Oral vitamin A supplements to prevent acute upper respiratory tract infections in children up to seven years of age

Xiao Cheng; Dan Li; Chunsong Yang; Bin Chen; Ping Xu; Lingli Zhang

doi:10.1002/14651858.CD015306.pub2

. 2024 May 13;2024(5):CD015306. doi: 10.1002/14651858.CD015306.pub2

Oral vitamin A supplements to prevent acute upper respiratory tract infections in children up to seven years of age

Xiao Cheng ^1,^2,^3,⁴, Dan Li ^5,⁶, Chunsong Yang ^1,^2,^3,⁴, Bin Chen ⁷, Ping Xu ⁸, Lingli Zhang ^1,^2,^3,^4,^9,^✉

Editor: Cochrane Acute Respiratory Infections Group

PMCID: PMC11089595 PMID: 38738639

Abstract

Background

According to global prevalence analysis studies, acute upper respiratory tract infections (URTIs) are the most common acute infectious disease in children, especially in preschool children. Acute URTIs lead to an economic burden on families and society. Vitamin A refers to the fat‐soluble compound all‐trans‐retinol and also represents retinol and its active metabolites. Vitamin A interacts with both the innate immune system and the adaptive immune system and improves the host's defences against infections. Correlation studies show that serum retinol deficiency was associated with a higher risk of respiratory tract infections. Therefore, vitamin A supplementation may be important in preventing acute URTIs.

Objectives

To assess the effectiveness and safety of vitamin A supplements for preventing acute upper respiratory tract infections in children up to seven years of age.

Search methods

We searched CENTRAL, MEDLINE, Embase, the Chinese Biomedical Literature Database, and two trial registration platforms to 8 June 2023. We also checked the reference lists of all primary studies and reviewed relevant systematic reviews and trials for additional references. We imposed no language or publication restrictions.

Selection criteria

We included randomised controlled trials (RCTs), which evaluated the role of vitamin A supplementation in the prevention of acute URTIs in children up to seven years of age.

Data collection and analysis

We used the standard methodological procedures expected by Cochrane.

Main results

We included six studies (27,351 participants). Four studies were RCTs and two were cluster‐RCTs. The included studies were all conducted in lower‐middle‐income countries (two in India, two in South Africa, one in Ecuador, and one in Haiti). Three studies included healthy children who had no vitamin A deficiency, one study included children born to HIV‐infected women, one study included low‐birthweight neonates, and one study included children in areas with a high local prevalence of malnutrition and xerophthalmia. In two studies, vitamin E was a co‐treatment administered in addition to vitamin A. We judged the included studies to be at either a high or unclear risk of bias for random sequence generation, incomplete outcome data, and blinding.

Primary outcomes

Six studies reported the incidence of acute URTIs during the study period. Five studies reported the number of acute URTIs over a period of time, but there was population heterogeneity and the results were presented in different forms, therefore only three studies were meta‐analysed. We are uncertain of the effect of vitamin A supplementation on the number of acute URTIs over two weeks (risk ratio (RR) 1.00, 95% confidence interval (CI) 0.92 to 1.09; I² = 44%; 3 studies, 22,668 participants; low‐certainty evidence). Two studies reported the proportion of participants with an acute URTI. We are uncertain of the effect of vitamin A supplementation on the proportion of participants with an acute URTI (2 studies, 15,535 participants; low‐certainty evidence). Only one study (116 participants) reported adverse events. No infant in either the placebo or vitamin A group was found to have feeding difficulties (failure to feed or vomiting), a bulging fontanelle, or neurological signs before or after vitamin A administration (very low‐certainty evidence).

Secondary outcomes

Two studies (296 participants) reported the severity of subjective symptoms, presented by the mean duration of acute URTI. Vitamin A may have little to no effect on the mean duration of acute URTI (very low‐certainty evidence).

