Prevalence of gestational diabetes mellitus and associated factors among women attending antenatal care at Gondar town public health facilities, Northwest Ethiopia

The study was conducted at selected Gondar town public health facilities. It included one tertiary facility called University of Gondar Comprehensive Specialized Hospital (UOGCSH) and four health centers namely, Azezo, Gondar polyclinic, Woleka, and Maraki. Gondar town is located in the Northwestern part of Ethiopia, 747 km from Addis Ababa (the capital of Ethiopia), and 170 km from Bahirdar (the regional capital). It has 12 sub-cities with 12 urban and 10 rural kebeles (the smallest administrative units). Based on the 2014 population projection and census of the Central Statistical Agency (CSA) of Ethiopia [25], the total population of the town was estimated to be 306, 246 of whom 149, 970 were men and 156, 276 women. The majority (84.2%) of the people were Orthodox Christians, 11.8% Muslims and 1.1% Protestants. The expected number of pregnant women in the town was 11,225, of which at least 8913 were living in urban kebeles in 2017/18. The town had one comprehensive specialized hospital, eight health centers, and more than 15 private clinics.

A cross-sectional study was conducted among a cohort of pregnant women receiving antenatal care at selected public health facilities from March 30, 2018 to January 4, 2019. The required sample size was determined using the single population proportion formula with the following statistical assumptions: a survey conducted by Management Sciences for Health (MSF) reported the prevalence of GDM among urban women in Tigray, Northern Ethiopia as 13% [12], a 95% confidence level, 3% marginal of error, a design effect of 2, and considering 15% of non-response and attrition rates.

Where, n = the sample size, Z = the desired level of the confidence interval, P = proportion of GDM, and d = margin of error. The minimum required sample size for significant result was 965, however we have added 15% by considered the non-response and attrition rates throughout the study period.

Therefore a total of 1110 were recruited for the study. A two-stage sampling technique used to invite pregnant mothers and include them in the cohort. On the first stage, one hospital and four health centers were selected by the simple random sample; on the second stage, pregnant women who fulfilled the inclusion criteria were chosen using the systematic random sampling technique.

During the study, 1110 new antenatal care attendants were invited and screened to participate in the study. Women were enrolled if they were aged 18 years or more with 20–23^+ 6 weeks of gestational age and singleton pregnancy. But pregnant women who had pre-existing diabetes mellitus or overt DM, chronic diseases, medications that may affect glucose metabolism such as steroids, β-adrenergic agonists, anti-psychotic drugs [26, 27] were excluded. On their first visit, participants were asked for written consent for their enrolment in the study.

Screening at first visit was carried out according to the recommendations of the International Diabetes Federation (IDF) [22], ADA [28], and WHO [29] in order to rule out pre-existing diabetes. However, most pregnant women did not fast in their first visits. For subsequent tests, participants were informed to come fasting on their next appointments. Universal screening for GDM using a two-hour 75 g OGTT was performed for all pregnant women at 24–28 weeks of gestational age. Besides, 352(34.3%) women had at least one type of risk factors for GDM (pre pregnancy BMI ≥ 30 Kg/m², MUAC ≥28 cm, age ≥ 35 years, previous macrosomia, glycosuria, history of GDM, family history of diabetes, previous poor pregnancy outcome or developed pregnancy-related complications) were advised to repeat the test at 32–36 weeks even if their OGTT results were negative at regular tests and GDM diagnosis ascertained by the second test.

A structured and pretested questionnaire was developed and interviewer-administered to all study participants The questionnaire was prepared in English and then translated to Amharic (the local and national language) and then translated back to English by language experts to check its consistency (English version questionnaire attached as Additional file 1).

A face to face interview was employed to gather information on residence, age, last normal menstrual period (LNMP), marital status, religion, ethnicity, level of education, occupational status, average monthly income, family history of diabetes, previous history of GDM and birth weight of the previous child, behavioral and lifestyle characteristics, such as current exposure to alcohol, coffee, antenatal depressive symptoms, physical activity, and dietary diversity. Gestational age estimations were based on a reliable LNMP combined with first-trimester ultrasonography (if available). Additionally, socio-demographic information (parity, gravidity, gestational age (early fetal ultrasound result), obstetric history, medical history, hemoglobin) were retrieved from ANC cards.

The short form International Physical Activity Questionnaire (IPAQ) was used to assessed the physical activities that women do as part of their everyday lives [30]. The IPAQ was suitable for adults between 15 and 69 years of age and implemented in different countries. It was designed to assess specific types of activity such as walking, moderate and vigorous intensity activities done at work, as part of house and yard work, to get place to place, and in spare time for recreation, exercise or sport. Women were asked to recall their activities of the last 7 days preceding the interview. Data was reported as metabolic equivalents (MET-minutes per week) using the IPAQ scoring protocol to considered women into high, moderate and low level of physical activity categories [30].

