Article Text
PDF
Abstract
Objective Pneumonia is associated with four times higher odds of death among children with severe undernutrition. However, there is an equipoise for the mortality of children without severe undernutrition and the impact of macronutrient interventions. We collated evidence on mortality, anthropometric outcomes and the effect of macronutrient interventions in the management of non-severely undernourished children (28 days–17 years) with pneumonia globally.
Design Systematic review and meta-analysis using a priori criteria developed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline.
Data sources PubMed, Medline, EMBASE, Web of Science, Google Scholar, Scopus, Cochrane Central and bibliographies were searched between January 2000 and July 2024.
Eligibility criteria We included articles conducted among children between 28 days and 17 years with pneumonia and non-severe malnutrition that reported on mortality and changes in anthropometric status or macronutrient supplementation. Studies conducted exclusively among adults, on micronutrient supplementation, case studies, commentaries and reviews were excluded.
Data extraction and synthesis Two reviewers independently screened, abstracted the data and conducted risk of bias (RoB) using standard criteria including the RoB in non-randomised follow-up studies of exposure for observational studies and the revised Cochrane RoB assessment tool for randomised studies (RoB 2.0). Heterogeneity was assessed using the I2 statistic, and subgroup analysis was done. Data were analysed using both narrative and quantitative synthesis. Quantitative synthesis was done using the maximum likelihood random-effects model in STATA V.18.0, with the ‘meta_es’ command.
Results A total of 15 articles were included (11 conducted in sub-Saharan Africa and four in Asia), with 169 901 participants overall. The mortality among non-severely undernourished children with pneumonia was 3.0% (95% CI 2% to 5%, I2=99.38%), with a range of 1–13% across studies. Children with moderate undernutrition had a higher overall mortality, 9.0% (95% CI 6% to 13%, I2=89.50%), than well-nourished children, with a range of 3–19% across studies. Only one of the 15 studies reported anthropometric outcomes during follow-up and compared mortality rates of those who did versus did not receive macronutrients. The study results were inconclusive.
Conclusions Mortality among non-severely undernourished children with pneumonia ranges between 1–13% globally. There is inadequate follow-up nutritional assessment and management for non-severely undernourished children with pneumonia.
PROSPERO registration number CRD42021257272.
- Nutrition
- Child
- Systematic Review
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
This review considered and included studies that assessed macronutrient interventions.
The systematic review was based on a comprehensive search.
There was a high level of heterogeneity and risk of bias among included studies.
Publication bias was present despite searching multiple databases (over 5) including bibliography screening.
There were few studies assessing the effect of macronutrient interventions on mortality and anthropometric outcomes.
Introduction
Pneumonia is the major cause of morbidity and mortality among children under the age of 5 years worldwide with an incidence of 37 million episodes in 2021.1 Pneumonia accounts for 725 000 deaths among children under 5 years, and 14% of all-cause mortality in this age group.2 3 Over 50% of pneumonia deaths occur in low- and middle-income countries.1 4 Hospitalisation with pneumonia is associated with a 2.5-fold higher risk of death compared with being hospitalised with other illnesses.5 6 Mortality among children with pneumonia has been widely attributed to poverty-related factors like undernutrition.5–7
Pneumonia increases the body’s nutritional requirements to support physiological responses to the infection such as hyperventilation and hyperpyrexia.8 Coupled with reduced food intake from poor appetite, infection increases catabolism and reduces anabolism.8 This results in weight loss and deterioration in anthropometric measurements manifesting as undernutrition.9 10 In addition, children with undernutrition have 4.5-fold higher odds of developing severe infections such as pneumonia,11 resulting in a vicious cycle of infections and malnutrition.8 12 In 2020, approximately 22% of children under 5 years were estimated to be stunted, and 6.7% were wasted globally.13 Moreover, those with moderate and severe malnutrition have significantly higher odds of death compared with well-nourished children (OR of 1.1–36.6 for moderate and 2.5–15.1 for severe malnutrition, respectively).14–16
Macronutrient interventions among moderately undernourished children have been associated with improvement in anthropometric measurements, as well as reduced likelihood of developing severe malnutrition and death.17 18 Some of the macronutrient interventions that have been explored include locally produced or processed ready-to-use supplementary food, ready-to-use therapeutic food (RUTF) and fortified superfoods like super cereals, among others.18 Prevention of malnutrition among children with pneumonia through early nutritional interventions in addition to pneumonia treatment could potentially break the vicious cycle by improving short-term and long-term clinical outcomes and anthropometric measurements.19 20 However, children managed for pneumonia without severe malnutrition (non-severely undernourished children with pneumonia) are currently not considered for adjunctive, potentially lifesaving nutritional interventions.20 The WHO and UNICEF, under the ‘treat’ element of the protect, prevent and treat framework for pneumonia treatment, recommend ‘continued’ feeding, but do not give any specific recommendations for nutritional support,21 largely due to inadequate evidence in the non-severely undernourished children.