Authors' conclusions

The evidence for the use of vitamin A supplementation to prevent acute URTI is uncertain, because population, dose and duration of interventions, and outcomes vary between studies. From generally very low‐ to low‐certainty evidence, we found that there may be no benefit in the use of vitamin A supplementation to prevent acute URTI in children up to seven years of age. More RCTs are needed to strengthen the current evidence. Future research should report over longer time frames using validated tools and consistent reporting, and ensure adequate power calculations, to allow for easier synthesis of data. Finally, it is important to assess vitamin A supplementation for preschool children with vitamin A deficiency.

Keywords: Child; Child, Preschool; Humans; Infant; Acute Disease; Administration, Oral; Bias; Dietary Supplements; Randomized Controlled Trials as Topic; Respiratory Tract Infections; Respiratory Tract Infections/prevention & control; Vitamin A; Vitamin A/administration & dosage; Vitamin A Deficiency; Vitamin A Deficiency/prevention & control; Vitamins; Vitamins/administration & dosage

Plain language summary

Oral vitamin A supplements to prevent acute upper respiratory tract infections in children up to seven years of age

Background

Acute upper respiratory tract infections (URTIs) are the most common acute infectious diseases in children, especially in preschool children. Vitamin A has a positive role in the immune system and may improve the host's defences against infections. Studies show that children with vitamin deficiency are more likely to suffer from respiratory tract infections. Therefore, we assessed the role of vitamin A supplementation in the prevention of acute upper respiratory tract infections in preschool children (up to seven years of age).

Review question

What is the role of vitamin A supplementation in the prevention of acute URTIs in preschool children (up to seven years of age), compared to no supplementation?

Search date

We searched for evidence up to 8 June 2023.

Study characteristics

All included studies were conducted in lower‐ and middle‐income countries (two in India, two in South Africa, one in Ecuador, and one in Haiti). Three studies included healthy children who had no vitamin A deficiency, one study included children born to HIV‐infected women, one study included low‐birthweight neonates, and one study included children in areas more likely to experience malnutrition and dry eye. In two studies, vitamin E was a co‐treatment administered in addition to vitamin A.

Key results

We included six studies involving 27,351 participants. We have very low to low confidence in the evidence provided by the studies.

Five studies reported the number of acute URTIs over a period of time. The included studies varied by population and the results were presented in different forms, so we could only combine three studies together (meta‐analysis). We are uncertain if vitamin A supplementation lowers the number of acute URTIs over a two‐week period (three studies, 22,668 participants). Two studies reported the proportion of participants with acute URTI. The effect of vitamin A supplementation on the proportion of participants with acute URTI is uncertain (two studies, 15,535 participants). Only one study (116 participants) reported adverse events and none occurred in infants in the placebo or vitamin A groups.

Two studies (296 participants) reported the severity of subjective symptoms, presented by the average duration of acute URTIs. Vitamin A may have little to no effect on the average duration of acute URTIs.

Authors' conclusions

From evidence in which we have very low to low confidence, we found that there may be no benefit in using vitamin A supplementation to prevent acute URTIs in children up to seven years of age. High‐quality studies focusing on preschool children with vitamin A deficiency are needed to identify whether vitamin A supplementation may be effective.

Summary of findings

Summary of findings 1. Summary of findings table ‐ Vitamin A supplements compared to placebo for preventing acute URTIs in children up to seven years of age.

