Key associations for hepatitis C virus genotypes in the Middle East and North Africa.

This study aimed to investigate the epidemiology of hepatitis C virus (HCV) genotypes in the Middle East and North Africa (MENA) through an analytical and quantitative meta-regression methodology. For the most common genotypes 1, 3, and 4, country/subregion explained more than 77% of the variation in the distribution of each genotype. Genotype 1 was common across MENA, and was more present in high-risk clinical populations than in the general population. Genotype 3 was much more present in Afghanistan, Iran, and Pakistan than the rest of countries, and was associated with transmission through injecting drug use. Genotype 4 was broadly disseminated in Egypt in all populations, with overall limited presence elsewhere. While genotype 2 was more present in high-risk clinical populations and people who inject drugs, most of the variation in its distribution remained unexplained. Genotypes 5, 6, and 7 had low or no presence in MENA, limiting the epidemiological inferences that could be drawn. To sum up, geography is the principal determinant of HCV genotype distribution. Genotype 1 is associated with transmission through high-risk clinical procedures, while genotype 3 is associated with injecting drug use. These findings demonstrate the power of such analytical approach, which if extended to other regions and globally, can yield relevant epidemiological inferences.

INTRODUCTION

The hepatitis C virus (HCV) is a major global health challenge.1 An estimated 20% of all individuals chronically infected with HCV are residing in the Middle East and North Africa (MENA).2, 3 Chronic HCV infection leads to various morbidities, such as liver fibrosis, cirrhosis, and cancer,4 all of which strains healthcare systems.5

HCV displays extensive genetic diversity and is categorized into seven genotypes (numbered from 1 to 7).6 Although genotypes 1 and 3 are common worldwide,6 genotype distribution can vary from one geographical area to another.6, 7 Specific HCV genotypes have been hypothesized to be associated with specific modes of acquisition, or with specific populations.6, 7, 8 However, any specific genotype circulating within a specific population tends also to be reflective of the circulating genotypes in the wider population of that country.7

Delineating the epidemiology of HCV genotypes is critical as it can convey inferences about the modes of transmission and their dynamics in a given population.6, 9 Despite this, existing studies tend to be descriptive and qualitative in nature.6, 7, 8 Against this background, we aimed to demonstrate, to our knowledge for the first time, an analytical and quantitative approach to investigate the epidemiology of HCV genotypes, through meta‐regressions, as applied to HCV genotype distribution in MENA. This study is part of the MENA HCV Epidemiology Synthesis Project, an ongoing endeavor to delineate HCV epidemiology and inform key public health research, policy, and programming priorities in MENA.7, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26

METHODS

Studies reporting HCV genotype data were retrieved from the MENA HCV Epidemiology Synthesis Project database,2 originally for an earlier study of HCV genotypes in MENA.7 This comprehensive database consists of several subdatabases on different HCV epidemiological measures, and was populated through a series of systematic reviews for HCV infection across MENA.10, 11, 12, 13, 14, 15, 16, 17, 26 The reviews followed the same methodology, informed by the Cochrane Collaboration handbook,27 and used the Preferred Reporting Items for Systematic reviews and Meta‐analyses (PRISMA) guidelines28 to report their findings.

In these reviews, genotype information were extracted from individual studies to populate the HCV genotype subdatabase. Participants with untypeable genotype were removed from the sample size of the study. Participants with mixed genotypes contributed separately to the quantification of each genotype. The population of each study was classified into six risk categories based on the exposure risk to HCV infection, as presented in Figure 1 and as informed by existing literature29, 30, 31 and our earlier studies.10, 11, 12, 13, 14, 15, 16, 17, 26 A total of 175 genotype studies on 15 960 participants were included.

Figure 1

Population classification into categories by risk of exposures to hepatitis C virus (HCV) infection

Univariable and multivariable random effects meta‐regressions were conducted for the proportion of each HCV genotype, based on established statistical methodology,27 to assess the association between genotype and each of country/subregion and population classification. Variables with a likelihood ratio test P < .2 in the univariable analysis qualified for inclusion in the multivariable analysis. Relative risks and adjusted relative risks (ARRs) with a P value between .05 and .10 in the multivariable model, for any association between genotype and a given factor, indicated good evidence for the association. P ≤ .05 indicated strong evidence for the association.

