situation and risk behaviors were collected. We identified 11 new hepatitis C virus infections and calculated incidence at 9.03 cases per 100 person-years. Only one new HIV infection and one new hepatitis B virus infection were detected. Factors associated with seroconversion were assessed. Given the high incidence of HCV and lack of HBV vaccination coverage, prevention and treatment resources need to be targeted to this population.

Introduction

Recent events in Scott County, IN, where 181 people who inject drugs (PWID) were diagnosed with HIV between December, 2014 and August, 2015,¹ have drawn attention to the expansion of injection drug use outside of urban areas. The outbreak has been linked to syringe sharing among those injecting the prescription opioid oxymorphone, and 114 (84.4%) of the first 135 people diagnosed with HIV were found to have been coinfected with hepatitis C virus (HCV).²

The high prevalence of HCV among PWID in rural Indiana mirrors trends in hepatitis C epidemiology nationwide, which has shown a shift towards increasing infection in suburban or rural PWID, many of whom are young.³ It is thought that these changing trends are responsible for the continuing HCV epidemic.⁴ We add to the literature on epidemic trends in HCV infection among PWID by presenting some of the first findings on the incidence of HCV in the nonurban population of injectors. Our study was conducted in the suburbs of southwestern Connecticut, which include some of the wealthiest communities in the US.

Methods

Between 2008 and 2012, we conducted a multiple-methods, longitudinal study of active PWID who presented evidence of injection drug use and residence of at least six months in the nonurban communities of Fairfield and New Haven Counties in Connecticut. Participants were recruited using respondent-driven sampling (RDS), a modified form of chain referral sampling. This method of sampling exploits existing social networks within hidden populations.^5,6 Individuals within the population known to the study were enrolled as “seeds” who, after completing all study procedures, were given four coupons to recruit others who met study eligibility criteria. Recruited individuals who were enrolled and completed study procedures were also given four coupons to recruit others, and the process continued until the recruitment period ended. Recruiters were provided modest remuneration for each recruit successfully enrolled.

Enrolled individuals completed a two-part baseline interview that included detailed questions on sociodemographics, injection drug use and injection risk behaviors, somatic and mental health, and interactions with the criminal justice and substance abuse treatment systems as described in earlier papers.⁷ At baseline, serological testing for HIV, hepatitis B virus (HBV), and HCV was conducted. Participants were then interviewed semiannually, and serological testing occurred on an annual basis. The study protocol was approved by the Human Investigation Committee at Yale University.

Incident cases of HIV, HBV, and HCV were identified, and incidence rates were calculated using the number of seroconversions and the total amount of person-time contributed. If a study participant had not seroconverted, the total amount of time they participated in the study was counted. If a participant had seroconverted, then their person-time contribution was calculated by dividing the amount of time between their enrollment in the study and their first positive test in half to account for not knowing exactly when seroconversion occurred.

Data were further analyzed using standard bivariate statistics to determine if we could identify any sociodemographic, behavioral, spatial, or service use factors associated with seroconversion events.

Results

During the study period of November 1, 2008 and January 31, 2012, a total of 462 eligible PWID were enrolled in the study. At baseline, 181 of 447 (40.5%) participants tested were serologically positive for HCV. Baseline seroprevalence for HIV-1 and HBV was 1.6% and 24.4%, respectively. Only 139 (31.2%) participant serologies revealed evidence of HBV immunization while 198 (44.4%) remained susceptible. Only 15 people (3.3%) had used a safe syringe program in the month prior to the baseline interview while 378 (81.8%) had engaged in some form of injection behavior that put them at risk for acquiring or transmitting a bloodborne virus. A summary of key characteristics of study participants appears in Table 1.

Table 1. Characteristics of Study Participants.

