Consanguineous marriages and their association with women’s reproductive health and fertility behavior in Pakistan: secondary data analysis from Demographic and Health Surveys, 1990–2018

Background Pakistan has been showing consistently the highest prevalence of consanguinity. The popularity of consanguineous marriages is not declining in the country, because of social, cultural, and religious beliefs as well as economic advantages. However, couples also face various health-related implications, such as poor pregnancy outcomes or multiple reproductive and fertility consequences, having adverse effects on mothers and their children. This research investigated the trend of consanguineous marriages and their association with women’s reproductive health and fertility behavior in Pakistan from 1990 to 2018. Methods This study is based on secondary data analysis, using all four waves of the Pakistan Demographic Health Surveys carried out from 1990 to 2018. The analysis is limited to women aged 15–49 years, who had given birth in the previous five years preceding each survey. Sampling weights were calculated and subsequently weighted analysis was conducted. Descriptive statistics, bivariable and multivariable logistic regression analysis were performed to determine the association of consanguinity with multiple characteristics related to socio-demographics, co-variates, and women’s reproductive health and fertility behaviors. Results The findings revealed a high but overall stable trend of consanguinity prevalence of about 63% during the last three decades. Consanguineous marriages were more prevalent amongst young and uneducated women, living in rural areas, with poorer wealth status and having less exposure to mass media to access information. A strong association of consanguinity was observed with women’s reproductive health and fertility behavior, particularly for women who gave first birth at a younger age, had multi-gravida pregnancies, multi-parity, pregnancy termination, ANC visits, and higher fertility. Conclusion Consanguineous marriages are predominant in the patriarchal society of Pakistan. Findings revealed that consanguinity contributes significantly to women’s reproductive health and fertility behaviors. Appropriate counseling, educational, and health promotional programs related to consanguinity should be designed and launched at the community level to raise awareness about risks towards women’s reproductive health and fertility.

Background Consanguinity is termed as wedlock or marriage between close blood relations or biological kin. Consanguineous marriages have been very common since the early existence of humanity. According to a rough estimate, nearly one billion (20%) of the global population live in communities with a preference for consanguineous marriages [1,2], predominantly in Muslim countries of the Middle East, Africa, and South Asia [3,4]. With 65%, Pakistan has one of the highest rates of cousin marriages globally, followed by India (55%), Saudi Arabia (50%), Afghanistan (40%), Iran (30%), Egypt, and Turkey (20%) [5]. The prevalence of consanguineous unions differs amongst countries due to socio-demographic factors, such as geography (urban-rural residential community, isolated area, and population), religion, education, socio-economic status, a familial pattern towards early marriages, or consanguinity between parents [4,[6][7][8][9][10][11][12]. Although the incidence of consanguineous unions somehow decreased with urbanization, modernization, and smaller/nuclear families, however, it is still in practice [13].
Pakistan, a multi-cultural country with diverse caste systems, has been shown consistently the highest prevalence of consanguinity [14,15]. Consanguineous marriages are encouraged in the country due to multiple reasons, e.g. to strengthen interfamily ties between close family members, a preference to the same caste and status, fear of incompatibility or difficulty in finding the right partner outside the family, security of being familiar with spouse and in-laws before marriage, restriction for socialization with the opposite gender, and financial constraints especially for dowry [16]. However, various socio-cultural and health-related implications have been identified for consanguineous couples [5]. Owing to shared alleles, consanguinity may lead to genetic disorders [17,18], poor pregnancy outcomes, or multiple reproductive and fertility consequences, having adverse effects on mothers, their children, family, and society as a whole [5,13]. A strong association of consanguineous marriages has been reported with increased rates of abortion, stillbirths, pregnancy terminations, low birth weights, increased mortality, and congenital malformations [13,17,[19][20][21][22][23]. Furthermore, a low preference for contraception, extended childbearing age, and higher fertility has also been observed in such unions [8]. Although consanguineous marriages are also linked with poor pregnancy outcomes [22] and higher reproductive risks in Pakistan, nonetheless these are also associated with increased fertility rates and larger family size [24,25].
Consanguineous unions have remained under continuous investigation by social scientists, medical researchers, biologists, and physicians. However, it received less attention in mainstream demographic research. Although multiple studies are available regarding the effects of consanguineous marriages on either reproductive health or fertility behaviors [22,24,25], nevertheless, there is a need to examine and explain the trends of consanguineous marriages and their association with women's reproductive health and fertility behavior in Pakistan. To our best knowledge, this is the first paper to examine trends in consanguineous marriages in Pakistan in recent decades. Therefore, this research is an attempt to investigate the differentials in reproductive health and fertility behaviors over almost three decades .

