Based on our analysis of nationally representative datasets in five South Asian countries, EBF is a protective factor for diarrhoea in infants aged 0–6 months in Afghanistan, India, and Nepal and for fever in Afghanistan and India. Our study also showed that infants who were EBF had lower odds of ARI in Afghanistan. These findings are consistent with data from other studies in LMICs which also estimated the beneficial association of EBF [29, 33,34,35,36, 40]. A large study [34] that included data from the 2015–2016 India DHS found that EBF was protective against diarrhoea in babies aged 0 to 5 months at a national level (AOR: 0.64, 95% CI 0.57, 0.72) as well as in the Central, Northern and Eastern regions of the country. Another large study [29] that pooled data from the DHS in nine sub-Saharan African countries with high rates of diarrhoeal morbidity reported that EBF was significantly associated with decreased likelihood of diarrhoea among babies aged 0–5 months (AOR: 0.50, 95% CI 0.43, 0.57). A cross-sectional study in Vietnam [41] suggested that babies who were predominantly or partially breastfed had a greater likelihood of having diarrhoea in comparison to babies who were EBF. Individual studies in Bangladesh [33, 35] and Pakistan [36] revealed that EBF reduces the risk of diarrhoea, ARI, and fever. Although our study found a protective effect of EBF in Bangladesh and Pakistan where the general direction of the results was similar with Afghanistan, India, and Nepal, these results were non-significant. This inconsistency in findings could be attributed to the considerable smaller sample size of the surveys in these two countries compared to Afghanistan and India.

Several biological mechanisms explain why EBF may have protective effects on infectious diseases such as diarrhoea and ARI. Breast milk contains numerous anti-inflammatory, antimicrobial, growth factors and bioactive elements such as oligosaccharides, immunoglobulin A (IgA), and lactoferrin that protect against childhood infections [4, 5, 7]. A baby’s immunological development and maturation are potentially aided by human milk, which has its own immune system and a variety of soluble and cellular components [7]. IgA stops bacteria and viruses from adhering to the mucosal epithelial cells, which might lead to infections [4, 5, 7]. Additionally, there is a theory that oligosaccharides may prevent respiratory infections and gastrointestinal illnesses in babies by preventing pathogens from attaching to their mucosa [4, 5, 7]. The primary protein in human milk is lactoferrin, which functions as a microbicidal agent to eradicate viruses and bacteria [4, 5, 7]. Furthermore, infant weaning foods and powdered infant formula have been shown to be contaminated with pathogens, putting non-breastfed newborns at an increased risk of exposure to these contaminants [42,43,44]. Powdered infant formula is not sterile and has been associated with significant illness from infections with the bacteria Cronobacter sakazakii and Salmonella spp [43, 45]. For example, a study in Nigeria found that Cronobacter sakazakii was found in 16 of the 360 powdered infant formula samples that were examined, representing an average prevalence rate of 4.4% [42]. Also, other sources of contamination during formula milk preparation include the addition of unclean water, improper handling or inadequate nipple and bottle cleaning [45].

Given its potential to have a significant negative impact on health outcomes, prevalence of bottle feeding is a IYCF indicators [32]. Our study found that babies who were bottle fed were more likely to have diarrhoea in India, ARI in Afghanistan, and fever in India and Afghanistan. The similar pattern across these South Asian countries suggests a high degree of consistency in the association between bottle feeding and diarrhoea, ARI, and fever. This pattern was also observed in earlier investigations in India [34] and Ethiopia [46] which have reported the impact of bottle feeding. The recent paper in Ethiopia [46] revealed that babies and young children aged 0–23 months who receiving bottle feeding were 1.36 times more likely to experience ARI relative to those who did not receive bottle feeding. Also, children who were bottle fed tested positive for rotavirus infection, the most prevalent cause of diarrhoea, according to an observational study [47] done in a hospital in India. This may be due to the higher risk of contamination from the water used to make breastmilk substitute, bottle, teat, or milk in babies who are fed bottle milk [5]. In Peru, an investigation [48] showed that 23% of bottles and 35% of bottle nipples were contaminated with Escherichia coli, which was greater than any other household item tested.

Another key IYCF indicator is early initiation of breastfeeding, within an hour of birth. Our analysis indicated that early initiation of breastfeeding was associated with lower diarrhoeal disease in India. This result was consistent with a past study conducted on the DHS data in India [34] and sub-Saharan African nations [29].

