Reproductive factors and the risk of incident dementia: A cohort study of UK Biobank participants.

BACKGROUND: Women's reproductive factors have been associated with the risk of dementia; however, these findings remain uncertain. This study aimed to examine the risk of incident all-cause dementia associated with reproductive factors in women and the number of children in both sexes and whether the associations vary by age, socioeconomic status (SES), smoking status, and body mass index (BMI) in the UK Biobank. METHODS AND
FINDINGS: A total of 273,240 women and 228,957 men without prevalent dementia from the UK Biobank were included in the analyses. Cox proportional hazard regressions estimated hazard ratios (HRs) for reproductive factors with incident all-cause dementia. Multiple adjusted models included age at study entry, SES, ethnicity, smoking status, systolic blood pressure, BMI, history of diabetes mellitus, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. Over a median of 11.8 years follow-up, 1,866 dementia cases were recorded in women and 2,202 in men. Multiple adjusted HRs ((95% confidence intervals (CIs)), p-value) for dementia were 1.20 (1.08, 1.34) (p = 0.016) for menarche <12 years and 1.19 (1.07, 1.34) (p = 0.024) for menarche >14 years compared to 13 years; 0.85 (0.74, 0.98) (p = 0.026) for ever been pregnant; 1.43 (1.26, 1.62) (p < 0.001) for age at first live birth <21 compared to 25 to 26 years; 0.82 (0.71, 0.94) (p = 0.006) for each abortion; 1.32 (1.15, 1.51) (p = 0.008) for natural menopause at <47 compared to 50 years; 1.12 (1.01, 1.25) (p = 0.039) for hysterectomy; 2.35 (1.06, 5.23) (p = 0.037) for hysterectomy with previous oophorectomy; and 0.80 (0.72, 0.88) (p < 0.001) for oral contraceptive pills use. The U-shaped associations between the number of children and the risk of dementia were similar for both sexes: Compared with those with 2 children, for those without children, the multiple adjusted HR ((95% CIs), p-value) was 1.18 (1.04, 1.33) (p = 0.027) for women and 1.10 (0.98, 1.23) (p = 0.164) for men, and the women-to-men ratio of HRs was 1.09 (0.92, 1.28) (p = 0.403); for those with 4 or more children, the HR was 1.14 (0.98, 1.33) (p = 0.132) for women and 1.26 (1.10, 1.45) (p = 0.003) for men, and the women-to-men ratio of HRs was 0.93 (0.76, 1.14) (p = 0.530). There was evidence that hysterectomy (HR, 1.31 (1.09, 1.59), p = 0.013) and oophorectomy (HR, 1.39 (1.08, 1.78), p = 0.002) were associated with a higher risk of dementia among women of relatively lower SES only. Limitations of the study include potential residual confounding and self-reported measures of reproductive factors, as well as the limited representativeness of the UK Biobank population.
CONCLUSIONS: In this study, we observed that some reproductive events related to shorter cumulative endogenous estrogen exposure in women were associated with higher dementia risk, and there was a similar association between the number of children and dementia risk between women and men.

Introduction

The dementia epidemic confronts the world as a major challenge, with extensive impact on individuals, carers, families, and societies at large [1,2]. Fifty million people live with dementia globally, and this number is projected to triple by 2050 [1]. There is no effective course-modifying treatment for dementia to date [2]; mitigation and modification of risk factors, therefore, present opportunities to reduce the burden associated with dementia at a population level [2].

The age-standardised global prevalence and death rates for dementia were estimated to be higher in women than men [3]. While the risk of developing dementia increases with age, the extent to which the female predominance is simply due to women’s longer life span remains far from conclusive, and female-specific reproductive factors may be able to explain these sex disparities [4,5].

Several endogenous estrogen changes occur throughout a woman’s reproductive life. Estradiol (E2) is the most predominant form of estrogen during reproductive life (from menarche to menopause) [6], and estriol (E3) is the primary estrogen during pregnancy [7]. Exogenous hormone use, such as oral contraceptives during reproductive years, and hormone replacement therapy (HRT) in later life can also influence estrogen level. Few studies evaluated the long-term effect of reproductive factors on dementia risk, such that conclusions remain uncertain, and the putative mechanisms are not well understood [4,8-13].

This study examined the reproductive factors and exogenous hormone use in relation to the risk of incident all-cause dementia in women in the UK Biobank. We assessed whether the associations between these factors and the risk of dementia vary by age, socioeconomic status (SES), smoking status, and body mass index (BMI). In addition, we included men to compare the association between number of children fathered and the risk of all-cause dementia, with the association in female counterparts.

Methods

Study design

No prospective analysis plan was specifically designed for the current study, although the framework for the design was drawn from the analyses carried out in a previous study, which examined the associations between reproductive factors and cardiovascular diseases in the UK Biobank [14], with prespecified subgroup analyses to assess effect modification.

The study design was further augmented by including broader selection and exposure categories, such as exogenous hormone use, and the timing of HRT in relation to menopause, as these are critical aspects to consider which may implicate the risk of dementia based on previous literature [4,8-13].

Post hoc sensitivity analyses included death as a competing risk in estimating the associations between reproductive factors and dementia risk, given that dementia requiring a follow-up measurement which death may preclude. Combinations of reproductive factors, rather than examined in isolation, were also included in the ancillary analyses. Last, the analyses were reweighted according to the social structure of the population in general [15].

Study population

The UK Biobank is a prospective population-based cohort, recruited over 500,000 (aged 40 to 69 years) women and men between 2006 and 2010 [16]. Individuals were invited to attend one of the 22 centres across the United Kingdom for baseline assessment, which included questionnaires soliciting information on lifestyle, medical history, and reproductive history. Physical measurements were collected, and a blood sample was taken. Written informed consent was obtained for all UK Biobank participants electronically. UK Biobank has obtained Research Tissue Bank approval from its governing Research Ethics Committee, as recommended by the National Research Ethics Service. This research has been conducted using the UK Biobank Resource (application No. 2495). Permission to use the UK Biobank Resource was approved by the access subcommittee of the UK Biobank Board.

Measurement of reproductive factors

Self-reported reproductive factors included in this study were age at menarche, pregnancy history, number of live births, age at first live birth, number of stillbirths, number of miscarriages, number of abortions, reproductive life span, (age at) natural menopause, (age at) hysterectomy, and (age at) oophorectomy. Exogenous hormone exposures included oral contraceptive pills use, age started using oral contraceptive pills, use of HRT, age at HRT initiation, and duration of HRT use. Early menarche was defined as age at the first menstrual period before 12 years of age. Early natural menopause was defined as the permanent absence of a menstrual period before 47 years of age. The reproductive life span was defined as the difference between the age at natural menopause and the age at menarche. Age at hysterectomy and oophorectomy was used to determine the timing of these procedures. Age at natural or artificial menopause, used to examine the risk associated in relation to the timing of HRT use, was defined as age at natural menopause, age at hysterectomy, or age at oophorectomy, whichever took place first. The number of children fathered was also recorded for men and was thus analysed here.

Study endpoint

The primary endpoint in this study was incident (i.e., fatal or nonfatal) all-cause dementia, as defined by the UK Biobank Outcome Adjudication Group, using the International Classification of Diseases-10th Revision (ICD-10) codes A81.0, F00, F01, F02, F03, F05, G30, G31.0, G31.1, G31.8, and I67.3 [17]. Hospital inpatient data from England, Scotland, and Wales, as well as the national death registers, were used to identify the date of the first known dementia after the date of baseline assessment. Follow-up for all participants started at the entry to the study, with data from the death registers and hospital inpatient data ended on November 30, 2020 or when fatal, nonfatal all-cause dementia, or death was recorded.

Covariates

Social deprivation was determined from the Townsend deprivation index [18]. Townsend deprivation index measures area deprivation, comprised of 4 domains including information about unemployment, car ownership, household overcrowding, and owner occupation, derived from the national census data, with higher scores indicate higher levels of social deprivation. The Townsend deprivation index was calculated for participant immediately prior to joining UK Biobank, based on the preceding national census output areas, in which the participant’s postcode is located. Smoking status was self-reported and categorised as never, former, or current smokers. Systolic blood pressure was taken at study baseline using the Omron HEM-7015IT digital blood pressure monitor as the mean of 2 sitting measures. BMI was calculated as the weight of the individual in kilogrammes, measured using the Tanita BC-418 MA body composition analyser, divided by the square of the individual’s standing height in metres. Diabetes status was self-reported: If the age at diagnosis was younger than 30, and the participant was using insulin, they were classified as type 1 diabetes, otherwise as type 2 diabetes. Total cholesterol was measured using the Beckman Coulter AU580. Self-reported medication use was also recorded.

Statistical analysis

The present analyses excluded participants with prevalent dementia at baseline (N = 263). Baseline characteristics are presented as mean with standard deviation (SD) for continuous variables and number with percentage for categorical variables.

Sex-specific crude incidence rates of dementia were estimated using Poisson regression models, with a log offset for person-years. We estimated the unadjusted and multiple adjusted rates for dementia per 10,000 person-years in all risk factor categories. Unadjusted and multiple adjusted models included age at study entry, SES, ethnicity, smoking status, systolic blood pressure, BMI, history of diabetes mellitus, total cholesterol, antihypertensive drugs, and lipid-lowering drugs.

