Relapse and regression to severe wasting in children under 5 years: A theoretical framework

Abstract Systematic reviews have highlighted that repeated severe wasting after receiving treatment is likely to be common, but standardised measurement is needed urgently. The Council of Research & Technical Advice for Acute Malnutrition (CORTASAM) released recommendations on standard measurement of relapse (wasting within 6 months after exiting treatment as per recommended discharge criteria), regression (wasting within 6 months after exiting treatment before reaching recommended discharge criteria) and reoccurrence (wasting after 6 months of exit from treatment as per recommended discharge criteria). We provide a theoretical framework of post‐treatment relapse and regression to severe wasting to guide discussions, risk factor analyses, and development and evaluations of interventions. This framework highlights that there are factors that may impact risk of relapse and regression in addition to the impact of contextual factors associated with incidence and reoccurrence of severe wasting more generally. Factors hypothesised to be associated with relapse and regression relate specifically to the nutrition and health status of the child on admission to, during and exit from treatment and treatment interventions, platforms and approaches as well as type of exit from treatment (e.g., before reaching recommended criteria). These factors influence whether children reach full recovery, and poorer nutritional and immunological status at exit from treatment are more proximate determinants of risk of severe wasting after treatment, although post‐treatment interventions may modify risks. The evidence base for many of these factors is weak. Our framework can guide research to improve our understanding of risks of relapse and regression and how to prevent them and inform programmes on what data to collect to evaluate relapse. Implementation research is needed to operationalise results in programmes and reduce post‐treatment severe wasting at scale.

(wasting within 6 months after exiting treatment before reaching recommended discharge criteria) and reoccurrence (wasting after 6 months of exit from treatment as per recommended discharge criteria). We provide a theoretical framework of post-treatment relapse and regression to severe wasting to guide discussions, risk factor analyses, and development and evaluations of interventions. This framework highlights that there are factors that may impact risk of relapse and regression in addition to the impact of contextual factors associated with incidence and reoccurrence of severe wasting more generally. Factors hypothesised to be associated with relapse and regression relate specifically to the nutrition and health status of the child on admission to, during and exit from treatment and treatment interventions, platforms and approaches as well as type of exit from treatment (e.g., before reaching recommended criteria). These factors influence whether children reach full recovery, and poorer nutritional and immunological status at exit from treatment are more proximate determinants of risk of severe wasting after treatment, although post-treatment interventions may modify risks. The evidence base for many of these factors is weak. Our framework can guide research to improve our understanding of risks of relapse and regression and how to prevent them and inform programmes on what data to collect to evaluate relapse. Implementation research is needed to operationalise results in programmes and reduce post-treatment severe wasting at scale. Wasting in children 6-59 months of age is a significant global public health problem. It represents a deterioration in nutritional status that is identified by anthropometric indicators such as a low weightfor-height z-score (WHZ) (or weight-for-length in children aged under 24 months), low mid-upper arm circumference (MUAC) and/or the presence of nutritional oedema. The most severe form of wasting, called 'severe wasting' , is diagnosed by having a WHZ < −3 or MUAC < 115 mm, whereas 'moderate wasting' is diagnosed by WHZ < −2 but > −3 or MUAC between 115 and < 125 mm (World Health Organization [WHO], 2013). Wasting is, in its moderate form, associated with a threefold increase in mortality and, in its severe form, with an 11-fold increase  Only about one third of severely wasted children receive treatment (UNICEF, 2020), and all forms of wasting account for an estimated 12.6% of all child deaths in the world .