Vitamin A supplements compared to placebo for preventing acute URTIs in children up to seven years of age
Patient or population: children up to the age of 7 at risk of upper respiratory tract infection Setting: hospital, community, or primary care population Intervention: vitamin A supplements Comparison: placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo	Risk with vitamin A supplements
Incidence of acute URTIs during the study period (the number of acute URTIs over a period of time)	501 per 1000	501 per 1000 (461 to 546)	RR 1.00 (0.92 to 1.09)	22668 (3 RCTs)	⊕⊕⊝⊝ Low^a,^b	Vitamin A supplements do not reduce or increase the incidence of acute URTIs. We compared the prevalence of acute URTIs in the vitamin A and placebo groups. The unit for the denominator was a 2‐week period. For the risk with placebo, the event number was 6436 and the total was 12,042. For the risk with vitamin A, the event number was 5322, and the total was 10,626.
Incidence of acute URTIs during the study period (proportion of participants with an acute URTI)	There were 56 and 60 children in the treatment and control group, respectively, and the numbers of children suffering from acute URTI were 43 and 46, respectively. At the 3‐month follow‐up, the numbers of children with acute URTI were 17 and 25, respectively. At the 6‐month follow‐up, they were 31 and 32, respectively. At the 9‐month follow‐up they were 39 and 44, respectively (Coutsoudis 2000). There were 7764 and 7655 children in the treatment and control group, respectively, and the numbers of children suffering from acute URTI were 4196 and 4026, respectively (Rahmathullah 1991).			15535 (2 RCTs)	⊕⊕⊝⊝ Low^b,^c	Vitamin A supplements do not reduce or increase the incidence of acute URTIs.
Adverse events	No infant in either the placebo or vitamin A group was found to have feeding difficulties (failure to feed or vomiting), a bulging fontanelle, or neurological signs either before or after administration of vitamin A (Coutsoudis 2000).			116 (1 RCT)	⊕⊝⊝⊝ Very low^b,^d,^e	‐
Severity of subjective symptoms (mean duration of acute URTI)	Coutsoudis 2000 reported that the mean duration of acute URTI was similar, 6.8 days and 6.9 days, respectively. The other study showed that the mean duration of acute URTI was 5.05 ± 4.1 (mean ± standard deviation) days versus 4.91 ± 3.6 days (Biswas 1994). The mean difference was 0.14 (95% confidence interval (CI) ‐1 to 1.28).			290 (2 RCTs)	⊕⊝⊝⊝ Very low^b,^c,^e	Vitamin A supplements do not decrease the mean difference in the mean duration of acute URTI episode in children up to 7 years of age.
Time to occurrence of acute URTI ‐ not reported	‐			‐	‐	Not reported
Number of cases with complications ‐ not reported	‐			‐	‐	Not reported
Rate of hospitalisation ‐ not reported	‐			‐	‐	Not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: confidence interval; MD: mean difference; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
See interactive version of this table: https://gdt.gradepro.org/presentations/#/isof/isof_question_revman_web_443608830227828470.

Open in a new tab

^a The I2 was 44%, there may be moderate heterogeneity, and one study focused on special populations (participants were recruited from among HIV‐infected women who had attended the prenatal clinic and who delivered infants at King Edward VIII Hospital, Durban).  ^b The certainty of evidence was downgraded by one level due to study limitations (high/unclear risk of bias for sequence generation, blinding, incomplete outcome data, and selective reporting).  ^c The certainty of evidence was downgraded by one level due to large differences in population, and dose and duration of interventions, between studies.  ^d The certainty of evidence was downgraded by one level, as most studies did not report this outcome.  ^e The certainty of evidence was downgraded by one level, as the sample size failed to meet the optimal information size, taken as < 300 events for a dichotomous outcome and < 400 participants for a continuous outcome.

Background

Description of the condition

Acute upper respiratory tract infections (URTIs) refer to acute infections of the nose, pharynx, or throat caused by viruses or bacteria, or both. Acute URTIs include the common cold, acute tonsillitis, and herpetic angina. The diagnosis of typical acute URTIs is usually made by clinical presentation, and confirmatory testing is not required. Common symptoms of acute URTIs are nasal congestion, rhinorrhoea, sneezing, sore throat, cough, and fever (Guibas 2017). The symptoms disappear within 15 days in 90% of children (Thompson 2013). The pathogens that cause acute URTIs include influenza A virus (FLUA), influenza B virus (FLUB), respiratory syncytial viruses A (RSV‐A), respiratory syncytial viruses B (RSV‐B), respiratory enteroviruses (EVs), rhinoviruses (RVs), respiratory adenoviruses (ADVs), human metapneumovirus (hMPV), parainfluenza viruses (PIVs) 1 to 4, and coronaviruses (CoVs) (NL63, OC43, HKU‐1, and 229E) (Charlton 2019). At least 200 viruses can cause acute URTIs (Lehtoranta 2020).

