Complex interactions between malaria and malnutrition: a systematic literature review

Malaria and malnutrition in children (refers to all forms of undernutrition) are reported independently to be major causes of morbidity and mortality in low- and middle-income countries. About 3.2 billion people remained at risk of malaria with an estimated 216 million new cases (95% confidence interval 196–263 million) and 445,000 deaths worldwide in 2016; the majority of deaths occur in children under 5 years in sub-Saharan Africa [1]. Approximately 3.1 million deaths in children under five are attributed to malnutrition each year, representing 45% of all childhood deaths [2]. The interaction between malaria and childhood malnutrition has been studied for many years and complex interactions between these high-burden conditions are now increasingly recognised. Understanding the consequences of malnutrition on malaria and vice versa is crucial and may help guide the choice of public health interventions and research priorities where significant co-morbidity exists.

Malnutrition is a complex phenomenon due to its multifactorial aetiology and diverse clinical presentation. Acute malnutrition manifests with wasting (low weight for height) and chronic malnutrition as stunting (low height for age). Being underweight (low weight for age) can result from either chronic or acute malnutrition or both. Assessment of malnourishment can be conducted using anthropometric indicators which compare child’s weight and height to the standardised age- and sex-specific growth reference derived from the international reference population of children between 6 and 59 months of age (World Health Organization (WHO) Child Growth Standards 2006) [3]. The anthropometric indicators are expressed as a number of standard deviations (SDs) below or above the reference mean or median value, Z-score. Cutoffs of − 3 are used to indicate severe malnutrition and values between − 2 and − 3 are considered to be moderate malnutrition. Weight-for-height Z-score (WHZ) is the indicator used to classify a child with wasting. Mid-Upper Arm Circumference (MUAC) is another frequently used indicator for wasting. Severe acute malnutrition (SAM) is defined as MUAC < 115 mm and/or WHZ < − 3 and/or bilateral pitting oedema. Stunting is defined by the measure of height-for-age Z-score (HAZ) and a child is considered as being underweight based on low weight-for-age Z-score (WAZ). These definitions do not take micronutrient malnutrition into account, which can occur even if the person is getting enough energy and they are not thin or short.

The exact relationship between childhood malnutrition and malaria remains complex, controversial, and poorly understood. One of the key questions is, to what extent the burden of malaria is attributable to wasting and stunting? In regard to the impact of malaria on malnutrition, some evidence suggests malaria has adverse effects on nutritional status of young children [4‐9]. On the other hand, whether and how malnutrition influences malaria morbidity and mortality is debated. Some studies have reported that malnutrition is associated with a higher risk of malaria [10‐13], others have suggested a protective effect [14‐17], or no differential risk [18, 19]. Similarly, there is very limited evidence on the relationship between nutritional status and antimalarial drug efficacy. Clinical efficacy of the current first-line malaria treatment, the artemisinin combination therapies (ACTs), in malnourished children has been rarely explored [20‐22]. Understanding the complex relationship of the immune response of individuals infected with malaria and suffering of malnutrition is crucial to guide specific antimalarial therapeutic approaches in this vulnerable sub-population. There are key knowledge gaps in defining the complex relationship between malnutrition and malaria, which need to be identified and addressed. We aimed to conduct a systematic review of the current understanding of interactions between acute or chronic malnutrition and the risk of developing malarial infection. A further objective was to explore published literature on the impact of malnutrition on the efficacy of antimalarial treatment.

We conducted a systematic literature review of manuscripts published between Jan 1, 1980, and Feb 19, 2018. PubMed, Global Health and Cochrane Libraries were searched using key terms (Additional file 1), and articles published in English, French or Spanish were accessed. Two reviewers identified relevant articles of interest, as per criteria listed below, by screening titles and abstracts of publications retrieved. The preferred reporting items for systematic reviews and meta-analyses (PRISMA) guideline [23] was followed. The PRISMA checklist is provided as Additional file 2 and the review is registered in international prospective register of systematic reviews (PROSPERO Registration No. CRD42017056934).

The literature search identified 2945 articles. A total of 32 articles identified through the search and 1 article obtained through citation tracking were included, describing cross-sectional surveys (n = 16), longitudinal studies (n = 12), interventional studies (n = 3), case-control study (n = 1) and individual patient data meta-analysis (n = 1) (Fig. 1). Details of the 33 studies included in this review are given in Table 1, while the study characteristics are summarised in Additional file 4.

The evidence on the effect of malnutrition on malaria risk remains controversial and in many instances contradictory. The current review highlights some key limitations in the way the interaction between malaria and malnutrition has been assessed and reported. First, differences in methodology, study populations, the variability in measures used to define malnutrition (e.g. different growth references, different cut off thresholds), and the heterogeneous malaria transmission intensities with different levels of host immunity within the different studies make the comparison challenging. Second, there is a paucity of information on the effect of malnutrition on therapeutic responses to ACTs and their pharmacokinetic properties in malnourished children in published literature. Generally, vulnerable populations with common co-morbidities such as malnutrition, obesity, HIV or tuberculosis co-infection are excluded from or under-represented in antimalarial drug efficacy trials [37]. Although weight is documented, height is rarely recorded in ACT efficacy trials (< 20% of 250 trials currently included in the WWARN repository, personal communication Kasia Stepniewska), restricting the possibilities for secondary analyses. Another useful metric, mid-upper arm circumference (MUAC) is also rarely documented in malaria clinical trials despite being relatively easy to measure and low MUAC shown to be associated with increased malaria risk [38] and decreased lumefantrine bioavailability [39]. Several confounding factors and effect modifiers have been suggested such as age, co-morbidities (e.g. HIV, tuberculosis co-infection and drug interactions), immunity, socio-economic status, or refeeding practices. However, these confounding factors are poorly documented and controlled for in most of the reported studies in this review.

