Survival Status and Its Determinants among Under-Five Children with Severe Acute Malnutrition Admitted to Inpatient Therapeutic Feeding Centers in South Wollo Zone, Amhara Region, Ethiopia

Background Under nutrition is one of the leading causes of morbidity and mortality in under-five children in developing countries including Ethiopia. In Ethiopia, many children with severe acute malnutrition (SAM) are treated at inpatient therapeutic feeding centers. However, the survival status and its determinants are not well understood. Therefore, the aim of this study was to estimate the survival status and its determinants among under-five children with severe acute malnutrition admitted to inpatient therapeutic feeding centers (ITFCs). Methods A record review was conducted on 414 under-five children who were admitted with severe acute malnutrition to ITFCs in South Wollo Zone, northeast Ethiopia, between September 11, 2014, and January 9, 2016. Data were entered into Epi-Info version 7.2 and analyzed using SPSS version 20. Life table analysis was used to estimate cumulative proportion of survival. The relationship between time to recovery and covariates was determined using Cox-proportional hazards regression model. p < 0.05 was used to declare presence of significant association between recovery time and covariates. Results Of the total children recorded, 75.4% of children were recovered and discharged, 10.3% were defaulters, 3.4% died, 7.4% were nonresponders, and 3.4% were unknown. The mean (±standard deviation) time to recovery was 12 (±5.26) days, whereas the median time to recovery was 11 (interquartile range of 8–15) days. Children's breastfeeding status at admission (AHR: 1.42, 95% CI: 1.10, 1.83) and children without comorbidities at admission (AHR: 1.44, 95% CI: 1.03, 2.00) had statistically significant effect on time to recovery from SAM. Conclusion All treatment responses in this study were within the recommended and acceptable range of global standards. Policy makers, health facilities, and care providers may need to focus on the importance of breastfeeding especially for those under two years of age and give emphasis for cases with comorbidities.


Introduction
Globally, under nutrition remains one of the most common causes of morbidity and mortality among children. Nearly 20 million children below 5 years of age suffer from wasting and are at risk of death or severe impairment of growth and psychological development [1]. Of these, over 90% are found in South and Southeast Asia and sub-Saharan Africa. About one million deaths occur annually among children under five years of age in developing countries [2]. Ethiopia has one of the highest child mortality rates in the world with 57% of all deaths in children who have stunting and wasting as the underlying cause [3]. e percentages of children with stunting, wasting, and underweight at national level were 40, 9, and 26, respectively [4]. e minimum international standard set for management of severe acute malnutrition (SAM) is a recovery rate of at least 75% and death rate less than 10% [5,6].
Studies conducted in Ethiopia revealed that the recovery rates and median survival times of under-five children with SAM were ranged from 83% to 89% and 14 days to 25 days, respectively [7][8][9][10][11]. Previous research studies conducted in the country identified determinants of children's survival. Sex of child, medical comorbidities, treatment and follow-up status, and clinical diagnosis of SAM were associated to survival of children [12][13][14][15][16].
e Ethiopian Ministry of Health established inpatient therapeutic feeding centers in the health facilities to reduce SAM-related mortalities of under-five children. Despite the existence of SAM management at hospital or health center level in every corner of the country, deaths due to SAM is indicated to be still high, and little is known about the recovery time and its determinants from SAM particularly in under-five children admitted to inpatient therapeutic feeding centers (ITFCs) [17].
is study aimed to estimate the survival status and its determinants among under-five children with severe acute malnutrition admitted to inpatient therapeutic feeding centers (ITFCs) in South Wollo Zone, northeast Ethiopia, over the past two and half years.
us, these findings would be essential to provide evidences for decision makers, health facilities, and care providers to take measures to target those children at highest risk of slow recovery.