In 2012, the World Health Assembly adopted the global target of reducing childhood wasting to less than 5% (WHO, UNICEF, & World Food Programme [WFP], 2014). This requires large-scale implementation of effective prevention strategies. A recent Child Health and Nutrition Research Initiative (CHNRI) prioritisation exercise identified research priorities for the prevention of wasting (Frison et al., 2020).
This exercise focused on the prevention of processes that lead to any degree of wasting in children and prevention of worsening of wasting from moderate to severe. In addition to primary prevention of wasting to occur in the first place, addressing relapse to wasting is a form of secondary prevention. The aim of this secondary prevention is to prevent a deterioration back to being wasted after initial recovery following completed treatment. This is also related to treatment, aiming to ensure initial recovery is reached and sustained. However, relapse was not considered in the research prioritisation exercise and may be an overlooked aspect of prevention of wasting.
A recently published systematic review on relapse to severe wasting (Stobaugh et al., 2019) and a systematic review of long-term treatment outcomes (O'Sullivan, Lelijveld, Rutishauser-Perera, Kerac, & James, 2018) suggest that severe wasting after receiving treatment is considerable, and so prevention of relapse is likely to be an important aspect of wasting prevention generally. Studies measured relapse across timeframes between 1 week and 18 months, with the proportion of discharged children relapsing to severe wasting within different time periods ranging between 0% and 37%. Four out of nine studies found more than 10% relapse within 1-5 months after exit from treatment and relapse to severe wasting tended to be more likely within 6 months of exit from treatment. However, the reviews identified large variations in definitions and measurements of relapse (including using different denominators), types of treatments, and criteria for admission and recovery, making comparisons of results across studies difficult. Developing a standardised definition of post-treatment severe wasting for research and programmes is, therefore, a priority in order to define standards for acceptable levels of relapse and evaluate interventions to reduce wasting after treatment.
The Council of Research & Technical Advice for Acute Malnutrition (CORTASAM) has recently recommended definitions and use of the following terms related to severe wasting (CORTASAM, 2020): relapse, regression, ongoing episode and reoccurrence (Table 1). Given that the systematic review on relapse (Stobaugh et al., 2019) found that the majority of relapse occurred within 6 months of exiting treatment, relapse to severe wasting is defined as an episode of severe wasting as per WHO definition (WHZ < -3, MUAC < 115 mm and/or presence of oedema; WHO, 2013) within 6 months of being discharged from treatment as recovered according to WHO guidelines (WHZ ≥ -2 or MUAC ≥ 125 mm and no oedema for at least 2 weeks; WHO, 2013). Cases of severe wasting exiting treatment before reaching recommended criteria should be considered ongoing episodes of wasting or regressions to severe wasting after incomplete recovery. Regression to severe wasting is defined as cases of severe wasting within 6 months of exiting treatment while still moderately malnourished and before reaching the recommended discharge criteria for recovery according to WHO guidelines (WHZ ≥ -2 or MUAC ≥ 125 mm and no oedema for at least 2 weeks; WHO, 2013).
An ongoing episode is defined as severe wasting cases that exit treatment while still severely wasted and before reaching the recommended discharge criteria for recovery according to WHO guidelines (WHZ ≥ -2 or MUAC ≥ 125 mm and no oedema for at least 2 weeks; WHO, 2013). Reoccurrence of severe wasting is defined as a separate episode of severe wasting that occurs after 6 months following discharge as recovered according to WHO guidelines (WHZ ≥ -2 or MUAC ≥ 125 mm and no oedema for at least 2 weeks).
All terms should be considered distinct from one another. This is an

Key messages
• We propose a theoretical framework to harmonise efforts for reducing relapse and regression to severe wasting after treatment, distinguishing contextual factors with broad influence on risks of wasting from additional factors hypothesised to impact risks of relapse and regression.
• Factors hypothesised to be associated with relapse and regression relate to the status of the child at admission, treatment interventions and type of exit from treatment.
These influence whether children reach full recovery, which determines relapse risk, although post-treatment interventions may modify the risk.
• Our framework can guide much needed research on risk factors for relapse and regression and how to prevent them and inform programmes on what data to collect.
important step towards standardised measurement of outcomes after treatment for severe wasting. However, a theoretical framework of relapse and regression to severe wasting to support more harmonised efforts is important to guide discussions around relapse and regression, the identification of factors associated with severe wasting after treatment, and interventions to reduce risks.
The systematic review on relapse to severe wasting (Stobaugh et al., 2019) found that children discharged before reaching WHO recommended discharge criteria tended to have higher risks of post-treatment severe wasting, representing ongoing episodes of or regressions to severe wasting after partial recovery. Furthermore, worse anthropometric measurements at admission and discharge were most consistently found to be associated with increased risk of post-treatment severe wasting. These cases may represent relapse despite nutritional recovery or a regression after incomplete recovery.
Some studies reported illness among children at time of relapse, suggesting that children discharged on the basis of anthropometric criteria may have remained immunologically susceptible to infection (Chevalier et al., 1998). There were mixed results regarding household-level factors, such as socio-economic status, feeding practices and sanitary living conditions, as well as seasonal patterns of food security and infectious diseases. However, studies commonly did not differentiate between factors associated with incidence of severe wasting and those specifically associated with relapse, regression and reoccurrence to severe wasting after exit from treatment, with widespread confusion between causality and association.
In this article, building on the recent push to standardise the definition, measurement and reporting of severe wasting after treatment, we aim to address the need for theoretical guidance by providing a framework for post-treatment severe wasting to facilitate discussions around relapse and regression. The intention is to highlight evidence gaps, provide a basis for hypothesis generation and guide researchers,   (Figure 1). These factors may be present before, during and/or after treatment for severe wasting and may be the same factors causing severe wasting in the first place while also contributing to relapse, regression or reoccurrence after treatment.
The relative importance of these factors differs across settings, and studies have linked risk of relapse to factors such as age and gender (Abitew, Yalew, Bezabih, & Bazzano, 2020;Adegoke et al., 2020;Chang et al., 2012;Stobaugh et al., 2018), HIV status (Bahwere et al., 2008;Chang et al., 2012), vaccination status (Somassè, Dramaix, Bahwere, & Donnen, 2016)  for these factors hypothesised to be associated with relapse and regression is limited, and our framework aims to illustrate these large gaps in the evidence on severe wasting following exit from treatment. We hope that this will guide hypothesis generation and future research to determine the importance of factors specific to relapse and regression beyond contextual factors associated with wasting.
Nutritional and immunological status at exit from treatment may also be associated with reoccurrence of severe wasting after 6 months.
It seems likely, however, that the longer the time interval between F I G U R E 1 A framework for relapse and regression to severe wasting after exit from treatment. The framework differentiates between contextual factors with broad influence on risks of wasting, including relapse, regression and reoccurrence, and factors hypothesised to impact risks of relapse and regression in addition to these contextual factors. Characteristics at exit from treatment are hypothesised to be more proximate determinants of relapse and regression through which other case characteristics and treatment-related factors influence risks of severe wasting after exit from treatment. Postdischarge interventions are hypothesised to influence risks of severe wasting after exit from treatment.
There are large gaps in the evidence base around relapse and regression, so this framework is meant to be a starting point for future research and should be constantly revised in light of new evidence. CHW, community health worker; CMAM, community-based management of acute malnutrition; HIV, human immunodeficiency virus; LBW, low birth weight; MUAC, mid-upper arm circumference; SES, socio-economic status; WASH, water, sanitation and hygiene; WHO, World Health Organization exit from treatment and reoccurrence of severe wasting is, the less influential these factors become, and the more influential broader contextual factors are. The focus in this article is on severe wasting within 6 months of exit from treatment.  (Frison et al., 2020). Cross-sectional studies found LBW to be associated with being wasted, stunted and underweight (Harding, Aguayo, & Webb, 2018;Ntenda, 2019;Rahman, Howlader, Masud, & Rahman, 2016) as well as concurrently wasted and stunted (Harding, Aguayo, & Webb, 2018