This review has several limitations. First, one third of the studies included in this review recruited individuals of all ages, and disaggregating observations by the age of individuals (below and above 5 years) was not possible. Methodologically, the temporal relationship between malnutrition and risk of malaria (and progression from infection to symptomatic malaria) could not be assessed because of cross-sectional study design (50% included studies). It is also limited by the extent to which important confounders (such as differential micronutrient deficiencies, ecological and genetic factors) are measured and reported in the included articles. Finally, the heterogeneity of the selected studies (presented as Additional file 5) including variations in measurement of nutritional status, definition of malaria, and statistical approaches adopted in deriving the risk estimates restricted plausible aggregated data meta-analysis in this review.

Interestingly, while no consistent association between risk of malaria and acute malnutrition was found, chronic malnutrition was relatively consistently associated with severity of malaria such as high-density parasitemia and anaemia [10, 26, 27]. The mechanism behind the higher risk of recurrent parasitemia could be explained partially by the impact of chronic malnutrition on the immune system and/or lower antimalarial bioavailability. Likewise, the apparent protection of wasted children from clinical malaria might be caused by their being administered a higher mg/kg antimalarial dose and/or a modulation of their immune response and thus an absence of symptoms, e.g. fever usually associated with malaria as opposed to an absence of malaria infection. Friedman et al. showed high-density parasitemia as a predictor for chronic malnutrition [26]. Nevertheless, the role of malaria in the aetiology of malnutrition remains unclear. The effect of malaria on nutritional status appeared to be greatest during the first 2 years of life and age acted as an effect modifier in the association between malaria episodes and malnutrition [6].

ACTs are now recommended for the treatment of uncomplicated falciparum malaria in almost all malaria-endemic countries and the number of children exposed to these antimalarial agents is increasing. A priority area is to identify gaps in our current knowledge in regard to the pharmacokinetic properties of artemisinins and partner drugs in malnourished paediatric populations to optimise dosing in order to ensure efficacy, safety and avoid the selection of parasite resistance [40]. Exposing pathogens to sub-therapeutic levels of active ingredients is a major driver of resistance. Protein-energy malnutrition, defined as insufficient calorie and protein intake, may have potential physiological effects on the absorption, distribution and metabolism of ACTs and subsequently affect the efficacy and safety of ACTs. Severe acute malnutrition can cause pathophysiological changes, including increasing total body water, leading to greater volume of distribution of drugs, which in turn may cause sub-optimal drug exposure when ACTs are given at standard doses [35, 41]. This could be further compounded by malnutrition in paediatric patients leading to dosing inaccuracies of ACTs when dose is calculated by age (over-dosing for actual body weight) or weight (under-dosing for age). Malnutrition can also be associated with intestinal malabsorption and villous atrophy of the jejunal mucosa which may cause impaired drug absorption [42]. The reduced absorption of lipids and fats has the potential to specifically affect the lipid-soluble ACTs [43]. The hepatic metabolism of the ACTs may be compromised in malnourished children. For instance, in case of quinine, hepatic metabolism is decreased in protein deficiency and increased in global malnutrition [44]. Thus, hepatic metabolised drugs should be carefully monitored in children with malnutrition. A recent individual patient data (IPD) meta-analysis conducted by the WWARN to investigate the effect of dosing strategy on efficacy of artemether-lumefantrine showed that the risk of treatment failure was greatest in malnourished children aged 1–3 years in Africa (PCR-adjusted efficacy 94.3%, 95% CI 92.3–96.3) [45]. Another large WWARN IPD meta-analysis of individual PK-PD data from patients treated with artemether-lumefantrine (AL) for uncomplicated malaria, showed day 7 concentrations adjusted for mg/kg dose were lowest in very young children (< 3 years), among whom underweight-for-age children had 23% (95% CI − 1 to 41%) lower concentrations than adequately nourished children of the same age and 53% (95% CI 37 to 65%) lower concentrations than adults [36]. This raises the question of whether an adapted dosing regimen is needed in malnourished young children. The PK-PD evaluation of artemisinins and longer-acting partner antimalarials for the treatment of malaria in paediatric populations and the effect of malnutrition on the pharmacological activity of ACTs is a priority area to identify and address key knowledge gaps.

A summary of the remaining knowledge gaps is presented to serve as the basis for prioritising future research strategies and highlights the need for standardising measures and reporting of nutritional status. Further analyses using individual patient data could provide an important opportunity to better understand the variability observed in publications by standardising both malaria nutritional metrics. In an era of emergence and spread of antimalarial drug resistance, it is imperative to improve our understanding of the pharmacodynamics and pharmacokinetics of ACTs in malnourished children to optimise antimalarial treatment of this very large vulnerable population. Pooled analysis, gap analysis and carefully designed prospective, randomised controlled clinical trials can provide strong evidence on the outstanding questions raised in this review related to malaria-malnutrition interactions.