In the present analysis, we investigated the frequency and risk factors related to RVA infection in undernourished children under 5 years old with diarrhoea enrolled under the National Diarrhoea Surveillance in Mozambique. Our findings show that over one-quarter (27.2%) of children were infected with RVA. However, the frequency was twice higher before widespread implementation of the RVA vaccine (42.7% versus 12.2%). It appears that the introduction of the vaccine is effective to prevent RVA infection among undernourished children as the frequency decreased in the post-vaccine period (12.2%).
A study conducted in the southern region of Mozambique before the RVA vaccine introduction, found a lower frequency of RVA infection (37.7%) in underweighted children from rural and urban areas, compared with our results (80.7% data not shown) [19]. This difference can be explained by divergent methodological approaches, while de Deus et al. only included children with moderate-to-severe underweight, the present analysis assessed mild-to-severe underweight [19]. Besides that, de Deus et al. only assessed children from southern Mozambique, while in this analysis we included children from the northern and central provinces (Nampula and Zambézia), where the prevalence of undernutrition is known to be higher [12].
A study conducted in Angola reported a lower prevalence of RVA infection (13.08–31.06%) in undernourished children before the introduction of the RVA vaccine compared with the present analysis (42.7%). This finding may be explained by the higher proportion of undernourished children in Mozambique (43%) compared with Angola (38%) [7, 12, 29].
Another study conducted in Zambia aiming to determine RVA infection in hospitalized children reported a higher rate of RVA infection in well-nourished children compared with undernourished (27.6% versus 19.3%) respectively [10]. The proportion of undernourished children with RVA positive was also lower than in the present analysis.
In Mozambique, the Rotarix® vaccine (GlaxoSmithKline Biologics, Belgium) was introduced in the National Expanded Program on Immunization in September 2015. In 2017, de Deus et al. determined the early impact of RVA vaccine through ViNaDia using the general sample without considering nutritional status. Comparing results by de Deus et al to our analysis, the previous one reported a lower RVA infection rate in 2015 (40.2% versus 42.7%), but not in 2016 (13.2% versus 12.2%) nor in 2017 (13.5% versus 9.1%) [20]. However, it is important to highlight that de Deus et al analysis was done until June 2017. In addition, it seems that undernourished children contributed significantly to the high rate of RVA infection observed by de Deus et al analysis as they represented more than 60% of the total population.
Our study suggests that children aged between 0 and 11 months are more likely to be infected by RVA when compared to children between 24 and 59 months of age. This finding probably shows that the infection in undernourished children commonly occurs during infancy and early childhood, as reported in Zambia and other low-in-come countries [10, 30]. This highlights that vaccination opportunity against RVA can not be missed in infants to avoid morbidity and mortality in this age group, since the double burden of diseases may happen before their first anniversary. Future studies are needed to understand the long-term impact of these double burden conditions on children’s lives.
Although this study did not show statistically significant differences between gender and RVA infection, several studies conducted in different countries have reported a higher proportion of RVA infection in males than in females [7, 18, 19, 31].
Interestingly, the risk of RVA infection in undernourished children was higher in Maputo the capital of the country, where most families have a higher wealth quintile compared to other provinces [12]. In addition, most of the undernourished RVA positive children families lived in brick houses. However, our surveillance has limited information about the conditions of water, sanitation and hygiene (WASH) in brick houses, leaving the question open, to what extent whether living in a brick house means having all the conditions to adequate access to WASH. According to a survey on a family budget (2015), Maputo has a higher number of family members (5.2) than Nampula (4.8) and Zambézia (4.7), which is in accordance with our analysis, which shows that overcrowded environments may provide a higher risk of RVA infection [32]. Also, most of the heads of households living in Maputo have a formal occupation, in contrast with those living in Nampula and Zambézia [32]. Based on this, our speculation is that most of the children from Maputo spend the day in kindergarten/home group care which increases the risk of diarrhoea and RVA infection when compared to children who spend the day at home [32, 33].
A cross-sectional study that included 71 households found a relationship between a high number of family members in household and risk for diarrhoea diseases [34]. In this study children’ living in households with more than four family members are at higher risk for diarrhoea RVA infection. A multi-country birth cohort study in countries with a high burden of diarrhoea and malnutrition (South Africa, Tanzania, Brazil, Peru, India, Nepal, Pakistan and Bangladesh) showed that children living in crowded households were more likely to have RVA diarrhoea [35]. Crowded households aligned with poor hygiene and sanitation can contribute significantly to the spread of enteric pathogens [36, 37].
