Introduction

Viral Hepatitis C (VHC) is an inflammation of the liver caused by the hepatitis C virus, a bloodborne virus. Although it can lead to short-term illness, most infected people develop a long-term, chronic infection which can result in cirrhosis and liver cancer. In the absence of a vaccine for hepatitis C, prevention is mainly based on reducing the risk of exposure to the virus by avoiding behaviors that can spread the disease, in health care settings and in higher risk populations [1,2].

VHC infection remains a leading cause of morbidity and mortality worldwide. The WHO estimates the global prevalence of chronic VHC infection to more than 70 million people in 2015 [3]. Three to four million new infections and nearly 400 000 attributed deaths occur each year, mostly from cirrhosis and hepatocellular carcinoma [4]. A recent meta-analysis estimated the overall global prevalence of hepatitis C in the general population to 1.8% [5].

The prevalence of VHC infection varies across and within the countries. It is low in north America, moderate in Europe and some African and South American countries, but it’s considered to be high in the Middle East and North Africa [6,7]. In fact, the Eastern Mediterranean region seems to be the most affected region with an overall prevalence of 2.3% and at least 23 million infected people [8,9]. However, this prevalence is heterogeneously distributed with rates ranging between 1 and 2% in most countries except Egypt with an estimated prevalence of 14.7% and the highest in the world [8,10].

In Tunisia, reliable VHC epidemiological data are limited to specific regions or populations such as blood donors [11] or dialysis patients [12,13]. The VHC burden in Tunisia has long been thought low. However, prevalence estimates in the general population are lacking. Several studies have attempted to understand the magnitude of VHC prevalence in Tunisia, but most have methodological limitations or focus on specific sub-populations rather than on the general population, with considerable variation in their estimates of prevalence rates. The last population-based study in 1996 showed a prevalence ranging from 1.7% in the North-Western region to 0.2% in the Southern region [14]. A literature review reported an overall prevalence of VHC infection of 0.7% [15].

Regarding VHC transmission, blood transfusion was the main mode of transmission but since the establishment of the National Program of Screening of VHC among Blood Donors in 1994, nosocomial transmission was more incriminated [16]. Recently, VHC treatment has been transformed with the advent of direct-acting antivirals (DAA), which are better tolerated than interferon-based treatment and offer high cure rates [17]. DAA agents has been available in Tunisia since 2016 as a pillar of the national plan for VHC elimination [18].

Our study aimed to estimate the national prevalence of VHC infection and to identify the associated risk factors immediately before the implementation of the national plan for VHC elimination (NPVHCE) in Tunisia. These estimations constitute the baseline to monitor the effectiveness of the NPVHCE in Tunisia.

Materials and Methods

Study design, study population and sampling

A national household cross-sectional population-based survey was conducted from January to December 2015 in Tunisia. A two-stage cluster design with probability proportional to size (PPS) was used with random sampling of districts and households: First, a random selection of districts was conducted in each of the 24 governorates, then, in each of the selected district, 20 households were randomly selected (Fig. 1). All eligible and consent individuals present during the investigators households visits were included.

Figure 1: Sampling procedure, VHC national serosurvey, Tunisia-2015.

The sampling and computation of sampling fractions, the sampling frame of districts and households were based on the data provided by the National Institute of Statistics (INS) according to the national census of population of 2014.

Sample sizes were calculated using EpiInfo software based on the given precision (d), the estimations of prevalence (p), the a error risk of 5% (za/2=1.96) and a correction factor DE equal to 1.5 (n= DE (za/2 p(1-p)) / d2) [19]. Table 1 shows the sample size calculation based on an estimated of VHC prevalence by region.

Considering a refusal rate of 15%, the final sample size of 22 275 individuals was necessary for this study.

Table 1: Sample size calculation by region.

Data and Blood Sample Collection

A face-to-face standardized, pre-piloted questionnaires were administered to households and individuals by trained 24 teams (1 Medical doctor, 1 nurse). Investigators interviewed all the members of the seleted households through door-to-door visits. Blood samples were taken in the nearest primary healthcare center (Supplementary file).

