Study design and setting
A case-control study targeting Saudi postmenopausal women was conducted in Jeddah, Saudi Arabia. Participants were recruited from multiple dental clinics, including seven primary healthcare facilities operated by the Ministry of Health (MOH), three private health facilities, and the Armed Forces Hospital. Data collection took place in a single private dental clinic to ensure consistency in examination tools and procedures.
Sample size calculation
The required sample size was calculated to detect a significant odds ratio using a case-control sample size formula [8,9,10]. The formula considered a two-sided type I error probability (α) of 5% and a power (β) of 80%, with estimated exposure probabilities of 45% in controls and 30% in cases, calculated from previous literature. A minimum of 308 participants was needed, with 154 in each group. This study included 372 participants (186 cases and 186 controls) to improve statistical reliability. Detailed calculations are provided in the supplementary material for clarity as S1: Sample Size Calculations.
Inclusion criteria
-
Saudi women aged 45 years and above.
-
At least 12 months post-menopause.
-
A minimum of 8 teeth (including third molars), If a tooth loses its crown and becomes a remaining root, it will still be counted as a tooth and included in the periodontal assessment.
-
Cases were defined as those who had generalized periodontitis, characterized by 30% or more sites with periodontitis in the oral cavity. Controls were defined as individuals with 0% periodontitis in the oral cavity, matched to cases by age and location.
Exclusion criteria
-
Women who had used HRT but discontinued it before the study were excluded to eliminate potential confounding effects on the study’s outcomes.
-
Early menopause onset participants (before 45 years) were excluded to focus on participants with a more typical age range for menopause onset, which may impact hormonal influences on periodontitis.
-
History of hysterectomy, as the absence of ovaries or the alteration of hormone production can prevent the use of HRT and affect the results regarding hormonal impacts.
-
Cases that couldn’t be matched to controls by age and clinic location. This ensures comparability between cases and controls, eliminating potential confounding factors related to demographics or regional influences.
-
Localized incisor-molar periodontitis or localized periodontitis (affecting 1–29% of the oral cavity). By excluding participants with localized periodontitis, the study is less likely to be influenced by factors like genetic predispositions, trauma, or misclassification, which could create bias in the interpretation of results. Focusing on generalized periodontitis, which is more closely related to systemic factors such as hormonal changes, ensures that the study directly examines the relationship between hormonal fluctuations and periodontal health. This reduces bias by narrowing the focus to a more homogenous group, making it easier to interpret the results without the confounding effects of localized, non-systemic influences.
Exclusion criteria were made to improve the reliability of the results and reduce bias. This transparency not only strengthens the validity of the study but also enhances its credibility, as it shows a thoughtful approach to controlling for confounding factors.
Sampling methodology
A systematic random sampling technique was employed. Cases were selected based on their periodontal status, following a pattern (e.g., 1st, 3rd, 5th). Controls were matched to cases by age and clinic location. If matching controls were unavailable, the corresponding cases were excluded.
Data collection
Data collection occurred between July 2023 and August 2023 and involved structured interviews, clinical examinations, and medical file reviews. A single researcher conducted all data collection to ensure consistency and minimize variability. To reduce potential interviewer bias, the researcher underwent standardized training to ensure uniformity in administering the structured questionnaire and performing clinical measurements. The questionnaire (see Supplementary Material S3: Questionnaire) was developed based on known periodontitis predictors and common chronic diseases observed in the elderly Saudi population, as outlined in previous research [1, 5, 6, 11, 12]. It consisted of three sections: (1) sociodemographic and medical information, (2) medical and menopausal history (including chronic diseases, age of menopause, and hormonal replacement therapy usage), and (3) periodontal outcome assessment using the AAP 2017 chart [13]. Clinical attachment loss (CAL) and pocket depths were measured using Williams 15 periodontal probes and radiographs, with precise measurements recorded to ensure accuracy.
Assessment Tools (can be found in supplementary material S2: clinical assessment).
-
1.
Periodontal Health:
-
Measured using the American Academy of Periodontology (AAP) 2017 classification [13].