Breast milk provides essential elements such as human oligosaccharides, secretory IgA, T and B lymphocytes to protect infants against enteric pathogens [38]. In this study, we found a high proportion of children infected with rotavirus fed exclusively breast milk or a combination of breast milk and formula. In Mozambique, national guidelines on breastfeeding practices comply with the WHO guidelines, which recommends exclusive breastfeeding until 6 months of age and continued breastfeeding until the child’s second anniversary with appropriate complementary solid food [12]. However, according to the last Demographic and Health Survey, only 43% of the Mozambican children were exclusively breastfed, with a median age of 1.3 months and for the complementary breastfeeding median age of 4.6 months [12]. This evidence shows that in Mozambique, younger children are exposed to complementary feeding earlier than the recommended. Our data confirmed that 45.5% (35/77) of children under 6 months of age received exclusive breastfeeding, while 36.5% (239/654) of children between 6 and 24 months of age were fed by breast milk and formula.
Early weaning and introduction of complementary food, has significant implication in the children’s health as their immune system is immature and may affect the intestinal function, increasing the risk of diarrhoea disease and stunting [39, 40].
Our binary logistic model seemed to show that different types of food are associated with RVA infection. However, the multiple logistic models, adjusted for age, suggests only a borderline significance when we compare other types of food with breast milk. This finding should be interpreted carefully. It is expected that children less than 6 months are more likely to have exclusive breastfeeding while others with more than 6 months are more likely to have a combination of breast milk and formula or other types of food. Thus, the association between the type of food and RVA infection tends to disappear when we introduced the age in the model. Some studies pointed out that the protective effects of breastfeeding seem to decrease with age [41,42,43]. In the literature, we found different variables to express the definitions of breastfeeding, age of children, and study designs, making comparisons and interpretations of the findings of different studies very difficult [43].
Most of the positive children had four to five episodes of diarrhoea in less than 24 h. Vomiting has been reported as one of the main symptoms of RVA infection, which is consistent with our results in undernourished children [19, 44, 45]. However, it is important to highlight that these studies were conducted without considering the children’ nutritional status.
The fact that 7.4% of the infected children were HIV positive indicates a triple burden to address (undernutrition, HIV and RVA infection). According to the Malaria, HIV/AIDS, and Immunization Indicator Survey in Mozambique women in reproductive age present the higher prevalence of HIV infection [46]. Which has a significant impact on increasing the rate of HIV infection in children due to vertical transmission.
While HIV contributes to the reduction of CD4+ T cells and the role of the immune system, which requires greater energy needs, undernutrition leads to immune dysfunction increases the risk of infections, poor response to vaccination and low antiretroviral treatment efficacy leading to high mortality [47]. RVA infection associated with this double burden diseases will generate synergism increasing the chances of developing Acquired Immunodeficiency Syndrome (AIDS) and chronic undernutrition, which can lead to the children’ mortality [48].
The majority of children who presented undernutrition, HIV and RVA infection were infants (< 12 months) from Maputo and Nampula provinces. Maputo was reported as having a high prevalence of HIV infection (16.9%) while Nampula province was reported as having a high prevalence of chronic undernutrition (55.3%) [12, 46]. Our data show that children living in both provinces are exposed to undernutrition and co-infections during the first year of life. Future studies are needed to understand this relationship and the long-term consequences of these conditions for the children.
Hospitalization of undernourished children is frequently associated with longer stay in healthcare facilities and a higher risk of death, mainly if associated with gastroenteritis [49]. However, in this study, most of the children were hospitalized for 1 to 4 days.
The findings of this study shows a high proportion of children undernourished with diarrhoea infected by RVA even after the vaccine introduction. This finding suggests that there is a need to evaluate if the treatment of the children affected by the above-mentioned conditions follows the recommended guidelines in order to understand the impact on the morbidity and mortality associated in children.
The major limitation of this study is the high number of missing data on the variables, even with the continuous training staff at the sentinel site to improve data quality collection. Additionally, the design of the study included only children who looked for health services care (inpatient and outpatient) and consequently more likely to have a severe infection and be undernourished. This is a cross-sectional study, which does not allow measuring the causality between variables and the RVA infection. Future studies should consider the possibility to include a community approach to understand better the real situation of RVA infection and the role of undernutrition in case-control or longitudinal studies.