Study design and data source
This research performed secondary data analysis, using all four waves of Pakistan Demographic and Health Surveys (PDHSs), carried out during 1990-1991 (wave 1) [26], 2006-07 (wave 2) [27], 2012-2013 (wave 3) [28], and 2017-2018 (wave 4) [29]. These PDHSs are characterized as nationally representative and large-scale cross-sectional household surveys, conducted under the international series of MEASURE Demographic and Health Survey (DHS) Program and funded by the United States Agency for International Development (USAID). These surveys were carried out by the National Institute of Population Studies (NIPS) with the technical assistance of ICF International and the Pakistan Bureau of Statistics. A series of PDHSs is the largest household and publicly available dataset, with information on women's reproductive health, fertility behavior, marital status, and other socio-demographic variables. Each PDHS applied a random two-stage cluster sampling design, wherein firstly rural and urban sampling units were chosen, followed by the selection of eligible households with evermarried women (aged 15-49 years) [26][27][28][29].
Within each wave of PDHS, varied numbers of field teams collected data, each comprised of one male and three female interviewers, field editor, and supervisor. All teams were closely monitored by quality controllers, provincial/regional field coordinators, as well as the NIPS and ICF core team. Simultaneously, data editing, processing, and double data entry of completed questionnaires were also completed. Each wave of PDHSs used various questionnaires at household, community, women, and men levels for data collection.
This research limited the analysis to the women of reproductive age (15-49 years), who had given birth in the previous years preceding each of the four PDHS waves from 1990 to 2018. Therefore, the sample size used for this study was 4,061, 5,677, 7,446, and 6,711, for 1990-1991, 2006-2007, 2012-2013, and 2017-2018, respectively.

Variables
Outcome variable: The outcome variable for this research is the marital status of respondents, i.e. nonconsanguineous marriage versus consanguineous marriage. Further, types of consanguineous marriages were divided into three categories, including married to paternal first cousins, married to maternal first cousins, and married to relatives other than first paternal or maternal cousins.

Statistical analysis
Data analysis was conducted using SPSS version 21. Initially, sample weights were calculated and weighted analysis was performed. Descriptive statistics were presented in the form of frequencies and percentages. For bivariate analyses, cross-tabulation and chi-square tests were applied. Afterward, bivariable and multivariable logistic regression analyses were performed to measure the association of consanguineous marriages with sociodemographic characteristics, as well as women's reproductive health and fertility behaviors. During regression analysis, odds ratios (OR) and adjusted odds ratios (AOR) were calculated with a 95% confidence interval (CI). A p-value ≤ 0.05 was considered statistically significant.

Ethical considerations
Procedures and questionnaires for all DHS surveys have been reviewed and approved by ICF Institutional Review Board. Before each interview, an informed consent statement was read to the respondent, who had the chance to accept or decline to participate. The informed consent statements emphasized that participation is voluntary and provided details regarding the purpose of the interview, the expected duration of the interview, interview procedures, potential risks and potential benefits to respondents, as well as the contact information who could provide the respondent with more information about the interview. Table 1 highlights respondents' socio-demographic characteristics and other co-variates for the four PDHS waves from 1990 to 2018. The majority of women were of 25-34 years of age, uneducated (but with a lower trend over time: 79.2%, 64.6%, 55.8%, and 47.9%), and resided in rural areas (70.9%, 69.8%, 59.9%, and 66.5%). About three quarter or even more of the women were unemployed. Contrary to respondents, most of their husbands acquired a secondary level of education and almost all of them were employed. Although, findings revealed that about 60-70% of women had exposure to mass media, about half of the participants had no autonomy in healthcare decision-making. Table 2 shows the proportion of consanguineous marriages amongst women of reproductive age from 1990 to 2018. Results indicate that about two-thirds of women were married to their cousins, more frequently with paternal first cousins than maternal ones. Findings revealed an overall stable proportion of consanguinity prevalence during the last three decades. However, even a slightly upward trend of consanguineous marriages was witnessed from 63.0% in 1990-1991 to 67.9% in 2007-2008, followed by a gradual downwards trend. Similarly, the pattern of marriages with paternal and maternal first cousins decreased during the period from 1990 to 2013, nonetheless, slightly increased in 2017-2018. ). The proportion of deliveries conducted by unskilled birth attendants during the last pregnancy constantly decreased from 81.5% in 1990-1991 to 28.0% in 2017-2018. Findings also revealed the gradual improvement in women's reproductive healthcare-seeking behaviors, particularly in availing at least 4 ANC visits and deliveries conducted by SBAs from 1990 to 2018. Furthermore, regarding women's fertility behavior, the analysis indicated that most of the women (68.8%) reported an ideal family size of 5 children and above in 1990, which decreased over time to 3-4 children in 2017-2018 (51.8%). Although about half of the respondents showed no more desire for children or fertility intention in all four waves, there was a high-but overall decreasingrate of not using contraception (88.3%, 70.6%, 62.7%, and 63.9%).

