A descriptive study of hepatitis C in people who inject drugs

Control of viral hepatitis is a global public health priority and is a chosen target of the 2030 agenda of sustainable development1. The National Programme for Surveillance of Viral Hepatitis has reported a higher seroprevalence of the hepatitis B virus (HBV) than that of hepatitis C virus (HCV)2. However, vaccination against HBV is part of the national immunization programme and its prevalence will decrease with time. On the other hand, there are no effective vaccines against HCV, and directly acting antivirals (DAAs) are expensive.

People who inject drugs (PWIDs) are at a high risk of contracting HCV. Globally, over one-third of the HCV burden is attributable to injection drug use3. In India, the estimated HCV seroprevalence in PWID is as high as 44-71 per cent4 compared to 0.32 per cent in the general population2.

The National Viral Hepatitis Control Programme (NVHCP) borrows from the National AIDS Control Programme (NACP) and identifies PWID as a key population for targeted interventions2. There are two notable differences between NACP and NVHCP. First, NACP draws on decades of research that characterized PWID at risk of contracting human immunodeficiency virus (HIV). Second, sustained Information Education and Communication (IEC) activities have raised awareness of HIV and its treatment, whereas HCV and DAA are relatively new.

Opioids account for 91 per cent of injecting drug use (IDU) in our country5; thus, opioid using PWIDs constitute a large at-risk population. Historically, the northern and north eastern parts of the country have had a high prevalence of opioid addiction. Recently, we reported an increased number of patients using unsafe injection practices to abuse a new pharmaceutical opioid6 in southern India. This research expands our earlier work by describing the risk factors, awareness and treatment continuum for HCV in opioid using PWID. Furthermore, we use geospatial clustering to identify localities needing urgent targeted interventions.

Material & Methods

This study was conducted at the Centre for Addiction Medicine, Department of Psychiatry, National Institute of Mental Health and Neurosciences, Bengaluru, Karnataka. A government, tertiary-level addiction treatment centre with inpatient and outpatient departments (IPD and OPD, respectively) and opioid substitution treatment after obtaining ethics clearance from the Institutional Ethics Committee.

Inclusion criteria: All individual contacts in OPD are routinely captured in an electronic health record (EHR) software in the study hospital. This software accommodates structured (sociodemographic, diagnostic, etc.) and unstructured (free-text notes entered by doctors) data. In the IPD, discharge summaries were entered in the EHR and include mandatory free-text fields where doctors entered details regarding the route of substance use. In this study all adult study participants presenting between January 1, 2016 and January 1, 2021 in the database and among them, identified all study participants who received opioid use-related diagnoses (F11 as per the International Classification of Diseases – version 10)7.

Since all individuals with opioid use were not injecting drugs, the study identified 160 keywords (including misspelt words and abbreviations) in the doctors’ notes signifying current or past use of injections. This was achieved by tokenizing (breaking into words) clinical notes of all participants and choosing specific terms related to IDU (Supplementary Table). Probable cases of PWID, thus identified, were manually reviewed for documented IDU. One-fourth of the cases with no keywords were randomly chosen and manually checked to ensure the validity of this process. The final sample consisted of individuals with documented IDU.

Study instruments: Sociodemographic and substance use-related data were collected from the clinical records. A structured interview schedule was used to assess HCV awareness telephonically. This interview was developed based on hepatitis C follow up questionnaire used by the Centre for Disease Control8. The modifications to the instrument were made based on one round of expert review by five psychiatrists working at the study centre. These experts were chosen as they were treating individuals included in this report and thus were best-suited to make the instrument locally relevant. The final questionnaire is provided as

Close

Figure

Study procedure. ICD, international classification of diseases; IDU, injecting drug use.

Export to PPT

At the first contact, blood investigations, including liver function tests, haemogram, HCV, HBV and HIV, were offered to all study participants. However, it was not mandatory for a participant to undergo these tests to access treatment. Aspartate aminotransferase to platelet ratio index (APRI) was calculated to assess HCV-associated hepatic fibrosis. The World Health Organization recommends using APRI in resource-poor settings9.