Several studies describe mortality among children with pneumonia and often focus on mortality among severely undernourished children, while non-severely undernourished children are mostly used as comparators.14 22 However, given that pneumonia potentially increases the risk of developing severe malnutrition and death among non-severely undernourished children, there is a need to collate evidence on mortality in this population globally. Collating evidence on anthropometric outcomes of pneumonia and the potential effect of macronutrient interventions on mortality and anthropometry in this population will help provide evidence to guide policy and decision-makers in establishing effective point-of-care interventions. In this systematic review and meta-analysis, we aimed to assess the all-cause mortality among non-severely undernourished children with community-acquired pneumonia globally, their anthropometric outcomes, as well as the effect of macronutrient interventions on their anthropometry and mortality.
Methods
Protocol development
The review was conducted using a priori criteria developed following the Methodological Expectations for Cochrane Intervention Reviews guidelines23 and reported in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.24 The protocol was registered in PROSPERO (CRD42021257272: https://www.crd.york.ac.uk/PROSPERO/%23myprospero) and published.24
Summary of modifications to the protocol
Population consideration: The protocol stipulated the age group of 2 months–17 years. However, in the final review, we considered studies that had categories that extended beyond the protocol-specified range. Where studies provided the number of participants in the ranges extending beyond the review age ranges, they were included in the analysis. Such studies included those by: (1) Hooli et al16 and (2) Lazzerini et al25. However, some studies did not provide the number of participants with review eligible age who were contained within the age ranges that were outside the review age range. These studies were also included in the review: (1) studies by Ngari et al5, Bokade et al26 and Eckerle et al27 which included children aged 1–59 months and (2) the study by Negash et al28 which included children aged 0–15 years, of whom only five were aged less than 1 month.
These were considered because the age range for the review was selected to eliminate neonates who have several other explanations for their respiratory symptoms beyond pneumonia. However, the neonatal age range is considered as <28 days by most scholars and <2 months by a few due to the similar physiology. In the final analysis, we included articles with children aged 28 days–17 years.
Primary review question
What is the all-cause mortality among non-severely undernourished children (28 days–17 years old) with community-acquired pneumonia globally?
Secondary review questions
The secondary review question included: (1) what is the mean change in anthropometric outcomes, that is, mid-upper-arm-circumference (MUAC), weight-for-height z-score (WHZ), weight-for-age z-score (WAZ), body mass index-for-age z score (BAZ) or height-for-age z-score (HAZ), among non-severely undernourished children with pneumonia? (2) What is the mean change in anthropometric status, that is, MUAC, WHZ, WAZ, BAZ and HAZ, among non-severely undernourished children with community-acquired pneumonia who received a macronutrient supplement compared to those who did not receive it globally? and (3) what is the all-cause mortality among non-severely undernourished children (28 days–17 years old) with community-acquired pneumonia who received a macronutrient supplement compared to those who did not receive?
Article search
Data sources
We conducted a comprehensive search of electronic databases: PubMed, Medical Literature Analysis and Retrieval System Online, EMBASE, Web of Science, Cochrane Central, Google Scholar and Scopus. Additionally, articles were obtained from bibliographies. The search was initially conducted in March 2023 and updated in July 2024.