RESULTS

The results of the meta‐regressions are shown in Table 1 with key results described in the following subsections.

Table 1

Univariable and multivariable meta‐regression models for hepatitis C virus (HCV) genotypes in the Middle East and North Africa (MENA)

		Number of studies	Univariable analysis				Multivariable analysis
			RR (95% CI)	P value	LR test P value	Variance explained (adjusted R 2 (%))	ARR (95% CI)	P value	Variance explained (adjusted R 2 (%))
Genotype 1
Population classification	General population	32	1	⋯			1	⋯
	High‐risk clinical populations	63	2.0 (1.4‐3.0)	<.001			1.2 (1.0‐1.5)	.098
	Populations at intermediate risk	16	0.8 (0.4‐1.6)	.605			1.0 (0.7‐1.4)	.927
	Populations with liver‐related conditions	28	0.6 (0.3‐0.9)	.018			1.1 (0.8‐1.4)	.619
	PWID	13	1.2 (0.7‐2.1)	.569			0.9 (0.6‐1.2)	.378
	Special clinical populations	13	1.5 (0.8‐3.1)	.210			1.3 (0.9‐1.9)	.229
	Mixed populations	10	1.1 (0.6‐2.1)	.784	<.001	22.0	0.9 (0.6‐1.2)	.367
Country/subregion	Egypt	38	1	⋯			1	⋯
	Afghanistan	3	5.5 (3.0‐10.3)	<.001			6.5 (3.3‐12.8)	<.001
	Gulf a	11	6.8 (4.7‐9.8)	<.001			6.1 (4.1‐9.0)	<.001
	Iran	33	7.7 (5.9‐10.2)	<.001			7.2 (5.4‐9.6)	<.001
	Fertile Crescent b	45	5.7 (4.3‐7.5)	<.001			5.3 (4.0‐7.1)	<.001
	Maghreb c	29	10.8 (8.3‐14.2)	<.001			10.2 (7.7‐13.5)	<.001
	Pakistan	16	1.4 (1.0‐2.0)	.075	<.001	83.1	1.3 (0.9‐1.9)	.111	84.4 d
Genotype 2
Population classification	General population	32	1	⋯			1	⋯
	High‐risk clinical populations	63	2.9 (1.2‐7.0)	.020			2.3 (1.0‐5.5)	.056
	Populations at intermediate risk	16	3.9 (1.2‐12.6)	.025			5.1 (1.7‐15.7)	.005
	Populations with liver‐related conditions	28	1.6 (0.6‐4.5)	.364			1.4 (0.5‐3.8)	.465
	PWID	13	2.9 (0.8‐10.5)	.110			2.9 (0.9‐10.0)	.086
	Special clinical populations	13	1.3 (0.3‐7.0)	.720			1.0 (0.2‐5.3)	.961
	Mixed populations	10	3.1 (0.9‐11.0)	.075	.176	3.0	3.0 (0.9‐9.6)	.069
Country/subregion	Egypt	38	1	⋯			1	⋯
	Afghanistan	3	0.6 (0.0‐25.4)	.774			0.4 (0.0‐21.6)	.668
	Gulf a	11	3.8 (1.2‐11.9)	.021			5.0 (1.4‐17.3)	.012
	Iran	33	1.4 (0.5‐3.5)	.512			1.4 (0.5‐4.0)	.576
	Fertile Crescent b	45	3.4 (1.3‐8.9)	.011			3.4 (1.3‐9.2)	.017
	Maghreb c	29	7.4 (3.0‐18.0)	<.001			7.1 (2.6‐19.2)	<.001
	Pakistan	16	3.6 (1.4‐9.5)	.009	<.001	23.2	4.8 (1.7‐13.6)	.003	28.8 d
Genotype 3
Population classification	General population	32	1	⋯			1	⋯
	High‐risk clinical populations	63	1.0 (0.5‐2.2)	.983			0.9 (0.6‐1.4)	.646
	Populations at intermediate risk	16	0.6 (0.2‐2.2)	.430			1.9 (0.8‐4.2)	.124