Study Design, Period, and Setting.
A facility-based retrospective record review was employed to estimate survival status and its determinants among under-five children with severe acute malnutrition admitted to ITFCs in South Wollo Zone, northeast Ethiopia, from September 11, 2014, to January 9, 2016. e study was conducted in South Wollo Zone, which is bordered on the south by North Shewa, on the west by East Gojjam, on the northwest by South Gondar, on the north by North Wollo, on the northeast by Afar Region, and on the east by the Oromia Zone and Argobba special district. It is one of the 11 Zones found in Amhara National Regional State. e zone has twenty-two districts. e zone has a total population of 2,518,862, of whom 1,248,698 are men and 1,270,164 women. It has a total area of 17,067.45 square kilometers. e population density of the zone is 147.58 per square kilometer. About 12% of the population inhabits urban areas. Children under five years of age constitute 352,642 of the population [18]. According to the South Wollo Zone Health Department report, the zone has seven governmental hospitals, one hundred thirty-five health centers, and four hundred ninety-six health posts. Most of these people are seasonal agriculturalists and prone to recurrent drought and food insecurity. e common health problems in this zone are diarrhea, under-five pneumonia, other communicable diseases, and malnutrition [19].

Study Participants.
Under-five children in the study area were screened for signs of SAM and diagnosed based on anthropometric measurements and examination of the feet for bilateral pitting edema. ey were admitted to ITFCs after fulfilling the criteria of admission. ey were admitted if the W/H was <70% of the median WHO growth reference, or if the mid upper arm circumference (MUAC) was less than 11 cm, or children with bilateral pedal edema [20]. On admission, malnourished patients were assessed for hydration status, anemia, and signs of infections. ey were given oral dose of vitamin A, mebendazole, folic acid, and a course of amoxicillin for five days. Rehydration solution for malnutrition (ReSoMal) was used for treating dehydrated cases, and drugs such as gentamicin, chloramphenicol, or quinine were used based on causes of infections. Treatment of severe malnutrition was divided into three phases: the first phase (phase I), transition phase, and phase II. In phase I, health workers resuscitated patients, treated for infections, restored electrolyte balance, and prevented hypoglycemia and hypothermia on indication. F75 milk (formula 75 that contains 75 kcal in 100 ml, minerals, and proteins) was used during phase I treatment; malnourished cases who responded on treatment by return of appetite, beginning of loss of edema, and no intravenous line or nasogastric tubes were transferred to transition phase to receive F100 (formula 100 that contains 100 kcal in 100 ml). Afterwards, cases were transferred to phase II after they gained good appetite (finish 90% of F100 prescribed for transition phase) and cleared edema. e discharge criterion for children with SAM is W/ H ≥ 85% [21].

Sample Size and Sampling Procedures.
Sample size was determined based on the assumptions with recovery rate of 93.3%, hazard ratio of 0.38 [9], 5% margin of error, 95% CI, and power of 80%. us, considering after 10% adding for incomplete records, the total number of study sample size was 414. ere are 46 public health facilities that have ITFCs in South Wollo Zone. Of these, 15 public health facilities were randomly selected and included in the study. After proportional allocation was employed for each health facility, a total sample size of 414 under-five children with SAM were selected for the study using systematic random sampling method ( Figure 1).

Data Analysis.
e data were collected by reviewing records from ITFCs' registration book and individual follow-up chart using standardized and pretested data collection tool. Data were entered into Epi-Info version 7.2 and analyzed using SPSS version 20. e outcome variable in this study was time to recovery (stabilized and discharge to outpatient therapeutic centers) following treatment for SAM. Children who defaulted from treatment, died, or became nonresponders were considered to be censored. Finally, the outcome of each subject was dichotomized into censored or survived/recovered.
For comparison, demographic characteristics were described in terms of mean (standard deviation) and median (interquartile range) for continuous data and frequency distribution for categorical data. Determinants were assessed by bivariate and multivariable screening models. Variables in the bivariate analysis of sociodemographic and admission information, medical comorbidities, treatment given, and follow-up with respect to survival status, which were found at p value <0.20, were further considered in to multivariable Cox-regression model. e enter method of selection for final model was used. Crude and adjusted hazard ratios with their 95% confidence interval (CI) were estimated, and p < 0.05 was used to declare presence of significant association between time to recovery and covariates. Life table analysis was used to estimate cumulative proportion of survival among children with SAM at different time points. Kaplan-Meier test was used to show nutritional survival time after initiation of inpatient treatment, and log rank survival curve was used to compare median survival time between groups.