Bivariate association of consanguinity with various factors
The association of consanguineous marriages with various factors-including sociodemographic characteristics, and reproductive health and fertility behaviors-amongst respondents of reproductive age in each of the four PDHS waves is presented in Table 3. Findings demonstrate the significant relationship of consanguinity with regions/provinces, urban/rural geographical classification, respondents' education, wealth quintile, mass media exposure, and healthcare decision-making autonomy in all waves. Respondents' age, employment status, and their husbands' educational status were also found statistically significant in most of the PDHS waves. Furthermore, results highlight the significant relationship of consanguinity with respondents' age at first birth, ANC visits, deliveries by SBAs, ideal family size, fertility intention, and current use of contraception in the majority of PDHS waves. In a few PDHS waves, a strong association of consanguineous unions was also observed with the number of living children, gravidity, and pregnancy termination. Table 4 depicts the results of the bivariable logistic regression analysis of consanguinity with respondents' sociodemographic characteristics, reproductive health, and fertility behaviors. Almost all sociodemographic variables in the four PDHS waves were significantly associated with consanguinity-at least one category per item. However, the husband's employment status was only significant for 2017-2018. Furthermore, gravidity and parity, and the number of children showed nearly no significant associations.

Bivariable and multivariable logistic regression
When interpreting the results of the multivariable logistic regression analysis (Table 5), one needs to keep in mind that not all variables and categories have been assessed in all four waves. However, the results highlight that respondent's age was not significantly associated with consanguinity. The region and geographical classification were significant predictors of consanguinity, except for 2017-2018. The impact of education was contrary between women and their husbands: Women with higher education had a lower likelihood for consanguineous marriages, whereas men with higher education were more likely to marry their relatives. The strength of the association for both variables reduced over time. Employment status both for women and their husbands showed inconclusive results-the only significant association was found for employed women showing a higher likelihood of consanguinity in 2012-2013 (AOR = 1.23, 95% CI 1.08-1.39, p < 0.01).All significant values (p < 0.05) are marked in bold In 1990-1991, the wealth index was not significantly associated with consanguinity, but for all other waves, a higher wealth quintile was linked with a lower likelihood of consanguineous marriages. For example, in 2012-2013 (AOR = 0.53, 95% CI 0.41-0.69, p < 0.01) and 2017-2018 (AOR = 0.47, 95% CI 0.35-0.62, p < 0.01) the likelihood was halved in the richest wealth quintile compared to the poorest one. There was no impact of mass media exposure and respondents' healthcare decision-making autonomy, except for 2012-2013, where women having decision-making autonomy were less likely to be married to relatives (AOR = 0.83, 95% CI 0.74-0.92, p < 0.01).
As already shown in the bivariable logistic regression models, respondents' age at first birth, gravidity and parity, and the number of children showed almost no or only in some years or categories significant associations. Nevertheless, there is a trend younger age at first birth and lower gravidity is linked with consanguineous marriages. In 1990-1991 and 2006-2007, a lower number of children was associated with a higher likelihood for  Table 3 Relationship of consanguinity with sociodemographic characteristics, reproductive health and fertility behavior

Consan-guineous
Non-consanguineous  Geographical classification       consanguineous marriages, ever having terminated pregnancy was only significantly associated with a higher likelihood for consanguinity in 2012-2013 (AOR = 1.29, 95% CI 1.15-1.44, p < 0.01). Visits of ANC and deliveries by SBAs were also no significant predictors, except for single exceptions. However, an ideal family size of less than five children was almost entirely significantly associated with not marrying a relative. The use of contraceptives was not significantly associated with consanguineous marriages.

