Factors associated with congenital anomalies in Addis Ababa and the Amhara Region, Ethiopia: a case-control study

The study was carried out in Addis Ababa and the Amhara region, Ethiopia.

Addis Ababa, the capital of Ethiopia, has 10 sub-cities with an estimated total population of 3,273,001(47.4% male, 52.6% female). The population of Addis is diverse by ethnicity. The majority of the people are Christians, followed by Muslims and some followers of other religions [25]. The city is subject to population increase through continuous migrations [25]. All of the hospitals in the city provide inpatient, outpatient, and delivery services, except Cure Ethiopia International Children’s Hospital, which does not provide delivery services.

The Amhara Region, the second largest northwest region of the country has a population of 20,399,004 (50.1% male, 49.9% female). The majority of the people live in rural areas and the minority in large and small towns. Most of the Amhara people are Christians, with a good number of Muslims and some followers of other faiths [25]. All of the hospitals in the regional state provide various inpatient and outpatient services including delivery.

Five public hospitals in Addis Ababa and three in the Amhara Region were purposively selected for this study on the basis of case load. All of the hospitals are referral with high patient flow and various specialized departments or units, which provide various services to neonates, infants, children, and adults.

A case-control design was used to investigate exposure status to risk factors for CAs in the case and control children.

A sample of 414 children (207 cases, 207 controls) was recruited for this study. The sample size was calculated by using a ratio of 1:1, power of 90%, a 5% significance level and 38.81% of likelihood of an event in controls (i.e. exposure in controls) [17], and assuming the Odds ratio to be 2. The study children were matched by age. Cases were neonates, infants, and children 0-11 months of age with external and internal major CAs. Major CAs are defined as anomalies that have partial or complete organ/organ system absence/defects, resulting in functional, health, and cosmetic effects, which require surgical repair and rehabilitations [4]. The cases were selected from live births at the study hospitals, neonate units and under five clinics. The type of CAs were classified as major by using WHO/CDC/ICBDSR manual 2014 [4]. Based on WHO/CDC/ICBDSR classification method, the cases with neural tube, orofacial region, musculoskeletal system and genitourinary system defects and less than 12 months of age were purposively selected and included in the study. Children aged 12 months and above and had major and minor CA were excluded. All cases had isolated CAs except one child who had multiple anomalies. The diagnoses of the cases were confirmed through physical examinations and echocardiography (i.e. for cases with heart defects) by pediatricians. The cases were selected purposively. The controls were neonates, infants, and children 0-11 months of age without external and internal CAs. The controls were randomly selected in the same hospitals where children with CAs were identified. All children aged 0-11 months constituted the source population.

The data were collected by a face to face interview of the children’s mothers/caregivers by using a structured questionnaire and checklists. The questionnaire prepared in English and translated to Amharic was pretested at other hospitals and amended according to the results. Data were collected on residence, sociodemographic characteristics, reproductive/pregnancy history and adverse outcomes, family history of CAs, socio-economic status, parental ethnicity, parental occupation, maternal educational level, maternal nutritional status/history, parental exposures to risk factors and/or teratogenic agents, such as radiation, chemicals, pesticides, etc. for CAs, maternal illness/disease, medication/illicit drug use, substance abuse (alcohol consumption, smoking), blood relationships among parents, antenatal care visits, folic acid and multivitamin use, and birth order of children. The data were gathered at delivery rooms/wards, neonatal units, and under five clinics. Prior to data collection, data collectors (nurses, midwifes) were trained for 2 days. The data were collected from January 1- June 30, 2015.