According to a global prevalence analysis study of 310 diseases, there were approximately 17.2 billion acute URTIs worldwide in 2015 (Vos 2016). Acute URTIs are the most common acute infectious disease in children (van Driel 2018). The average incidence of acute URTIs in preschool children is five to seven per year (Turner 2015). The attendance of preschool children in daycare increases the risk of respiratory infections (Laursen 2018; Schuez‐Havupalo 2017). The definition of preschool age differs across countries and regions. In the United States, children under the age of five are preschoolers, as defined by the World Health Organization (WHO) (De Onis 2015; Leeb 2020). Most European countries define children up to six years as preschoolers. The "Compulsory Education Law of the People's Republic of China" stipulates that children over the age of seven must receive primary education. To include as many preschool children as possible globally, our review therefore focuses on children up to seven years of age.

Acute URTIs are associated with high costs, including effects on work or study, visits to healthcare providers, and drug consumption, as well as the development of antibiotic resistance. In general, URTIs result in an economic burden on families and society (Thomas 2022). The annual economic cost of coughing due to acute URTIs per child under five years old is approximately USD 991 (Lovie‐Toon 2018). Acute URTIs are self‐limiting diseases, but they significantly affect the patient's quality of life and can cause serious complications (such as pneumonia and otitis media), with children being more susceptible to complications (Charlton 2019; Incze 2018).

Description of the intervention

Vitamin A refers to the fat‐soluble compound all‐trans‐retinol, and also represents retinol and its active metabolites, including retinal, retinyl ester, and retinoic acid. Humans and animals must obtain vitamin A from the diet or through supplemental sources (Debelo 2017). Vitamin A is arguably the most multifunctional vitamin in the human body (Tanoury 2013). It is essential for maintaining epithelial function, including the respiratory epithelium, and vitamin A deficiency (serum retinol 0.70 µmol/L or lower) is associated with respiratory infections (Simkin 2016).

Vitamin A deficiency affects about 190 million preschool‐aged children. High‐dose vitamin A supplementation is recommended where the prevalence of vitamin A deficiency is 20% or higher in infants and children aged 6 to 59 months. The WHO recommends oral administration of 100,000 IU (30 mg retinol equivalent) of vitamin A for infants aged 6 to 11 months, and 200,000 IU (60 mg retinol equivalent) of vitamin A every four to six months for children aged 12 to 59 months (WHO 2011). Vitamin A supplements are associated with a significant reduction in the mortality and morbidity of children under five years of age (Mayo‐Wilson 2011). Zinc combined with vitamin A has been shown to reduce the incidence of URTIs in preschool children (Kartasurya 2012).

There are many types of vitamin A supplements, including vitamin A in oil‐based or water‐soluble formulations, and β‐carotene, usually in gelatinous capsules.

In this review, we will consider the effectiveness and safety of vitamin A taken as oral supplements, in addition to the usual diet, for preventing acute upper respiratory tract infections in children.

How the intervention might work

Vitamin A interacts with both the innate immune system and the adaptive immune system and improves the host's defences against infections (WHO 2009). It also plays an anti‐inflammatory role in repairing protective mucosal epithelium damaged by infection (Wiseman 2017).

Vitamin A regulates cytokine expression in respiratory epithelial and macrophage cell lines and is a key regulator of lung development (Fernandes‐Silva 2020; Penkert 2017). Chronic vitamin A deficiency is related to histopathological changes in the lining of the lung epithelium, which destroy normal lung physiology and predispose the host to respiratory diseases (Timoneda 2018). There is evidence to suggest a possible association between vitamin A deficiency and asthma (Allen 2009; Reifen 2015). Children with recurrent respiratory infections are more prone to severe vitamin A deficiency (Zhang 2020). Optimal vitamin A levels and vitamin A supplementation treatment have been shown to be beneficial for reducing the severity of bronchopulmonary dysplasia in extremely preterm infants (Rakshasbhuvankar 2017).

In this review, 'prevention' refers to oral vitamin A supplements in children with no signs and symptoms of acute URTI.