Search terms
The search combined controlled vocabulary and natural language searching reflecting the concepts of pneumonia or lower respiratory tract infection, nutritional supplementation and children. A complete search string is available in online supplemental appendix 1.
Supplemental material
Search strategy
AAK conducted the article search from different sources. A hand search of reference lists of included articles was done. There was no language restriction applied during the article search. Non-English articles were translated using Google Translate.
Eligibility criteria for inclusion of articles in the review
Included articles fulfilled the following criteria: (1) articles published from January 2000 to July 2024 in peer-reviewed journals. This is because developing uniform ways for the diagnosis of malnutrition started in the early 2000s29; (2) articles that report on mortality and change in anthropometric status (anthropometry), such as MUAC, WHZ, WAZ or HAZ, among non-severely undernourished children with pneumonia; (3) articles that report macronutrient supplementation such as RUTF or enriched food given for any duration. We excluded articles from studies done exclusively among severely undernourished children defined using z-scores <−3 and/or MUAC <11.5 cm, neonates (children under 28 days) and among children with hospital-acquired pneumonia. Studies involving adults only, categorised as case studies, commentaries, editorials or reviews, as well as those that did not explore the relationship between nutritional status and treatment outcomes, were also excluded. Furthermore, studies that focused on micronutrient nutritional interventions were excluded.
The definition of non-severe undernutrition in this review was based on the WHO classification for (1) wasting; WHZ z-score >−3 SD or MUAC >11.5 cm, 2) stunting; HAZ z-score >-3 SD, or 3) underweight; WAZ z-score >−3 SD for the reference population. Children with z scores between −3SD and −2SD and MUAC between 11.5 and 12.5 were considered moderately undernourished.30 31 Children with z scores above −2SD and MUAC above 12.5 cm were considered well-nourished. Pneumonia diagnosis was considered in all children who had clinical symptoms and signs of acute lower respiratory tract infections as defined by WHO: the presence of fast breathing, with or without chest indrawing, and general danger signs20 with or without radiological or aetiological confirmation.
Article selection criteria
The articles identified through the search were imported into Endnote reference manager V.20 (Thomson Reuters, 2015), and duplicates were removed. The endnote file of the articles was then transferred to EPPI-Reviewer Software (UK) V.4.0. Title/abstract and full-text screening was then done. Three pairs of reviewers (DN, LN, TK, GK, RA and MO) independently screened the articles using predetermined criteria, and any disagreement between the reviewers was resolved through discussion and consensus.
Data extraction
The tool for data abstraction was developed in an Excel spreadsheet and pretested.24 Data were extracted on the following areas: (1) administrative information, (2) the study design, (3) study population and comparators if any, (4) anthropometric status and its assessment method, (5) intervention used if any, (6) mortality proportions and (7) anthropometric status outcome. For each randomised control trial, overall mortality was considered if the intervention did not affect mortality. The mortality in the control arm was considered if the intervention affected mortality. Abstraction was done in pairs and discrepancies were resolved through discussion and consensus. Corresponding authors were contacted directly to request further information on articles that were missing relevant information.
Outcomes
The primary review outcome
All-cause mortality among non-severely undernourished children with community-acquired pneumonia.
The secondary outcomes
Mean change in anthropometric outcomes among non-severely undernourished children with pneumonia globally.
Mean change in anthropometric outcomes among non-severely undernourished children with community-acquired pneumonia who received a macronutrient supplement compared with those who did not receive it globally.
Effect of macronutrient supplementation on all-cause mortality among non-severely undernourished children with community-acquired pneumonia who received a macronutrient supplement compared with those who did not receive it globally.
Risk of bias assessment
The risk of bias was assessed independently by RA and MO using the risk of bias in non-randomised follow-up studies of exposure32 for observational studies and the revised Cochrane risk of bias assessment tool for randomised studies (RoB 2.0).33
Heterogeneity assessment
The level of heterogeneity in the articles was established using the χ2 test and Cochran’s Q statistic, with a level of significance of 5%.34 The I2 statistic was used to indicate percentage (%) heterogeneity that is attributed to between-study variance.35 Interpretation of the I2 statistic is: I2 = 25% (small heterogeneity), I2 = 50% (moderate heterogeneity) and I2 = 75% (large heterogeneity).36 Subgroup analysis was conducted. The subgroups included: (1) region where studies were conducted, (2) nutritional status (no malnutrition vs moderate malnutrition) and (3) macronutrient intervention received.
