Trends in global and national infertility and factors associated with primary infertile couples in recent middle-aged Chinese.

Infertility is defined as an inability to conceive without contraception for more than 1 year. The disease has aroused extensive concerns. Globally, the disease burden has increased by 0.37% and 0.29% during the past three decades for females and males, respectively [ 1 ]. In China, the epidemiology of infertility among the reproductive population remains poorly understood, except the the population-based study conducted in 2010 [ 2 ]. The study showed that the overall prevalence of infertility was 15.5% [ 2 ]. However, due to the long elapsed time and a lack of infertility information at a national level, an urgent need exists to explore the prevalence of infertility, particularly primary infertile couples in middle-aged Chinese, which is lacking in the literature. The reproductive population in China, especially the middle-aged, faces a heavy workload and a large amount of stress to fight and live for their life. An overwhelming burden might pose a negative impact on an individual’s reproductive system, resulting in a childless life because of primary infertility [ 3 – 6 ]. Hence, we tried to investigate the global infertility rate and associated factors in the Chinese middle-aged childless population through both the Global Burden of Disease (GBD) study and the Chinese Health and Retirement Longitudinal Study (CHARLS).

From 1990 to 2019, the global infertility population has shown a steady annual increase ( Fig 1A ). The situation was also applied to DALYs and YLDs ( Fig 1B , 1C ). In the female group, all three parameters were approximately double those in the male group ( Fig 1 ). As for countries falling into different SDI subgroups, we found the prevalence of female infertility in high SDI countries was slightly higher than the prevalence of male infertility after adjusting for age ( Fig 2A ). While in other subgroups, the discrepancy was more obvious. Both high-middle SDI and middle SDI subgroups had the largest number of infertile people per 100000 population ( Fig 2A ). In addition, based on the calculated SDI in China, the country has been considered to be among the middle SDI subgroup ( S1 Fig ), where infertility prevailed and ranked high among subgroups ( Fig 2A ). A. Global prevalence number of infertility. B. Global DALYs of infertility. C. Global YLDs of infertility. GBD: Global Burden of Disease; DALYs: disease-adjusted life years; YLDs: years lived with disability. A. Global age-standardized prevalence rate of infertility among countries with different SDI for male (left) and female (right). B. The age-standardized prevalence rate of infertility around the world (left) and in China (right). GBD: Global Burden of Disease; SDI: socio-demographic index. In China, the age-standardized prevalence rate of infertility has remained relatively stable over the past three decades ( Fig 2B ). However, this rate was notably higher than the global age-standardized infertility prevalence rate, particularly for female infertility ( Fig 2B ). According to the GBD study, we retrieved data from the 2015 CHARLS Wave 3. After filtering ineligible samples, we finally included eligible couples (population size = 56, 892, 517) and individuals (population size = 110, 166, 036) for further analysis, among which we identified 947, 953 potential infertile couples ( Fig 3 ). The rough prevalence rate of primary infertile couples was 1.7% (947, 953/56, 892, 517, Fig 3 ). CHARLS: Chinese Health and Retirement Longitudinal Study. Table 1 demonstrated demographic information, physical examination, history of disease, regular habits, mental health, and individual data. Initially, we found that infertile couples usually had bad life habits, like less moderate activities and drinking a lot. And they also showed psychological disorders, like anxiety and depression ( Table 1 ). CHARLS: Chinese Health and Retirement Longitudinal Study; BMI: body mass index; HDL: high-density lipoprotein; LDL: low-density lipoprotein. # : Average hours per night during the last month. ## : Average hours for a nap during the last month. ### : Do moderate physical activities that make breathing somewhat harder than normal and may include carrying light loads, bicycling at a regular pace, or mopping the floor. Do at least 10 minutes continuously during a usual week. #### : Defined as used to smoke and have not quit till now. ##### : Defined as used to drink more than once a month and have not quit till now. ###### : According to the feelings and behavior during the last week. ####### : According to the last year. Taking the baseline characteristics into account, we enrolled related factors for further logistic regression analysis. We discovered that after adjusting with several confounders, younger age (OR: 0.68, 95% CI: 0.50–0.92, Table 2 ), longer nap time (OR: 1.01, 95% CI: 1.00–1.02, Table 2 ), and presence of anxiety (OR: 9.99, 95% CI: 2.86–34.92, Table 2 ) were significantly associated with infertile couples. Multicollinearity tests indicated that all included variables in the logistic regression model were far from collinearity (VIF < 10, S1 Table ). OR: odds ratio; CI: confidence interval. The data were generated from the CHARLS database. * : We enrolled individuals with available variables into logistic regression analysis. We finally included year, location, hypertension, dyslipidemia, diabetes mellitus, sleeping time, nap time, moderate activities, smoke, drink, anxiety, and depression into the logistic regression model. In the logistic regression model, the parameters’ codes were as follow: infertile couples (1 = yes, 0 = no); year (continuous variable); location (1= from main city zone, 0 = not from main city zone); hypertension (1 = yes, 0 = no); dyslipidemia (1 = yes, 0 = no); diabetes mellitus (1 = yes, 0 = no); sleeping time (continuous variable); nap time (continuous variable); moderate activities (1 = yes, 0 = no); smoke (1 = yes, 0 = no); drink (1 = yes, 0 = no); anxiety (1 = yes, 0 = no); depression (1 = yes, 0 = no). ** : Not available because individuals with dyslipidemia, diabetes mellitus, and moderate activities were all within fertile couples. # : Average hours per night during the last month. ## : Average hours for a nap during the last month. ### : Do moderate physical activities that make breathing somewhat harder than normal and may include carrying light loads, bicycling at a regular pace, or mopping the floor. Do at least 10 minutes continuously during a usual week. #### : Defined as used to smoke and have not quit till now. ##### : Defined as used to drink more than once a month and have not quit till now. ###### : According to the feelings and behavior during the last week.

In summary, infertility continued to be a pressing issue on both global and national levels. The situation should arouse widespread attention among Chinese authorities. In China, the overall prevalence of primary infertile couples among the middle-aged group was around 1.7%. The findings might guide future policy-making and medical interventions in China, to put an emphasis on supporting the reproductive needs of middle-aged individuals.

The study utilized public databases, and institutional review board (IRB) approval was exempted. The GBD study was conducted by the Institute for Health Metrics and Evaluation (IHME) and was available for non-commercial use ( https://www.healthdata.org ). The study focused on numerous diseases and injuries around the world and in various countries with different socio-demographic indices (SDI) to calculate measures including prevalence, incidence, disease-adjusted life years (DALYs), years lived with disability (YLDs), and years of life lost (YLLs). To determine SDI, the geometric mean of lag-distributed income, average years of schooling among individuals aged 15 years and above, and total fertility rate is calculated [ 1 ]. More information on how these estimates were generated could be accessed in related articles [ 7 – 9 ]. CHARLS aimed to collect representative samples of Chinese middle-aged and elderly people with high quality [ 10 ]. A multi-stage cross-sectional survey was applied in 150 counties/districts, each with 3 randomly selected primary sample units (PSUs), resulting in 450 villages/resident committees [ 11 – 12 ]. More information was able to be reached at http://charls.pku.edu.cn . Infertility is mainly classified into two types: primary infertility and secondary infertility. The critical distinction between them is the existence of reproductive history. For the GBD database, total infertile (including primary and secondary) prevalence was initially estimated among surveyed married respondents, and then the surveyed population was proportioned to the overall population. The gender factor was taken into account, and a prediction model was established utilizing DisModMR 2.1, a software implementing the Bayesian meta-regression method to calculate non-fatal outcomes based on sparse and heterogeneous epidemiological data [ 7 – 9 ]. However, in our analysis of the Chinese population (CHARLS data), we were only able to identify potential primary infertility populations in an age subgroup, as a result of survey design. We included participants who were cross-sectioned and had been married. According to the report of the National Bureau of Statistics in 2018 ( http://www.stats.gov.cn/zt_18555/ztfx/ggkf40n/202302/t20230209_1902601.html ), the mean age of marriage and childbirth in China was 25.7 and 26.8, respectively, from which we assumed that most of the fertile couples had attempted to conceive and not utilized contraception for at least 1 year in their thirties. As CHARLS surveyed and represented the Chinese middle-aged population, we restricted the age to 35–50 years old, a group where most Chinese have been married for more than 1 year and taken the responsibility for fostering children if possible. Infertile couples would experience their life without babies during this period. In other words, we considered the childless population as a proxy for the potential primary infertile population. To illustrate the global trend of infertility, we initially downloaded data from the GBD website with various parameters, including location data (global, low SDI, low-middle SDI, middle SDI, high-middle SDI, and high SDI), year data (1990–2019), context data (cause), age data (all ages and age-standardized), metric data (number, percent, and rate), measure data (prevalence, DALYs and YLDs), sex data (male and female) and cause data (male infertility and female infertility). Furthermore, we explored the national trend of age-standardized infertility in China, utilizing corresponding parameters. To demonstrate infertility prevalence in China, we searched the 2015 CHARLS