Why it is important to do this review

Children with URTIs account for more than 60% of paediatric outpatients, which puts a demand on many medical resources (National Bureau of Statistics of China 2018).

The acquisition of pathogenic viruses and bacteria in the upper respiratory tract can lead to lower respiratory tract infections (LRTIs). These viruses and bacteria replicate, spread to the lower respiratory tract, and then invade the mucosa, leading to inflammation and clinical disease. Therefore, URTIs can evolve into LRTIs (Tsolia 2004). Severe acute LRTIs can be a heavy burden on health services and are the leading cause of hospital referrals and admissions for young children (Nair 2013). Consequently, there is a need for methods to prevent acute URTIs.

Published Cochrane reviews suggest that vitamin A does not have a positive effect on respiratory disease (Imdad 2016; Imdad 2017). Vitamin A appears to have no significant effect on respiratory infection‐related mortality (Mayo‐Wilson 2011). However, correlation studies show that serum retinol deficiency is associated with a higher risk of respiratory tract infections (Ahmed 2006; Merera 2021; Merera 2022).

Randomised controlled trials have explored the relationship between vitamin A supplements and the prevention of acute URTIs, but there has been no systematic review investigating this issue. We have therefore conducted a systematic review to evaluate the effectiveness and safety of vitamin A supplements for preventing acute URTIs in children up to seven years of age.

Objectives

To assess the effectiveness and safety of vitamin A supplements for preventing acute upper respiratory tract infections in children up to seven years of age.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs), including cross‐over trials. We included cluster‐RCTs where participants were randomised at the cluster level, and outcomes were measured at the individual level. We included studies reported as full text, studies published only as an abstract containing key data, and unpublished studies. We excluded non‐RCTs or quasi‐RCTs because these designs introduce a high risk of bias.

Types of participants

We included children aged up to seven years, with no restriction on setting, gender, or ethnicity. We reviewed whether vitamin A is beneficial in preventing acute URTIs for both healthy and non‐healthy children without acute URTI. 'Acute URTI' refers to acute infections of the nose, pharynx, or throat caused by viruses or bacteria, or both. Acute URTIs include the common cold, acute tonsillitis, and herpetic angina. The diagnosis of typical acute URTI is usually made on clinical presentation. Sometimes a study includes some 'eligible' participants and some 'ineligible' participants. We included studies with mixed populations when information on children with acute URTIs was available separately.

Types of interventions

We included trials comparing oral vitamin A administration (any dose and any frequency) with placebo, or no supplementation, irrespective of usual dietary intake. We described the nature of the placebo used because it can be challenging to produce a placebo that looks and tastes similar to the intervention. We included co‐treatments in which micronutrients such as folic acid and zinc were present in the intervention and control groups but were not part of the randomised treatment. We considered all time frames for the duration of prevention.

Types of outcome measures

Primary outcomes

Effectiveness: incidence of acute URTIs during the study period (expressed as a proportion of participants with an acute URTI or the number of acute URTIs over a period of time, or both).
Safety: adverse events such as diarrhoea, nausea, vomiting, heartburn, abdominal symptoms, headache, insomnia, flushing, etc.

Secondary outcomes

Time to occurrence of acute URTI.
Severity of subjective symptoms such as overall symptoms, nasal congestion, sneezing, rhinorrhoea, headache, sore throat, and fever. Severity may be measured using Likert scales, visual analogue scales (VAS), or other validated instruments. We also considered measures that combined the duration and severity of specific symptoms (e.g. area under the curve of severity over time).
Number of cases with complications, e.g. acute otitis media, acute LRTIs, etc.
Rate of hospitalisation.
Number of cases of antibiotics used for URTIs.
Time to return to school.
Rate of school absenteeism.
Guardian satisfaction, as measured by validated instruments.

Search methods for identification of studies

Electronic searches

We searched the following databases from inception to 8 June 2023.

MEDLINE Ovid (1946 to 8 June 2023, Appendix 1).
Embase Ovid (1974 to 8 June 2023, Appendix 2).
Cochrane Central Register of Controlled Trials in the Cochrane Library (CENTRAL) (searched 8 June 2023, Appendix 3).
Chinese Biomedical Literature Database (searched 8 June 2023, Appendix 4).