-
Clinical attachment loss (CAL) was assessed using the formula: CAL = Pocket Depth (PD) + Distance from Cementoenamel Junction (CEJ) to Gingival Margin.
-
CAL and pocket depth were measured at four sites per tooth (mesiobuccal, distobuccal, mesiolingual, and distolingual).
-
Bitewing radiographs were used to assess bone loss, defining bone loss as > 15% of bone around the root.
-
2.
Secondary Outcomes:
-
CAL was measured at four sites per tooth (mesiobuccal, distobuccal, mesiolingual, and distolingual). Then the average attachment loss was measured by dividing the total loss by/ number of sites.
-
Salivary Secretion: Stimulated saliva flow was measured using a 5-minute wax-chewing test, with results recorded in ml/min.
-
Bone Loss: Measured using radiographs and categorized as < 15% (no bone loss) or > 15% bone loss.
Pilot study
A pilot study involving 11 participants was conducted to refine the questionnaire and ensure the feasibility of the methodology. Adjustments included switching to interview-based questionnaires due to participants’ limited literacy and adding questions about unlisted illnesses common in the population, not found in previous research. Data collection per participant took approximately 25 min.
Ethical considerations
Ethical approval was obtained from the Ministry of Health’s Institutional Review Board (MOH IRB log A0168, registration number NCBE-KACST, KSA: H-02-J-002). The study adhered to the principles outlined in the Declaration of Helsinki, ensuring the protection of participants’ rights, safety, and well-being. Participants provided informed consent through verbal agreement and fingerprint acknowledgment. Data collection and analysis were anonymous, and questionnaires were destroyed after data entry.
Statistical analysis
The primary outcome, Periodontitis, was coded as a dichotomous variable, then further stratified as ordinal categorical variables arranged according to severity. The sample’s binary socioeconomic and medical data were displayed as percentages, while continuous data were expressed as mean and standard deviation. It was also divided between cases and controls to display any significant difference in the socioeconomic and medical traits between the two groups. The association between exposure (HRT usage) and outcome (periodontitis presence) was measured through odds ratio and Chi-square with significance at 95% CI. The effect of HRT on the severity of periodontitis was measured using a multinomial categorical regression coefficient, with healthy being the reference. A logistic model was created that includes all the variables, showing the significance of each variable in association with periodontitis. It was followed by creating a forward stepwise logistic regression to see significant predictors of periodontitis presence (binary) with entry at (0.05) and exit at (0.051). The primary exposure, which is HRT usage, was lock-termed. The following variables were added to all the logistic model: HRT usage (Binary), duration of HRT usage (continuous), age (continuous), BMI (continuous), employee status (binary), smoking status (binary), monthly income (continuous), an education level (ordinal categorical), age of menopause (continuous), hypertension (binary), diabetes (binary), osteoporosis (binary), vitamin d deficiency (binary), hypothyroidism (binary), malignant tumors (binary), arthritis (binary), asthma (binary), liver diseases (binary), kidney diseases (binary), Salivary flow rate (continuous), cortisone usage (binary), aspirin usage (binary), antidepressant usage (binary). The final model contained: User of HRT (Binary), Age of menopause (continuous), Smoking status (binary), Education level (Categorical ordinal), Diabetes (binary), kidney diseases (binary), with p-value < 0.05 and 95% CI significance. Height and weight were excluded from the model due to BMI and collinearity. A ROC graph for the model area under the curve (AUC) was created to measure the effect of all significant variables in association with HRT. The graph also measures the model’s sensitivity, specificity, and predictability. Secondary outcomes were also measured in association with HRT usage. Clinical attachment loss was measured using a correlation coefficient. A forward stepwise linear regression was created to see possible predictors of CAL (continuous) with entry at (0.05) and exit at (0.051). The primary exposure, which is HRT usage, was lock-termed. All previously mentioned variables (included in the logistic model) were added to the linear regression model with a p-value < 0.05 and 95% CI significance, and R2 was used to assess the model. The Association between HRT salivary secretion was measured once with saliva as a continuous variable using a coefficient and again with saliva as a binary variable (Normal saliva secretion) and (Low saliva secretion) using the odds ratio. The association of bone loss and HRT was also measured using OR.