Data were checked on daily basis by collectors and further overseen by the principal investigator and supervisors for completeness and correctness during collection periods. It was also rechecked during data entry. The data were coded and entered into Epi-Info version 3.5.1 and then transferred to SPSS version 21 for analysis. Data cleaning, error checking, and analysis were conducted by using SPSS version 21. Data were calculated by using frequency, cross tabulation, and binary logistic regression. Significance level was considered at less than 0.05 p-value. Crude Odds ratio and confidence level were calculated. After the crude Odds ratios were identified, adjusted Odds ratios were computed to assess risk factors associated with the occurrence of CAs by using backward stepwise logistic regression. Exposure variables with P-value ≤0.2 in the crude Odds ratio were entered into the multivariable logistic regression model to observe their relationships and CAs (i.e. to observe associated variables with CAs in the adjusted Odds ratio). The characteristics of cases and controls were explained in percentages and numbers. The crude and adjusted Odds ratios, confidence intervals, p-values and the overall findings are presented in the result section in the form of texts and tables.

Ethical clearance was obtained from Addis Ababa University, College of Health Sciences Institutional Review Board, the National Research Ethics Review Committee, Addis Ababa City Administration Health Bureau Ethical Clearance Committee, and the Amhara National Regional State Health Bureau Regional Health Research Laboratory Center. Supportive letters were written to all zonal health departments and study hospitals by the health bureaus. Ethical and supportive letters were submitted to all hospital administrators. The purpose of the study was explained to the participant children’s mothers/caregivers. Data collection began after permissions were obtained from hospital managers/medical directors and after a written consent was obtained from children’s mothers/caretakers. Data collected from the participants were kept in secured and locked cabinets in order to maintain confidentiality.

The sample consisted of 414 (207 cases, 207 controls) children. The age of children ranged from 1 day to 11 months. About 87.4% of the children were below 6 months and 12.6% between 6 and 11 months. The majority (59.9%) of the children were male and 40.1% female with a sex ratio of 1.49, respectively. The mean age of the children’s mothers was 26 years (range 16-45). Among the participants, 63.3% were urban and 36.7% rural dwellers. In terms of income, 47.8% had middle and 37.2% low economic status as stated by the mothers/respondents. The majority (93.5%) of the children’s mothers were married; about 37.4% had no formal education; 31.9% had primary education, and 30.7% secondary and higher level of education. By religious affiliation, 72.2, 27.3, and 0.5% of the parents were Christian, Muslim, and others, respectively. Among the mothers, 53% were housewives, 17.9% government, private, or self-employed, 14.3% farmers, and 14% jobless. Similarly, 49.3, 38.9, and 11.8% of the children’s fathers were employees, farmers, and jobless, respectively. The major (62.1%) ethnic group in this study were Amhara, 23.2% Oromo, 5.6% Guragie, 3.6% Tigre, and 3.6% Seltie. The other socio-demographic characteristics of the participants are shown in Table 1.

About 12 (5.8%) and 13 (6.3%) of mothers whose children had and had no CAs, respectively, were in the 15-19 years of age group. Besides, 49 (23%) and 68 (32.9%) of mothers whose children were exposed and not exposed to CAs, respectively, were 20-24 years of age. Similarly, 69 (33.3%) and 68 (32.9%) of the mothers who had children with and without CAs, respectively, were in the 25–29 years of age group, whereas 49 (23.7%) and 31 (15%) mothers who had and free children, respectively, were in the 30–34 years of age group. Likewise, 28 (13.5%) and 27 (13%) of the mothers who had affected and non-affected children, respectively, were 35–39 years of age.

About 99.5% of the CAs identified were isolated/single, while 1 (0.5%) was multiple. The frequency of CAs by organ/organ system is shown in Table 2.

As far as the birth order of children was concerned, 32.9, 19.8, and 47.3% were first, second, third and above (3+) children to their families, respectively. Among the controls, about 48.3, 22.7, and 29% were first, second, third and above (3+) children, respectively, to their parents. About 16.4% of mothers whose children had CAs and 9.2% of the mothers who had children without CAs, had no antenatal care (ANC) visits. The total number of pregnancy history, preterm, miscarriage, stillbirth, gravidity, gestational age at first ANC, number of live births, and neonate/infant/child deaths are presented in Table 3.

There were maternal/medical illnesses (i.e. infections or chronic diseases such as diseases with fever, chronic disorders) in 24 (11.6%) of the mothers who had children with CAs and 13 (6.3%) of mothers whose children were not affected with CAs. Passive smokers were observed in 37 (17.9%) and 28 (13.5%) of mothers who had children with and without CAs, respectively. Among the participants, unidentified medication use in early pregnancy by mothers whose children had and had no CAs were 16.9 and 3.9%, respectively. Contraceptive (both pill and injection) use around conception was 11.6% in the mothers who had children with CAs and 2.4% in the mothers who had no children with CAs. In addition, anemia was observed in 55 (26.6%) of the affected children’s mothers and 37 (17.9%) of mothers whose children had no CAs (Table 4).

Blood relationships among parents were reported in 4.8% of children’s the exposed mothers and 1.0% of not exposed children’s mothers. Family history of CA as seen in the children was also reported in 4.3% of the affected children’s parents and 1.4% of the parents whose children had no CAs. Exposure to chemicals among mothers who had children with and without CAs was 6.3 and 0.5%, respectively. On the other hand, two (0.97%) of mothers whose children had CAs used fertility enhancing drugs. Again, one (0.5%) of the affected child mother had diabetes during pregnancy. On the contrary, mothers who had children without CAs reported no use of fertility enhancing drugs around conception and had no diabetes during and before pregnancy. In addition, none of the affected children’s mothers took multivitamins before they got pregnant, while one (0.5%) of the mother who had child without CA used multivitamin before pregnancy. Moreover, about 25 (12.1%) of mothers whose children had CAs and 17 (8.2%) of the mothers who had children without CAs, had no food shortage before and during pregnancy (Table 4). However, the above-mentioned factors were not statistically significant in the multivariable analysis, the adjusted Odds Ratio. In addition, maternal age, parental ethnicity, mothers educational level, parental occupation (Table 1) and parity, birth order, preterm, still birth, and miscarriage (Table 3) were not statistically significant in the crude and adjusted Odds Ratio analysis.

Thirteen variables were entered into the COR analysis. Of these, unidentified medication used during the first 3 months of gestation (COR = 5.062; 95% CI: 2.286-11.206, P-value = < 0.001), alcohol consumption during early/throughout pregnancy (COR = 3.256; 95% CI: 1.823-5.816, P-value = < 0.001), anemia during early pregnancy (COR = 1.663; 95% CI: 1.038-2.662, p-value = 0.034), contraceptive use around conception (COR = 5.298; 95% CI: 1.980- 14.175, P-value = 0.001), and chemical exposure during early pregnancy (COR = 13.804; 95% CI; 1.789-106.522, P-value = 0.012), and blood relation among parents (COR = 5.203; 95% CI:1.126-24.046, P-value = 0.035) were associated with CAs in the crude Odds ratio analysis. On the other hand, folic acid/iron folate (COR = 0.082; 95% CI: 0.019-.354, P-value = 0.001) and vegetable servings/use during early pregnancy (COR = 0.536; 95% CI: 0.344-0.835, P-value = 0.006) was protective against the occurrence of CAs (Table 4). However, vegetable serving was not associated with the occurrence of CAs in the adjusted Odds ratio analysis.

Twelve variables (alcohol drinking, anemia, iron folate/folic acid, unidentified medication, contraceptive, exposure to chemicals, vegetable serving, blood relationship between child parents, exposure to radiation, maternal illness, family history, and food shortage during early pregnancy) whose value ended up with 0.2 and below in the crude Odds ratio were entered into the multivariable logistic regression model to observe exposure variables associated with CAs. As indicated in Table 4, unidentified medication use (AOR = 4.595; 95% CI: 1868-11.301, P-value = 0.001), maternal alcohol drinking in early pregnancy (AOR = 2.394; 95% CI: 1.212-4.726, P-value = 0.012), and exposure to chemicals during early pregnancy (AOR = 9.964; 95% CI: 1.238-80.193, P-value = 0.031) were significantly associated with the occurrence of CAs. On the contrary, iron folate/folic acid supplementation (use) before and during early pregnancy (AOR = 0.051; 95% CI: 0.010-0.260, P-value = < 0.001) had a protective effect against CA.