| Treatment interventions, platforms and approaches
We argue that our framework is relevant for all treatment settings, including inpatient treatment and community-based management of   (Guesdon & Roberfroid, 2019). In another study in India, severely wasted children who defaulted from treatment were more likely to be found severely wasted during follow-up, with higher risks of being severely wasted at follow-up for those who defaulted with lower MUAC and WHZ (Burza et al., 2016).
Exiting treatment before meeting recommended criteria, due to discharge criteria not following global recommendations or defaulting, is likely to lead to poorer nutritional status at discharge. Children may never recover from severe wasting, so should be considered an ongoing episode when readmitted for further treatment at a later point.
Children may also partially recover but then regress to severe wasting.
Moreover, children may reach nutritional recovery despite premature exit from treatment but may be at increased risk of relapse. In practice, it may be difficult to differentiate between these different trajectories after exit from treatment, but in all cases, children are at an increased risk of continuing to suffer from, regressing to or developing severe wasting due to the poorer physiological status at treatment exit.

| Nutritional and immunological status at exit from treatment
Case characteristics and aspects of treatment programmes, including the type of and criteria for exit from treatment, are linked to risks of relapse and regression due to varying nutritional and immunological characteristics among children at exit from treatment, which most proximately determine risks of severe wasting after treatment. In Burkina Faso (Somassè, Dramaix, Bahwere, & Donnen, 2016), lower MUAC at discharge from treatment for severe wasting was associated with increased postdischarge severe wasting risk, but the study did not differentiate between relapse and regression. Similar trends for lower anthropometric measures at discharge (lower MUAC, WHZ and height-for-age z-score [HAZ]) and increased relapse risks were found in a study in India (Burza et al., 2016), although results were not statistically significant. Moreover, studies involving moderately wasted children are likely to be relevant for understanding relapse and regression to severe wasting given that moderate and severe wasting represent different levels of severity of a single condition rather than distinct pathologies. In Malawi, two studies enrolling children at discharge from treatment for moderate wasting found lower risks of relapse among children with higher MUAC, WHZ and HAZ (Chang et al., 2012;Stobaugh et al., 2017). In a cohort study in Ethiopia, in a setting without a supplementary feeding programme (SFP), an approach of providing additional food (such as fortified flours or ready-to-use supplementary food [RUSF]) to vulnerable children and households, moderately wasted children with lower MUAC and WHZ had higher risk of relapsing to moderate wasting after recovery and of developing severe wasting (James et al., 2016).
Some studies have noted high prevalence of illnesses, particularly diarrhoea, among children at the time of follow-up after exit from treatment, suggesting that infections could have contributed to relapse and regression (Ashraf et al., 2011;Begashaw, 2013;Chang et al., 2012;Dani et al., 2016;Stobaugh et al., 2018). Wasting is often characterised by impaired immune function (Rytter, Kolte, Briend, Friis, & Christensen, 2014) in addition to nutritional and physiological deficits. It has been hypothesised that immune recovery in wasted children may take longer than nutritional recovery (Chevalier et al., 1998), so children may not be immunologically recovered when exiting treatment even after meeting recommended criteria for 'recovery'. However, there has been limited formal analysis of risk of infections and immune function between recently discharged children (who may be at higher risk of infections) and those not recently discharged from treatment. Still, a study in Ethiopia found higher incidence of reporting of fever, diarrhoea and coughing among children discharged as recovered from severe wasting compared with nonwasted community control children (Bahwere et al., 2017). Few studies have evaluated immunological status at exit from treatment for wasting. In Malawi, blood serum complement component 3 (C3) was found to be in the normal range in nearly all samples from children discharged from treatment for moderate wasting (Stobaugh et al., 2017).
However, C3 as a proxy measure for immune recovery does not capture immune function more broadly, and immune function may deteriorate less in moderate wasting compared with severe wasting (Rytter, Kolte, Briend, Friis, & Christensen, 2014). A study in Ethiopia concluded that tuberculin skin tests were inappropriate to assess immune functioning in the study setting, possibly due to high levels of enteric dysfunction and incidence of infections (Bahwere et al., 2017). Subclinical levels of inflammation have also been found to be common in children with moderate wasting (Cichon, 2017), but the impact on relapse and regression risks is unclear.

| Post-treatment interventions and follow-up systems
Interventions following exit from treatment could modify risks of relapse and regression by addressing risk factors for severe wasting.
Such interventions could include supplementary feeding following exit from treatment, although few studies have demonstrated effects on relapse. In Malawi, extended duration of supplementary feeding following treatment of moderate wasting reduced risk of relapse (Trehan et al., 2015). Studies in Ethiopia noted that lack of linkage to SFPs following treatment for severe wasting may have contributed to high rates of severe wasting after exit from treatment ( givers to detect changes in nutritional status early and intervene to prevent worsening (Alé et al., 2016;Bliss et al., 2018). FamilyMUAC approaches may be more relevant for preventing relapse for children discharged as recommended but may also reduce risks of regression if caregivers were trained prior to exiting treatment before meeting recommended criteria. Nevertheless, evidence on the impact of this intervention on relapse or regression is lacking.

| CONCLUSIONS AND THE WAY FORWARD
Recent systematic reviews have highlighted that incidence of relapse and regression to severe wasting after receiving treatment is likely to be substantial and standardised definitions, measurement and reporting are needed urgently. CORTASAM has published recommendations for defining and reporting severe wasting after exit from treatment (CORTASAM, 2020). We aimed to further support research and programmes with a theoretical framework of post-treatment relapse and regression to severe wasting as guidance for discussions, hypothesis generation, risk factor analyses, and development and evaluations of interventions to reduce risks of developing severe wasting after exit from treatment. This framework highlights that there may be additional factors that impact risks of relapse and regression to severe wasting beyond the impact of contextual factors associated with severe wasting more broadly but also highlights the limited evidence on these factors.
The evidence base for many factors hypothesised to be associated with relapse and regression to severe wasting in the framework is weak. Wasted children are not a homogenous group. Case characteristics, such as multiple nutritional deficiencies, including concurrent wasting and stunting, and birth weight and growth trajectories, may be linked to relapse and regression but are rarely considered. Treatment components, for example, providing antibiotics, may influence recovery but longer term outcomes are not commonly measured in research or programmes, so effects on risk for severe wasting after exit from treatment are poorly understood. Even for anthropometric status at exit from treatment, most consistently found to be associated with relapse, evidence is limited. Across studies (Bahwere et al., 2017;Burza et al., 2016;Chang et al., 2012;James et al., 2016;Somassè, Dramaix, Bahwere, & Donnen, 2016;Stobaugh et al., 2017;Tadesse, Worku, Berhane, & Ekström, 2018)

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.