	Populations with liver‐related conditions	28	1.4 (0.5‐3.7)	.486			1.3 (0.7‐2.4)	.330
	PWID	13	3.1 (1.1‐8.5)	.029			1.7 (1.0‐3.0)	.074
	Special clinical populations	13	0.8 (0.2‐3.4)	.745			1.4 (0.5‐3.7)	.518
	Mixed populations	10	0.5 (0.1‐1.6)	.228	.055	13.2	0.8 (0.4‐1.6)	.525
Country/subregion	Egypt	38	1	⋯			1	⋯
	Afghanistan	3	60.9 (21.5‐172.3)	<.001			45.4 (14.3‐143.8)	<.001
	Gulf a	11	7.5 (2.5‐22.9)	<.001			7.5 (2.3‐24.6)	<.001
	Iran	33	33.8 (15.6‐73.0)	<.001			43.9 (18.9‐101.8)	<.001
	Fertile Crescent b	45	9.2 (4.0‐21.2)	<.001			11.2 (4.6‐27.0)	<.001
	Maghreb c	29	4.4 (1.9‐10.3)	<.001			5.4 (2.2‐13.1)	<.001
	Pakistan	16	71.7 (32.4‐158.9)	<.001	<.001	77.0	69.6 (31.2‐155.5)	<.001	77.9 d
Genotype 4
Population classification	General population	32	1	⋯			1	⋯
	High‐risk clinical populations	63	0.3 (0.2‐0.7)	.004			0.8 (0.6‐1.1)	.250
	Populations at intermediate risk	16	1.3 (0.5‐3.5)	.580			1.0 (0.7‐1.5)	.915
	Populations with liver‐related conditions	28	0.8 (0.4‐1.9)	.654			1.0 (0.7‐1.3)	.957
	PWID	13	0.4 (0.1‐1.2)	.104			1.0 (0.6‐1.7)	.851
	Special clinical populations	13	1.3 (0.5‐3.9)	.583			1.0 (0.7‐1.5)	.985
	Mixed populations	10	0.6 (0.2‐1.8)	.341	.007	9.9	0.8 (0.5‐1.3)	.348
Country/subregion	Egypt	38	1	⋯			1	⋯
	Afghanistan	3	0.0 (0.0‐0.5)	.023			0.0 (0.0‐0.5)	.021
	Gulf a	11	0.5 (0.4‐0.7)	<.001			0.5 (0.3‐0.8)	.002
	Iran	33	0.0 (0.0‐0.0)	<.001			0.0 (0.0‐0.1)	<.001
	Fertile Crescent b	45	0.6 (0.5‐0.8)	<.001			0.6 (0.5‐0.8)	<.001
	Maghreb c	29	0.1 (0.1‐0.1)	<.001			0.1 (0.1‐0.1)	<.001
	Pakistan	16	0.0 (0.0‐0.1)	<.001	<.001	91.1	0.0 (0.0‐0.1)	<.001	90.3 d
Genotype 5
Population classification	General population	32	1	⋯			1	⋯
	High‐risk clinical populations	63	0.3 (0.1‐1.0)	.055			0.3 (0.1‐1.2)	.087
	Populations at intermediate risk	16	0.6 (0.1‐3.7)	.540			0.7 (0.1‐4.6)	.694
	Populations with liver‐related conditions	28	0.7 (0.2‐2.7)	.594			1.0 (0.2‐4.2)	.956
	PWID	13	1.1 (0.2‐5.7)	.875			1.9 (0.3‐10.9)	.478
	Special clinical populations	13	6.6 (1.4‐30.2)	.015			12.2 (2.4‐62.7)	.003
	Mixed populations	10	0.1 (0.0‐0.8)	.031	<.001	45.9	0.2 (0.0‐1.2)	.080
Country/subregion	Egypt	38	1	⋯			1	⋯
	Afghanistan	3	1.7 (0.0‐86.2)	.778			1.0 (0.0‐59.0)	.995
	Gulf a	11	3.3 (0.6‐18.9)	.187			1.1 (0.2‐5.7)	.945
	Iran	33	0.5 (0.1‐2.1)	.323			0.9 (0.2‐4.5)	.929
	Fertile Crescent b	45	2.3 (0.6‐8.5)	.227			4.7 (1.2‐17.6)	.024
	Maghreb c	29	0.5 (0.1‐2.1)	.313			1.1 (0.2‐5.5)	.881
	Pakistan	16	0.9 (0.2‐4.5)	.940	.157	11.5	1.2 (0.3‐5.8)	.789	57.3 d

