Sex differences in hepatitis A incidence rates–a multi-year pooled-analysis based on national data from nine high-income countries

Background Possible sex differences in hepatitis A virus (HAV) incidence rates in different age groups are not well documented. We aimed to obtain stable pooled estimates of such differences based on data from a number of high-income countries. Methods We obtained data on incident cases of HAV by sex and age group over a period of 6–25 years from nine countries: Australia, Canada, Czech Republic, Finland, Germany, Israel, Netherland, New Zealand and Spain. Male to female incidence rate ratios (IRR) were computed for each year, by country and age group. For each age group, we used meta-analytic methods to combine the IRRs. Meta-regression was conducted to estimate the effects of age, country, and time period on the IRR. Results A male excess in incidence rates was consistently observed in all age groups, although in the youngest and oldest age groups, where the numbers tended to be lower, the lower bounds of the 95% confidence intervals for the IRRs were less than one. In the age groups <1, 1–4, 5–9, 10–14, 15–44, 45–64 and 65+, the pooled IRRs (with 95% CI) over countries and time periods were 1.18 (0.94,1.48), 1.22 (1.16,1.29), 1.07 (1.03,1.11), 1.09 (1.04,1.14), 1.46 (1.30,1.64), 1.32 (1.15,1.51) and 1.10 (0.99,1.23) respectively. Conclusions The excess HAV incidence rates in young males, pooled over a number of countries, suggest that the sex differences are likely to be due at least in part to physiological and biological differences and not just behavioral factors. At older ages, differential exposure plays an important role. These findings, seen in the context of the excess incidence rates in young males for many other infectious diseases, can provide further keys to the mechanisms of the infection.


Introduction
There is an expanding literature on sex differences in the incidence rates of various infectious diseases [1][2][3]. The type and extent of the differences frequently vary by disease and age group. The mechanisms underlying these differences have not been fully elucidated and cannot be explained entirely by differences in exposure. The pattern of male to female ratios in the incidence rates of different infectious diseases can make an important contribution to understanding the underlying mechanisms of the diseases.
Despite the availability of an effective vaccine, hepatitis A virus (HAV) infection remains a common disease, particularly in low-income countries with overcrowding and poor sanitation, where the incidence rates of the disease are particularly high in infancy and childhood [4,5]. In countries with high hepatitis A vaccine coverage, the incidence of cases and outbreaks have decreased in children and the infection has shifted significantly to other risk groups, such as men who have sex with men (MSM) [6][7][8][9].
There are reports in the literature on sex differences in the incidence rates of hepatitis A, but they are inconsistent and poorly documented by age group [10][11][12][13]. While some report higher incidence rates of viral hepatitis A in males [6,10], there are inconsistencies. For example, in a report from Germany, during 2018-2020, no sex differences were observed in the incidence of the disease [11]. One report from South Korea found a change in the sex differences, possibly due to increased immunization in the military [13].
In this study, we aimed to obtain pooled estimates of the age-specific male to female ratios in the incidence rates of HAV infection based on data from a number of developed countries over extended time periods.
Monitoring System [25], for Israel from the Central Bureau of Statistics [26], for Netherland from Netherlands' database (StatLine) [27], for New Zealand from Statistics New Zealand [28] and for the Spain from the Department of Economic and Social Affairs, Population Division [29].

Ethics and informed consent
National, open access, sex-and-age disaggregated, anonymous data were used and there was no need for ethics committee approval.

Statistical analyses
Data analysis. HAV incidence rates (IR) per 100,000 were calculated by age group and sex, for each country and calendar year using the number of reported cases divided by the respective population size and multiplied by 100,000. The age groups considered were <1 years (infants), 1-4 (early childhood), 5-9 (late childhood), 10-14 (puberty), 15-44 (young adulthood), 45-64 (middle adulthood) and 65+ (senior adulthood). Surveillance systems in Canada and New Zealand used similar age-groups except for 15-39, 40-59 and 60+. For Australia, data for infants and age 1-4, disaggregated by sex and age, are missing. The male to female incidence rate ratio (IRR) was calculated by dividing the incidence rate in males by that of females, by age group, country and time period.
Pooled analysis. As in previous studies of sex differences in infectious diseases [1][2][3], we used meta-analytic methods to establish the magnitude of the pooled sex differences in the incidence of HAV infection, by age group, across different countries and over a number of years. The outcome variable was the male to female IRR. For each age group, the IRRs for each country were pooled over time periods and then the pooled IRRs for each country were combined. Forest plots with the pooled IRRs, over countries and years of reporting, were prepared separately for the seven age groups. Heterogeneity was evaluated using the Q statistic and I 2 was calculated as an estimate of the percentage of between-study variance. If the p-value for the Q statistic was less than 0.05, or I 2 exceeded 50%, the random effects models was used to estimate pooled IRRs and 95% confidence intervals (CI). Otherwise, the fixed effects model was considered, although due to the low power of the Q statistic, the more conservative random effects model was preferred. In order to explore the contribution of countries and the reported years to the variability in the IRRs, meta-regression analyses were performed. To evaluate the effect of individual countries and years on the male to female incidence risk ratio, we performed leave-one-out sensitivity analysis and recomputed the pooled IRRs. The meta-analytic methods and meta-regressions were carried out using STATA software version 12.1 (Stata Corp., College Station, TX).