Treatment Response and Length of Stay for Recovery.
Of the total 406 children admitted to ITFCs, the recovery rate from SAM was 306 (75.4%), whereas the remaining 14 (3.4%) children died, 42 (10.3%) defaulted, 30 (7.4%) were nonresponders, and 14 (3.4%) were unknown. e overall average (±SD) length of stay for the edematous children with SAM was 12.5 (±5.6) days, and children with SAM that stayed almost similar in the ITFCs had severe wasting (12.0 (±5.1)). e median (IQR) length of stay for children who were diagnosed with edematous and severely wasted was 12.5 (8, 16) and 11.0 (8, 15) days, respectively.

Determinants of Nutritional Recovery Time.
After minimizing the confounders, only two variables were associated with time to recovery. Children who had breastfeeding status and children who had no comorbidities at admission had shown significant association with time to recovery from SAM. Children who had been breastfed at admission were 1.42 times (AHR: 1.42, 95% CI: 1.10, 1.83) more likely to have shorter recovery time as compared to the nonbreastfed at any time. Children with the absence of comorbidities at admission were 1.44 times (AHR: 1.44, 95% CI: 1.03, 2.00) more likely to have shorter recovery time than who had comorbidities at any time (Table 3).

Life Table and Kaplan-Meier Analysis.
e life table showed that cumulative probability of staying in the program was 92% at third day, 72% at sixth day, 47% at ninth day, 30% at twelfth day, and zero percent at 33 days of admission (Table 4). e overall Kaplan-Meier curve ( Figure 2) showed a step function and was used to estimate the survival function from lifetime data. Time to recovery of 286 children with breastfeeding and 120 children without breastfeeding was compared.
e Kaplan-Meier (KM) survival curve for breastfeeding status at admission illustrated that the treatment recovery time of children with breastfeeding was better than that of children without breastfeeding at admission. e significance of the observed differences of the Kaplan-Meier survival curves (times) among the two groups of children was assessed using log rank test. As a result, breastfeeding status of children was found to have statistically significant association (X 2 � 7.995, p < 0.005) (Figure 3).   e time to recovery of 355 children with medical complications and 55 children without complications at admission was compared. e KM survival curve for children without complications at admission illustrated that the treatment outcome of children without complications was better than that of children with comorbidities status at admission. e significance of the observed differences of the    KM survival curves (times) among the two groups of children was assessed using log rank test. As a result, children with no comorbidities were found to have statistically significant association (X 2 � 4.923, p < 0.0271) (Figure 4).

Discussion
Among 414 eligible study subjects, 406 under-five children were considered in this study and eight were missed due to incomplete data. More than 75% of them were recovered from SAM, while the remaining were censored. e survival rate was 75.4%. e mean time-to-cure was 12 (±5.26) days, whereas the median time-to-cure was 11 (IQR: [8][9][10][11][12][13][14][15] days. e mean (±SD) age of children with SAM was 17 (±11) months, which was consistent with similar studies conducted in developing countries [9,19,22]. Majority (85%) of admissions in ITFCs were between 1-24 months of age due to higher risk factors for malnutrition [19]. In the current study, survival (recovery) rate was 75.4%, defaulter was 10.3%, and death rates were 3.4%. Ethiopia is currently using a protocol for the management of severe acute malnutrition that is an update of existing guideline. As mentioned in this protocol, the minimum acceptable reference values which have been developed by Sphere project are >75% recovery, <15% defaulter, and <10% death rates [23]. us, our finding was within the acceptable range of this protocol. is is also consistent to other studies in the country [8,10]. is might be due to proper use of clinical management protocols which is capable of reducing case fatality rates [19]. In this study, the clinical classifications of SAM reported as overall proportion of recovery were 76.1% with marasmus, 14.7% with kwashiorkor, and 9.2% with marasmic-kwashiorkor. is finding is in line with other studies conducted in north Ethiopia [24], whereas it is dissimilar with other study conducted outside the country [11]. It also contrasted with the national report [4]. is might be due to differences in feeding culture and socioeconomic status.
is study indicated that the mean time taken to recover from SAM for both children diagnosed with edematous (12.5 days) and severe wasting (12.03 days) was almost similar. It was found to be contrasting with other findings in the country in which those children who were diagnosed with marasmus at admission stayed longer before recovery than their kwashiorkor counterparts [10].
is study revealed that the overall average length of stay in South Wollo Zone in ITFCs was 12.27 days (12.5 days for edematous and 12.03 days for severe wasting children). It is consistent with the minimum international standard set for management of severe acute malnutrition which is an average length of stay less than 30 days [23]. e mean and median time-to-cure from SAM for patients in the zone were 12 (±5.26) days and 11 days (IQR: 8-15 days) with minimum and maximum cure time of 1 and 33 days, respectively. A program is well functioning and considered acceptable if the length of stay in a hospital is less than 28 days. is implies that the program was acceptable as per the national standard [19]. is was consistent with another study conducted in the country [7]. Majority of the children recovered from SAM at first week were 86 (28%), at second week 265 (58.5%), and at third week 299 (11%). is was in line with findings of Jimma University Specialized Hospital [25].
Of the total study subjects, severe wasting (marasmus) was the most common one (75.9%) which was similar with other studies [11,14,26]. Fever, pneumonia, and diarrhea were the most common comorbidities.
is was almost similar with another study conducted in Zambia University Teaching Hospital [27].
In final analysis model (Cox regression), breastfeeding status and absence of medical complication at admission were found to be independent predictors of shorter recovery time. Other covariates did not have an impact. Children with breastfeeding at admission had shorter recovery rate compared to their counterparts at any time. In KM analysis, the two survival curves were significantly different. It implied that breastfeeding may always be the preferred option in most contexts to increase recovery rate of children from severe acute malnutrition. is was in line with the national severe acute malnutrition guidelines [19,23]. Even advice on breastfeeding for infants is a recommended treatment plan for uncomplicated SAM and moderate acute malnutrition [28]. It should also be offered to malnourished children before the diet and always on demand [19]. Children with the absence of comorbidities at admission had shorter recovery rate than who have comorbidities at any time [9]. Comorbidities may prolong the duration of stay in the ITFCs and may increase death rate [26].
is finding was in agreement with other studies conducted in Ethiopia and other developing countries [11,13,14,29]. Since the study was based on secondary data, incomplete registration was observed in some anthropometric measurements at discharge, and even measurement errors had been introduced. Among treatment responses, "unknown cases" were categorized as "censored", which may compromise the children's recovery rate from SAM. us, it may affect the findings through under or over reporting. Further prospective study is recommended to overcome these threats.

Conclusions
Our findings showed that the overall survival/recovery rate, death rate, and time to recovery from SAM were as per the international and national minimum standards set for SAM management. Children with breastfeeding status at admission had a shorter time of recovery from SAM. Children with medical complications at admission had a longer time of recovery from SAM. erefore, enhancing breastfeeding for children up to at least two years of age and early detection of under-five children's morbidities are largely important to improve time of recovery from severe acute malnutrition. Additionally, in the community, breastfeeding promotion should also prevent SAM, and treatment of nonbreastfeeding children with SAM should include breastfeeding support for mothers and relactation where possible.