Publication bias
Publication bias, which is the influence of research findings on the probability of a study being published, was assessed by plotting funnel plots and using the symmetry of the plots to detect the likelihood of publication bias among the included articles. The articles were adjusted for publication bias using the trim-and-fill method.37
Data analysis and synthesis
Data were analysed using narrative and quantitative synthesis methods. Statistical software STATA V.18.0 (StataCorp LLC, College Station, Texas, USA) was used for quantitative synthesis. Categorical data from the individual studies were summarised as frequencies and percentages and numerical continuous data as mean and SD or median and IQR. Quantitative synthesis was done using the maximum likelihood random-effects model in STATA V.18.0. For the calculation of overall effect estimates, we considered a CI=95% and α=0.05. A p value <0.05 was considered statistically significant for all hypothesis testing. The review data were pooled and the outputs were displayed in forest plots. Heterogeneity in the included studies was assessed. Where there was high heterogeneity, subgroup analysis was conducted to establish the source of the heterogeneity. The effect size measure was a proportion, which was assessed as a logit-transformed proportion using the ‘logitprop’ and ‘meta_es’ commands.
Patient and public involvement statement
Patients and the public were not involved in the design or conduct of this study.
Results
Description of included studies
The search of electronic databases identified 3726 citations, while bibliography searches and researcher/author recommendations identified 12 more citations. Of these, 642 duplicates were identified and removed, 2860 were eliminated at title-abstract screening due to not meeting the review criteria and 236 were eligible and included for full-text review. The full-text screening identified 15 eligible citations which were retrieved for data abstraction (figure 1).
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. ICU, intensive care unit.
Characteristics of included studies
The review included a total of 15 articles including prospective or retrospective cohort studies (n=11)5 16 25–28 38–42 and cross-sectional (n=2)43 44 and randomised control trials (n=2).45 46 The majority of the studies (11/15, 73%) were conducted in sub-Saharan Africa,16 25 27 28 38 39 41 43–45 47 while the rest were conducted in Asia (4/15, 27%).26 40 42 46 All the studies except one42 recruited participants in hospital, and the majority of the longitudinal studies followed up participants until death or discharge (11/15, 73%).16 25–28 38–41 43 44 Four articles reported that study participants were followed up after hospital discharge. Duration of follow-up was reported as 2 weeks,42 3 months,46 6 months45 or 1 year.5 The majority of studies (12/15, 80%) were conducted exclusively among children under 5 years.16 25–27 38 40–44 46 47 Some studies (3/15, 20%) reported providing nutritional interventions to participants.38 45 46 The majority of the studies (10/15, 67%), used WAZ in isolation or along with other anthropometric indices to assess nutritional status among included participants.5 16 25 26 28 39 40 42 45 46 MUAC and WHZ were used in isolation or along with other anthropometric indices in some (8/15, 53%) of included studies5 16 27 39 41–43 45 (table 1). Of the 15 studies, 8 reported comorbidities among the study participants.16 27 38 40 41 43–45 The comorbidities reported include malaria,38 41 diarrhoea,38 41 congestive heart disease,40 41 HIV,16 41 43–45 anaemia,27 45 sepsis41 and neurological conditions.41
Characteristics of included studies
Characteristics of excluded studies
A total of 219 articles were excluded at full text. The main areas for exclusion included irrelevant exposure (n=38), irrelevant outcomes (n=33), study design (n=45) and no full text (n=47) (figure 1).
Heterogeneity in the included studies
Studies were combined and assessed for heterogeneity. There was high heterogeneity among all included studies (I2=99.24%, p<0.000). A high heterogeneity was observed among studies conducted in sub-Saharan Africa (I2=98.02%, p<0.000) as well as those conducted in Asia (I2=98.41%, p<0.000). Similarly, high heterogeneity was observed in studies where children received macronutrient supplementation (I2=95.80%, p<0.000) and in studies among children with moderate undernutrition (I2=89.50%, p<0.000).