Descriptive statistics
The summary of the male to female IRRs per 100,000 populations in different countries for each age group is presented in Table 1.
Significant differences in incidence rates were observed between the countries, with the highest incidence rates in all ages and both sexes in Czech Republic. Higher incidence rates were observed in Israel and Spain up to age 44 and in Germany in the group of adults (age [45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60][61][62][63][64]. Forest plots. The forest plots for the IRRs by age group, are shown in Figs 1-7. The forest plot for infants is shown in Fig 1. The pooled male to female IRR was 1.18 (95% CI 0.94-1.48) with I 2 = 0.0% and varied between 0.86 in Canada and 2.26 in Spain. Other analyses. Meta-regression analysis showed that almost all the variance in the incidence RRs was contributed by the age groups, with small differences between countries and time periods. To evaluate the effect of individual countries on the male to female incidence ratios, we performed leave-one-out sensitivity analysis and recomputed the pooled IRRs (presented in Tables 2 and 3).
After omitting each country (one country at a time, Table 2) or a group of years at a time (Table 3), the pooled IRR's remained very similar.
Thus, no single country or group of years substantially affected the pooled IRRs. This confirms that the results of this pooled analysis are stable and robust.

Discussion
In this study, we found that the incidence rates of clinically manifested HAV, pooled over a number of years, for various high-income countries, are consistently higher in males in all age groups. In the youngest and oldest age groups, where the numbers were small, the confidence intervals included unity. Based on the pooled analysis of national data from nine countries, over a period of 6-25 years, we found that the incidence rates of clinical hepatitis A were higher in males by 22%, 7%, 9%, 46%, 32%, and 10% in the age groups 1-4, 5-9, 10-14, 15-44, 45-64 and 65+ respectively.

PLOS ONE
While sex differences in the incidence of HAV have been examined in a number of studies, they have usually been conducted in individual countries or selected groups of patients. For example, in a national study in Israel in 1992, there was a male predominance of HAV incidence rates [30]. This sex differential was especially pronounced among infants. In a 15-year nationwide epidemiological study in Taiwan, there were higher hospitalization rates in males while male sex and age over 40 years were significant factors associated with mortality [31]. In the study of HAV patients in Saudi Arabia, no sex differences were among hospitalized patients [12]. In a hepatitis A outbreak in Chiba, Japan, in 2011, 40.7% of the 27 patients were male [32], and in another, 65% of the 60 patients were male [33]. However, these figures may simply represent gender differences in exposure to the virus. In addition, the impact of vaccines on sex differences in HAV incidence rates is not clear. There is evidence that females may respond with up to 2-3 times higher anti-HAV antibody levels than males after the priming and after the booster dose and has been observed at different ages [34][35][36][37].
The incidence of both viral and bacterial diseases have frequently been reported to be higher in males [1][2][3]. In addition, there are reported sex differences in the severity of different infections, suggesting that males are more prone to suffer from clinical manifestations of infections than females [38,39]. While in excess morbidity in males is most common for infectious diseases [1][2][3], pertussis is a prominent exception, where there is a female excess in morbidity [40]. It is of interest that in the COVID-19 pandemic, there has been no clear evidence of sex differences in incidence rates, although case-fatality rates have consistently been reported to be higher in males [41,42], even after controlling for other variables.
It has been shown that the male to female IRRs differential will be most evident where there is a low proportion of clinical disease [30]. Since children more commonly suffer from asymptomatic HAV infection [43,44] and the clinical to subclinical ratio for HAV increases with age, one might expect that the male excess in disease would be less evident at older ages.