All family members available during the investigators visits, were included regardless of their age and sex. Those unavailable during the first visit were called for a second appointment and interviewed within the same month.

Were excluded those not permanent resident in the visited district, those having a contraindication to blood sampling and those didn’t give their consent to participate

Questionnaires included information on socio-demography and individual history of VHC exposure and risk factors and were divided in two parts. The first part included information on the family (family size, socio-economic level, social security, house type…) and the second part refered to the individual questionnaire for each family member and provided socio-demographic data (date of birth, gender, occupation, education level, insurance coverage…) as well as data on risk factors for VHC transmission (nosocomial risk, sexuality, drug use, etc...). It also contained information of whether the subjects had serology before and if they were known VHC positive, details of medical care and follow-up.

Risk factors for seropositivity identified a priori demographic variables, spatial variables, medical variables (history of blood transfusion and blood donation, history and location/provider type for medical injections, surgery and delivery, dental and gum treatment, type of contraception, miscarriage and abortion) and behavioral variables (tattoos, piercing, IV drug use, pedicures, manicures and frequenting of barbershops).

After being codified to ensure anonymity and confidentiality, all the blood samples were sent to the reference laboratory; the Microbiology-Biochemistry laboratory of Aziza Othmana Hospital in the capital Tunis. All samples were tested for antibodies antiVHC IgG using (electro) chemiluminescence technique on the «Cobas e411» automate Search for the viral HCV RNA was performed by RT-PCR in real time in subjects diagnosed positive for the screening test for VHC antibodies to distinguish between chronic VHC and healed subjects (Fig. 2).

Figure 2: Algorithm for analysis of VHC infection markers

Serological validation of the results was conducted for all samples by a group of biologists and epidemiologists of the steering committee.

Data Entry and Analysis

After initial validation in each region by the regional coordinator, all questionnaires were sent to the National Observatory of New and Emerging Diseases for final validation and data entry using EpiData 1.3. Laboratory results were added in the same database when received.

Statistical analyses were conducted using probabilities and sampling weights calculated for each stage of the sampling (District, household and individual) and analysis considered the finite population correction factor. The prevalence of VHC infection [95%CI] was estimated globally, by region and sociodemographic factors.

Univariate analysis was performed using the Pearson chisquare test and the Student t-test to compare the qualitative and quantitative variables, respectively and a binary logistic regression to estimate crude Odds Ratios [95%CI] associated to different VHC studied risk factors. 

We conducted a multivariate analysis, accounting for the sampling design, to identify risk factors for VHC infection among the population. The association between the seroprevalence and the explanatory covariates was quantified by fitting multivariate logistic regression model. The multivariate analysis retained variables from the univariate analysis with p-value less than 0.2. Estimates of the regression coefficients of the model and odds ratios with their confidence interval were presented. All statistical analysis were conducted using SPSS.20 software.

Accounting for the sampling design, the survey package (Analysis of Complex Survey Samples, Thomas Lumley) version

3.34 estimated parameters, including standard errors (HorvitzThompson-type standard errors) were used everywhere in the survey package. Confidence interval calculations usually used the scaled Chi squared distribution for the log likelihood from a binomial distribution.

Community Engagement

Prior to the start of the survey, a national workshop was organized to present the study protocol and data collection tools and needs to all investigators, coordinators and local authorities. Meetings were also organized with local authorities at all levels to introduce the objectives of the survey and to discuss the timeline and request for support. Mobilisers (identified by the chief of each village) visited selected households prior to the data survey to request the household’s presence.

Ethical Considerations

Our study had the approval of the Higher Council of Statistics, the National Ethic Committee and the National Instance of Personal Data Protection. Before participating to the study, all individuals had information note on the objectives and modalities of the survey and had to sign informed consent. These two documents were both in Arabic, native language.