The associations between each reproductive factors and dementia were assessed using Cox proportional hazard regression models that estimated the hazard ratios (HRs) with accompanying 95% confidence intervals (CIs) and p-values. When more than 2 groups were compared, the 95% CIs were estimated using floating absolute risks [19]. Covariate adjustments were the same as those made in the Poisson models. The association between the number of children fathered and dementia was assessed in men, fitted with the same set of covariates, to make a direct comparison with the number of live births in women, and the interaction term between the number of children and sex was used to obtain the women-to-men ratio of hazard ratios (RHRs) [20].

A series of models to explore combinations of reproductive factors were also constructed. These were (A) a combination of pregnancy related factors: number of live births, stillbirths, miscarriages, and abortions; and (B) a combination of factors throughout the life span including age at menarche, parous versus not, hysterectomy and or oophorectomy, HRT use, and contraceptive pill use.

Multiple adjusted restricted cubic splines (with kernel density plots) were constructed to assess the shape of continuous reproductive factors associated with dementia risk. The top and bottom 2.5% of the distributions, where precision is poor, were excluded, with the median value of the distribution taken as the reference.

Predefined subgroup analyses were conducted by age group (categorised as ≥65 versus <65 years, to yield an approximately equal number of events in each group), social deprivation (determined using the Townsend deprivation index at or below versus above the national median (−0.56)), smoking status (ever versus never smoker), and BMI (>25 kg/m2 versus ≤25 kg/m2), to examine the effect modifications by these characteristics. The interaction term was fitted between the exposure of interest and the prespecified subgroup to obtain the p-value. We examined social deprivation in finer detail by grouping the Townsend deprivation score into fifths based on nationally derived cutoffs, given the heterogeneity observed in subgroup analysis by the predefined 2-level SES. The lowest fifth contained the 20% least socially disadvantaged, and the highest fifth contained the 20% most socially disadvantaged, with the interaction term fitted between the exposure of interest and Townsend fifths, and p for heterogeneity were presented. In addition, the results were weighted according to the social structure of the population in general: Weighted HR were calculated according to equal weights of each stratified HR by fifths of Townsend score, since the national Townsend fifths, by definition, contain an equal number of people.

We excluded women who underwent hysterectomy or oophorectomy for sensitivity analysis and examined the risk of dementia associated with age at menarche, reproductive years, and age at natural menopause.

We also constructed Fine and Gray competing risk models [21] and multinomial regression models [22], to compare with results from Cox proportional hazards regression models, as sensitivity analysis. These methods will enable death to be accounted for as a competing risk for dementia, given that death may preclude dementia from occurring. The competing risk models, which estimated subdistribution HRs, were conducted for all-cause dementia accounting for all-cause mortality as a competing risk, incorporating time to event data. For multinomial regression models, the odds ratios were produced, with outcomes specified as (0) had neither dementia nor died by the last follow-up; (1) all-cause dementia by the last follow-up; and (2) all-cause mortality preceding all-cause dementia by the last follow-up. Time to event is not specified in multinomial models.

All main analyses were performed on complete case data using R version 4.1.0 (RStudio Team (2021). RStudio: Integrated Development for R. RStudio, PBC, Boston) and Stata 17.0 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC).

Further, missing data were imputed using Multivariate Imputation by Chained Equations (MICE), with 5 iterations. The multiple adjusted results from Cox regression models with imputed data were subsequently compared with the complete case analyses.

This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist).

Results

Over a median of 11.8 years follow-up, 1,866 cases of incident dementia were recorded among 273,240 women. At the study baseline, the mean age of women was 56 years; the mean age at natural menarche and age at first live birth was 13 and 26 years, respectively; 85% reported they have been pregnant at least once, and 44% reported having 2 children. For natural menopause, 61% of the women were postmenopausal, and the mean age at natural menopause was 50 years. For surgical-induced menopause, the percentage of women who reported a history of hysterectomy and oophorectomy were 19% and 8%, respectively; 81% reported ever used oral contraceptive pills, and 38% reported ever used HRT, with a mean age of 47 years for HRT initiation and a mean duration of 6.3 years.

Among 228,957 men, 2,202 incident cases of dementia were recorded. The mean age for men at baseline was 57 years, with 41% reported fathering 2 children (Table 1).

Table 1

Baseline characteristics of study participants in the UK Biobank.

	Women (n = 273,240)	Men (n = 228,957)
Dementia, n %	1,866 (0.7)	2,202 (1.0)
Age, years	56.3 (8.0)	56.7 (8.0)
Social deprivation, %
Higher (≤−0.56 on Townsend deprivation index)	183,388 (67.1)	151,860 (66.3)
Lower (>−0.56 on Townsend deprivation index)	66,740 (24.4)	58,259 (25.4)
Townsend fifths, %
First (<−2.938 on Townsend deprivation index (least disadvantaged))	100,950 (36.9)	84,842 (37.1)
Second (≥−2.938, <1.531 on Townsend deprivation index)	56,567 (20.7)	46,147 (20.2)
Third (≥−1.531, <0.170 on Townsend deprivation index)	41,229 (15.1)	33,314 (14.6)
Fourth (≥0.170, 2.448 on Townsend deprivation index)	36,888 (13.5)	30,431 (13.3)
Fifth (≥2.448 on Townsend deprivation index (most disadvantaged))	37,279 (13.6)	33,927 (14.8)
Ethnicity, %
White	257,304 (94.2)	215,104 (93.9)
Other	14,672 (5.8)	12,344 (5.4)
Smoking status, %
Never smoker	161,965 (59.3)	111,401 (48.7)
Former smoker	85,407 (31.3)	87,534 (38.2)
Current smoker	24,356 (8.9)	28,590 (12.5)
Blood pressure
Systolic blood pressure (mean (SD))	135.3 (19.2)	140.9 (17.5)
Diastolic blood pressure (mean (SD))	80.7 (10.0)	84.1 (10.0)
BMI (mean (SD))	27.1 (5.2)	27.8 (4.2)
Diabetes, %
Type 1 diabetesa	564 (0.2)	652 (0.3)
Type 2 diabetes	9,945 (3.6)	15,514 (6.8)
Total cholesterol (mean (SD))	5.87 (1.1)	5.48 (1.1)
Antihypertensive drugs	38,405 (14.1)	47,965 (20.9)
Lipid-lowering drugs	29,502 (10.8)	45,730 (20.0)
Age at menarche, years	13.0 (1.6)	-
Ever pregnant, %	231,352 (84.7)	-
Number of children, %
None	51,079 (18.7)	47,098 (20.6)
1	36,457 (13.3)	28,635 (12.5)
2	119,113 (43.6)	94,263 (41.2)
3	48,270 (17.7)	38,128 (16.7)
4 or more	17,493 (6.4)	16,562 (7.2)
Age at first live birth, years	25.9 (5.1)	-
Number of miscarriages, %
None	171,511 (62.8)	-
1	40,042 (14.7)	-
2 or more	15,876 (5.8)	-
Number of stillbirths, %
None	220,585 (80.7)	-
1	6,072 (2.2)	-
2 or more	963 (0.4)	-
Number of abortions, %
None	189,105 (69.2)	-
1	30,536 (11.2)	-
2 or more	7,349 (2.7)	-
Number of reproductive years	37.3 (4.8)	-
Menopause
Natural menopause
Postmenopausal, %	165,856 (60.7)	-
Age at menopause, years	50.3 (4.5)	-
Surgical menopause
History of hysterectomy, %	51,226 (18.7)	-
Age at hysterectomy, years	43.9 (8.0)	-
History of oophorectomy, %	21,935 (8.0)	-
Age at oophorectomy, years	47.4 (7.8)	-
History of both hysterectomy and oophorectomy, %	20,901 (7.6)
Exogenous hormone use
Oral contraceptive pills use
Ever used oral contraceptive pills, %	220,344 (80.6)	-
Age first taken oral contraceptive pills, years	21.5 (4.7)	-
HRT use
Ever used HRT, %	104,133 (38.1)	-
Age initiated HRT, years	47.4 (5.4)	-
HRT duration, years	6.3 (5.3)	-

aDefined as diagnosis before the age of 30 and receiving insulin treatment.

SD, standard deviation; HRT, hormone replacement therapy.

Missing data for each reproductive variable of interest were reported in the Supporting information (S1 Table).

Dementia rates

The crude incidence rate for dementia was 5.88 (95% CI (5.62, 6.16)) for women and 8.42 (8.07, 8.78) for men per 10,000 person-years.

The multiple adjusted rates of dementia per 10,000 person-years (95% CI) were the highest among those with shorter reproductive span (<33 years: 8.15 (6.87, 9.42)) and earlier age at natural menopause (<47 years: 8.85 (7.66, 10.04)) (Table 2).

Table 2

Unadjusted and multiple adjusted rates of incident dementia for reproductive risk factors in women.

Reproductive factor	Unadjusted rates/10,000 person-years (95% CI)	Multiple adjusted rates/10,000 person-years (95% CI)a
Age at menarche
<12	6.27 (5.64, 6.90)	6.31 (5.65, 7.00)
12	5.69 (5.07, 6.30)	5.64 (4.98, 6.29)
13	4.96 (4.46, 5.47)	5.26 (4.70, 5.83)
14	5.51 (4.92, 6.10)	5.13 (4.54, 5.72)
>14	6.58 (5.89, 7.28)	6.31 (5.60, 7.01)
Ever been pregnant
No	5.12 (4.48, 5.77)	6.69 (5.80, 7.58)
Yes	6.00 (5.71, 6.30)	5.72 (5.42, 6.01)
Number of live births
0	4.80 (4.24, 5.36)	6.46 (5.67, 7.26)
1	5.16 (4.47, 5.84)	6.01 (5.18, 6.84)
2	5.58 (5.19, 5.97)	5.52 (5.11, 5.93)
3	6.99 (6.30, 7.68)	5.72 (5.11, 6.33)
4 or more	9.41 (8.07, 10.75)	6.35 (5.37, 7.33)
Parous
No	4.80 (4.24, 5.36)	6.46 (5.67, 7.26)
Yes	6.12 (5.81, 6.42)	5.73 (5.43, 6.03)
Age at first live birth
<21	8.31 (7.37, 9.24)	7.15 (6.28, 8.03)
21 to 22	8.32 (7.32, 9.31)	6.17 (5.39, 6.95)
23 to 24	7.87 (6.98, 8.76)	6.32 (5.56, 7.09)
25 to 26	5.33 (4.62, 6.04)	5.01 (4.29, 5.72)
27 to 29	4.87 (4.26, 5.47)	5.79 (5.04, 6.55)
>29	3.75 (3.25, 4.25)	5.54 (4.74, 6.34)
Number of miscarriages
0	6.12 (5.78, 6.47)	5.99 (5.64, 6.35)
1	5.35 (4.68, 6.01)	5.38 (4.66, 6.10)
2 or more	5.77 (4.67, 6.87)	6.12 (4.88, 7.36)
Number of stillbirths
0	5.86 (5.57, 6.16)	5.88 (5.56, 6.19)
1	9.31 (7.05, 11.58)	6.72 (4.95, 8.48)
2 or more	11.01 (4.78, 17.24)	7.41 (2.56, 12.25)
Number of abortions
0	6.41 (6.08, 6.74)	6.02 (5.69, 6.36)
1	4.35 (3.67, 5.04)	5.79 (4.82, 6.77)
2 or more	1.54 (0.70, 2.37)	2.04 (0.77, 3.31)
Reproductive years
<33	8.60 (7.36, 9.83)	8.15 (6.87, 9.42)
33 to 35	7.56 (6.45, 8.67)	8.01 (6.77, 9.26)
36 to 37	6.10 (5.17, 7.03)	6.38 (5.36, 7.41)
38 to 39	5.20 (4.40, 6.01)	5.47 (4.58, 6.35)
40 to 42	5.82 (4.99, 6.64)	5.52 (4.68, 6.36)
>42	7.72 (6.43, 9.01)	6.43 (5.31, 7.56)
Age at natural menopause
<47	8.98 (7.85, 10.11)	8.85 (7.66, 10.04)
47 to 49	6.38 (5.39, 7.36)	7.20 (6.01, 8.38)
50	6.99 (5.97, 8.02)	6.73 (5.69, 7.76)
51 to 52	5.34 (4.54, 6.13)	5.38 (4.51, 6.25)
53 to 54	4.90 (3.97, 5.83)	5.12 (4.09, 6.14)
>54	7.66 (6.58, 8.74)	6.21 (5.29, 7.13)
Hysterectomy
No	5.64 (5.31, 5.96)	5.64 (5.31, 5.96)
Yes	6.32 (5.75, 6.89)	6.32 (5.75, 6.89)
Oophorectomy
No	5.52 (5.24, 5.79)	5.69 (5.39, 5.98)
Yes	8.61 (7.47, 9.76)	6.04 (5.18, 6.90)
Ever taken oral contraceptive pills
No	11.23 (10.38, 12.09)	6.79 (6.23, 7.36)
Yes	4.60 (4.34, 4.87)	5.37 (5.04, 5.70)
Ever used HRT
No	4.33 (4.04, 4.62)	5.83 (5.40, 6.25)
Yes	8.28 (7.77, 8.79)	5.77 (5.39, 6.15)

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs.

CI, confidence interval; HRT, hormone replacement therapy.

Among men, the multiple adjusted rates of all-cause dementia per 10,000 person-years (95% CI) were 8.65 (7.68, 9.62), 7.69 (6.63, 8.75), 7.95 (7.39, 8.51), 8.56 (7.70, 9.43), and 10.04 (8.67, 11.42) and for those who fathered none, 1, 2, 3, and 4 or more children, respectively. These rates were higher among men when compared with number of live births reported in women across all the categories (Table 2).

Age at menarche

Overall, the relationship between age at menarche and dementia appeared to be U shaped (Table 3, Fig 1A): The multiple adjusted HRs (95% CI) of the age at menarche <12 associated with dementia was 1.20 (1.08, 1.34) (p = 0.016) and at the age of >14 was 1.19 (1.07, 1.34) (p = 0.024), compared to women who had their menarche at 13.

Table 3

Unadjusted and multiple adjusted HRs for the risk of dementia associated with reproductive factors in women.

Reproductive factor	No. of events	Unadjusted HR (95% CI)	p-Value	Multiple adjusted HR (95% CI) a	p-Value
Age at menarche
<12	385	1.26 (1.14, 1.40)	<0.001	1.20 (1.08, 1.34)	0.016
12	331	1.15 (1.03, 1.28)	0.072	1.07 (0.95, 1.20)	0.389
13 (ref)	372	1.00 (0.90, 1.11)	-	1.00 (0.90, 1.11)	-
14	334	1.11 (1.00, 1.24)	0.167	0.97 (0.87, 1.09)	0.718
>14	341	1.33 (1.19, 1.48)	<0.001	1.19 (1.07, 1.34)	0.024
Ever been pregnant	1,613	1.17 (1.02, 1.34)	0.022	0.85 (0.74, 0.98)	0.026
Number of live births
0	284	0.86 (0.75, 0.98)	0.338	1.18 (1.04, 1.33)	0.027
1	218	0.93 (0.79, 1.06)	0.312	1.09 (0.95, 1.25)	0.276
2 (ref)	773	1.00 (0.93, 1.07)	-	1.00 (0.93, 1.08)	-
3	392	1.25 (1.15, 1.35)	<0.001	1.03 (0.93, 1.15)	0.641
4 or more	190	1.69 (1.55, 1.83)	<0.001	1.14 (0.98, 1.33)	0.132
Parous versus not	1,573	1.27 (1.12, 1.44)	<0.001	0.88 (0.77, 1.01)	0.061
Age at first live birth
<21	305	1.56 (1.40, 1.75)	<0.001	1.43 (1.26, 1.62)	<0.001
21 to 22	270	1.56 (1.39, 1.76)	<0.001	1.23 (1.08, 1.40)	0.034
23 to 24	302	1.48 (1.32, 1.65)	<0.001	1.26 (1.12, 1.42)	0.015
25 to 26 (ref)	217	1.00 (0.88, 1.14)	-	1.00 (0.87, 1.15)	-
27 to 29	251	0.91 (0.81, 1.03)	0.324	1.16 (1.02, 1.32)	0.129
>29	213	0.70 (0.62, 0.81)	<0.001	1.11 (0.96, 1.29)	0.296
Per additional year of age at first live birth	-	0.94 (0.93, 0.95)	<0.001	0.98 (0.97, 1.00)	0.006
Number of miscarriages
0 (ref)	1,220	1.00 (0.95, 1.06)	-	1.00 (0.94, 1.06)	-
1	249	0.87 (0.77, 0.99)	0.050	0.90 (0.78, 1.03)	0.147
2 or more	106	0.94 (0.78, 1.14)	0.571	1.02 (0.83, 1.25)	0.847
Miscarriage versus not	355	0.89 (0.79, 1.00)	0.060	0.93 (0.82, 1.06)	0.274
Per miscarriage	-	1.03 (0.97, 1.10)	0.308	1.01 (0.94, 1.08)	0.816
Number of stillbirths
0 (ref)	1,503	1.00 (0.95, 1.05)	-	1.00 (0.94, 1.06)	-
1	65	1.59 (1.25, 2.03)	<0.001	1.15 (0.88, 1.49)	0.312
2 or more	12	1.90 (1.08, 3.35)	0.027	1.27 (0.66, 2.45)	0.472
Stillbirth versus not	77	1.63 (1.30, 2.06)	<0.001	1.16 (0.91, 1.49)	0.234
Per stillbirth	-	1.28 (1.12, 1.45)	<0.001	1.10 (0.92, 1.31)	0.319
Number of abortions
0 (ref)	1,409	1.00 (0.95, 1.05)	-	1.00 (0.94, 1.07)	-
1	154	0.68 (0.58, 0.80)	<0.001	0.97 (0.82, 1.14)	0.735
2 or more	13	0.24 (0.14, 0.42)	<0.001	0.34 (0.18, 0.64)	<0.001
Abortion versus not	167	0.60 (0.51, 0.70)	<0.001	0.87 (0.73, 1.03)	0.101
Per abortion	-	0.62 (0.54, 0.71)	<0.001	0.82 (0.71, 0.94)	0.006
Reproductive years
<33 (ref)	186	1.00 (0.87, 1.15)	-	1.00 (0.86, 1.17)	-
33 to 35	178	0.88 (0.76, 1.02)	0.214	0.98 (0.84, 1.15)	0.878
36 to 37	166	0.71 (0.61, 0.83)	0.001	0.78 (0.67, 0.92)	0.031
38 to 39	160	0.60 (0.52, 0.71)	<0.001	0.67 (0.57, 0.79)	<0.001
40 to 42	192	0.68 (0.59, 0.78)	<0.001	0.68 (0.58, 0.79)	<0.001
>42	137	0.90 (0.76, 1.07)	0.362	0.80 (0.67, 0.95)	0.056
Age at natural menopause
<47	242	1.28 (1.13, 1.46)	0.011	1.32 (1.15, 1.51)	0.008
47 to 49	161	0.91 (0.78, 1.06)	0.382	1.07 (0.91, 1.26)	0.573
50 (ref)	180	1.00 (0.86, 1.16)	-	1.00 (0.86, 1.17)	-
51 to 52	173	0.76 (0.66, 0.88)	0.010	0.80 (0.68, 0.94)	0.048
53 to 54	107	0.70 (0.58, 0.84)	0.003	0.76 (0.62, 0.93)	0.035
>54	193	1.10 (0.96, 1.27)	0.358	0.93 (0.80, 1.08)	0.496
Hysterectomy versus not	537	1.78 (1.61, 1.96)	<0.001	1.12 (1.01, 1.25)	0.039
Oophorectomy versus not	217	1.57 (1.36, 1.81)	<0.001	1.07 (0.92, 1.24)	0.413
Ever taken oral contraceptive pills	1,180	0.41 (0.37, 0.45)	<0.001	0.80 (0.72, 0.88)	<0.001
Age started oral contraceptive pills (per year)	-	1.11 (1.10, 1.12)	<0.001	1.01 (1.00, 1.03)	0.143
Ever used HRT	997	1.91 (1.75, 2.10)	<0.001	0.99 (0.90, 1.09)	0.828
Age started HRT (per year)	-	0.99 (0.98, 1.01)	0.306	0.96 (0.95, 0.98)	<0.001
Duration of HRT use (per year)	-	1.03 (1.01, 1.04)	<0.001	1.00 (0.98, 1.01)	0.526