PLOS ONE
However, the higher male to female IRRs in the older age groups is most likely due to larger differences in exposure in high risk groups such as in the men who have sex with men (MSM) or people who are HIV positive [7,8,[45][46][47][48][49]. Thus, behavioral factors can partially explain sex differences in HAV incidence rates in the older age groups. For the youngets age group, there may be protection from maternal HAV antibodies on short-term immunity [50]. However, we have not found evidence that it impacts male and female infants differently.
The exact mechanisms underlying the excess HAV incidence rates in males found in this study are not clear and probably multi-factorial. This study was not designed to address the mechanisms. In addition to behavioral differences, genetic and hormonal factors could be important. In infants and early childhood, and based on the seroprevalence studies, it is unlikely that the sex differences in incidence rates are due to differences in exposure [51]. A study of kindergarten children showed that females had higher anti-HAV antibodies than

PLOS ONE
males [52]. In adults, the results are varied. In a study of blood donors in the US in 2015, [53] no sex differences were observed in the prevalence of anti-HAV IgG antibodies (61% and 60% for males and females, respectively). In a study of ambulatory patients in Portugal between 2002 and 2012, no significant differences between sexes were observed [54]. In a study of refugees and asylum seekers in Germany, HAV seroprevalence rates were higher in adult males than females [55].
Although liver injury in hepatitis A is known to be caused by immune-mediated events, the exact biological mechanisms are not clarified. It is plausible that immune-related mechanisms of liver injury are common to the pathogenesis of all types of hepatitis [56]. Virus-specific CD8 + T cells from hepatitis A patients are considered as a major cause of liver damage. Natural killer cells are also involved and contribute to liver damage [57,58]. In hepatitis A patients, serum levels of cytokines and chemokines, including interleukin (IL)-6, IL-8, IL-18, IL-22, CXC-chemokine ligand (CXCL)9, and CXCL10 are increased [59] and contribute to liver injury. Many studies have shown that the overall inflammatory response, innate and adaptive immune systems are stronger in females than males, with greater CD4+ T-cell counts a higher CD4+ /CD8+ ratio in females but higher CD8+ T and NK frequencies in males [60].
Sex differences in the clinical expression of hepatitis A may be related to the imbalance in the expression of genes encoded on the X and Y-chromosomes of a host. X chromosome-associated biological processes and X-linked genes are responsible for the immunological advantage of females due to the X-linked microRNAs related processes. The phenomenon of X chromosome inheritance and expression is a cause of immune disadvantage of males and the enhanced survival of females following immunological challenges [61].
The increase in sex hormone levels in infancy that mimics sex steroid levels during puberty ('minipuberty') could affect immune cells differently in boys and girls. Testosterone levels predominate in boys at 1-3 months of age and decline at 6-9 months of age, whereas in girls, estradiol levels remain elevated longer [62]. This phenomenon of ''mini-puberty'' with sex differences in gonadal hormone levels could influence the maturation of the immune system [63]. This transient rise in sex steroid levels may also influence immune cells differently between boys and girls at later ages [64]. Before any physical signs of puberty, girls had higher levels of estrogens than boys at age 5-9. These higher estradiol levels or lower testosterone levels in young girls may play a part in protection against clinical disease and should be investigated further.

Strengths and limitations
This current study has several strengths and limitations. The inclusion of nine countries, each evaluated over a number of years, allowed us to evaluate the consistency of the findings over different populations and many years. The analyses are based on national data where both the numbers of cases and denominators are large. Selection bias has been minimized by using national data, which should be representative of each country. However, the countries evaluated in this study are classified as high-income, so the results may not be directly generalizable to low-and middle-income countries. Differential underreporting between countries is likely and may contribute to the variability in the incidence of reported cases of HAV. However, there does not appear to be any reason to believe that the reporting differs between males and females. In the countries examined, there is no evidence that male infants and children are more likely to receive health care. Thus any information bias in the underreporting of incidence rates will most likely be non-differential by sex and the IRRs should not be materially affected. In adults, there could be gender differences in the utilization of medical care, although reports suggest that females in some countries tend to make greater use of health services [65], which would operate in the opposite direction of our observations.

Conclusions
This study provides stable estimates of the excess male incidence rates in hepatitis A incidence rates in most age groups. While much of the excess in older males may be attributed to differential exposure, the excess in young males, while not large, is remarkably consistent over a number of high-income countries and for extended periods of time. The mechanism is largely unknown. A better understanding of the gender differences can help to elucidate genetic and hormonal determinants of HAV infection and contribute to the role of sex as a biological variable.