The number and proportion (for categorical variables) or mean and standard deviation (for continuous variables) for each characteristic is presented for all participants (n = 462) and those who were HCV-negative at baseline (n = 266)

		All Participants	HCV-Neg Participants
Sociodemographic Characteristics
Women		174 (37.7%)	93 (35.0%)
Age ± s.d.		35.6 ± 11.0	32.5 ± 10.0
Race/Ethnicity	African-American	29 (6.8%)	17 (6.4%)
	Hispanic	40 (8.6%)	27 (10.1%)
	White, non-Hispanic	389 (84.2%)	219 (83.3%)
	Other/Bi-racial	4 (0.9%)	3 (1.1%)
Educational achievement	Less than high dchool	87 (18.8%)	47 (17.7%)
	High dchool	195 (42.2%)	115 ($3.2%)
	More than high dchool	180 (40.0%)	104 (39.1%)
Employment status	Full-time	61 (13.2%)	39 (14.7%)
	Part-time	72 (15.6%)	47 (17.7%)
	Unemployed	329 (71.2%)	179 (67.3%)
Monthly income	< $500	148 (32.0%)	93 (35.0%)
	$500 – $999	99 (21.4%)	51 (19.2%)
	$1000 – $1999	133 (28.8%)	72 (27.1%)
	≥ $2000	82 (17.8%)	50 (19.5%)
Health insurance	None	97 (21.0%)	65 (24.4%)
	Private	40 (8.6%)	26 (9.8%)
	Government	277 (60.0%)	145 (54.5%)
	Missing	48 (10.4%)	30 (11.3%)
Health
Self-reported health status	Excellent/very good	154 (33.3%)	93 (35.0%)
	Good	174 (37.7%)	110 (41.3%)
	Fair/poor	134 (29.0%)	63 (23.7%)
Experiencing chronic pain1		152 (32.9%)	85 (32.0%)
Depression2	None/mild	244 (52.8%)	156 (58.6%)
	Moderate/severe	218 (47.2)	110 (41.4%)
Anxiety3	None/mild	345 (74.7%)	203 (76.3%)
	Moderate/severe	117 (25.3%)	63 (23.7%)
Experience of opioid overdose		143 (31.0%)	63 (23.7%)
Drug Use and Injection Practices
Drug of choice, past 30 days	Heroin	417 (90.3%)	242 (91.0%)
	Cocaine	26 (5.6%)	20 (7.5%)
Injection frequency, past 30 days4		50.6 ± 53.6 mode = 30	53.5 ± 55.6 mode = 31
Usual site for acquiring syringes, past 30 days	Pharmacy	343 (74.2%)	194 (72.9%)
	Syringe exchange program	16 (3.5%)	8 (3.0%)
	Other	100 (21.6%)	62 (23.3%)
Shared syringes, past 30 days5	At least once	101 (21.9%)	61 (22.9%)
	Never	360 (77.9%)	205 (77.1%)
Any form of unsafe injection, past 30 days	At least once	378 (81.8%)	210 (78.8%)
	Never	84 (18.2%)	56 (21.1%)
Interactions with Substance Abuse Treatment and Criminal Justice Systems
Any substance abuse treatment, ever		358 (77.5%)	185 (69.8%)
Short-term detoxification	Ever	307 (66.4%)	159 (59.8%)
	Last 6 months	79 (17.1%)	51 (11.0%)
Longer-term abstinence6	Ever	290 (62.8%)	150 (56.4%)
	Last 6 months	70 (15.2%)	54 (20.3%)
Opioid agonist treatment7	Ever	294 (63.6%)	155 (58.3%)
	Last 6 months	27 (5.8%)	17 (6.4%)
Ever arrested		415 (89.8%)	230 (86.5%)
Ever jailed or imprisoned8		332 (71.9%)	171 (64.3%)
Ever convicted		284 (61.5%)	140 (52.6%)
Number of times jailed		7.23 ± 9.0	5.9 ± 6.9

NOTES

^1.Chronic pain was measured using the Brief Pain Inventory.¹⁴

^2.Depression was measured using the Center for Epidemiological Studies Depression Scale (CES-D).¹⁵

^3.Anxiety was measured using the Beck Anxiety Index.¹⁶

^4.Number of injections per month was capped at 200.

^5.Unsafe injection practices include sharing of syringes, other injection paraphernalia, or drugs once dissolved.

^6.Refers to both in-patient and out-patient abstinence-based programs.

^7.Refers to both methadone and buprenorphine treatment.

^8.Jailed refers to pre-conviction incarceration and imprisoned refers to post-conviction incarceration.

