Male- and female-specific reproductive risk factors across the lifespan for dementia or cognitive decline: a systematic review and meta-analysis

We followed the updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) statement’s instructions [9]. Systematic searches were conducted through PubMed, EMBASE, and Cochrane Library principally up to January 1st, 2023, for studies focusing on the associations of female- and male-specific factors with dementia or cognitive decline. As there is already extensive literature on studies of hormone replacement treatment (HRT) in females, this paper will not discuss that topic [10]. Leveraging prior research [11] and guidance from experienced gynecologists and andrologists, we identified female-specific reproductive risk factors using the following search strategy: menarche, placental bed disorders, menstrual cycle, polycystic ovarian syndrome, reproductive period, menopause, climacteric symptoms, reproductive history, estradiol, sex hormone–binding globulin, pregnancy, gravidity, parity, breastfeeding, intrauterine growth retardation, preterm delivery, stillbirth, induced abortion, gestational diabetes, ectopic pregnancy, hyperemesis gravidarum, gestational hypertension, ovarian hyperstimulation, in vitro fertilization, oophorectomy, hysterectomy, subfertility, cesarean section, natural labor, dystocia, fetal death, multiple pregnancies, postpartum hemorrhage, amniotic fluid embolism, puerperal infection, postpartum depression, hyperprolactinemia, amenorrhea, premature ovarian failure, hypothalamus-pituitary-ovary axis, test tube baby, premature rupture of membrane, intrahepatic cholestasis of pregnancy. The following strategy was used for male-specific reproductive risk factors: erectile dysfunction, testosterone, dihydrotestosterone, androstanediol glucuronide, androgen, androgen deprivation therapy, orchiectomy, haplotype Y chromosome cryptorchidism, prostatic hyperplasia, cryptorchidism, varicocele. The literature search strategy included the following terms for outcomes: Alzheimer, dementia, cognitive, and cognition. The main outcome was dementia or cognitive decline. When multiple outcome types were reported for a factor, findings were categorized into the subgroups cognitive decline, dementia, and AD.

The inclusion criteria were as follows: (1) Studies were required to be cohort studies, however, for female reproductive factors (menstrual factors, parity, and gynecological operations), besides cohort studies, case–control studies were also included given the expectation of modest recall bias; (2) The measurement of the outcomes had to be described in detail, with cognitive decline assessed using standard and full-scaled cognitive tests, and dementia or AD diagnosed using objective and globalized diagnostic criteria such as Diagnostic and Statistical Manual of Mental Disorders criteria, International Classification of Diseases codes, and National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association criteria; and (3) For dose–response analysis, the pooled related factors needed to be separated into at least three levels, with specific or calculable person-years and case numbers within each level.

Two researchers independently screened the included studies. If a discrepancy arose, the third author was asked to determine whether to include or exclude the study. When two or more studies originated from the same database, the study with the biggest sample size and/or most detailed information was retained. Additionally, we combed through the bibliographies of qualified studies to avoid overlooking any potentially pertinent studies.

Predesigned templates were used to extract data from each article, including first author, publication year, study design, population resource, cognitive status at the baseline, mean age, gender composition, follow-up duration, attrition rate, total sample size for analysis, incident case, type of outcome, outcome measurement, type of exposed factor, measurement of exposed factor, adjusted confounders, and risk estimates. If any required data was not reported in the publication, we contacted the authors to obtain it. Two authors with extensive experience extracted the data, and any disagreements were resolved with assistance from a third reviewer.

The Newcastle–Ottawa Scale (NOS) was used to evaluate potential bias. In order to more accurately measure potential bias in studies, a modified version of the NOS was utilized [12] (Additional file 1 Appendix A and Appendix B). The NOS can fully evaluate a single study, incorporating various criteria such as representativeness, comparability, objectivity, and reliability (Additional file 1: Appendix C).

Above all, the risk estimates and 95% confidence intervals (CIs) for a series of risk factors were acquired for further analysis. When odds ratios (ORs) were provided in some articles instead of relative risks (RRs) or hazard ratios (HRs), we used the following algorithm to convert ORs to RRs [13]:

P₀ represents the incidence of dementia or cognitive decline in the non-exposed group. If P₀ cannot be calculated, the overall incidence rate of the entire sample can be used instead.

To begin, we specified definitions of exposures to facilitate comparison of pooled data across studies: (a) early menarche was defined as age at menarche ≤ 13 years; (b) late menarche was ≥ 16 years; (c) early menopause was ≤ 45 years; (d) late menopause was ≥ 54 years; (e) short reproductive period was ≤ 34 years; (f) long reproductive period was ≥ 38 years; (g) early childbearing was ≤ 20 years; and (h) late childbearing was ≥ 30 years. The fixed model combined risk estimates of the same category within a study, while the random model pooled estimates across studies [13]. Heterogeneity was evaluated by the Q test and quantified by the I² metric [14]. For factors with ≥ 10 studies, subgroup analysis and meta-regression were conducted.

Dose–response analysis for eligible components was undertaken using the inverse variance weighted least squares regression with cluster robust error variances (REMR model) [15]. For studies that did not use the lowest category as the reference group, we reassigned the reference group and recalculated the effect sizes using Orsini’s method [12]. When a range was provided, the midpoint represented the average exposure level. For open-ended categories, the exposure level was set to the boundary limit plus/minus the interval length of adjacent groups [16]. Figures and analyses were performed using GraphPad Prism 9.0 and Stata Version 12.0.