Wave 3 database and emphasized participants’ age between 35–50, an age group that is usually considered to have been married and attempted to conceive for at least 1 year ( http://www.stats.gov.cn/zt_18555/ztfx/ggkf40n/202302/t20230209_1902601.html ). Then, we relied on the item “xchildnum: total number of respondent’s children”, “xchildnum_alive: number of respondent’s living children”, “cb051_w3_1: number of respondent’s unrecorded living children”, and the question “whether cross-section sample” to identify the possible childless population. Additionally, we excluded participants who had separated, divorced, and not lived with their spouses for reasons like work (item be001). After all these selections, we were able to identify our targeted primary infertile population. Moreover, to discover associated factors of primary infertility in China, we also collected demographic background data (birth date, gender, and location), health status, and functioning data (history of hypertension, dyslipidemia and diabetes mellitus, sleeping time, nap time, frequency and length of moderate activities, history of smoke and drink, anxiety and depression), individual income data (income per year), biomarker data (height, weight, and waist circumference), weights data (individual weight with household and individual response adjustment and biomarker weight with household and individual response adjustment), sample information data (interview year), PSU data (province English names, city English names, strata, and PSUs) and blood data (blood weight, fasting triglycerides, fasting high-density lipoprotein cholesterol, fasting low-density lipoprotein cholesterol, fasting total cholesterol, fasting blood glucose, and glycated hemoglobin). Categorical variables were classified as follows: location (main city zone, combination zone between urban and rural areas, and village and other places), hypertension (with and without), dyslipidemia (with and without), diabetes mellitus (with and without), moderate activities (yes and no, do at least 10 minutes continuously during a usual week), smoke (still smoke, and quit or never smoke), drink (drink more than once a month, and quit or never drink or drink a little), anxiety (with and without) and depression (with and without). All the analysis was completed on Stata SE 16.0 (Stata Corp) and GraphPad Prism 9 (GraphPad Software Inc., San Diego, CA). CHARLS data were derived from a multi-stage cross-sectional study in China, and participants were surveyed according to previous stratified regions (strata = 150) with probability proportional to population size. For each region, 3 randomly selected PSUs were surveyed with probability proportional to population. Hence, we decided to utilize sampling weight, strata, and PSUs parameters to comprehensively delineate infertility prevalence among the Chinese elderly population. We would report the calculated population size in our analysis. To explore associated factors in primary infertile couples, we enrolled individuals with available variables in logistic regression analysis. A logistic regression model based on the design of sampling weight was applied to explore the relationship between infertility and the associated factors. Due to the presence of missing values, we finally were only able to include year, location, hypertension, dyslipidemia, diabetes mellitus, sleeping time, nap time, moderate activities, smoke, drink, anxiety, and depression in the logistic regression model. To avoid correlation among variables, multicollinearity tests were performed to examine their independence from one another. Generally, a variance inflation factor (VIF) less than 10 was considered to be without significant multicollinearity. Continuous variables were shown as mean±standard deviation (SD), and categorical variables were represented as percentages. The results were reported as odds ratio (OR) with 95% confidence interval (95% CI). The two-sided test level α was set to 0.05.

GBD: Global Burden of Disease; SDI: socio-demographic index. (PDF) (DOCX)