We used the search strategy described in the appendices. We imposed no language or publication restrictions.

Searching other resources

We supplemented the electronic searches with manual searches for published, unpublished, and ongoing RCTs on the following clinical trial registration platforms.

World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (trialsearch.who.int/) (searched 8 June 2023, Appendix 5).
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov/) (searched 8 June 2023, Appendix 6).

We checked the reference lists of all primary studies and review articles for additional references.

Data collection and analysis

Selection of studies

Two review authors (XC, DL) independently screened the titles and abstracts of all studies identified by the search for potential relevance. We retrieved the full‐text article/study reports for all studies deemed potentially relevant. Two review authors (XC, DL) independently read each article/study report to identify studies for inclusion and identified and recorded reasons for the exclusion of ineligible studies. We resolved disagreements by consulting a third review author (CSY). We excluded duplicates and collated multiple reports of the same study. We documented the selection process in enough detail to complete a PRISMA flow diagram and Characteristics of excluded studies table (Moher 2009).

Data extraction and management

We designed and used a structured data extraction form for study characteristics and outcome data. Two review authors (XC, DL) extracted study characteristics from the included studies. For RCTs with a covariate adjustment, we extracted adjusted results. We extracted the following study characteristics.

Methods: study design, total duration of the study, details of any 'run‐in' period, number of study centres and location, study setting, withdrawals, and date of the study.
Participants: number, mean age, age range, gender, ethnicity, the severity of the condition, diagnostic criteria, baseline vitamin A status, inclusion criteria, and exclusion criteria.
Interventions: intervention (dosage, dosage form, manufacturer, etc.), comparison, and concomitant medications.
Outcomes: primary and secondary outcomes specified and collected, and time points reported.
Notes: trial funding, and notable conflicts of interest of trial authors.

Two review authors (XC, DL) independently extracted outcome data from the included studies. If outcome data were not reported in a useable manner, we noted this in the Characteristics of included studies table. Two review authors (XC, DL) input data into RevMan Web (RevMan Web 2022), and double‐checked the data entry. A third review author (CSY) spot‐checked the accuracy of study characteristics based on the trial reports.

Assessment of risk of bias in included studies

Two review authors (XC, DL) independently assessed the risk of bias in the included studies using the Cochrane risk of bias tool for randomised trials (RoB 1) as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022). We resolved disagreements by consulting a third review author (CSY). We assessed the risk of bias according to the following domains.

Random sequence generation.
Allocation concealment.
Blinding of participants and personnel.
Blinding of outcome assessment.
Incomplete outcome data.
Selective outcome reporting.
Other bias. In cluster‐randomised and cross‐over trials, there are biases of particular concern, such as incorrect analysis.

We classified each potential source of bias as low, high, or unclear risk. We provided a citation from the study report with a rationale for our judgement in the risk of bias table. We summarised the risk of bias judgements for each of the domains listed in the studies. We considered blinding separately for different key outcomes, where necessary. Where information on the risk of bias was related to unpublished data or communications with investigators, we indicated this in the risk of bias table. When considering treatment effects, we took into account the risk of bias from the studies that contributed to that outcome.

Assessment of bias in conducting the systematic review

We performed the analysis according to the published protocol and reported any deviations from it in the Differences between protocol and review section of this review (Cheng 2023).

Measures of treatment effect

We entered the outcome data for each study into the data tables in RevMan Web to calculate the treatment effects (Higgins 2022; RevMan Web 2022). We calculated risk ratios (RR) with a 95% confidence interval (CI) for dichotomous outcomes to assess the association between interventions and outcomes. We calculated the mean difference (MD) with 95% CI for continuous outcomes to estimate the differences between the intervention group and the control group (Higgins 2022).

The included studies reported data in different forms. Some studies reported the number of acute URTI episodes, and some studies reported the number of participants with acute URTIs. During the follow‐up period, acute URTIs may recur. Therefore, we standardised the data. