Factors associated with undernutrition among children with sickle cell disease attending the sickle cell clinic in Mulago National Referral Hospital, Uganda

Background Sickle cell disease (SCD) is among the neglected non-communicable diseases, which significantly contributes to early childhood mortality. In Uganda, over 20,000 children are estimated to be sicklers. Undernutrition is common among children with SCD and contributes to increased morbidity and mortality. There is paucity of data on prevalence of undernutrition and associated factors in Uganda. Objective To assess the extent of undernutrition and related factors among children aged 5-12 years with SCD attending the sickle cell clinic at Mulago hospital, Uganda. Methods A total of 263 children with SCD attending the sickle cell clinic at Mulago National Referral hospital were recruited consecutively between May and June 2017. The nutritional status of the children was assessed by weight-for-age, BMI-for-age, and height-for-age z-scores calculated using STATA in accordance with WHO 2007 growth standards. Binomial regression was conducted to assess the predictors of undernutrition. Results About 20.2%, 11.4%, and 13.7% of the children were underweight, wasted and stunted respectively. Wasting was significantly associated with older age (10-12 years) (AOR=4.20, CI=2.18-8.10) and living in a female headed household (AOR=0.43, CI=0.19-0.99). Stunting was significantly associated with older age (10-12 years) (AOR=2.90, CI=1.39-6.06). Underweight was significantly associated with older age (10-12 years) (AOR=2.23, CI=1.05-5.16). Conclusion Underweight, wasting and stunting were prevalent among children with SCD attending Mulago hospital. The factors associated with undernutrition were older age and living in a female headed household.


Abstract
Background Sickle cell disease (SCD) is among the neglected non-communicable diseases, which significantly contributes to early childhood mortality. In Uganda, over 20,000 children are estimated to be sicklers. Undernutrition is common among children with SCD and contributes to increased morbidity and mortality. There is paucity of data on prevalence of undernutrition and associated factors in Uganda. Objective To assess the extent of undernutrition and related factors among children aged 5-12 years with SCD attending the sickle cell clinic at Mulago hospital, Uganda. Methods A total of 263 children with SCD attending the sickle cell clinic at Mulago National Referral hospital were recruited consecutively between May and June 2017. The nutritional status of the children was assessed by weight-for-age, BMI-for-age, and height-for-age z-scores calculated using STATA in accordance with WHO 2007 growth standards. Binomial regression was conducted to assess the predictors of undernutrition. Results About 20.2%, 11.4%, and 13.7% of the children were underweight, wasted and stunted respectively. Wasting was significantly associated with older age (10-12 years) (AOR=4.20, CI=2. 18-8.10) and living in a female headed household (AOR=0.43, CI=0.19-0.99). Stunting was significantly associated with older age (10-12 years) (AOR=2.90, CI=1.39-6.06). Underweight was significantly associated with older age (10-12 years) (AOR=2.23, CI=1.05-5. 16).
Conclusion Underweight, wasting and stunting were prevalent among children with SCD attending Mulago hospital. The factors associated with undernutrition were older age and living in a female headed household.

Background
Sickle cell disease (SCD) remains a major neglected tropical disease in Africa (1). The disease is a genetic disorder that affects the shape of haemoglobin in the red blood cell leading to formation of a sickle shape (2). The sickle shape contributes to co-morbidities through the life time including pulmonary hypertension, stroke, organ damage, gall bladder disease and premature mortality (3). Globally, 250,000 out of 300,000 infants born with haemoglobin disorders every year have SCD(1). Furthermore, an estimated 240,000 children in low and middle income countries are born with SCD annually of which 50%-80% die before the age of five years (2). In Uganda, about 20,000 children are born with SCD every year(4). This could be an under estimate since majority of the children are not tested in most districts in the country (5).
Several nutrient deficiencies have been reported among children with SCD and this greatly affects their quality of life (6).Undernutrition(stunting, underweight and wasting) are a common occurrence among children with SCD in Africa (6)(7)(8).
SCD has been known to cause undernutrition by increased metabolic demands, chronic and acute vaso-occlusions, poor appetite and compromised absorption (8).Other factors that influence undernutrition among children with SCD include: being a male, older age, household income, duration of exclusive breastfeeding, average weight during childhood, average number of hospitalization, anaemia severity, geographical location (urban vs. rural),history of hand and foot syndrome, more than two sickle cell crises a year, medical history of severe infections and presence of hepatomegaly (7).
Previous studies on undernutrition in children with SCD have focused on the prevalence of the condition however data on the factors associated in developing countries is scarce (7,(9)(10)(11). The objective of this study was to assess the prevalence and factors associated with undernutrition among children with sickle cell disease attending the sickle cell clinic in Mulago National Referal Hospital (MNRH), Uganda.

