An analytical cross-sectional study was conducted across eight study sites to collect data on socio-demographics, sexual and reproductive health, the prevalence of UFs and associated risk factors among women seeking health services at gynaecology outpatient departments or admitted to gynaecological wards. Women with signs or symptoms associated with UFs (abnormal menstrual bleeding, heavy, prolonged bleeding), pelvic pain, pelvic pressure, enlarged abdomen, lower abdominal pain and back pain or ultrasound-confirmed or unconfirmed UFs diagnoses or those who had undergone UFs-related surgery (hysterectomy, myomectomy, laparoscopic procedure or uterine embolisation) were targeted. To determine the prevalence of a disease, one requires a study population of both patients with the disease and those without.^44,45

This study was conducted in the Kingdom of Eswatini, a sovereign state in Southern Africa bordered by South Africa and Mozambique. It is one of Africa’s smallest countries and is classified as a lower-middle-income country. Eswatini has four geographical regions: Hhohho, Manzini, Lubombo and Shiselweni, with an estimated population of 1.2 million. Human development is medium, according to the 2022 Human Development Index (0.60). The study was conducted within a hospital-based sample of women from the four geographic regions between 21 August 2021 and 21 August 2022. The women were recruited from eight study sites, namely Piggs Peak Government Hospital, Mbabane Government Hospital, Mbabane Private Clinic, Raleigh Fitkin Memorial Hospital, Women and Children Private Clinic, Mankayane Government Hospital, Hlatikulu Government Hospital and Good Shepherd Mission Hospital, representative of the country’s main hospitals.

To estimate the prevalence of UFs, three considerations informed the optimal sample size for this study: a 95% desired confidence interval, an accepted 5% margin of error and the estimated prevalence of the outcome of interest. The prevalence of the outcome of interest was assumed unknown, with a maximum variability of 50%. A sample size of 384 participants was required; this was increased by 13% to account for potential rejections, then multiplied by a design effect (D) of 1.5, yielding a final desired sample size of 652 women. However, 645 women were eventually enrolled, as some screened women did not meet the age criterion for inclusion. Our approach to determining the required sample size was in line with established guidelines for health research, as sample size calculations are directly related to the variance. Assuming a 50% prevalence yields the most conservative estimate for the sample size, meaning the largest sample size is obtained. This approach minimises the risk of underestimating the required sample size and ensures adequate statistical power for the study.^46,47 Increasing the sample size reduced the type I and type II errors, as well as the effects of known and unknown confounders. The power (1-β) (the % chance of detecting a difference) of the study was set at 80%.⁴⁸

Admitting hospitals across the four Eswatini regions were stratified by facility type (government, mission or private). From these, eight hospitals were purposively selected based on the presence of a gynaecologist, a radiology department with fully functioning equipment (e.g. scans) and an obstetric department. Participants were recruited using purposive sampling. Women attending the gynaecology outpatient departments or admitted to the wards were screened and purposively enrolled as they presented to the gynaecology clinics, provided they met the inclusion criteria. These included women aged 25–64 years with signs and symptoms of UFs (pelvic pain, abnormal menstruation), an unconfirmed or confirmed UFs diagnosis or a history of any UFs-related surgical procedure (hysterectomy, myomectomy, laparoscopic or uterine embolisation), who had voluntarily consented. Women younger than 25 years or older than 64 years, without a confirmed UFs diagnosis or UFs-related surgery and not attending the gynaecology outpatient departments or admitted to the wards were excluded from the study. Enrolment was offered equally to participants, irrespective of culture, religion, race and social class. Trained research assistants (RAs) individually explained the study aims and procedures to women willing to participate. We also discussed the potential risks and benefits of participating in the study with each participant before obtaining their written consent. To protect their privacy, we assigned them an identification number unrelated to their real names and surnames.

Data were collected through face-to-face interviews using structured questionnaires. A pilot study was conducted at Manzini Clinic, which was not among the selected study sites. The pilot informed refinements to the tools and additional training for the RAs, who were registered nurses. The RAs administered the questionnaires to participants to collect sociodemographic data (age, marital status, nature of work); complete health history; history of chronic diseases such as diabetes mellitus and hypertension; family history of UFs; lifestyle history (exercise, alcohol use and smoking) and clinical presentation data. All participants underwent a pelvic ultrasound scan to confirm UF diagnoses. Each participant was assigned a unique study identification number to link the questionnaire to the electronic database and prevent data errors. Data were then entered into Epi Info software (7.2.2.5) and stored in a central location. Access to data was restricted to the data manager and the principal investigator to ensure privacy, safety and confidentiality.

Before data collection, ethical approval was obtained from the University of KwaZulu-Natal Biomedical Research Ethics Committee (BREC) (BREC/00002571/2021) and the Eswatini Health and Human Research Review Board (EHHRRB023/2021). Permission was also obtained from the study site gatekeepers. After eligible participants were identified, the research aims and objectives were explained to them individually with assistance from the RAs and registered nurses. Potential participants were given time to read and understand the consent form. They were encouraged to ask questions, and both written and verbal consent were obtained. Participation in the research study was completely voluntary. Each participant was assigned a unique number to guarantee their anonymity.