Figure 1A depicts the process of selecting literature. After evaluating 10,153 publications, a total of 94 studies were identified, with 9,839,964 females and 3,436,520 males. The included studies consisted of 82 cohorts and 12 case–control studies (all case–control studies were included in analyses of female reproductive factors). Of these, 68 of them were from the community and 26 were from the hospital. Basic study information was provided in Additional file 1: Appendix E. Additional file 1: Appendix F summarizes the outcome measurements used in the included studies. Additional file 1: Appendix G contains the risk estimates from each study for each sex-specific reproductive factor.

Female-specific reproductive factors were classified into 5 categories according to life stages: (1) puberty-related factors, (2) latency period-related factors, (3) peri-menopause-related factors, (4) sex hormones in post-menopause, and (5) gynecological operations across the lifespan. Male-specific reproductive factors were sorted into 2 categories: (1) male-specific diseases and related factors and (2) late-life sex hormones. Figure 1B summarizes the number of included studies investigating each risk factor in relation to three outcomes: cognitive decline, dementia, and AD (Fig. 1).

In puberty, late menarche could elevate the risk of dementia or cognitive decline by nearly 15% (RR = 1.15, 95% CI = 1.14 to 1.47, pooled studies = 4, I² = 0%). In the latency period, the pooled RRs were 1.14 (95% CI = 1.05 to 1.24, pooled studies = 8, I² = 72.3%) for the short reproductive period and 0.91 (95% CI = 0.83 to 0.99, pooled studies = 7, I² = 0%) for the long reproductive period. For parity, increased risks were seen for both nulliparous (RR = 1.11, 95% CI = 1.06 to 1.16, pooled studies = 10, I² = 0%) and grand parity (≥ 5) (RR = 1.28, 95% CI = 1.15 to 1.44, pooled studies = 4, I² = 0%), with a 3% increase per parity (95% CI = 1.15 to 1.44, pooled studies = 3, I² = 0%). One study reported a decreased risk of dementia with infertility (RR = 0.80, 95% CI = 0.68 to 0.96) (Fig. 2).

In the peri-menopause period, early menopause was associated with a 22% higher risk of dementia or cognitive decline (RR = 1.22, 95% CI = 1.11 to 1.34, pooled studies = 7, I² = 68.5%), while late menopause was protective, lowering risk by 7% (RR = 0.93, 95% CI = 0.91 to 0.96, pooled studies = 5, I² = 0%). Dose–response analysis showed an inverse linear relationship between age at menopause and dementia/cognitive decline (p < 0.0005) (Additional file 1: Appendix H). For postmenopausal hormone levels, pooled RR was 1.46 (95% CI = 1.15 to 1.85, pooled studies = 4, I² = 0%) for increased total estradiol. Findings from an individual study suggested that increased free estradiol levels could reduce the risk of cognitive decline (RR = 0.30, 95% CI = 0.10 to 0.90). Higher sex hormone–binding globulin (SHBG) levels increased the risk of dementia in one study (RR = 1.30, 95% CI = 1.10 to 1.70). For gynecological surgery, five studies on bilateral oophorectomy gave a pooled adjusted RR of 1.08 (95% CI = 1.02 to 1.15, I² = 0), while one study on hysterectomy with bilateral oophorectomy had a RR of 0.85 (95% CI = 0.75 to 0.97) (Fig. 3).

The pooled RRs for dementia or cognitive decline with ADT were 1.18 (95% CI = 1.08 to 1.29, pooled studies = 13, I² = 94.5%) compared to prostate cancer patients not receiving ADT, and 1.17 (95% CI = 1.09 to 1.25, pooled studies = 2, I² = 68.7%) versus those without prostate cancer. By ADT type, RRs versus prostate cancer controls were 2.03 for antiandrogens (95% CI = 1.76 to 2.35, pooled study = 1) and 1.47 for orchidectomy (95% CI = 1.11 to 1.95, pooled study = 1). Versus healthy controls, RRs were 1.60 for orchidectomy (95% CI = 1.32 to 1.93, pooled study = 1) and 1.15 for gonadotropin-releasing hormone (GnRH) agonists (95% CI = 1.07 to 1.23, pooled study = 1). Subgroup analysis was conducted by race, follow-up time, study quality, and sample size (Additional file 1: Appendix I). Meta-regression analysis was performed on the pooled sample sizes, mean follow-up years, years of publication, and mean ages, but the degree of heterogeneity among the aforementioned factors could not be distinguished. Dose–response analysis showed an increased risk of dementia or cognitive decline with longer ADT duration (Additional file 1: Appendix H).

Erectile dysfunction (ED) (RR = 1.67, 95% CI = 1.34 to 2.08, pooled study = 1) and benign prostatic hyperplasia (BPH) (RR = 1.21, 95% CI = 1.17 to 1.24, pooled study = 1) were associated with increased risk of dementia. Additionally, higher SHBG levels were linked to an elevated risk of dementia or cognitive decline (RR = 1.22, 95% CI = 1.06 to 1.39, pooled studies = 5, I² = 0%) (Fig. 4).

Based on GRADE scores, all characteristics included in the meta-analysis were rated as a certainty of “extremely low” (Additional file 1: Appendix J). The low level of certainty was mainly driven by low scores on risk of bias, with 58% of studies graded as high concern. In addition, the limited indirectness and public of bias also contributed to the low GRADE scores.

According to the results of index S divergence, 40% of female-specific reproductive factors had divergence > 50%. Ranging from − 44 to 72%, parity was the most controversial factor (index S divergence =  − 0.04). For male-specific reproductive factors, 28.5% had divergence > 50%, with ADT being the most disputed result (index S divergence =  − 0.02) (Additional file 1: Appendix K and Appendix L).