Dietary diversity was assessed using a 24-h food recall method by the Food and Nutrition Technical Assistance (FANTA) 2016 version woman’s minimum dietary diversity measurement tool [31]. It contained a list of ten food groups (starchy staples, nuts and seeds, pulses, dairy, meat, eggs, poultry and fish, dark green leafy vegetables, other vitamin-A rich fruits and vegetables, other vegetables, and other fruits). The minimum dietary diversity score (MDDS) was dichotomized on the basis of whether or not women have consumed the list of defined food groups the previous day or night. The MDDS of five and more was categorized as adequate dietary diversity [31].

A mother was asked, “How often have you drunk coffee since your pregnancy?” If the answer was “daily” or “sometimes in a week” the mother was categorized as exposed to coffee. She was also asked, “How often have you drunk alcohol since your pregnancy?” If the answer was “daily” or “sometimes in a week” the mother was labelled as exposed to alcohol.

The mid-upper arm circumference (MUAC) was measured on the left arm using a non-stretchable measuring tape. As the result of most of the pregnant women could not recall their weight before conception, it was difficult to determine BMI. MUAC was as reliable measure due to quite stable during the course of pregnancy and highly correlated to the BMI before conception [32, 33]. Pregnant women with MUAC of ≥28 cm were considered as having overweight and/or obesity [34].

For measuring blood pressure (BP), the pregnant women were asked to take rest for at least 5 min in sitting positions if they were exerted. Then, the pressure was measured on the right arm using normal cuffs fitted for adults with a standard sphygmomanometer placing the stethoscope bell lightly over the brachial artery. The average systolic blood pressure (SBP) and diastolic blood pressure (DBP) were recorded in mmHg after two readings were taken with an intermission of 5–10 min. Hypertension was considered to be present if the systolic and diastolic blood pressures were greater than or equal to 140 mmHg and 90 mmHg, respectively.

Antenatal depression symptoms were measured by using the Edinburgh Postnatal Depression Scale (EPDS) screening tool developed [35] and validated in urban Ethiopia [36]. The tool was used to measure the feelings a mother had experienced in the past week. The tool contains ten specific questions with four Likert scale response options (most of the time, sometimes, not often, never), scored from 0 to 3 (a higher score indicated more depressive symptoms), which is simple to use, can be scored by simple addition. An EPDS score of 13 and more used like similar studies conducted in Ethiopia and abroad [37‐39] to categorize the presence of antenatal depression.

Pregnant women scheduled for 2 h-75 g OGTT were called and reminded to complete an overnight fast for 8–12 h. Blood glucose level was analyzed using 5 μl capillary whole blood with HemoCue Glucose B-201⁺ (A¨ngelholm AB, Sweden). (The specifications of the HemoCue Glucose B-201⁺ glucometer attached as Additional file 2). Pregnant women were asked to sit and take rest before the first sample was taken using a finger prick at the side of the fingertip with a sterile lancet after cleaning in disinfectant and allowed to dry. A drop of blood was allowed to fill the cuvette after wiping away the first 2 or 3 drops of blood. Wipe off excess blood on the outside of the cuvette tip and place the filled cuvette in the cuvette holder. After 40–240 s the glucose value of the sample was displayed. The necessary precautions were taken during sample collection. Then, they were given a 75 g of glucose dissolved in 250 mL of water which they drank within 5 min under supervision. Capillary blood samples were taken again at 1 h and 2 h. The whole blood capillary values were converted to plasma venous values by multiplying a constant factor of 1.11 [40]. GDM diagnosis was made by using the 2017 American Diabetes Association (ADA) [28] or 2013 WHO [41] or modified International Association of the Diabetes and Pregnancy Study Groups (IADPSG) [42] diagnostic criteria. Similar values have been accepted by the WHO though slightly modified as a range [29]. The diagnosis of GDM is made when one or more of the values was met (fasting: ≥ 92 mg/dL, 1 h: ≥180 mg/dL; 2 h: ≥ 153 mg/dL). Hemoglobin analysis was also carried out by laboratory technologists in their respective health facilities. Pregnant women with hemoglobin concentration below 11 g/dl were considered as having anaemia [43].

All data were entered into Epi Info™ 7 software, then exported to statistical package for the social sciences (SPSS) version 22.0 (IBM, NY, USA) for analysis. Descriptive statistics, like frequency, percentage, and mean, SD, and the range were used for the presentation of variables. Tables and figure were also used for data presentation. Categorical variables were expressed as proportions and chi-square analysis was performed to compare proportions. Student t-test was used to compare means for normally distributed variables. Binary logistic regression model was used to identify factors associated with GDM. Variables with P-value of ≤0.20 in the bivariate analysis were exported to the multivariate analysis to control the possible effect of confounders. Model goodness of fit test was checked by the Hosmer-and Lemeshow test (P-value = 0.953). Multicollinearity was checked by using the variance inflation factor (VIF). Adjusted Odds Ratio (AOR) with a 95% confidence level was estimated to show the strength of association, and a P-Value < 0.05 was used to declare statistical significance in the multivariate analysis.

Out of the 1110 pregnant women invited to participate in the study, 83 mothers (52 did not return for OGTT, 15 did not complete the tests, 9 were diagnosed with overt diabetes, 5 have a medical emergency, and 2 had an abortion before OGTT) were excluded. The characteristics of pregnant women who of included and excluded in the final analysis were similar in terms of key socio demographic variables such as residence (P = 0.835), maternal age (P = 0.0691), parity (P = 0.544), gravidity (P = 0.109), marital status (P = 0.734), MUAC (P = 0.08), employment status (P = 0.056). However women with complete data were on average attended secondary educational level and above and had higher income level than those who were excluded in the analysis.

The remaining 1027 women who participated and completed OGTT with overall response rate of 92.5% were included in the analysis. The mean age of the women was 27.22 (SD ± 5.24) years. Most of the participants (91%) lived in urban areas, 92% were married, and 60.5% unemployed. Five hundred and ninety-eight (58.2%) of women had secondary education and above. The mean of the MUAC was 24.77 (SD ± 3.08) cm, systolic blood pressure was 112.07 (SD ± 11.91) mmHg, and diastolic blood pressure was 71.91 (SD ± 10.22) mmHg. The mean hemoglobin level of the pregnant women with GDM and normal glucose level was 12.39 (SD ± 1.80) g/dl and 12.64 (SD ± 1.74) g/dl, respectively. Nearly half, 504 (49.1%) of the women were nulliparous. Sixty-two (6%) had a family history of DM (Table 1).

Out of total participants, a moderate level of physical activity was reported by 44.8%. Normal glucose level was more common among women having high physical activity level (33.4% vs 17.6%) and adequate dietary diversity (51.5% vs 27.5%) than women with GDM. Concerning exposure to drinking alcohol and coffee, no statistically significant difference was observed among women with normal glucose profile and GDM at P = 0.476 and P = 0.311 respectively. Eighty-nine (8.7%) of the pregnant women had antenatal depression symptoms (Table 2).

Of the total of 588 pregnant women who had prior pregnancy history, 71(12.1%) had macrocosmic babies, 166 (28.2%) history of abortion, 38 (6.5%) history of stillbirth, 42 (7.1%) history of preterm labor, 97(16.5%) history of caesarean delivery, and 54 (9.2%) previous GDM (Table 3).

A total of 131 women were diagnosed as having GDM according to 2017 ADA and IADPSG diagnostic criteria which resulted in a GDM prevalence of 12.8% (95% CI: 10.7–14.8%); 118 (90%) of them were diagnosed at regular OGTT and the rest 13 (10%) at late OGTT. At regular OGTT test (24–28 weeks), the mean fasting, 1 h, and 2-h glucose levels of the pregnant women were 81.17 (SD ± 12.96) mg/dl, 140.69 (SD ±24.55) mg/dl, and 119.10 (SD ± 19.75) mg/dl, respectively. The plasma glucose levels taken at the three time points of the OGTT were significantly different for women with GDM and normal glucose level (p < 0.001). Pregnant women with GDM had the mean fasting, one-hour and two-hour glucose level of 100.81 (SD ±15.88) mg/dl, 163.82 (SD ± 28.42) mg/dl, and 138.76 (SD ±23.40) mg/dl, respectively. Whereas, women with normal glucose profiles had the mean fasting, one-hour and two-hour glucose levels of 78.3 (SD ± 9.55) mg/dl, 137.31 (SD ±22) mg/dl, and 116.22 (SD ±17.4) mg/dl, respectively (Table 1). More than half (56.5%) of the women with GDM were diagnosed on a fasting plasma glucose status only, while 21.4% had high fasting glucose level and one or two extra abnormal levels (Fig. 1).

Results of the unadjusted binary logistic regression showed that advanced maternal age, marital status, level of education, employment status, MUAC ≥28 cm, parity, previous history of GDM, family history of DM, history of having macrocosmic baby, history of abortion, history of stillbirth, SBP, DBP, anemia (Hb < 11 g/dl), level of physical activity, dietary diversity, and antenatal depression were associated with GDM. However, on multivariate logistic regression, variables GDM was independently associated with being overweight and/or obese women (MUAC ≥28 cm) AOR = 2.25; 95% CI: 1.18–4.26), previous history of GDM (AOR = 5.82; 95% CI: 2.57–13.18), family history of diabetes (AOR = 4.03; 95% CI: 1.57–10.35), low physical activity (AOR = 3.36; 95% CI: 1.60–7.04), inadequate dietary diversity (AOR = 1.9; 95% CI: 1.02–3.53), and antenatal depression (AOR = 4.12; 95% CI: 1.85–9.20) were significantly associated with GDM (Table 4).