Table 1 summarises the characteristics of the study participants. A total of 645 women participated in the study from August 2021 to December 2022. The mean age (± standard deviation) was 40.4 years (± 8.9); weight, 78.9 kg (± 13.6); height, 2.10 cm (± 8.9) and body mass index (BMI), 30.9 (± 5.13). Among the study participants, 20.3% were aged 35–39 years, 36.4% were from the Manzini region, and 57.8% resided in rural areas. Approximately 47.9% of the participants were employed, and 47% had tertiary-level education, which is not representative of educational trends in Eswatini. This could be influenced by the purposive sampling method used in the study and other socioeconomic factors. More than half of the participants were married (56.3%), and 76.9% had experienced at least one pregnancy. About 90% of participants menstruated before the age of 15, 35% used the Depo Provera (DMPA) contraceptive method, 51.6% were classified as obese and 36.6% were categorised as overweight. Some participants had chronic diseases; 22.5% were hypertensive, and 21.9% had a positive human immunodeficiency virus (HIV) status.

Table 2 shows the prevalence of symptomatic UFs in Eswatini. The hospital-based, symptom-defined prevalence of UFs is calculated as the number of participants with confirmed UFs at a given time divided by the total number of participants in the sample.

The overall prevalence of UFs was 67.8% (95% CI: 61.5–74.4). The prevalence among the women with symptomatic UFs increased with age: from 33.3% (95% CI: 21.9–48.5) among women aged 25–29 years to 53.8% (95% CI: 39.9–70.9) for those aged 30–34 years, 71.8% (95% CI: 57.9–87.8) for those aged 35–39 years and 79.5% (95% CI: 64.8–96.6) for those aged 40–44 years, respectively (Figure 1). The prevalence then declined for those aged 45–49 years to 73.3% (95% CI: 57.9–91.7) and thereafter increased to 81.4% (95% CI: 63.4–100.3) for those aged 50 years. The prevalence was significantly higher among women aged 50 years and older than among women aged 25–29 years, as shown in Figure 1.

Those with a primary level of education had a higher observed prevalence of UFs, 79.2% (95% CI: 59.9–102.6), than those with secondary education, 74.6% (95% CI: 60.7–90.8) (Table 2). The observed prevalence of UFs for participants residing in the Hhohho (76.3%, 95% CI 64.3–89.9) and Manzini regions (69.8%, 95% CI: 59.5–81.3) was higher than that of those who reside in Lubombo (57.0%, 95% CI: 44.0–72.7) and Shiselweni (60.0%, 95% CI: 46.4–76.3; Table 2). The observed prevalence of UFs was high across all four chronic diseases among the women with confirmed UF diagnoses, ranging from 75.2% to 88.5% (hypertension: 75.2%, 95% CI: 61.7–90.7; diabetes mellitus: 88.5%, 95% CI: 56.1–132.7 and positive HIV status: 71.6%, 95% CI: 58.3–87.0), as shown in Table 2. The prevalence was higher among participants who were not currently using any form of contraception (70%, 95% CI: 62.8–77.5) than among those who were currently using contraceptives (59.7%, 95% CI: 43.7–74.3). The prevalence was noticeably higher among participants who used Depo-Provera (75.7%, 95% CI: 64.7–87.9) than among those using other contraceptive methods (combined oral contraceptives: 73.1%, 95% CI: 57.6–91.5; Norplant: 44.3%, 95% CI: 30.1–62.9).

There was a higher prevalence among participants who had been pregnant in their lifetime at 70.0% (95% CI: 62.8–77.7). According to the marital status of the women diagnosed with UFs, the prevalence was higher among widowed and married participants: 80.0% (95% CI: 62.8–77.7) and 71.6% (95% CI: 63.1–80.9), respectively. The observed prevalence of UFs was higher among obese and overweight participants, 76.3% (95% CI: 67.2–86.3) and 61.9% (95% CI: 52.2–72.8), respectively.