Risk of bias assessment
In the majority of observational studies, 62% (8/13) had a high risk of bias, 31% (4/15) had a moderate risk of bias and 7% (1/13) had a low risk of bias. Following the risk of bias assessment criteria used, most observational studies had a potential risk of confounding (8/13), missing data (4/13), postexposure interventions (2/13) and selection of studies into the study (2/13) (online supplemental appendix 2).
Supplemental material
The randomised control trials had an unclear risk of bias in the blinding of participants and personnel and in the outcome assessment (online supplemental appendix 2).
Publication bias assessment
There was evidence of publication bias on visual inspection of the funnel plot for the primary outcome (figure 2).
Publication bias assessment.
Mortality among non-severely undernourished children with community-acquired pneumonia
The prevalence of pneumonia among children in the included studies was 90.8% (154 233/169 901). Of the 154 233 participants with pneumonia evaluated in the included studies, 8377 (5.4%) died. The overall mortality for non-severely undernourished children with pneumonia from the included studies was 3.0% (95% CI 2.0% to 5.0%), with a 1–13% range (figure 3). The all-cause mortality was higher among children in studies conducted in sub-Saharan Africa, 4% (95% CI 2 to 6), with a range of 1–13%, compared with those conducted in Asia, 2% (95% CI 0% to 6%), with a range of 1–7% (figure 3). The overall in-hospital all-cause mortality from the included studies was 4% (95% CI 2% to 7%) (figure 4) with a 1–13% range. For studies that reported both in-hospital and postdischarge mortality between 2 weeks and 1 year, the overall mortality was 1% (95% CI 0% to 2%), figure 4, with a mortality range between 1–4%.
Overall and regional all-cause mortality among children with non-severe malnutrition and pneumonia. ML, maximum likelihood.
The overall and regional all-cause mortality in-hospital and following discharge among children with pneumonia who are not severely undernourished. ML, maximum likelihood.
Overall mortality among children with moderate malnutrition and pneumonia was higher compared with well-nourished children at 9% (95% CI 6% to 13%) (figure 5) with a range of 3–19%. Mortality in Asia was 7% (95% CI 2% to 22%), with a range of 3–16%, while that in Sub-Saharan Africa was 9% (95% CI 6% to 10%) (figure 5) with a range of 3–19%.
The overall and regional all-cause mortality among children with moderate malnutrition and pneumonia. ML, maximum likelihood.
Mortality among non-severely undernourished children with community-acquired pneumonia who received macronutrient supplementation
In three of the included studies, children received nutritional interventions during hospitalisation or at home. The interventions included an enriched milk diet (50), a high-energy food containing 100 kcals per 100 mL (36) or RUTF for 2 months.45 However, the intervention was randomised in only one of the studies,45 while all participants in the two observational studies received the nutritional intervention. In the randomised control trial, RUTF did not reduce mortality among children 6 months–12 years with severe pneumonia (6-month mortality of 15 children in the intervention vs 18 in the control arm). The all-cause mortality among children with non-severe malnutrition who received a macronutrient intervention was 3% (95% CI 1% to 11%), which was no different from that among those who did not receive a nutritional intervention, 3% (95% CI 1% to 5%) (figure 6). Mortality in the studies with nutritional interventions ranged between 1–10%, while the range for studies without nutritional intervention was 1–13%.
The all-cause mortality among children with non-severe malnutrition and pneumonia who received macronutrient interventions versus those who did not receive interventions.