Discussion
With the advanced research and expansion of knowledge in public health and social sciences, the topic of consanguineous unions has received higher importance.
All significant values (p < 0.05) are marked in bold

Reproductive health and fertility behaviors
Respondents' age at first birth  From the beginning of mankind, consanguinity or closekin marriages were socially and culturally deeply rooted.
Although it is presumed that the rates of consanguineous marriages decline with modernization and literacy, this is not transferable to all countries [30]. Presently, consanguinity is widely popular and respected in many communities, particularly in Muslims [2,3]. Pakistan ranks amongst those countries, where the highest prevalence of consanguinity is still in vogue [3,22,25]. This article examined the trends of consanguineous marriages over approximately three decades, from 1990 to 2018, and their association with women's reproductive health and fertility behavior in Pakistan. It is an effort to bridge the gap in existing literature, documenting the relevance All significant values (p < 0.05) are marked in bold This highlights the fact that the popularity of consanguineous unions is not declining in the country, because of social, cultural, religious, and economic advantages, which outweigh the disadvantages given the population [31]. In particular, consanguinity promotes family stability, inheritance, and spouse compatibility, nonetheless lessens hidden financial risks [6,7,16,32]. These results are similar to other studies, carried out in many subpopulations within Pakistan [33], such as northern Punjab [34,35], southern Khyber Pakhtunkhwa [36], Balochistan [37], Kashmir [38], and also in other Arab countries [1,39], Yemen [40], Qatar [41] and Algeria [42]. Contrary, these results are not consistent with some of the research, where a decreasing trend in consanguineous unions was reported over time [43,44].
This research also reiterated that consanguinity is associated with sociodemographic characteristics, as results demonstrated that consanguineous marriages are more prevalent amongst uneducated women, living in rural areas, and with poorer wealth status. These findings are comparable to other studies, where less-educated women get married to their cousins at a younger age, particularly in poor traditional rural areas [10,[45][46][47][48]. This highlights the need to educate and empower young girls, enabling them to make better informed decisions for their reproductive life to ensure their well-being. Previous empirical results found a strong association of consanguineous unions with women's reproductive health and fertility behaviors. Findings demonstrated that those women who married their cousins were more likely to give first birth at a younger age (between 20 and 34 years). Although not entirely significant in our analysis, we can also confirm this result. Our findings correspond to the previous studies, showing that consanguinity is associated with higher fertility rates and larger family sizes, which affects the health of both mothers and children [5,8,13,22,24,25], particularly in the case of younger women [49]. Thus, there is a need to educate communities about linkages of consanguinity with poor reproductive health, adverse impact on fertility outcomes, and overall family health. This research emphasizes educating families regarding implications of consanguinity and associated health risks, through increasing public awareness, providing informational material, promoting health education, and enhancing capacities of primary healthcare and outreach workers to counsel communities effectively on health and social issues related to consanguineous marriages. It is pertinent to actively engage all key stakeholders in the public and private sector, particularly healthcare providers, outreach workers, and social mobilizers to elucidate the health and social effects of consanguineous marriages and promote healthy mothers, children, and communities.

Limitations
Since this research analyzed the four waves of PDHS from 1990 to 2018, few variables were not uniform and found missing, particularly in PDHS 1990-1991 and 2006-2007, such as regions/provinces, mass media exposure, respondent's healthcare decision-making autonomy, and pregnancy termination. Due to the cross-sectional design, we are not able to draw any causal conclusions. When interpreting the results, one needs to consider that some of the variables might be predictors of consanguineous marriages (such as low education), whereas others are effects (such as visits of ANC and deliveries by SBAs) or both (such as ideal family size).

Conclusion
This research concludes that consanguineous marriages are predominant in Pakistan, particularly in the context of the large power structure and patriarchal society. Findings revealed that consanguinity is associated with sociodemographic characteristics and women's reproductive health and fertility behaviors in Pakistan. The high prevalence of consanguineous marriages and their implications on women's health is essential to be considered in health policies. Owing to dilute these prevailing socio-cultural practices, a nationwide public education program has to be conducted, engaging key stakeholders (including health managers, healthcare providers, and outreach workers) and highlighting the risk factors associated with consanguinity to minimize the adverse health outcomes. Further, there is also a dire need to actively engage public health and reproductive health professionals to promote the health and wellbeing of the female population. Healthcare providers, outreach workers, and social mobilizers may play a critical role in this regard, particularly in identifying the consanguineous couples within their serving community, counselling them and providing information on potential risk factors, and enabling them to make informed choices regarding their reproductive health. Appropriate counselling, health educational, and promotional programmes related to consanguinity should be designed and launched at health facilities and community level to build capacities of healthcare providers and raise awareness amongst the general population on danger signs. Though there are multiple socio-cultural and economic benefits of consanguinity perceived by women in Pakistan, improvements in health literacy and behavior change will endorse an attitudinal change in society.
Abbreviations ANC: Antenatal care; AOR: Adjusted odds ratio; CI: Confidence interval; OR: Odds ratio; PDHS: Pakistan Demographic and Health Survey; SBA: Skilled birth attendant; SPSS: Statistical Package for the Social Sciences.