Table 3 summarises the risk factors associated with UFs. All covariates were included in the multivariate analysis: those that were significant and those that lost significance in the univariate analysis. However, all variables that lost significance in the univariate analysis have been removed, and only significant socio-demographic variables have been presented in the table. The research methodology used in this study emphasised including numerous risk-related factors, utilised a large sample size and employed logistic regression analysis to account for multiple confounders.⁴⁸ Age, educational status, ever being pregnant and BMI were associated with UFs in univariate analyses (Table 3). Based on the univariate analysis, ages 30–34 years (OR: 2.33; 95% CI: 1.26–4.35, p < 0.007), 35–39 years (OR: 5.08; 95% CI: 2.82–9.36, p < 0.001), 40–44 years (OR: 7.77; 95% CI: 4.19–14.9, p < 0.001), 45–49 years (OR: 5.50; 95% CI: 2.96–10.5, p < 0.001), 50+ (OR: 8.80; 95% CI: 4.58–17.6, p < 0.001), primary level of education (OR: 2.43; 95% CI: 1.27–4.84, p = 0.009), secondary level of education (OR: 1.88; 95% CI: 1.12–3.18, p = 0.017), ever been pregnant (OR: 1.53; 95% CI: 1.04–2.23, p = 0.029) and BMI (obesity (OR: 3.39, 95% CI: 2.02–5.70, p < 0.001), overweight (OR: 1.71; 95% CI: 1.02–2.89, p = 0.044) were statistically significantly associated with UFs (Table 3). However, alcohol consumption (OR: 0.98; 95% CI: 0.67–1.45, p = 0.937), exercise (OR: 0.81; 95% CI: 0.56–1.16, p = 0.238), sexually transmitted infections (gonorrhoea: OR: 0.72, 95% CI: 0.27–1.73, p = 0.476 and syphilis: OR: 0.82, 95% CI: 0.30–2.10, p = 0.686), Manzini region (OR: 0.72; 95% CI: 0.46–1.11, p = 0.135), marital status (single: OR: 0.87, 95% CI: 0.31–2.82, p = 0.803), combined oral contraceptives pills (OR: 0.90; 95% CI: 0.04–7.40, p = 0.932), Norplant (OR: 0.26; 95% CI: 0.01–2.18, p = 0.260), non-hormonal contraceptive methods (OR: 0.57; 95% CI: 0.03–467, p = 0.633), hypertension (OR: 0.67; 95% CI: 0.19–1.95, p = 0.498), HIV (OR: 0.56; 95% CI: 0.15–1.62, p = 0.322), tertiary level of education (OR:1.19; 95% CI: 0.78–1.80, p = 0.422), diabetes mellitus (OR: 1.70; 95% CI: 0.33–9.59, p = 0.519), being widowed (OR: 1.58; 95% CI:0.79–3.46, p = 0.26), urban residence (OR: 1.15; 95% CI: 0.82–1.61, p = 0.421), self-employed (OR: 1.31; 95% CI: 0,77–2.28, p = 0.328), employed (OR: 1.26; 95% CI: 0.88–1.79, p = 0.208) and age of menarche (OR: 1.35; 95% CI: 0.77–2.48, p = 0.307) (Table 3) were inversely associated with UFs and not statistically significant.

After the multivariable adjustment, ages 35–39 years (aOR: 4.42; 95% CI: 2.16–9.28, p < 0.001), 40–44 years (aOR: 7.86; 95% CI: 3.52–18.2, p < 0.001), 45–49 years (aOR: 6.14; 95% CI: 2.55–15.2, p < 0.001) and ≥ 50 years (aOR: 11.3; 95% CI: 4.06–32.8, p < 0.001) remained significantly associated with UFs while being from Lubombo and Shiselweni indicated a protective effect (OR = 0.28; 95% CI: 0.15–0.53; p < 0.001 and OR = 0.41; 95% CI: 0.21–0.79; p < 0.007, respectively). Risk factors such as alcohol consumption, caffeine consumption, cigarette smoking, exercise, marital status, region, residence, employment status, level of education, ever being pregnant, age of menarche, family history of UFs, sexually transmitted infections, chronic diseases and contraceptive use methods (Table 3) were inversely but not significantly associated with UFs.

Our study’s key strength was the use of primary data to determine study outcomes. This is the first study looking at the burden of UFs in the country; other researchers may generate hypotheses from the identified relationships for future research. Our study was able to provide local data to inform policymakers and strengthen SRH policy/guidelines to incorporate UFs into the Sexual Reproductive Health (SRH) programme.

The research was hospital-based in gynaecology departments; thus, some participants could have been missed in other healthcare settings or those who did not come to the hospital. The inclusion criteria focused on participants with signs and symptoms, with or without a confirmed UFs diagnosis, yet UFs are asymptomatic at times, so some could have been missed. This study also used self-reported data, which can lead to over- or underestimation of the prevalence. The methodology used to select covariates in the multivariate analysis has the potential to introduce bias. Hence, we are proposing future longitudinal and clinical studies, as this was the first study in the country on the prevalence and risk factors associated with UFs. The study was done during the coronavirus disease 2019 (COVID-19) pandemic. The restrictions on the movement of people and hospital admissions criteria may have biased the estimation of the UFs’ prevalence. This bias is nearly impossible to quantify and account for in the estimate.

This study indicated a high prevalence of UFs in the country. The findings indicated high prevalence among participants with a higher BMI, with chronic diseases such as hypertension, diabetes mellitus and HIV/acquired immunodeficiency syndrome, widows and use of certain contraceptives such as Depo-Provera.

In Eswatini’s health facilities, there are currently no standard guidelines for managing UFs. We therefore recommend that guidelines be formulated for the care and management of the disease. These guidelines should incorporate educational programmes on weight management strategies introduced in gynaecological clinics, communities and hospitals to reduce the risk of UFs. Additionally, they include programmes for the screening and monitoring of chronic diseases (blood pressure treatment, diabetes mellitus, HIV/AIDS) and care among women attending gynaecological clinics.

Further investigations are required to broadly explain the complex interplay of factors influencing fibroid formation and the precise relationship with marital status. Consequently, we advocate for forthcoming longitudinal studies that will incorporate the age at which contraceptives are initiated and the duration of their usage to elicit excellent results.