Follow-up serologies were obtained from 33% – 43% of susceptible participants. Eleven people seroconverted to HCV positive, one person seroconverted to HIV positive, and one person seroconverted to HBV positive on core antibody testing. As shown in Table 2, there was an incidence rate of 9.03 HCV cases per 100 person-years (95% CI: 5.14 – 15.29), with a total of 121.80 person-years of exposure time.

Table 2.

Incidence Rates by Serological Test Result

	HIV	Hepatitis B	Hepatitis C
Number of susceptible individuals at baseline	442	199	267
Number (and percent) of susceptible individuals with follow-up	190 (43.0%)	67 (33.7%)	99 (37.1%)
Number of seroconversions	1	1	11
Total exposure time (months)*	3062.4	962.5	1461.7
Incidence rate (per 100 person-years)	0.39	1.25	9.03
95% Confidence interval	0.01 – 2.16	0.22 – 6.69	5.14 – 15.29

^*Exposure time was measured as the number of months between baseline and follow-up visits for study participants negative for a given viral infection at baseline. For those who seroconverted, half of the period between baseline and first seropositive visit was used in calculating total exposure time.

We explored the association of HCV incidence with items concerning sociodemographics, health, drug use items, and self-reported interactions with criminal justice and substance abuse treatment systems, but with only 11 incident cases, no statistically significant associations with HCV incidence were detected. Spatial analysis of the cases by residential address revealed no detectable clustering of incident infections. Nor could we detect clustering of new infections within recruitment chains, which if detected would be indicative of an outbreak. Adding the two other incident infections failed to produce any significant associations for any of the demographic, spatial, or recruitment variables.

Discussion

Our results add to the evidence that nonurban injection drug use and associated infections are a subject of increasing concern. The HCV incidence rate of 9.03 cases per 100 person-years indicates a relatively high burden of disease transmission in this population. It confirms the need for targeted prevention and expansion of treatment for HCV beyond urban regions to previously neglected populations of PWID. This is especially true since most of our study participants did not know of their status until informed by the study. There is also substantial need for target catchup HBV vaccination in this population. Although the incidence was low, previous analysis of our data revealed that 43% of our sample remained susceptible (neither previously infected nor vaccinated) and 5% carried the virus.^7,8

Achieving prevention goals will be complicated by factors we have also identified in our studies on this population of PWID. First, since most of those we identified as HCV seropositive were unaware of their status, HCV testing has to be made more readily available. Second, harm-reduction programs that can reach PWID and vaccinate them against HBV and prevent HCV transmission by providing safe injection training and sterile syringe access are lacking outside of urban centers in Connecticut. We found that that more than 30% of participants engaged in at least one injection practice that put them at risk for a bloodborne infection and only 3.3% of study participants had used one of the urban syringe exchanges in the month prior to being surveyed.⁵ These levels of risky injection and uncertain syringe access are true not only for Connecticut, but also in most other states.⁹ Third, directly acting antiviral medications to treat HCV, especially the NS3, NS5a, and NS5b inhibitors that in combination achieve close to 100% sustained virological response, have a high cost that could leave patients such as those in our study without access to these more effective medications.¹⁰ Fortunately, three-fifths of our sample relies on Medicare or Medicaid for medical insurance,¹¹ which in Connecticut has recently been expanded to cover HCV treatment for many in our population.¹² But uncertainties in federal support for Medicaid (and even Medicare) may have a future negative impact on future access.

Our study has several major limitations. The low number of incident infections (as shown in Table 1) and a significant loss to follow-up reduced our statistical power to identify factors associated with HCV transmission. It is not possible to extrapolate our results to other nonurban PWID populations or demonstrate that our sample is representative of the population of people who inject drugs in nonurban municipalities in southwestern Connecticut. Nevertheless, the data reveal high rates of ongoing HCV incidence that in the context of ongoing injection risk and very limited access to prevention programs require enhanced prevention efforts.

For prevention to achieve HCV epidemic control, key components are prevention education, access to sterile injection equipment, targeted screening, and, for those who test positive, access to health care, and treatment with appropriate medications.¹³ We believe that control of the HCV epidemic among nonurban PWID is possible but only if resources are allocated to expand HCV testing and harm-reduction programs outside urban cores and to prioritize treating actively infected PWID to reduce the rate at which they transmit HCV to others. The expansion of the HCV epidemic beyond the traditional urban core necessitates nation-wide efforts to obtain these resources.