In this study a point-of-care anti-HCV antibody test [4^th Generation HCV-Tridot – HCV antigens for CORE, NS3, NS4, NS5 (J. Mitra and Co. Pvt. Ltd.)] was used to detect HCV infection.

Statistical analysis: Descriptive statistics are provided for sociodemographic and clinical data. Patients’ residential addresses obtained from EHR were converted to latitudes and longitudes using Google^® Geocode API. We wanted to test the hypothesis that HCV and its associated risk behaviour, i.e. IDU, occurs in geographical clusters. However, there were two challenges in testing this hypothesis. First, our centre receives cases from across the country, and cluster analysis on such a large geographical area will show states with a high prevalence of HCV. Therefore, we restricted the geospatial analysis to individuals within a 100 km radius of the Bengaluru Municipal Boundary in the southern State of Karnataka. Second, since human habitations are not uniformly distributed, a control was needed for geospatial clustering. While population density is a commonly chosen control, individuals seeking help for alcohol use as a comparator to guard against possible bias were chosen due to the location of the treatment facility. That is, if the location of our facility (tertiary level addiction care centre) spuriously inflates the prevalence of addictive behaviours in the vicinity, then this effect must operate equally for alcohol and opioid use disorders.

Formal hypothesis testing for the presence of clusters included Kernel Density Estimation (KDE) and the Diggle and Chetwynd test10. KDE can be understood as the number of neighbours around a point at a given distance. The hypothesis to be tested was that the KDE of cases and controls were not significantly different and the hypothesis testing was done by simulation. Analysis was carried out using R11; the R package-spatstat was used12.

Results & Discussion

The sampling frame comprised 2629 unique cases, 391 meeting the inclusion criteria.

Risk factors related to transmission: Most study participants in the sample were male (n=370, 95%). There was early onset of substance use other than nicotine [median=17yr, interquartile range (IQR)=15-20yr] with minimal separation between the onset of opioid use (median age at first opioid use=21yr, IQR=18-25yr) and first injection (median age at first injection=21yr, IQR=19-26yr). Study participants who reported sharing needles or paraphernalia at any time (n=167) differed substantially from the whole group. They had an earlier onset of any substance use (x₁=16.1, s₁=2.8; x₂=18.6, s₂=7; t=−11, df=166, P<0.001) and first injection (x₁=20.8, s₁=4.3, x₂=23.4, s₂=7.6, t=−7.9, df=166, P<0.001).

Treatment continuum for hepatitis C virus: Of the total study participants (n=391), 220 (56.26%) were tested for HCV, 109 (27.87%) were positive for HCV antibodies and 93 (85.3%) of them were referred for HCV treatment immediately. It must be noted that no participant had received a diagnosis of HCV or taken treatment for HCV in the past. All seropositive individuals received post-test counselling, but 16 individuals refused to go for further testing and treatment. The projected seropositivity in this group was between 27.9 per cent (best case scenario, all untested assumed negative) and 71.6 per cent (worst case scenario, all untested deemed positive). The high seroprevalence in these young individuals was of concern for two reasons; first, it showed a shift in the epidemiology of HCV away from older males2 into younger populations. Acquiring infection at a younger age and not receiving treatment would translate to a longer duration of infection, with greater likelihood of resulting in hepatic cirrhosis, a risk factor for hepatocellular carcinoma13. Second, this group would be more likely to transmit HCV to non-drug-using populations through sexual route.

APRI could be calculated for 248 out of 391 participants. The percentage of participants with APRI values suggestive of cirrhosis (16/94, 17%) or significant fibrosis (27/94, 28.7%) was significantly higher in seropositive individuals as compared to seronegative (5/89, 5.6%; 15/89, 16.9%, respectively) and the individuals who were not tested (4/65, 6.2%; 8/65, 12.3%, respectively). This difference was significant (χ²= 18.4, df=4, P=0.002).