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs.

CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy.

Fig 1

Multiple adjusted restricted cubic splines (with kernel density plots) showing HRs for the risk of dementia associated with reproductive factors.

The blue line represents the hazard function, and the blue shaded area represents the 95% CIs. The shaded grey region represents the kernel density plot for the distribution of data. After excluding the values from the top and bottom 2.5% of the distribution, with the median value being the reference. Splines adjusted for age, Townsend deprivation index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. (A) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age at menarche. (B) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age at first birth. (C) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with reproductive years. (D) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age at natural menopause. (E) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age at hysterectomy. (F) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age at oophorectomy. (G) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age started contraceptive pills. (H) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with age started HRT. (I) Restricted cubic spline plot with multiple adjusted HRs (95% CI) for all-cause dementia associated with years using HRT. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy.

Parity-related factors

The HR for dementia who had ever been pregnant was 0.85 (0.74, 0.98) (p = 0.026) compared with never pregnant. Younger age at first live birth was associated with a higher dementia risk, with the HR for <21 years at first live birth was 1.43 (1.26, 1.62) (p < 0.001)) compared with first live birth at 25 to 26 years (Table 3, Fig 1B). Compared with women who never had an abortion, the HR for dementia in women who had 2 or more abortions was 0.34 (0.18, 0.64) (p < 0.001). Stillbirth and miscarriage were not associated with dementia risk.

Number of children

Compared with those who had 2 children, the associations between the number of children and dementia were similar for women and men and appeared to be U shaped (Fig 2, S2 Table): for instance, for those who had no children, the adjusted HR was 1.18 (1.04, 1.33) (p = 0.027) for women and 1.10 (0.98, 1.23) (p = 0.164) for men, with the women-to-men ratio of HRs being 1.09 (0.92, 1.28) (p = 0.403); for those who had 4 or more children, the adjusted HR was 1.14 (0.98, 1.33) (p = 0.132) for women and 1.26 (1.10, 1.45) (p = 0.003) for men, with the women-to-men ratio of HRs being 0.93 (0.76, 1.14) (p = 0.530). Unadjusted results by predefined subgroups were also presented in the Supporting information (S3 Table).

Fig 2

Multiple adjusted HRs for the risk of dementia associated with number of children for women and men.

The HRs are plotted on a floating absolute scale. The squares represent the HRs, and the bars represent the 95% CIs. Analyses were adjusted for age, Townsend deprivation index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HR, hazard ratio.

Menopause-related factors

A longer reproductive life span and an older age at natural menopause had inverse log-linear associations with dementia risk (Table 3, Fig 1C and 1D). The HR of dementia associated with menopause before the age of 47 was 1.32 (1.15, 1.51) (p = 0.008) compared to women who had their menopause at the age of 50. For women who reported a history of hysterectomy, the HR for dementia was 1.12 (1.01, 1.25) (p = 0.039) compared with women who never had a hysterectomy. Oophorectomy was not significantly associated with dementia; the HR was 1.07 (0.92, 1.24) (p = 0.413). Younger age at hysterectomy was associated with increased dementia risk (Fig 1E), whereas the relationship for age at oophorectomy and dementia risk appear to be U shaped (Fig 1F).

Regarding the timing of hysterectomy and oophorectomy, in comparison to women who never underwent hysterectomy or oophorectomy, women who had hysterectomy after oophorectomy had an increased risk (HR, 2.35 (1.06, 5.23), p = 0.037), while the other variations for the timing of the procedures did not appear to be associated with dementia risk (S4 Table).

For hysterectomy and oophorectomy, there was some evidence of heterogeneity by SES (p = 0.013, p = 0.002, respectively), such that women of relatively lower SES had increased dementia risk relative to women of higher SES (Table 4).

Table 4

Multiple adjusted HRs for the risk of dementia associated with reproductive factors, by age, SES, smoking, and BMI.

Risk factors	Age			SES
	<65 years	≥65 years	p-Value	Higher SES	Lower SES	p-Value
Early menarchea versus not	1.14 (0.96, 1.36)	1.15 (0.98, 1.36)	0.905	1.17 (1.00, 1.36)	1.08 (0.87, 1.35)	0.820
Age at first live birth per year	0.98 (0.96, 1.00)	0.99 (0.97, 1.00)	0.206	0.99 (0.98, 1.01)	0.97 (0.95, 0.99)	0.081
Each childb
Women	0.97 (0.92, 1.04)	0.99 (0.94, 1.05)	0.684	0.93 (0.87, 0.98)	1.06 (1.00, 1.13)	0.002
Men	1.05 (1.01, 1.09)	1.01 (0.96, 1.06)	0.205	1.01 (0.96, 1.06)	1.05 (1.01, 1.08)	0.299
Stillbirth versus not	1.02 (0.66, 1.58)	1.25 (0.92, 1.69)	0.531	1.16 (0.82, 1.64)	1.16 (0.78, 1.73)	0.896
Miscarriage versus not	0.94 (0.78, 1.14)	0.92 (0.78, 1.09)	0.880	0.90 (0.76, 1.06)	1.05 (0.84, 1.32)	0.219
Abortion versus not	0.76 (0.59, 0.98)	0.98 (0.77, 1.24)	0.179	0.84 (0.66, 1.07)	0.92 (0.70, 1.21)	0.400
Early menopausec versus not	1.61 (1.28, 2.02)	1.34 (1.09, 1.64)	0.146	1.27 (1.04, 1.56)	1.75 (1.34, 2.29)	0.053
Hysterectomy versus not	1.16 (0.98, 1.38)	1.09 (0.95, 1.26)	0.450	0.96 (0.84, 1.11)	1.31 (1.09, 1.59)	0.013
Oophorectomy versus not	1.18 (0.94, 1.48)	0.99 (0.81, 1.21)	0.220	0.82 (0.66, 1.02)	1.39 (1.08, 1.78)	0.002
Oral contraceptive pill use versus not	0.64 (0.54, 0.76)	0.92 (0.81, 1.05)	<0.001	0.87 (0.76, 1.00)	0.69 (0.57, 0.83)	0.176
HRT use versus not	1.12 (0.96, 1.30)	0.92 (0.81, 1.05)	0.069	0.95 (0.84, 1.08)	1.13 (0.94, 1.35)	0.374
Risk factors	Smoking status			BMI
	Never	Ever	p-Value	≤25 kg/m 2	>25 kg/m 2	p-Value
Early menarchea versus not	1.07 (0.91, 1.27)	1.23 (1.03, 1.46)	0.347	1.25 (1.01, 1.55)	1.10 (0.95, 1.27)	0.455
Age at first live birth per year	0.98 (0.96, 0.99)	0.99 (0.97, 1.01)	0.318	0.98 (0.96, 1.00)	0.98 (0.97, 1.00)	0.876
Each childb
Women	0.95 (0.90, 1.00)	1.02 (0.97, 1.09)	0.098	1.00 (0.93, 1.07)	0.98 (0.94, 1.03)	0.794
Men	1.01 (0.95, 1.06)	1.04 (1.01, 1.08)	0.340	1.06 (1.00, 1.12)	1.03 (0.99, 1.07)	0.565
Stillbirth versus not	1.03 (0.71, 1.48)	1.31 (0.93, 1.84)	0.340	1.27 (0.82, 1.97)	1.11 (0.82, 1.50)	0.553
Miscarriage versus not	0.82 (0.68, 0.98)	1.07 (0.89, 1.28)	0.044	0.76 (0.61, 0.96)	1.02 (0.88, 1.19)	0.036
Abortion versus not	0.88 (0.68, 1.14)	0.86 (0.68, 1.09)	0.820	0.93 (0.70, 1.22)	0.83 (0.66, 1.03)	0.521
Early menopausec versus not	1.40 (1.12, 1.74)	1.49 (1.20, 1.84)	0.711	1.45 (1.13, 1.87)	1.43 (1.18, 1.74)	0.735
Hysterectomy versus not	1.21 (1.05, 1.39)	1.02 (0.87, 1.20)	0.113	1.03 (0.85, 1.25)	1.17 (1.03, 1.34)	0.276
Oophorectomy versus not	1.14 (0.93, 1.39)	0.98 (0.78, 1.24)	0.313	1.02 (0.78, 1.34)	1.08 (0.90, 1.30)	0.710
Oral contraceptive pill use versus not	0.79 (0.69, 0.91)	0.80 (0.68, 0.95)	0.942	0.80 (0.67, 0.95)	0.80 (0.70, 0.91)	0.558
HRT use versus not	0.92 (0.81, 1.05)	1.09 (0.94, 1.26)	0.096	0.85 (0.72, 1.01)	1.09 (0.96, 1.23)	0.043