Around 27% of all deaths in children under five occurred in the South Asian region according to UNICEF in 2020 [14]. A meta-analysis [49] undertaken estimated that babies who were predominantly breastfed, partially breastfed, and non-breastfed had a greater risk of death (relative risk of 1.48, 2.84, and 14.4 respectively) relative to babies 0–5 months of age who were EBF respectively in low-income countries. One of the most cost-effective child survival interventions is EBF during the first six months of life, which significantly lowers the chances of a child dying from pneumonia or diarrhoea [6, 16]. Breastfeeding promotes development gains at all levels, from decreased illness incidence to economic returns, and will be a significant factor in reaching the Sustainable Development Goals (SDG) 2 and 3 on avoiding child deaths, attaining food security, and improving nutrition in South Asia [50]. To successfully establish a more supportive environment for mothers who choose to breastfeed, we require investments from governments at all levels of society [4]. This involves providing women with the information they need to make informed choices, as well as the support they require from their families, communities, work environments, and healthcare systems, in order to ensure that EBF for the first six months is possible [4]. Increased protection, promotion, and support for EBF would offer a cost-effective approach towards reaching the SDGs.

In South Asia, various factors have an influence on EBF. For instance, compared to boys, girls are less likely to EBF in India [19]. Moreover, in contrast to some low caste groups in India, the Tajik and other smaller four ethnic groups in Afghanistan are less likely to EBF, which implies that sociocultural norms may have an impact on EBF and are context specific [19]. A qualitative study in Pakistan in 2018 revealed that negative attitudes towards colostrum, poor social support, influence of social and family decision-makers, perceived inadequate milk supply, mother’s heavy workload, and advertising of infant formula were all factors that were barriers to optimal EBF practices [51]. Therefore, these highlight the need for context-specific and integrated interventions tailored to each South Asian country to support mothers in overcoming the barriers to optimal breastfeeding practices. The promotion of laws and rules governing the marketing of breast milk substitutes, supporting breastfeeding in the workplace and households and paid maternity leave in LMICs is crucial [28].

Potential limitations of the study should be taken into consideration when interpreting the results. The statistical findings revealed significant effects for Afghanistan and India, which are, respectively, large sample populations, and highly deprived environments where the impacts of EBF are large. It is likely that the absence of significant results for the three nations with smaller samples (Bangladesh, Nepal, and Pakistan) may be due to a lack of statistical power, or Type II error. A paper published in 1994 showed the extent of underpowered studies that led to null trials in the literature [52]. Only 36% of the null trials included in the survey had sufficient power (80%) to identify a relative difference of 50% [52]. The “absence of evidence” in these studies should not be regarded as “evidence of absence,” and underpowered studies should be interpreted with caution [53].

Given that the data were gathered via self-reports, recall bias may have had an impact on the findings. Nevertheless, we restricted the study sample to the youngest living baby who resided with the mother to lessen the possible impact of recall bias. Also, our study used cross-sectional data, which makes it challenging to establish a causal relationship. Another important limitation is that there may have been potential misclassification bias, as measurement of ARI, fever, and diarrhoea was determined in the two weeks before the survey. This implies that mothers could have inaccurately stated that their babies had symptoms of ARI, fever, or diarrhoea resulting in an underestimation or overestimation of the measure of association between EBF and other IYCF indicators and morbidity outcomes. Another limitation is that we were not able to adjust for all the confounding variables such as seasonal and cultural variations which could have affected the association between the exposure and outcomes. An important limitation to consider is that dichotomous feeding exposures in nations where breastfeeding is practiced differently, EBF and not EBF have varied meanings; because of this, when comparing EBF and not EBF in nations such as Bangladesh, India, and Nepal with high rates of both exclusive and partial breastfeeding, the magnitude of the benefits of EBF relative to commercial milk formula or bottle feeding is understated. For nations like Nepal with a relatively small population size in the surveys, there may be very few bottle-fed babies in comparisons with the non-EBF group, which could provide challenges for the statistical analysis. Additionally, for similar reasons, cross-country comparisons of EBF and not EBF results may be inaccurate for nations like Pakistan that have relatively low breastfeeding rates.

A key strength is that our study uses the latest data from nationally representative sample from each South Asian country included to enable sufficient generalizability of study findings. Finally, due to the high response rate in the surveys and the use of consistent standardized questionnaires, the study results are less likely to be impacted by selection bias.