This descriptive analytical cross-sectional study used the data obtained from the enrollment phase of the Azar cohort study which was a part of the Prospective Epidemiological Research Studies (PERSIAN cohort) for Iranian adults [15]. In 2014, Azar cohort sample was selected from permanent residence in Shabestar city in East Azerbaijan province, Iran About 15,000 participants with the age range of 35 to 70 years were included, which about 55% of whom were women. Azar cohort, a large population-based cohort study, has conducted for adults to assess risk factors for non-communicable diseases [14]. The assessors performed the interviews using predefined and validated research protocols and questionnaires of Persian cohort [14]. All eligible adults of the city, aged 35 to 70 years, were invited to participate in Azar cohort. Anyone who accepted the aim and steps of the study, filled a written informed consent and was free to leave the study at any time. The objectives of Azar cohort, sampling method, variables of the Azar cohort, and data collection tools and methods were explained in previous published papers [14, 15]. In our cross-sectional study, the inclusion criteria were the data of all women participating in the Azar cohort with a history of reproduction and oral health, while the exclusion criteria were the data of women lacking a history of reproduction and oral health. Consequently, the data of all 8294 women from the enrollment phase of the Azar cohort were entered into the analysis using the census method. This cross-sectional study was approved by the Ethics Committee of the Tabriz University of Medical Sciences, Tabriz, Iran (IR.TBZMED.REC.1400.1017).

Data collection in the Azar cohort was performed through 4 forms. We extracted the necessary data for the present study from three of those forms. First, the “General information” form that covered sociodemographic characteristics. Socioeconomic (SES) status and educational level as the qualitative variable were extracted. In terms of SES, the participants were categorized into 5 groups (very poor, poor, medium, good, very good). Socioeconomic status was determined on the basis of job title, car ownership, number of trips made (per year), type of travel, having a personal computer, home ownership and having multiple jobs. In terms of educational level, the participants were categorized into 4 groups (illiterate, primary, diploma, university). Age at interview and BMI as the continuous data were reported by mean and standard deviation.

The second was medical history form which the participants were asked about history of different non-communicable diseases in their life. Self-report history of diabetes mellitus, hypertension, lung disease, depression, and cardiovascular disease were extracted from this form. The third was the “Oral health” and “reproductive history” form. “Oral health” information had two parts. First, the participant answered the questions about oral hygiene habits. We extracted the answer of this question for our study: “How often do you brush your teeth?” and divided the answers into two groups: brushing at least once a day and brushing less than once a day. Oral examination was done according to the World Health Organization (WHO) Oral Health Surveys Basic Methods [16] by a single examiner (a general practitioner who was trained by a skilled dentist), and the DMFT index was recorded properly using headlight, intraoral mirror, an explorer probe, and a piece of sterile gauze to clean teeth surfaces [17]. In this form, the participants were also asked about age at the onset of menstruation, age of onset of menopause, age of first pregnancy, frequency of pregnancy. The number of pregnancies was evaluated as a categorical variable in five groups with zero, one, two, three, four and more pregnancies. The age at the onset of menstruation was evaluated as a categorical variable in three groups: early (< 12), normal (12–14 years old), late (14<) [18]. The normal group was considered as the reference group. The age at the onset of menopause was evaluated as a categorical variable in three groups: Premature (< 40), early (40–45), normal (> 45). The normal group was considered as the reference group and the premature and early groups were compared with the normal group. The age of the first pregnancy was categorical variable in four age groups: 10–19, 20–29, 30–39, ≥40 which the reference group was 20–29. In our study, the variables related to the reproductive history were the exposure variables, and DMFT was the outcome variable. The other variables were considered as the possible confounding variables.

For analysis, negative binomial analysis with loglink was applied. To adjust the effect of possible confounding variables, such as age at interview, educational level, socioeconomic status, frequency of tooth brushing, chronic diseases and body mass index along with the main variable were analyzed in several different regression models. Unadjusted model analyzed without considering the confounding variable. Model 1 adjusted for educational level, socio-economic status, age at interview, chronic diseases (diabetes mellitus, hypertension, cardiovascular, depression and lung diseases), body mass index, and frequency of tooth brushing. Model 2 adjusted for educational level, socioeconomic status, age at interview, chronic diseases (diabetes mellitus, hypertension, cardiovascular, depression and lung diseases), and body mass index. Model 3 adjusted for educational level, socio-economic status and age at interview. Statistical tests were carried out using SPSS 20.0 software (IBM Company, Chicago, IL, USA). Statistical significance was set at p < 0.05.