Treatment was initiated for one-third of the seropositive participants (36/109=0.33). Most study participants who commenced treatment completed it (27/36; =75%); however, this was only one-fourth (24.8%) of all seropositive individuals. This study compared the tested participants with those not tested to evaluate for a possible systematic bias in testing. Participants reporting unsafe injection practices were more often tested than those who did not (p1=0.58, p2=0.23, χ²=44.9, df=1, P<0.01).

Awareness of hepatitis C virus among people who inject drug: Half the study participants could not be contacted for telephonic interviews (208/391;=53%). Among the 183 who could be interviewed (125 had been tested, 63 were seropositive), only 27 (14.75%) reported being aware of HCV before contact with treatment facilities. Overall, awareness of HCV was low. However, among those who were aware of HCV, most were aware of the route of spread (96%) and the availability of treatment (70.4%), but only 41 per cent knew about the risk of reinfection after treatment.

Spatial cluster analysis of PWID and hepatitis C virus in Bengaluru: The present study compared the spatial distribution of seropositive cases and controls residing within a 100 km radius of the Bengaluru Municipal Boundary (77 seropositive cases, 6744 controls). Seventy seven controls were randomly selected for further analysis. This showed clustering of seropositivity (D=2316.05, P=0.001, 999 simulations) (

This study was not without its limitations. First, this hospital-based study may not reflect the ground reality. However, a community-based multi-city survey by Solomon et al15 reported similarly low awareness and treatment uptake for HCV in PWID. Second, the study could not estimate the true seroprevalence in the sample as about half of the PWIDs were not tested. This lack of testing could have also biased the clustering analysis. Nevertheless, the present study estimated a best (27.9%) and a worst (71.6%) case scenario prevalence, both of which deemed alarming. Third, there was a substantial non-response bias for the awareness assessment part of this study. For example, participants who spoke to the investigator may generally be more aware and compliant, giving an inflated estimate of awareness about HCV in PWIDs. Since even this presumably biased estimate was only 14.75 per cent, it was inferred that the awareness in the study population was low. Fourth, we cannot rule out the presence of many more geographical clusters of HCV infection away from our facility. Finally, the gold-standard test for HCV infection, i.e. detection and quantification of HCV ribonucleic acid were not used in this study. However, the 4^th generation HCV-Tridot has high concordance with the gold-standard test (sensitivity and specificity of HCV-Tridot are 98.2% and 98.4%, respectively)16.

Implications for national viral hepatitis control program: Our findings have important implications for the NVHCP as it enters the 5^th yr of its implementation. For example, the awareness about HCV was low in this sample from a metropolitan city, indicating low awareness in the community and high-risk groups. Similarly, only one-third of the seropositive cases could be started on DAAs. Systematic assessment of barriers to treatment will require a detailed interview, which was not possible in the current study. However, we assessed the ease of accessing definitive treatment of HCV in this group. First, the individuals have to visit the government-designated facility twice at an interval of one week to receive the full course of DAAs. On the first visit, they must pay approximately ₹ 6000 INR for viral load testing. The treatment, however, is provided free of cost. Most of these individuals would typically undergo acute withdrawal during this period, are financially constrained, do not have family members to accompany them for hospital visits and thus, do not receive proper treatment and care. Therefore, there is need to integrate the treatment for IDU and HCV so that both are available in the same place. For the first 3.5 yr of the study, Bengaluru did not have a facility to provide free DAAs. Currently, there are 30 treatment centres under NVHCP in Karnataka; however, number of individuals receiving free DAAs from them is not known with disegregated information on PWID on DAA therapy. The wide and easy availability of free DAAs is the most crucial step in controlling HCV transmission and importantly the provision of generic, and subsidized antiretroviral therapy has been the cornerstone of NACP17.

Supplementary Material

Centre for Addiction Medicine, NIMHANS

(To be administered by a trained mental health professional, followed by an information-sharing session)