All subjects had the right to refuse to participate and were respected in their decision. Confidentiality and anonymity were ensured throughout the study as well as the feedback and support of positive people. The serological results were sent to the concerned subjects through the regional coordinator and positive people were sent to the gastroenterology service for further exploration and proper medical supervision. Patients with VHC positive result were referred to reference VHC elimination program to measure viral load and initiate treatment.

Results

Description of the Study Population

From the expected sample size of 22 275 subjects, a total of 21 720 subjects were surveyed, (participation response rate of 97.5%). Of these, we collected 19 693 blood samples (90.6% response rate).

Male to female sex-ratio was 0.67 with 59.9% females (Fig. 3). Adults aged 30 to 40 years and 40 to 50 years were the most represented (14.9% and 14.7% respectively) (Fig. 4).

Most of the participants were living in the north of the country (48.8%) followed by the center (36.6%) and the south (14.5%) with a predominance of urban areas (67.8%). The highest participation rates were in Tunis (9.7%), Sfax (8.7%) and Nabeul (7.2%).

Figure 3: Distribution of study population by gender, VHC national serosurvey, Tunisia-2015.

Figure 4: Distribution of study population by age groups, VHC national serosurvey, Tunisia-2015

Prevalence of anti-HCV positive sera

National HCV Prevalence was 0.88% (95%CI [0.78% – 0.99%]) estimating 88130 VHC positive subjects in the general population. There was no significant difference in the prevalence of VHC infection between males and females (0.88% vs 0.89% respectively). However, rates increased significantly with age (Fig. 5) with the highest prevalence among participants aged 70 years and more (2.82%). The mean age of positive subjected was 21.19 (SD ± 1.31) years. Prevalence ranged from 0.13% in the center-east to 2.57% in the north-west region (p<0.001) (Fig. 6).

Figure 5: Distribution of anti-HCV prevalence by age groups, VHC national serosurvey, Tunisia-2015

Figure 6: Distribution of anti-HCV prevalence by region/governorate, VHC national serosurvey, Tunisia-2015

VHC infection associated risk factors

Univariate analysis showed that age over 40 years was associated with HCV prevalence with a significant increase of OR; from 5.16 for the age group 40-50 years to 20.72 for those aged =70 years. VHC infection prevalence was higher in the North region (OR=8.06). People living in rudimentary houses were more likely to be positive (p=0.01) but there was no significant difference between rural and urban areas. Besides, VHC infection was higher among married subjects (OR=2.08) and divorced or separated or widowed (OR=5.14) comparing to single participants. Prevalence was significantly higher among those without a professional activity and varied significantly with profession type as well as education level (table 1).

Regarding hygiene, sharing personal items such as towels, toothbrushs or scissors wasn’t found to be an associated risk factor for VHC infection as well as sharing razor blades among men. However, the risk of being infected was significantly lower among men shaving at the barber (OR=0.67) and when the barber is using new razor blades each time (OR=0.39) (table 2).

Table 2: Socio-demographic factors associated with HCV infection

Unlike blood transfusion and blood donation, history of surgical intervention, invasive procedures, hospitalization, medical injections were significantly associated with VHC prevalence, as well as dental procedures. The OR associated with diabetes and hypertension was 3.06 and 4.20 respectively. Other risk factors found significantly associated with VHC infection were intravenous and intranasal drug use, traditional scarification and tattoos in addition to multiple sexual partners. Prevalence was also higher among those who had a relative infected with VHC (sister) (table 3 and 4).

Table 3: Life-related factors associated with HCV infection

Table 4: Medical history, healthcare-associated factors and HCV infection

By multivariate analysis, factors independently associated with VHC infection (Table 5) were divided in socio-demographic factors, healthcare-related factors and lifestyle habits. Socio-demographic factors included age and area, with older participants and those living in rural areas more at risk of being VHC infected; the corresponding AOR were 1.04 (95%CI [1.03–1.06]) and 1.57 (95%CI[1.15–2.14]) respectively. In addition, the geographic region was also significantly associated with VHC and the most associated category was the north region (AOR was 12.53 (95%CI[4.56 - 34.45]). 

Table 5: High-risk behaviors and HCV infection

Table 6: Multivariate analysis