Abbreviations: ARR, adjusted relative risk; CI, confidence interval; LR, likelihood ratio; PWID, people who inject drugs; RR, relative risk.

Countries include Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and United Arab Emirates.

Countries include Iraq, Jordan, Lebanon, Palestine, and Syria.

Countries include Algeria, Libya, Mauritania, Morocco, and Tunisia.

The adjusted R‐squared for the full model.

Genotype 1

There was strong evidence for variation in genotype 1 distribution by country/subregion, and good evidence for variation by population. Relative to Egypt, all countries/subregions had much higher presence of genotype 1, apart from Pakistan. Genotype 1 was also more present in high‐risk clinical populations (ARR of 1.2 [95% confidence interval [CI]:1.0‐1.5]) than in the general population. Country/subregion and population explained the vast majority of the variation in genotype 1 distribution (84.4%), mostly through the country/subregion variable.

Genotype 2

There was strong evidence for variation in genotype 2 distribution by country/subregion and by population. Relative to Egypt, the Gulf, Fertile Crescent, Maghreb, and Pakistan had higher presence of genotype 2. Genotype 2 was also more present in high‐risk clinical populations, populations at intermediate risk, and people who inject drugs (PWID), relative to the general population. Country/subregion and population explained only 28.8% of the variation in genotype 2 distribution, mostly through the country/subregion variable, with most variation remaining unexplained.

Genotype 3

There was strong evidence for variation in genotype 3 distribution by country/subregion, and good evidence for variation by population. Relative to Egypt, all countries/subregions had substantially higher presence of genotype 3. This was particularly so for Afghanistan, Iran, and Pakistan, that had much higher presence of this genotype. Genotype 3 was also more present in PWID (ARR of 1.7 [95% CI: 1.0‐3.0]) than in the general population. Country/subregion and population explained the vast majority of the variation in genotype 3 distribution (77.9%), mostly through the country/subregion variable.

A sensitivity analysis (Table S1) was performed in which the meta‐regressions were performed excluding countries in which genotype 3 was the most dominant genotype (Afghanistan and Pakistan but not Iran7). The analysis was conducted to assess the purported global association between this genotype and PWID.6 It was found that there is indeed strong evidence for an association between this genotype and PWID (ARR of 2.7 [95% CI: 1.2‐5.8]). The sensitivity analysis also confirmed the strong evidence for variation by country/subregion—country/subregion and population explained the majority of the variation (66.4%), mostly through the country/subregion variable.

Genotype 4

There was strong evidence for variation in genotype 4 distribution by country/subregion, with most countries having substantially lower (Gulf and Fertile Crescent), or much lower (Afghanistan, Iran, Maghreb, and Pakistan) presence of this genotype compared with Egypt. No significant evidence was found for variation in genotype 4 distribution by population, probably reflecting the broadly disseminated nature of the genotype 4 epidemic in Egypt.7, 10 Country/subregion and population explained the vast majority of the variation in genotype 4 distribution (90.3%), nearly all of which through the country/subregion variable.

Genotype 5

Unlike the other genotypes, most of the variation in the distribution of genotype 5 was explained by the population variable, rather than the country/subregion variable. Paradoxically also, compared with the general population, there was strong evidence for higher presence of this genotype in special clinical populations, but lower presence in high‐risk clinical populations. Moreover, of all countries/subregions, only the Fertile Crescent had higher presence of genotype 5 relative to Egypt. Population and country/subregion explained 57.3% of the variation in genotype 5 distribution. These results, which are epidemiologically not easy to interpret, could be an artifact of the very low presence of this genotype in the MENA region.7 Genotype 5 was identified in only 3.9% of all genotype studies, with only 0.1% of infections being with this genotype.