aEarly menarche was defined as age at first menstrual period before the age of 12 years.

bEach live birth in women and each child fathered in men.

cEarly menopause was defined as the permanent absence of menstrual periods before the age of 47 years.

Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. p-Values are the interaction between subgroups.

BMI, body mass index; HR, hazard ratio; HRT, hormone replacement therapy; SES, socioeconomic status.

Exogenous hormone use

The HR for dementia in women who reported oral contraceptive pill use was 0.80 (0.72, 0.88) (p < 0.001) (Table 3). There was no clear evidence of an association between age started using oral contraceptive pills (Fig 1G). There was some evidence of heterogeneity by age, such that the lower risk was only statistically significant in women younger than 65 years at study baseline (p < 0.001) (Table 4). There was no evidence of HRT use (0.99 (0.90, 1.09, p = 0.828) associated with dementia compared with those never used HRT, although there was evidence of a lower risk in older age at HRT initiation (0.96 (0.95, 0.98) per year, p < 0.001) (Table 3, Fig 1H), but no evidence of HRT duration associated with dementia risk (1.00 (0.98, 1.01) per year, p = 0.526) (Table 3, Fig 1I).

Further, there was no evidence of dementia risk varied by the timing of HRT initiation in relation to menopause, although those who had menopause with unknown HRT initiation were at an increased risk of dementia compared to those who had menopause without HRT use (HR, 1.49 (1.31, 1.69), p < 0.001) (S5 Table).

Sensitivity analysis

After excluding women who underwent hysterectomy or oophorectomy, the associations for age at menarche, reproductive years, and age at natural menopause in relation to dementia risk were similar to the main results (S6 Table).

Results from the competing risk models (subdistribution HRs) and multinomial regression models (odds ratios), which considered the competing risk of all-cause mortality, showed broadly similar results as the results from the Cox proportional hazards regression models (HRs) (S7 Table).

There was some evidence of heterogeneity across Townsend score fifths (S8 Table), including number of children, abortion, oophorectomy, and oral contraceptive pill use in women. After weighting the estimates by Townsend fifths according to the national distribution, the pooled estimates were similar to those in the main analysis but with attenuated the statistical significance of associations as indicated by wider CIs, caused by giving relatively higher weights to the fifths with the least precise estimates (S9 Table). After the results were weighted by Townsend fifths, there was extra imprecision in the weighted HR estimates for some exposures, which occurred where there is considerable difference in estimates (HR and standard errors) across the Townsend fifths; however, this may be a result of small numbers in some exposure categories.

When reproductive factors in the combined model compared with models included these factors individually, there were no major differences, suggesting that there is no substantial effect of confounding (S10 Table).

After imputing missing data, the multiple adjusted results were broadly similar to the results from the complete case analyses (S11 Table).

Discussion

In this large population-based cohort study, we found several associations between reproductive factors and exogenous use of hormone with dementia risk later in life. Younger age at first live birth, hysterectomy, specifically hysterectomy without concomitant oophorectomy or with a previous oophorectomy, were associated with greater dementia risk. Age at menarche, hysterectomy, and oophorectomy appear to be U shaped for the association with dementia risk. Ever been pregnant, ever had an abortion, longer reproductive span, and older age at natural menopause were associated with lower dementia risk. For exogenous hormone exposures, the use of oral contraceptive pills was associated with a lower risk of dementia. There was some evidence of heterogeneity by SES for hysterectomy, and oophorectomy, such that the elevated dementia risk associated with these risk factors was confined to women of lower SES. U-shaped associations were found for the number of children and dementia risk, similar for both sexes.

Surrogates for endogenous hormone exposures

Reproductive events indicating shorter cumulative exposure to estradiol, including later menarche, early natural menopause, shorter reproductive span, and hysterectomy, were all associated with an elevated risk of dementia in our study. However, previous studies on the relationship between these risk factors have reported mixed results. Consistent with our findings, the Kaiser Permanente (KP) study showed that reproductive events contributing to shorter estradiol exposure were associated with elevated dementia risk [10]. A nationwide study from South Korea also reported these comparable findings [13]. These findings may be driven by the effects of estradiol on brain health; in experimental studies, estradiol has been shown to be correlated with neuronal dendritic spine density [23] as well as reducing apoptosis and inflammation [24]. In contrast, the Gothenburg H70 [9] and Rotterdam study [11] reported that a longer reproductive span and later menopause were associated with greater dementia risk, while the 10/66 study reported no association for reproductive span and dementia [12]. The discrepancy in findings between our study and the Gothenburg H70 and Rotterdam study [9,11] was not due to the exclusion of women who reported hysterectomy or oophorectomy in the Gothenburg H70 and Rotterdam study, as indicated in our sensitivity analyses. However, the studies that reported null or opposite results included older women at study baseline (mean age around 70 years) [11,12] or the women were followed into their late life [9]. Notably, the Gothenburg H70 [9] only found significant associations with dementia for longer reproductive span and older age at menopause, among those with the older onset of dementia (75 years and above) [9]. As such, we hypothesise that the risk exposure in midlife and older life may be different.

Surgically induced menopause (hysterectomy and oophorectomy), when performed before the onset of natural menopause, can cease the secretion of endogenous sex hormones prematurely [25]. The KP study found the dementia risk is greater among women who underwent a hysterectomy [10], which was consistent with our findings. Similarly, a pooled analysis of 2 cohorts found that the risk of cognitive impairment and dementia was higher in women who underwent a hysterectomy, and the risk was even greater in those who had a hysterectomy and bilateral oophorectomy [26]. Moreover, consistent with findings on the timing of surgical menopause procedures in association with cardiovascular disease [14], our study also showed some evidence that there was a greater risk of dementia in those with hysterectomy with a previous oophorectomy. A meta-analysis did not find an overall association between surgical menopause and dementia [27], but surgical menopause before the age of 45 was associated with greater dementia risk [27]. We similarly demonstrated that younger age at hysterectomy and oophorectomy were inversely associated with dementia risk, providing further support that early cessation of hormones may be associated with poorer cognitive outcomes.

When disaggregated by SES, early natural menopause, hysterectomy, and oophorectomy were only associated with a greater risk of dementia in women of relatively low SES. Previous studies reported SES might adversely influence the age at entry to perimenopause [28,29]. Further, social disadvantage can modulate the level of cortisol [30]. During the menopausal transition, increased cortisol level has been associated with vasomotor symptoms and depressed mood [31], which are key determinants for cognitive function [32,33].

Pregnancy induces marked changes in endogenous estrogen levels [7,34], and estrogen can be both neuroprotective or neurotoxic, depending on the concentration [34,35]. A pooled study from the Cohort Studies of Memory in an International Consortium (COSMIC) found that the risk of Alzheimer disease doubled for women who had 4 or more completed pregnancies [34]. Another COSMIC analysis showed that having 5 or more children was associated with increased dementia risk, while nulliparity and having 2 to 4 children showed similar associations compared to primiparous women [36]. In our study, the number of children was similarly associated with dementia risk for women and men. As such, the risk variation in women appears to be more related to social and behavioural factors involved in parenthood rather than biological factors associated with childbearing. A plausible explanation for this could be related to the additional expenditures and responsibilities associated with the number of dependents, which could lead to economic hardships and increase psychological distress in parents [37]. In particular, mothers are more likely to bear the brunt of childcare than fathers in a low-income household; the impact of parenthood on mothers of lower SES may be more adverse [37].

Our study showed that abortion was associated with a lower risk of dementia, while we did not find any link for stillbirth or miscarriage. The COSMIC study by Jang and colleagues also found the risk of Alzheimer disease in women who had incomplete pregnancies was half that of those who never experienced an incomplete pregnancy [34]; however, incomplete pregnancies in this study encompassed surgical- or medical-induced abortion and spontaneous miscarriage. A Danish register–based cohort study that excluded women who had a surgical and medical abortion found stillbirth was associated with 86% greater risk of dementia, while miscarriage was not associated with dementia [38]. Both studies [34,38] had limitations that precluded the effect of spontaneous miscarriage and abortion from being differentiated. The course of pregnancy and childbirth can have a considerable influence on lifestyle and health [34], although we did not find any effect modifications to explain some of the findings in pregnancy-related factors. Further clarification for the mechanism which underpins these observations is needed.

The link between premenopausal hormone use and dementia risk has hardly been characterised [39]. A previous study suggested that women who reported hormonal contraceptive use performed better in the visuospatial ability and speed and flexibility domains of the neuropsychiatric tests than those who had never used hormonal contraception [39]. On the other hand, the potential benefits of HRT to prolong estrogen supply in older women have not been corroborated by interventional studies [40,41], while the observational evidence remains conflicting [4,42-45]. The Women’s Health Initiative Study (WHIS), the only clinical trial that evaluated postmenopausal hormone therapy on preventing dementia, concluded that the risk of dementia was doubled in women randomised to estrogen-progestin based HRT [40]. In a case–control study in Finland, long-term use of systemic HRT was associated with an increased risk of Alzheimer disease [45]. It is still largely contentious whether HRT can potentially prevent or increase the risk of dementia. The timing of HRT use may be crucial, such that there may be a critical window which exogenous hormone use can confer cognitive benefits in postmenopausal women [35,46]. Our study findings on HRT do not support associations between HRT and dementia risk, nor the aforementioned “timing hypothesis” for HRT initiation in relation to menopause for dementia risk. Consistently, a recent nested case–control study using a UK general practice cohort also found no evidence of an increased risk of dementia associated with menopausal hormone therapy; no evidence of different time of hormone therapy initiation may pose different risk of dementia [47].

Strengths and limitations

The strengths of our study were the large sample size, with linkage to national health records and death registers. Further, our study included a comprehensive list of reproductive factors and exogenous hormone use through the life course. The limitations included retrospective and self-reported measures of reproductive factors, which may be inherently subject to recall bias and misclassification. Socioeconomic position has been associated with several reproductive factors, such as age at first birth and parity [48]. After adjusting for Townsend deprivation index and weighting according to social structure in our study, the associations between reproductive factors and the risk of dementia attenuated. The limited representativeness of the UK Biobank population, which is a cohort predominantly of relatively healthy and affluent people of white ethnic background, means that it is unlikely to produce reliable estimates of either the prevalence of female reproductive factors or the risk of dementia in the UK population at large. Our finding of an interaction between social deprivation and certain reproductive factors suggests that the main results for these factors need to be treated with caution, although any error in estimated effect size appears to be minor. Further, the adjudication of incident dementia in the UK Biobank used the date recorded in the death registry and hospital admission as the incidence date. Age and SES, among other factors, might influence whether dementia is being recorded. It is also possible that some dementia cases were not picked up from hospital admission and death registry data, resulting in underreporting of dementia cases. Dementia subtypes were not differentiated due to the currently small number of events in the UK Biobank. Last, although multiple adjustments were made to account for confounders, there may still be other unmeasured factors that can lead to residual confounding.

Implications and next steps for research, clinical practice, and public policy

Our research supports a life course approach for dementia prevention, particularly around the design of risk reduction strategies pertaining to reproductive factors which are unique to women. It is necessary to validate our findings on exogenous hormone use through rigorous clinical trials, and our findings may be helpful for identifying high-risk women to participate in future trials. Further, social deprivation is likely to be an important determinant of dementia risk and other aspects of women’s health, given that the elevated dementia risk associated with early (natural and artificial) menopause were more strongly associated with dementia in women of lower SES.

Conclusions

This study highlights that the reproductive and endocrine milieu in women may be involved in dementia risk, although the physical experience of childbearing is unlikely to account for the risk variation in women, given the similar associations observed for number of children and dementia in women and men.

STROBE statement—Checklist of items that should be included in reports of observational studies.

STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

(DOCX)

Click here for additional data file.

Number of missingness for each reproductive factor of interest in the UK Biobank.

aCollected in women who indicated that they ever had given birth to more than 1 child (N = 184,876). bCollected in women who indicated that they ever had been pregnant (N = 231,352). cCollected in women who indicated that their periods had stopped (had natural menopause) (N = 165,301). dCollected in women who indicated that they had taken the contraceptive pill (N = 220,344). eCollected in women who indicated that they had ever used HRT (N = 104,133). fCollected in women who indicated that they had ever used HRT and not currently using HRT (N = 87,413). HRT, hormone replacement therapy.

(DOCX)

Click here for additional data file.

Unadjusted and multiple adjusted HRs and women-to-men ratio of HRs (95% CIs) for the risk of dementia associated with number of children for women and men.

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drug. BMI, body mass index; CI, confidence interval; HR, hazard ratio; RHR, ratio of hazard ratio.

(DOCX)

Click here for additional data file.

Unadjusted HRs (95% CIs) for the risk of dementia associated with reproductive factors, by age, SES, smoking, and BMI.

aEarly menarche was defined as age at first menstrual period before the age of 12 years. bEach live birth in women and each child fathered in men. cEarly menopause was defined as the permanent absence of menstrual periods before the age of 47 years. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy; SES, socioeconomic status.

(DOCX)

Click here for additional data file.

Unadjusted and multiple adjusted HRs for the risk of dementia associated with history and timing of hysterectomy and oophorectomy.

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HR, hazard ratio; SES, socioeconomic status.

(DOCX)

Click here for additional data file.

Unadjusted and multiple adjusted HRs (95% CIs) for the risk of dementia associated with the timing of HRT use in relation to age at (natural or artificial) menopause in postmenopausal women.

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy.

(DOCX)

Click here for additional data file.

Unadjusted and multiple adjusted HRs (95% CIs) for the risk of dementia associated with age at menarche, reproductive years and age at menopause, after excluding those had hysterectomy or oophorectomy.

aAnalyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HR, hazard ratio.

(DOCX)

Click here for additional data file.

Multiple adjusted risk ratios (95% CIs) for the risk of dementia associated with reproductive factors, comparing Cox proportional hazards regression (HRs), competing risk model (HRs), and multinomial regression (odds ratios).

Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. aThese CIs were calculated without the floating absolute risk, to provide comparable results to the competing risk models and the multinomial regression models. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy; OR, odds ratio.

(DOCX)

Click here for additional data file.

Multiple adjusted HRs (95% CIs) for the risk of dementia associated with reproductive factors stratified by Townsend fifths.

Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. aEarly menarche was defined as age at first menstrual period before the age of 12 years. bEach live birth in women and each child fathered in men. cEarly menopause was defined as the permanent absence of menstrual periods before the age of 47 years. BMI, body mass index; CI, confidence interval; HRT, hormone replacement therapy.

(DOCX)

Click here for additional data file.

Townsend-weighted multiple adjusted HRs (95% CIs) for the risk of dementia associated with reproductive factors.

Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. aWeighted HR are weighted according to equal weights of each stratified HR, to represent that the national Townsend fifths contain equal number of people. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy.

(DOCX)

Click here for additional data file.

Multiple adjusted HRs (95% CIs) for comparing individual and combined associations between reproductive factors and dementia.

(A) A consideration of pregnancy related factors: number of live births, stillbirths, miscarriages, and abortions, although there were no significant change in the model coefficients in this model compared with the individual associations. Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. (B) A consideration of factors throughout the life span including age at menarche, parous versus not, hysterectomy and or oophorectomy, HRT use, and contraceptive pill use. Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HRT, hormone replacement therapy.

(DOCX)

Click here for additional data file.

Multiple adjusted HRs for the risk of dementia associated with reproductive factors in women, after imputed for missing data using MICE, compared with complete case analyses.

Analyses were adjusted for age, Townsend index, ethnicity, smoking status, systolic blood pressure, BMI, diabetes, total cholesterol, antihypertensive drugs, and lipid-lowering drugs. BMI, body mass index; CI, confidence interval; HR, hazard ratio; HRT, hormone replacement therapy; MICE, Multivariate Imputation by Chained Equations.

(DOCX)

Click here for additional data file.

13 Apr 2021

Dear Dr Gong,

Thank you for submitting your manuscript entitled "Reproductive factors and the risk of incident dementia: Results from the UK Biobank" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Apr 15 2021 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

13 Oct 2021

Dear Dr. Gong,

Thank you very much for submitting your manuscript "Reproductive factors and the risk of incident dementia: Results from the UK Biobank" (PMEDICINE-D-21-01592R1) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to four independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we will consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Nov 03 2021 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. Please address the following comment from the Academic Editor: The paper gives a detailed look at this issue, but there are a couple of major areas of concern – one is the challenge of interpretation of the results given the low response rate for UK Biobank and the likelihood that those who took part from lower SES seem unlikely to fully represent these populations. The findings suggest residual confounding with these factors.

2. Title: Please revise your title according to PLOS Medicine's style. Your title must be nondeclarative and not a question. It should begin with main concept if possible. "Effect of" should be used only if causality can be inferred, i.e., for an RCT. Please place the study design ("A randomized controlled trial," "A retrospective study," "A modelling study," etc.) in the subtitle (ie, after a colon).

3. Abstract Background: Provide the context of why the study is important. The final sentence should clearly state the study question.

4. Abstract: Methods and Findings: Please quantify the main results presented in the abstract with both 95% CIs and p values.

5. Abstract: Methods and Findings: Please mention the important dependent variables that are adjusted for in the analyses.

6. Abstract: Methods and Findings: In the last sentence of the Abstract Methods and Findings section, please describe the main limitation(s) of the study's methodology.

7. Abstract: Conclusions: Please address the study implications without overreaching what can be concluded from the data; the phrase "In this study, we observed ..." may be useful.

8. Author Summary: At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary

9. Main text: For in-text citations, please note reference numbers within square brackets, placed before the sentence punctuation, rather than superscript numbers. Where multiple references are listed, please do not include spaces within brackets.

10. Introduction: Line 63: Please expand the description of the study’s objective to include that men were also included in the investigation.

11. Methods: Please ensure that the study is reported according to the STROBE guideline. Please add the following statement, or similar, to the Methods: "This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist)."

12. Methods: Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis-- including those made in response to peer review comments-- should be identified as such in the Methods section of the paper, with rationale.

13. Methods: Line 101: Please provide more information on the Townsend Deprivation Index for determining SES and how this was categorized.

14. Results: For all the main results presented in the text, please note both 95% CIs and p values.

15. Discussion: Please present and organize the Discussion as follows: a short, clear summary of the article's findings; what the study adds to existing research and where and why the results may differ from previous research; strengths and limitations of the study; implications and next steps for research, clinical practice, and/or public policy; one-paragraph conclusion.

16. Discussion: Line 263: Please check if this should read “dendritic spine”.

17. Conclusions: Line 350: Here and throughout, please avoid language that implies causality, such as “This study highlights that the reproductive and endocrine milieu in women can have a significant impact on their dementia risk…” as this study is observational and therefore causality cannot be inferred.

18. Tables and Supporting information tables: For all analyses presented, please include both adjusted and unadjusted analyses. Please report p values as well as 95% CIs.

19. Table 1: Please note all abbreviations used in the legend (SES, HRT, etc).

20. Table 3: Please also present the unadjusted analyses. Please present p values in addition to 95% CIs.

21. Table 4: Please also present the unadjusted analyses. Please define all abbreviations in the legend (HRT, BMI).

22. S1 Table: The rates of dementia by number of children for men could be moved to the main text (the adjusted and unadjusted rates for women are presented in Table 2).

23. Figure 1: Please describe each panel (A, B, C, etc) in the legend. Please also note in the legend the meaning of the line/shaded region.

24. Figure 2: Please note in the legend that the bars represent the 95% CIs.

25. References: Please use the "Vancouver" style for reference formatting (particularly journal title abbreviations), and see our website for other reference guidelines https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

26. STROBE Checklist: Thank you for including the checklist. Please refer to locations in the text using section and paragraph numbers, rather than page numbers.

Comments from the reviewers:

Reviewer #1: This paper investigates how reproductive factors are related to incident dementia

The analyses are based on the UK biobank

This a a well written paper but there are several potential methodological issues that are not addressed

- Firstly as authors are modeling age at dementia they need to make sure that the information used is accurate. In this case they use age at dementia based on the first time dementia is recorded and that includes death

This is very likely to be biased as dementia is not fatal. This definition would need to be challenged in sensitivity analyses

- the authors report higher incidence of dementia in men than in women which is in disagreement with most reports. How do they explain that ?

- The authors use multiple indicators of reproductive life and this poses two problems : first a DAG would be useful to understand the underlying mechanisms that are investigated and this will help the reader to understand the analytical strategy , second the authors might condider correcting for multiple comparisons if they don't explain their strategy better

- the authors would need to take into account death as a competing vent

Reviewer #2: This is a very interesting and important manuscript examining reproductive risk factors in women and men using UK Biobank data. The authors report that several aspects of reproduction in women including age at menarche, age at menopause, hysterectomy, oral contraceptive use, and hormone replacement therapy are associated with risk of dementia. Further, there were U-shaped associations between number of children and risk of dementia for both women and men. Overall, the manuscript is well-written. My comments are enumerated below.

1. A major concern is the lack of consideration of race/ethnicity. It is well known that age of menarche and menopause tend to occur earlier in women who are black as compared to white. Thus, race/ethnicity must be considered in the analyses. It should be reported in the tables and stratification is warranted.

2. The Townsend deprivation index is included as a covariate, but education is not. Although these should be related, it is also important to assess the effect of the number of years of education separately. Women with age at first birth who were aged 24 or less were at greater risk of developing dementia - this could partially be influenced by education. Further, women with an abortion were at reduced risk of dementia - could this also be related to education?

3. Premenopausal bilateral oophorectomy, but not unilateral oophorectomy, has been associated with risk of dementia. Is it possible to determine whether the oophorectomy was unilateral or bilateral?

4. There is no discussion of missingness. Presumably with this large of study, some data was missing. How was this handled? Were individuals excluded (in which case how did they compare to those included), or was missing data imputed?

5. What is meant by incident 'fatal' dementia? I have not heard this term before. Does this mean it was only reported on the death certificate but not the medical record?

6. Was the duration of contraceptive use associated with risk of dementia?

7. Given HRT and contraceptive use was ascertained for women, could the authors also look at testosterone use for men?

8. Timing of HRT in relation to menopause is important. Is there information regarding when HRT was started in relation to menopause? Several studies of women who start HRT around the time of menopause do not report an increased risk of cognitive decline or dementia.

9. Each of the reproductive factors were examined in isolation but that is not how they occur over the lifespan for women. It would be helpful to show the combination of factors or to put multiple in a model to determine which aspects might be most important. For example, is hysterectomy associated with dementia because it often co-occurs with bilateral oophorectomy and results in an earlier age at menopause?

Reviewer #3: The overall significance of this paper is very high. Female-specific reproductive factors may explain sex differences in dementia prevalence. Further, the effects of oral contraceptives on dementia risk are unknown. Thus, the current study brings exciting new data to the field while appropriately addressing limitations of the study. Only minor revisions are suggested to consider additional factors in data interpretation and expand on a few discussion points.

Main Comments:

1. It is stated several times through the abstract, results, and discussion that for early menopause hysterectomy or oophorectomy that "women of lower SES had a greater dementia risk, but not for women of high SES." This statement could be misleading. Table 4 shows that even in high SES women that early menopause was still associated with increased risk (1.19 HR). Thus, it seems more accurate to convey that early menopause increased dementia risk, but that low SES women had increased risk relative to high SES women.

2. Please discuss how HRT age of initiation could have played a role (if at all) in the SES difference? Did women with higher SES go into menopause earlier or later? Did they initiate HRT earlier or did a greater percentage of them undergo HRT? Could that have contributed to risk?

3.In Figure 1 it looks like some women started HRT very early - at 35 years of age and these were the women at highest risk (2.0 HR). Did these women also have premature menopause (either natural or surgical) that could have increased their risk? How was this accounted for in the interpretation of the data? For example, if women went through HRT early due to premature menopause (surgical or natural) then how was the risk due to early menopause vs. HRT separated statistically?

Minor Comments:

1. In the abstract, the statement "although female biological factors involved in childbearing are unlikely to account for risk variation" was unclear on it is own. It would be helpful to say why.

3. In the final paragraph it is unclear what is meant by "Further, deprivation is likely to an important determinant of dementia". Deprivation of what? Estrogen?

4. The final sentence it is unclear what the "Beijing Declaration" is. This ending to the manuscript felt odd.

Reviewer #4: See attachment

Michael Dewey

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: gong.pdf

Click here for additional data file.

18 Nov 2021

Submitted filename: Responses to Editor and referees_PMEDICINE-D-21-01592R1.docx

Click here for additional data file.

23 Dec 2021

Dear Dr. Gong,

Thank you very much for submitting your manuscript "Reproductive factors and the risk of incident dementia: a prospective cohort study of UK Biobank participants" (PMEDICINE-D-21-01592R2) for consideration at PLOS Medicine.

Your revised paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to three of the original reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of the remaining points mentioned by Reviewer 4, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Jan 13 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1. From the Academic Editor: In terms of discussing the limitation of the representativeness of UK Biobank, please describe specific evidence on socioeconomic and other factors that might covary with each of the variables, e.g. abortion, stillbirth, early pregnancy, etc. It seems as if the adjustment for Townsend Deprivation Index and weighting according to social structure has in fact attenuated the statistical significance of associations. Please provide additional comment and interpretation for this result. Please discuss differential responses by deprivation and within different communities that could be related to some of the factors.

Please discuss the limitations of the methods for determining dementia outcomes in terms of bias (including SES and age-related biases) and under-reporting.

