Malnutrition can be defined as a state of nutrition in which deficiency or excess of energy, protein, and other nutrients causes measurable adverse effects [1]. Although uncommon in developed countries, malnutrition in children remains major public health problem in many developing countries [2]. It continues to be the most important risk factor for growth retardation, illness and death with huge masses of young children particularly affected [3].

Malnutrition can be of the acute, chronic or mixed type [1]. Severe Acute Malnutrition (SAM) is one form of acute malnutrition which refers to weight for height ratio of less than − 3 standard deviations or weight for height ratio of below 70% or Mid Upper Arm Circumference (MUAC) < 110 mm or presence of nutritional edema [4].

Today, 52 million (about 8.3%) children under five in the world are suffering from acute malnutrition. Off those children affected, the majority (> 90%) are found in South and Southeast Asia and sub-Saharan Africa. Despite to communal belief, acute malnutrition (also known as wasting) does not occur only in emergency conditions, it is also common in stable settings in countries like India, Indonesia, Kenya and Ethiopia [5].

Till now, acute malnutrition was also perceived as a condition of humanitarian emergencies rather than a development and public health importance [5]. This is in the face of the long-term economic and social costs associated with the condition. But currently, more attention is started to be given to the problem and recognized as public health and development agenda [5].

While there is a well-established and evidence-based management protocol for treatment of SAM [6, 7], integrating it into essential health packages and achieving desired outcomes has proven to be challenging. This might be due to weaknesses in health systems and challenges in availability of treatment commodities at local levels [8]. Generally, effective management of SAM remains as a huge challenge in low resource healthcare settings [2].

Ethiopia, particularly the study area is not an exception in this regard and the successes of treatment outcomes of SAM in healthcare settings should be tested. However, there are only limited studies on this topic in the country and there was no any study particularly in the study area. The aim of this study was therefore, to determine recovery time from SAM and identify predictors of recovery time among children of 6–59 months of age using the inpatient program at Felegehiwot referral hospital, Bahir Dar city administration, Northwest Ethiopia.

In this study, the overall recovery time from SAM and differences of recovery time between different groups of children were estimated. The association between recovery time from SAM and independent predictors was also assessed.

Accordingly, the median nutritional recovery times was estimated to be 16 days (95%CI: 14.233–17.767) and it was within the acceptable maximum international standards set at < 28 days [10, 11]. This finding was consistent with other institution-based study in Mekelle, Ethiopia which reported 17 days [12]. However, it was higher than the study done in Zambia that reported 13 days [13]. This might be related to differences in treatment and caring practices, health care settings and other socioeconomic factors among the study areas. Studies indicated that it is only by complying with the standard protocol for management of SAM that better program outcomes could be assured [14].

However, the median nutritional recovery time was lowest compared to the study reports from Kamba District, South West Ethiopia that indicated recovery time of 50 days [15] and Karat and Fasha stabilization Centers, Southern Ethiopia that reported 26 days [16].

Further analysis comparing the median recovery time between different groups of predictor variables showed that there were significant differences in median nutritional recovery time between children with SAM who had cough at admission and those who didn’t have; children with human immunodeficiency virus (HIV) and without HIV and who had tuberculosis (TB) and those who didn’t have. Likewise, there were significant difference in the median recovery time with regard to variables like presence of anemia, fever > 39 °C, provision of plumpy nut, fail to regain appetite on day 4, fail to lose edema on day 4, edema still present on day 10, fail to enter phase-2 on day 10, failure to gain more than 5 g/kg/d for 3 successive days. The highest median recovery time difference was observed between children who received plumy nut and those who didn’t which were 14 days and 28 days respectively.

But, significant differences in median recovery time were not observed between other groups of predictor variables including residence, sex, age, admission status, diagnosis at admission, diarrhea, breastfeeding history (P-Value < 0.05). For some predictor variables, this is in consistent with a study conducted in therapeutic feeding centers, Southern Ethiopia, where median recovery time was not significantly different for residence, sex, age [16].

Concerning predictors of recovery time from SAM; presence of anemia at admission, provision of plumpy nut, fail to enter in to phase-2 on day 10 and mean weight gain at discharge had significant association with the dependent variable.

Children who had no anemia at admission were 1.6 times (AHR = 1.552; 95%CI: 1.134, 2.124) more likely to recover earlier compared to those who had. However, a study in Burkina Faso indicated that with a strong respect of current inpatient SAM management, anemia did not have negative impact on nutritional recovery during hospitalization [17] which means even though children are anemic at admission if treated according to the protocol for management of SAM, it doesn’t have negative effect on a nutritional recovery time. For this study, the reason might be due to the fact that children were not receiving iron supplement even when they were anemic. It is because the finding indicated that while 64.8% of severely malnourished children were anemic, it is only 21.2% of the total children enrolled received iron supplement.

Children who received plumpy nut during their treatment were 2 times (AHR = 2.063; 95%CI: 1.356, 3.139) more likely to recover earlier compared to their counterparts. This is in line with a meta-analysis which included 14 studies in low and middle income settings and indicated that children in the ready to use therapeutic food group (RUTF) or plumy nut group were significantly more likely to recover and less likely to be non-responders [18]. According to that meta-analysis, children who received plumpy nut were 1.5 times more likely to recover than those receiving normal therapy. This might be due to the fact that children who received plumy nut might achieve rapid weight gain so that fulfilled the discharge criteria (as cured) early compared to those who didn’t get the chance to consume plumpy nut [16]. A study carried out in Tigray, northern Ethiopia also indicated that plumpy nut had a positive effect to the recovery rate and revealed that as a child consumed one more sachet of plumpy nut, the recovery rate from SAM increased by 4% [14] and this might shorten the recovery time among those who received plumpy nut for this study.

Correspondingly, children who entered in to phase 2 on day 10 were about 3 times (AHR = 2.938; 95%CI: 1.635–5.279) more likely to recover in shorter days compared to those who failed to enter. This might be due to the reason that unless children enter in to phase 2, they will not be given therapeutic foods (F-100 & plumy nut) which can promote weight gain as rapid weight gain at stage one is dangerous. That is why F75 is formulated so that patients do not gain weight during this stage. In Phase-2; they are given plumpy nut or F100. Those formulas are designed for patients to rapidly gain weight (more than 8 g/kg/day). Therefore, children who enter phase 2 early (≤10 days) will enjoy this advantage and recover early compared to their counterparts [16].

Mean weight gain was also significantly associated with median recovery time by which children who gained more than 8 g/kg/day were 1.2 times (AHR = 1.200; 95%CI: 1.014–1.422) more likely to recover earlier compared to those who gained less than 8 g/kg/day. This is consistent with recommendations of the Ethiopian protocol for the management of severe acute malnutrition [9] as well as the sphere standard [10] which state that if children gain more than 8 g/kg/day starting from phase-2 (for wasted children) and after loss of edema (for edematous), the children can recover early within the acceptable minimum standard of days.

The median recovery time ranged in the accepted maximum international standards set for management of SAM. There were significant differences in the median nutritional recovery time between different groups of predictor variables. From the multivariate Cox proportional hazard regression, it was proved that children had a lower chance of recovering early when they had anemia, not provided plumpy nut, failed to enter phase 2 on day 10 and failed to gain more than 8 g/kg/day.

Health care providers (TFU care staff) are strongly advised to comply with inpatient SAM treatment and management protocols in early diagnosis and treatment of anemia, provision of plumpy nut, regular monitoring of weight gain and phase transition. To complement the limitations of this study, further study using prospective design should be conducted for better information including other factors not included under this study such as parental socio-demographic and socioeconomic characteristics, perception of care givers on severe undernutrition and therapeutic feeding programs.