Genotype 6 and genotype 7

There were too few studies in which genotype 6 was identified (1.7%), resulting in very broad confidence intervals of limited implications, thus no analyses are reported for this genotype. No study identified the presence of genotype 7 in MENA.

DISCUSSION

We used a quantitative analytical approach to investigate epidemiological associations for HCV genotypes, with application to the MENA region. The utility of this approach was demonstrated—the study yielded estimates for the differences in genotype distributions by country/subregion and by population type. Overall, country/subregion explained most of the variation, highlighting geography as the principal determinant of genotype distribution. Interestingly, we found evidence for an association between genotype 1 and high‐risk clinical populations, possibly explaining, with the global presence of these populations, the common presence of this genotype across countries.6 This finding also supports existing literature suggesting such an association.6 We further found an association between genotype 3 and PWID, thereby supporting an apparently global role for this genotype in HCV transmission through injecting drug use.6 While these findings were generated for MENA, they may reflect generic results of global relevance.

Genotype epidemiology in the region may have been influenced by several factors. The importation of contaminated blood products, before the onset of blood screening, primarily from Western countries, has been linked to a fraction of reported HIV cases,32, 33 and presumably helped disseminate HCV genotype 1 across the region.6 Genotype 1 was ubiquitous across most countries and subregions of MENA, which may also be attributed to population movement links to countries outside of MENA, as genotype 1 is the most common genotype globally.6, 30 For example, the presence of genotype 1 was highest in the Maghreb subregion, almost 11‐fold higher than in Egypt, a subregion which has strong migration links with western and southern Europe.34, 35

The emerging HIV epidemics among PWID in MENA, and specifically in Afghanistan, Iran, and Pakistan, have been linked to overlapping injecting networks across these countries.36, 37, 38 This may have (partially) contributed to the relatively larger presence of genotype 3 among PWID in these countries. A strong presence of genotype 4 was found in Egypt. Phylogenetic evidence has shown that genotype 4 in this country originated from central Africa,39 and circulated endemically until the mass expansion of the HCV epidemic through the parenteral antischistosomal therapy (PAT) campaigns and other healthcare practices.10, 11, 39, 40, 41, 42, 43 Genotype 4 was also found to be present in countries that host large migrant labor populations from Egypt,44 such as in countries of the Fertile Crescent13 and Gulf26 subregions of MENA.7, 18

This study had limitations. HCV genotype data were not available for several countries. The number of studies also varied by population type and country/subregion, and the sample size of genotyped individuals was small for a number of studies—we may not have had sufficient statistical power to discern specific associations. This may explain the lack of a clear interpretation/identification of epidemiological associations for genotype 5. We assessed associations for specific key factors that were available in extracted data, but we were unable to assess the role of a broader set of factors. This may clarify why most variation in genotype 2 distribution remained unexplained. Despite these limitations, we were able to utilize a breadth of genotype data that was systematically gathered and that allowed us to conduct such analysis, yielding relevant inferences about genotype distribution and HCV transmission dynamics.

In conclusion, geography appears to be the principal determinant of HCV genotype distribution. Genotype 1 is associated with transmission through high‐risk clinical procedures, such as blood transfusions, hemodialysis, and medical injections; while genotype 3 is associated with transmission through injecting drug use. These findings demonstrate the power of such an analytical approach, which if extended to other regions and globally, may yield relevant epidemiological inferences that can inform control and treatment programs.

CONFLICT OF INTERESTS

The authors declare that there are no conflict of interests.

AUTHOR CONTRIBUTIONS

SM conducted data extraction and analyses and wrote the first draft of the paper. LJA conceived and led the design of the study, analyses, and drafting of the article. All authors contributed to the extraction of data and writing of the manuscript.

Supporting information

Click here for additional data file.