2. Response to reviewers: Please completely address the remaining concerns of Reviewer 4, including the points regarding multiple imputation as well as the inconsistencies with the interpretation of observed U shaped relationships.

3. Title: Please capitalize the first word of the subtitle, and please remove “prospective” from the title: “Reproductive factors and the risk of incident dementia: A cohort study of UK Biobank participants”

4. Data Availability statement: Please omit the part of the statement indicating that UK Biobank resources are available from the authors, but do include the most direct link possible for data access requests (e.g. https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access) and also include any relevant (non-author) contacts (e.g. access@ukbiobank.ac.uk).

5. Abstract: Please combine the Methods and Findings sections into one section, “Methods and findings”.

6. Abstract: Line 37: Please change to “1866 dementia cases were recorded in women” if this is intended.

7. Abstract: Please report exact p values, except where p<0.001.

8. Author summary: What do these findings mean? Please format this section into 2-3 bulleted points.

9. Introduction: Line 95: Please avoid use of “effects” as this implies causality.

10. Methods: Line 100: Please mention that there was no prospective analysis plan designed specifically for this study, though the framework for the design was drawn from the analyses carried out in a previous study.

11. Methods: Line 153: Please clarify if this should be “including information about unemployment…” or similar.

12. Results: When reporting p values, please report the exact p value in the text, except please report p<0.001 where applicable. Please provide the exact p value where p<0.01.

13. Results: Line 330-331: Please revise to avoid casual implications “...no evidence of HRT duration affecting dementia risk…” and please avoid causal language throughout.

14. Discussion: Line 381-383: Please clarify this sentence as the meaning is not clear. “The discrepancy in findings cannot be explained by the exclusion of women who reported hysterectomy or oophorectomy in the two studies reported opposite findings”

15. Discussion: Line 401: Please revise to avoid describing causal interpretations of the study findings.

16. Discussion: Line 459: Please remove “...and the prospective design…” from the sentence.

17. Page 32: Please remove the sections Availability of data and materials, Financial disclosure statement, Competing interests, and Author contributions from the main text of the manuscript. Please make sure the information is accurately and completely entered in the relevant fields of the manuscript submission system.

18. Table 3, Table S2: Please report exact p values, unless p<0.001.

19. STROBE Checklist: Thank you for including the checklist. Please revise, using only section and paragraph numbers to refer to locations within the text (e.g. Methods, paragraph 1). Please do not include page or line numbers.

20. S3 Table, S4 Table, S6 Table: Please report p values along with the 95% CIs for these analyses.

21. S5 Table, S2 Figure: Please report exact p values, unless p<0.001.

22. S1 Figure and S2 Figure: Please also include a column with the HR, 95% CIs and p values from the unadjusted models.

23. Reference list: Please double check the reference formatting, and please see our website for additional guidance: https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

For example, it seems as if the journal title for reference 7 should be Am J Obstet Gynecol.

Comments from the reviewers:

Reviewer #2: The authors have addressed all of my comments.

Reviewer #3: All comments were addressed.

Reviewer #4: See attachment

Michael Dewey

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: gong2.pdf

Click here for additional data file.

18 Jan 2022

Submitted filename: Response document_PMEDICINE-D-21-01592R2.docx

Click here for additional data file.

10 Feb 2022

Dear Dr. Gong,

Thank you very much for re-submitting your manuscript "Reproductive factors and the risk of incident dementia: A cohort study of UK Biobank participants" (PMEDICINE-D-21-01592R3) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by one of the reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Feb 17 2022 11:59PM.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

1. Throughout (Abstract Line 45-46, Line 52-53; Author Summary Lines 69-70 and 73-74; Results Lines 287-288; Discussion Lines 425-428 , Conclusion Lines 497-499): A similar U-shaped association between number of children and the risk of dementia for both sexes is mentioned as a finding. The evidence in support of an overall relationship between number of children and risk of dementia was not entirely clear from the results. An apparent U-shaped pattern can be observed in Figure 2. The results seem to suggest a significant difference between women with no compared to 2 live births in the adjusted analyses, while there is no significant association with parity, or per live birth (Table 2). It is not completely clear, but from Figure 2, this may not be the case in men, and this would seem to suggest the relationship may differ between sexes to some extent. Also, Table 3 and Table S7 seem to suggest there may be different relationships between number of births and risk when SES/deprivation are examined, for women and men. Please clarify this throughout the manuscript, and revise conclusions as appropriate.

2. Abstract: Line 27: Please change “equivocal” to “uncertain” in this sentence.

3. Abstract: Line 29: Please revise to avoid the word “effects” as this may imply causality. We suggest: “...whether the associations vary by age, socioeconomic status…”

4. Abstract: Line 40: We suggest referring to either >14 or >/= 15 consistently throughout the manuscript. In the results, >14 yrs is used.

5. Abstract: Please combine the “Methods and Findings” and “Results” sections into a single “Methods and Findings” section.

6. Abstract: Line 45-46: For this sentence, please report whether there is a significant association between number of children and dementia risk, provide the results (HR and 95% CIs, p values) to support the mentioned associations between number of children and dementia risk. “The U-shaped associations between the number of children and the risk of dementia were similar for both sexes,”

7. Abstract: Line 46-47: Please mention that this association did not reach statistical significance for early menopause.

8. Abstract: Line 48-50: We suggest also mentioning the representativeness of UK Biobank population as one of the main limitations of the study.

9. Abstract: Line 52-53: “...and there was a similar association between the number of children and dementia risk between women and men.” We suggest removing or clarifying this point to avoid misinterpretation (please see point #1 above).

10. Author summary: Line 59: We suggest “uncertain” rather than “equivocal” in this point.

11. Author summary: Line 69: We suggest revising to: “The U-shaped associations between number of children and dementia appeared similar…”

12. Author summary: Line 72: We suggest revising to “In this study we observed that certain reproductive factors are associated with greater risk of dementia. Future work is needed to understand whether this is related to the fact that these factors may be associated with shorter cumulative exposure to endogenous estradiol.” or similar.

13. Author summary: Line 73-74: Please see point #1 above. We suggest revising to: “Findings suggest that risk variation in women may not be associated with factors associated with childbearing because a similar U-shaped pattern was observed between number of children fathered and dementia risk among men.” or similar.

14. Throughout: In-text citations: Please separate the square brackets from the preceding word with a space.

15. Introduction: Line 91: We suggest “uncertain” rather than “equivocal” in the sentence.

16. Results: Line 248: Please define the values in parentheses at the first point of use in the text, e.g. rate (95% confidence interval).

17. Results: Line 287-288: Please revise the sentence to: “Compared with those who had two children, the associations between the number of children and dementia were similar for women and men and appeared to be U-shaped (Figure 2).” Please describe the HR with 95% CIs and p values for associations overall, and separately for men and women.

18. Discussion: Line 364: We suggest revising to: “...with dementia risk later in life.”

19. Discussion: Line 384-385: We suggest: “...estradiol has been shown to be correlated with neuronal dendritic spine density [22]...” or similar, as it seems the reference does not describe regeneration.

20. Discussion: Line 408-409: We suggest revising to: “...providing further support that early cessation of hormones may be associated with poorer cognitive outcomes.” or similar.

21. Discussion: Line 424-425: Please clarify if this should be: “...while nulliparity and having two to four children showed similar associations compared to primiparous women…”

22. Discussion: Line 425-428: It is not clear from the results whether a statistically significant association between number of children and dementia risk was identified for both men and women. We suggest tempering this sentence accordingly. “In our study, the number of children was similarly associated with

dementia risk for women and men. As such, the risk variation in women appears to be more related to social and behavioural factors involved in parenthood rather than biological factors associated with childbearing.”

23. Discussion: Line 460: We suggest revising to: “...do not support associations between HRT and dementia risk…” or similar.

24. Discussion: Line 474: We suggest adding an additional sentence describing the limited representativeness of the UK Biobank, and a clear description of why this is important to consider in terms of interpretations of your results.

25. References: Please update the reference information for Referenes # 40 and # 46.

26. Figure 2: Please provide the results from both adjusted and unadjusted analyses for men (e.g. similar to the results for women are provided in Table 3).

27. S9 Table: Please provide p values in addition to 95% CIs.

Comments from Reviewers:

Reviewer #4: The authors have now addressed my remaining points.

Michael Dewey

Any attachments provided with reviews can be seen via the following link:

[LINK]

16 Feb 2022

Submitted filename: Response document_PMEDICINE-D-21-01592R3.docx

Click here for additional data file.

23 Feb 2022

Dear Dr Gong,

On behalf of my colleagues and the Academic Editor, Carol Brayne, I am pleased to inform you that we have agreed to publish your manuscript "Reproductive factors and the risk of incident dementia: A cohort study of UK Biobank participants" (PMEDICINE-D-21-01592R4) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

Please also address the following editorial requests:

-Abstract Line 49-50: Please correct the typo to read: “There was evidence that hysterectomy…”

-Abstract: Line 55-56: Please change to: “In this study, we observed that some reproductive events related to shorter cumulative endogenous estrogen exposure…”

-Discussion: Line 487: Please change “...Caucasian ancestry…” to “white ethnic background” if this is what is intended.

-Reference List: Please update reference 46 with the complete citation information.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine