Efficacy of mid-upper arm circumference in identification, follow-up and discharge of malnourished children during nutrition rehabilitation

Studies show that malnutrition affects the lives of many children, especially those from developing regions, and contributes to over 50% of child deaths globally [1, 2]. Severe acute malnutrition (SAM) is the most lethal form of malnutrition and is normally defined by a very low weight-for-height (below WHZ < -3SD of median WHO standard) or visible severe wasting or by the presence of oedema [3]. Nevertheless, for children aged 6-59 months, Mid-Upper Arm Circumference (MUAC) < 115 mm is the most commonly used indicator of severe acute malnutrition and in essence most nutrition rehabilitation interventions usually admit only those children with moderate acute or mild malnutrition while the severe cases are given specialised treatment before admission to the rehabilitation programmes [4, 5].

MUAC is gaining popularity over "weight-for-height based diagnosis (WH)" as a tool to follow-up children during nutrition rehabilitation [6, 7]. This is because, MUAC has only one measurement, while WH needs two, it uses lighter and cheaper materials, and no reference tables are required (unlike WH based diagnosis). These elements further explain the potential value of MUAC for community screening in large populations [7, 8]. WH seems to only confirm cases of severe acute malnutrition in children presenting a very low MUAC. Besides the advantages, MUAC also became an important tool for regions with high illiteracy levels, i.e. a characteristic that fits the requirements of Community Therapeutic Care (CTC) of acute malnutrition [9, 10]. In addition, it has been adopted by many agencies and community therapeutic care centres (CTCs) in recent years [11, 12, 13].

Furthermore, undernutrition in children can be a lethal condition which requires immediate nutrition attention and medical support [2, 8]. Therefore, it is very important to identify frail children as early as possible in order to adequately treat them through nutrition support programmes [7, 12, 14]. However there is no standard indicator for such identification and there is not enough evidence for using one indicator in the follow-up of children during rehabilitation. Therefore, this study has two objectives related to the types of identification where MUAC could be beneficial: First, Comparison of MUAC, WH and a multi-component indicator (MUAC adjusted for age and gender) to identify malnourished children for nutrition rehabilitation. In the presence of pathognomonic symptoms, during marasmus and/or kwashiorkor, rapid diagnosis is possible [15]. Conversely in children with no clinical signs WH and MUAC have been the two anthropometric indices most often used in identification of malnourished children [10, 16, 17, 18, 19]. WH and MUAC are proxies of an individual's nutritional status which reflects indirectly the catabolism of lean tissue and fat which occurred in undernourishment [12, 14]. Nevertheless their use as screening tools for undernourished children differs considerably. The WH definition of undernourishment relies on the normal distribution of the anthropometry in a reference population. In this case, severe acute malnutrition (SAM) is defined by WHZ < -3 and moderate acute malnutrition (MAM) by WHZ -3 to < -2 [3, 7, 8], with these cut-offs being age and sex specific. On the other hand, for MUAC based diagnosis, a single cut-off is applied independently of the age and sex. Here, severely malnourished is defined as 110 mm while moderately malnourishment is defined by > 110 to 120 mm, in 6-59 months old children. Recently these cut-offs have been adjusted to 115 mm and 125 mm respectively [11]. Nevertheless, using MUAC and WH globally provides similar prevalence rates of acute malnutrition [7, 20]. Besides, it is surprising to note that only a proportion of about 40% of malnourished cases identified by one indicator will also be identified by the other [7, 20, 21]. This is an interesting observation as these two indicators are assumed to assess the same problem and should identify children for a common treatment. However, in practice, the use of one indicator will lead to misdiagnosis of some cases while using the two independently should increase the case load. Consequently, since they detect different populations it can be hypothesized that these populations require different management approaches. In order to address this, we analyze the discrepancy between MUAC, WH and a multi-component indicator (MUAC adjusted for age and gender) in their identification of malnourished children for nutrition rehabilitation.

Second, Comparison of MUAC and WHZ in follow-up and discharge of malnourished children during nutrition rehabilitation. Admission and follow-up of undernourished children on the basis of MUAC alone is a relatively new strategy that has been considered by many agencies [12]. Although there exists a pool of evidence on the response of undernourished children selected on basis of WH, a similar response is less documented for undernourished children admitted on the basis of MUAC. As MUAC and WH identify different populations [7, 20, 21, 22], we hypothesize that undernourished children identified by MUAC alone, and not by WH, respond differently to nutrition rehabilitation. Previous research shows that MUAC responds to nutrition therapy even when home-based [23, 24] or even in the case of rehabilitation of moderately malnourished children [7, 25]. In addition, the global daily MUAC gain is between 0.2-0.4 mm during nutrition rehabilitation and this is significantly influenced by the initial nutrition status and stature of the child. Although in many studies the MUAC gain is consistent with this global range, as nearly all studies assessed MUAC during rehabilitation by using WHZ as the inclusion criteria [4, 6, 14]. Still there is limited evidence on the nutritional rehabilitation of malnourished children selected on basis of low MUAC alone. During rehabilitation of malnourished children the relationship between changes in MUAC and changes in weight are a function of age, sex, and height at inclusion and conditions that are associated with increased mortality risk [7, 16, 20]. Too our knowledge, there is no study describing the utility of MUAC as a monitoring tool during nutrition rehabilitation [4]. Hence, we address this concern by analyzing the responses of children admitted on basis of MUAC throughout the rehabilitation in relation to the generic admission and follow up using WHZ.

Given these two objectives, the current study attempts to respond to the following five questions: 1) Is MUAC useful in clinical identification and follow-up of malnourished children?; 2) What is the difference between MUAC and WH in the identification and follow up of malnourished children?; 3) How effective is MUAC single cut-off for all?; 4) What is the rate of recovery of children selected on the basis of MUAC alone compared to those selected using WH and multi component indices?, and 5) What is the discharge criterion for the children selected on the basis of MUAC alone?

The various advantages and simplicity positions MUAC as a very attractive tool for detection and follow-up malnourished children and eventually discharging them during nutrition rehabilitation. Nevertheless the use of MUAC single cut-off (< 12.5 cm) for admission, monitoring and discharging malnourished children during rehabilitation is not supported in literature.

The current study was carried out on moderately malnourished (MAM) children (Table 2) undergoing a single intervention module. And the results, expressed using changes in different diagnostic indices, show that there was a systematic change in nutritional status from admission through to discharge. Above and beyond MUAC gain (mm/day) and weight gain (g/day) was substantial which is consistent with earlier findings [31].

The capacity of MUAC to detect malnourished children was therefore, when compared with other anthropometric indices, promising for admission, follow-up, and eventually discharging the children. And this scenario is not only as regards the many advantages of MUAC, but also its ability to respond to treatment or improved dietary intervention. This finding is corresponds with earlier studies examining the changes in MUAC during rehabilitation [18, 21, 29]. Furthermore, the findings highlight a considerable discrepancy between the prevalence of malnutrition as defined by MUAC single cut-off (< 12.5 cm) and that defined by WAZ < -2, MUACZ < -2 and WHZ < -2 which was also the case in previous research works [7, 18, 29].

The single MUAC cut-off produced higher values of sensitivity, specificity when compared with MUACZ < -2 than when related to WHZ < -2 and WAZ < -2, while specificity increased at the expense of sensitivity between admission and discharge (Table 3). This points towards a higher capacity to detect the frail children for admission and those children still in need of extra attention before discharge. But, on the contrary, WHZ resulted in significantly lower values of specificity and sensitivity (Table 4) when compared with WAZ < -2, MUACZ < -2 and MUAC (< 12.5 cm). This is indicative of the appropriateness of using MUAC in identification and discharge of children during rehabilitation than WHZ < -2. These findings are in consonance with the findings by Le Thi Hop et al. [29] which demonstrated the relation of different MUAC cut-offs and other indices.

From the results, MUAC single cut-off (< 12.5 cm) produced higher values of positive likelihood ratio (+ LR) when compared with MUACZ. Correspondingly, the positive likelihood values (+ LR) for WHZ < -2 were higher when compared with WAZ < -2. Nonetheless, there were no significant differences between MUAC and WHZ which demonstrates that the two indices are influenced by changes in age and gender. Furthermore, negative likelihood ratio (-LR values) obtained were closer to unit when MUAC single cut-off (< 12.5 cm) was compared with WHZ and WAZ. But -LR values were far from unit when WHZ was compared to MUACZ, WAZ and MUAC (< 12.5 cm). In other words, the confirmatory power of a positive MUAC for a single cut-off (< 12.5 cm) and WHZ < -2 was indifferent but the negative test for MUAC gives more information than for WHZ. This information contradicts the assumption from previous studies that MUAC single cut-off is independent of age and gender for children below 60 months [4, 6, 21].

Furthermore, the average MUAC increment for the children was 0.98 cm, albeit higher in females (1.03 cm) than males (0.90 cm). This reflect the changes in the lean tissue during rehabilitation owing to the supplementary feeding as reported, also, in earlier findings [29, 30]. Also, the rate of MUAC and weight gain during rehabilitation of children was 0.14mm/day and 13.06 grams/day respectively. However, children admitted on the basis of low MUAC using a single cut-off (< 12.5 cm) gained a significantly higher MUAC but their gain in weight was not different from children admitted on the basis of low WHZ (< -2SD). These changes reflect the daily responses in MUAC and weight which need to be realized to obtain a meaningful recovery of malnourished children as demonstrated in a review by Myatt et al. [7].

Regarding the duration of stay in rehabilitation program, the study results show that children took an average of 86 days to recover. There were also no differences in duration of stay if children are admitted using different criteria. But generally females stayed for longer in the rehabilitation to recover than their male counterparts, albeit more if admitted with single MUAC cut-off (< 12.5 cm) than WHZ < -2SD. This indicates that there are no significant differences in the cost of rehabilitation for children admitted on basis of MUAC single cut-off (< 12.5 cm) and those admitted on basis of WHZ, WAZ and MUACZ (< -2). However, since females are likely to stay longer to recovery than male when admission is based on MUAC single cut-off (< 12.5 cm), the overall cost of rehabilitation might change depending on the proportion of each gender admitted.

This study also highlights the unique evolution of MUAC, weight, height and the WHZ, MUACZ and WAZ (Fig. 2). The polynomial graphs show that during rehabilitation MUAC increases steadily upon admission and during the first few visits then stabilizes and starts to decrease. However, weight increases at the initial stages and levels off after some time while height remains constant right from admission and start increasing after a number of visits. This demonstrates the capacity of MUAC to immediately respond to the treatment during rehabilitation malnourished children. This means that MUAC can be used as a good indicator of recovery and, in the short term, the weight based indices can be used in the rehabilitation while in the long term, height can be used as it takes longer to respond. As they evolve differently, all three indices are important for monitoring the trend of recovery of malnourished children in rehabilitation programs.Nevertheless, there was no significant difference in KAPPA coefficient when comparing MUAC single cut off and WHZ < -2 the degree of agreements/accuracy was higher (90.72%) compared to WHZ < -2 (85.1%) in relation to MUACZ and WAZ.

It is important to note that the sample size, sampling and data selection criteria may have influenced the findings. Besides, data was obtained from a specific population and in a small geographical area with specific needs which is obviously likely to be the strongest limitation to the findings of this study and validation using bigger study and careful sample selection and examination of the nutritional status of the participants is inevitable.

In conclusion, the MUAC single cut-off (< 12.5 cm) was found to yield higher sensitivity and specificity than WHZ when admitting and follow up of malnourished children during nutrition rehabilitation. An average stay of 86 days as well as a MUAC and weight gain of 0.14mm/day and 13.06grams respectively, is necessary for all children to reach full recovery when admitted to rehabilitation programmes. Moreover, a higher MUAC gain per day of 0.19mm/day is recorded when children are admitted on basis of MUAC single cut-off (< 12.5 cm). In principle these findings positions the MUAC single cut-off (< 12.5 cm) as a strong candidate for admission and monitoring malnourished children as well as discharging the recovered children during nutrition rehabilitation. Similar attempts have also been presented in earlier studies that point to the appropriateness of MUAC as a tool for nutrition rehabilitation. This is in line with the requirements attached to the limited-resource settings in the developing world, i.e. the key beneficiaries of these tools, particularly the Community Therapeutic Care (CTC) of acute malnutrition.