Methods
This was a hospital based cross-sectional study conducted among 270 children with SCD who were attending the sickle cell clinic in MNRH from May to June 2017. MNRH is a tertiary hospital and serves as a national center for managing cases with SCD. On average it manages 800 sickle cell patients from across the country each month. We consecutively sampled confirmed cases of SCD aged 5-12 years attending the outpatient clinic whose caretakers provided consent to participate in the study. Recruitment of study participants was conducted as they waited for their monthly routine medical care. Very ill eligible children who needed hospitalization were excluded.

Sample size
The sample size was determined using the formula for single proportion by Kish Leslie,
A structured interview administered questionnaire was used to collect child weight, height, socio-demographic, co-morbidities, food intake, and household characteristics information from caretakers.
Weight of the children was measured in kilograms using a calibrated Seca 874 electronic weighing scale (9).The children were requested to take off their shoes and heavy garments and asked to stand in the middle of the platform to evenly distribute the weight to both feet(10). The weight was recorded to the nearest 0.1kg.
Height was measured in centimeters using a stadiometer. Children were asked to remove their shoes, hair ornaments and stand erect with the back to the stadiometer. With the back of the head, upper back, buttocks, back of the legs and heels in contact with the board, relaxed arms and the eyes looking horizontally, the head piece of the stadiometer was lowered so that the hair (if present) was pressed flat, height was recorded to the nearest 0.1cm.

Study variables
The dependent variable in this study was under nutrition (wasting, underweight and stunting) measured by <-2 Z-scores of BMI-for-age, weight-for-age and height-for-age anthropmetric indices respectively.
The independent variables were categorized into child characteristics and caretaker / household characteristics.
The child characteristics included: age (5-9 years and 10-12 years), sex (male, female), birth order (1 st , 2 rd -3 rd , 4 th and above), morbidities experienced in the past two weeks before the study (Acute respiratory infection, fever and diarrhea) and food intake. Food intake was assessed using the validated dietary diversity score which was based on the 10 food groups consumed in the past twenty-four hours prior to the study (11). For each of the food groups consumed, a child was awarded a score of one or zero if the food group was consumed or not consumed respectively. A dietary diversity score (DDS) cut-off was constructed to describe the dietary diversity with a sum of 0-3 as inadequate DDS and 4-10 as adequate DDS score (11).  9) small farm animals like poultry, 10) a manufactured bed, and 11) the nature of the house (12). Household wealth index was ranked into two quintiles from the lowest to the highest (1=Low and 2=Highest).

Data analyses
Data was exported from excel to stata14.0 for analysis (Statacorp, 2015). Socio demographic data was summarized using descriptive statistics like proportions and frequencies for categorical variables and mean and standard deviations for continuous variables as appropriate. Anthropometric measurements of height and weight were used to determine undernutrition (9). Height and age were converted from centimeters to meters and years to months respectively. Body Mass Index (BMI) was calculated then the BMI for age z-scores, weight for age z-scores and height for age z-scores were computed using the WHO reference growth standards. Normal nutritional status was defined as zscores ≥-2 and undernutrition (underweight, wasting and stunting) was defined as z-sores <-2. The prevalence of underweight stunting and wasting was expressed in percentages.
In order to ascertain the factors that were associated with undernutrition we started with univariable and bivariable analysis. The factors that had a potential for being considered in the multivariable model were variables found to be statistically significant at bivariable level, factors with a P-value of < 0.2 and the insignificant variables which were deemed to be important in influencing undernutrition according to literature. We later performed multivariable analysis using binary logistic regression approach controlling for factors identified under bivariable analysis and literature. The strength of the association between variables was determined using odds ratios, P-values (α set at <0.05) and 95% confidence interval. Models were built using the stepwise regression method.
Confounding was checked by observing whether variables included in the model caused a change in the odds ratio of the main exposure by at least 10%. Also confounding was checked by assessing whether variables that were insignificant at bivariable stage became significant at the multivariable level. Variables in a model that were found consistently insignificant and did not add any value in terms of goodness of fit were eliminated. The model goodness of fit was tested using Hosmer-Lemeshow goodness of fit test.

Results
From the total of 270 child care-taker pairs recruited, only 263 child-caretaker pairs were included in the final analysis because some were outliers according to the WHO anthropometric indicator cut-offs. The mean age of the children was eight years (SD±2) and majority 66.5% (175/263) were in the age group 5-9 years (Table 1)
Wasting was significantly associated with age of the child and sex of household head (

Factors associated with stunting of study participants
In the present study stunting was independently associated with age of the child as shown in Table 5. Children aged 10-12 years were more likely to be stunted compared to their counterparts (AOR=2.90, CI=1.39-6.06). Factors associated with underweight of study participants Underweight was associated with age of the child and household wealth index (Table 6).

Discussion
The aim of the study was to assess the factors associated with undernutrition in children with SCD aged 5-12 years attending the sickle cell clinic in MNRH. The findings revealed that older age (10-12 years) and living in a female headed household were significantly associated with undernutrition.
In this study about 20.2%, 11.4%, and 13.7% of the children were underweight, wasted and stunted respectively. This shows that undernutrition was prevalent among children with SCD according to the WHO categorisation (13). Undernutrition has a negative impact on the quality of life of children with SCD (14). Evidence from a cohort study conducted in Tanzania reported an increased risk of hospitalisation with undernutrition among patients with SCD particularly with wasting (15).
A prevalence of 20.2% was reported for underweight in this study. Similar observations were made in a recent study conducted in Accra (16). On the other hand high prevalence's have been reported for example in Tanzania (36.2%) (15). This might be explained by the difference in age groups of the study participants that is five to twelve years in the present study versus 0.5 to 48 years in the Tanzanian study which did not look at the prevalence for the children separately. Wasting was found in 11.4% of the children with SCD. The prevalence of wasting among children with SCD in other studies ranges between 18.4% to 50.3% (15,17,18).Height for age was low in 13.7% of the children with SCD. A similar study conducted in Democratic Republic of Congo among children aged 0-12 years reported stunting at 10.5% (17). This is comparable to the finding of the present study probably because both studies were conducted in the major sickle cell centers in the main cities.

Factors associated with undernutrition
Results for this study showed that older children aged between ten to twelve years were more likely to be underweight, stunted and wasted compared to the younger ones. Among children with SCD delay in growth starts early in childhood but becomes more obvious as the child grows up (19). This is likely attributed to increased demand for nutrients for growth, fighting off the recurrent infections and a lack of appetite(14

Limitations
The study was cross sectional in nature so no casual relationships can be concluded on due to lack of a temporal relationship between the predictors and nutritional status.
Additionally some predictors related to the management of SCD and other comorbidities were not included in this study.

Generalisability
The generalisability of the results in this study is limited to children with SCD attending a health service setting.

Conclusion
In conclusion, undernutrition was prevalent among the children with SCD. Details of the study were explained to the Head of Department and caretaker's written consent was obtained from the participating caretakers. Care takers were interviewed individually for privacy and personal information was kept confidential by using identification numbers.

Consent for publication
Not applicable

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.