Mean change in anthropometric outcomes among non-severely undernourished children with pneumonia
Of the children with pneumonia, 32% (49 411/154 233) had undernutrition based on one or more anthropometric measures: 82% were underweight, 17% had low MUAC and 0.4% had low weight-for-height. The overall baseline prevalence of non-severe malnutrition among children with pneumonia was 73.5% (114 996/154 233). Only one study assessed the children’s anthropometry beyond admission.45 In this study, there was minimal mean change in anthropometric outcomes following a pneumonia episode (in the absence of nutritional intervention, ie, in the control arm), (MUAC-for-age: 0.0±1.1, 0.2±1.1 and 0.2±1.1, WHZ: −0.0±1.3, 0.2±1.3 and 0.3±1.2 and Skinfold-for-age: 0.0±1.1, 0.2±1.1 and 0.3±1.0 at 28 days, 90 days and 180 days of follow-up, respectively).45
Discussion
This systematic review and meta-analysis assessed the all-cause mortality and anthropometric outcomes among non-severely undernourished children with pneumonia, as well as the effect of macronutrient interventions on these outcomes. The findings suggest that 1–13% of non-severely undernourished children die of pneumonia globally and that moderately undernourished children have higher all-cause mortality at 3–19% compared with well-nourished children. Anthropometry of children with non-severe malnutrition and pneumonia is not routinely assessed at or after discharge from the hospital. These findings have important clinical and public health implications in highlighting the vulnerability of non-severely undernourished children with pneumonia.
Although non-severely undernourished children with pneumonia have better outcomes than severely undernourished children, among whom the reported in-hospital mortality is 40.9% (range 14.7–69.9%),22 a significant number of them die. The summary mortality in non-severely undernourished children with pneumonia reported in our meta-analysis is lower than that documented in other studies (11–18.3%),14 22 probably because only mortality in the moderately undernourished group was reported, while we report mortality in both moderately undernourished and well-nourished children with pneumonia. The higher all-cause mortality in moderately undernourished (3–19%) compared with that of all non-severely undernourished children with pneumonia (1–13%) found in this review is in keeping with a recent systematic review and network meta-analysis by Kirolos et al which demonstrated that the risk of pneumonia mortality increased with increasing severity of undernutrition with an OR of 2.0 (95% CI 1.6 to 2.6) and 4.6 (95% CI 3.7 to 5.9) in the moderately and severely underweight children when compared with the well-nourished children.22
Pneumonia increases children’s risk of developing moderate and severe forms of undernutrition via cachexia because of increased metabolic demands to meet physiological and immunological responses such as increased work of breathing, fever and production of immune cells, antibodies and acute phase proteins.8 12 48 49 Irrespective of nutrition status at the time of hospitalisation with severe pneumonia, the downstream probability of the development of moderate and severe forms of undernutrition significantly also increases the risk of acquiring more pneumonia episodes.11 14 16 22
Most of the studies included in the review followed up outcomes of children until discharge or death (in-hospital mortality), while four studies followed up the children for longer episodes, that is, 2 weeks, 3 months, 6 months and 1 year. In-hospital mortality is often attributed to delayed presentation to the hospital, poor quality of care received and severity of disease.50 Undernutrition increases the odds of acquiring severe forms of pneumonia 4.5-fold,11 which may be reflected in the in-hospital mortality. However, the long-term risk of mortality following acute illness is best captured by assessing postdischarge mortality and is often missed by in-hospital mortality data.50 Ngari et al found a higher mortality among children previously admitted for severe pneumonia compared with other diagnoses, HR of 2.5 (95% CI 1.2 to 5.3). The 1-year postdischarge mortality following a pneumonia episode was 3.1%, and 52% (95% CI 37% to 63%) of postdischarge mortalities were attributable to low mid-upper arm circumference.5 In a recent systematic review, the 6-month postdischarge mortality among children with pneumonia was 4.3%.51 This suggests that the currently reported pneumonia deaths attributable to undernutrition are likely underestimated.
In this review, only one of the included studies assessed anthropometric measurements following a pneumonia episode.45 Weight loss is expected to follow an episode of severe acute illness like pneumonia.9 10 Anthropometric assessments, especially length and mid-upper arm circumference, are often missed in routine clinical care for children without obvious signs of wasting.22 This implies that children with moderate forms of undernutrition may only be identified on subsequent visits when they have progressed to more obvious and severe forms of malnutrition, at which point the risk of mortality has increased exponentially. Routine nutritional supplementation to children with severe pneumonia could potentially mitigate this shortfall and reduce the risk of death in this population. Establishing the extent to which children develop undernutrition following pneumonia episodes is a critical step in the consideration of nutritional interventions to avert preventable mortality from pneumonia. Studies documenting nutritional (anthropometric) outcomes among non-severely undernourished children with pneumonia are recommended.
This meta-analysis did not find any significant difference in mortality between studies where children received nutritional interventions and those where they did not. This could be due to the small number of studies that provided nutritional supplementation and the fact that interventions were given for a very short period (only during hospitalisation or up to 2 months), which may be inadequate for complete recovery of the child’s diminished reserves. Additionally, two of the studies with macronutrient interventions were not designed to show an effect of the nutritional intervention on outcome. Since all children received the nutritional intervention, there was no comparison group. The study from the Democratic Republic of Congo compared mortality before and after an improvement in dietary management and showed a significant reduction in mortality in children with pneumonia.38 However, it is not clear if this effect was present as well, in children with moderate malnutrition, as they were not reported separately. In the randomised controlled trial by Kiguli et al, the macronutrient intervention did not reduce mortality among children hospitalised with severe pneumonia.45 This suggests that perhaps macronutrient interventions could be reserved for higher-risk children with pneumonia, that is, those with moderate malnutrition. During hospitalisation, children often have anorexia and cannot ingest the energy required to match metabolic needs and replenish reserves.12 Furthermore, restoration of lost fat and muscle reserves is likely to take longer following discharge and require more energy than the daily recommended allowance available in the readily available food consumed at home following discharge. Macronutrient interventions that suit the unique nutritional requirements of this population should therefore be further explored.
Limitations
This review had some potential limitations including the following: (1) missing data, which could have affected the results. However, all the authors of the articles with missing data were contacted. Additionally, the missing data were not on the primary outcome. (2) There was a high level of heterogeneity and risk of bias in the included articles. This still remained considerably high in the subgroup analysis (regions, nutritional status and nutritional supplementation), which was done using a random-effects model. The use of narrative synthesis, including a range in addition to the pooled value, provides a more accurate reflection of the effect measure in the included articles. (3) Publication bias was addressed through searching in multiple databases (over 5), including bibliography screening. (4) Only English articles were identified from the searches, so relevant articles could have been left out. However, the impact of this was minimised through the comprehensive search of the articles which was employed in the review. (5) Only one study compared mortality rates of those who did versus did not receive macronutrients, and its results were inconclusive. The other studies where nutritional interventions were used did not have control groups. As such, the foods more suited for children with severe infections like pneumonia to optimise outcomes, as well as the dose, duration, effectiveness, efficacy and cost-effectiveness remain in question. (6) In this review, the findings would have been better understood if mortality was segregated by aetiology. However, the current WHO guidelines for the diagnosis of paediatric pneumonia do not allow for disaggregation by aetiology.
Conclusion
In this systematic review and meta-analysis, we found that 1–13% of non-severely undernourished children with pneumonia die globally. There is inadequate follow-up nutritional assessment and management of non-severely undernourished children with pneumonia. There are limited data on macronutrient supplementation. The only interventions considered are milk-based diets and high-energy food for varying durations. Future prospective studies evaluating anthropometric outcomes and the impact of macronutrient interventions on mortality and anthropometry, especially among moderately undernourished children, are recommended.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. Not applicable.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the School of Medicine Research Ethics Committee (REC REF No. 2021-138).
Acknowledgments
We acknowledge the support of other members of the Africa Centre of Systematic Reviews and Knowledge Translation, including Geoffrey Kanalwa (GK), Robert Apunyo (RA) and Teddy Kabajurizi (TK), who participated in the review.
References
Footnotes
X @NalwangaDamalie, @MosesOcan
Contributors DN, KM, CK, AB, VM and MO conceptualised the study. AAK performed the article search. DN and LN led the screening and data extraction, quality assessment and data analysis and DN drafted the manuscript. All authors read, reviewed and approved the final manuscript. DN is the guarantor.
Funding Research reported in this publication was supported in part by the government of Uganda through the Makerere University Research and Innovations Fund (RIF1/CAES/030) and the EDCTP2 programme supported by the European Union (RIA2016S-1636-COAST-Nutrition, TMA2020CDF-3198). The content is solely the responsibility of the authors and does not necessarily represent the official views of EDCTP.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.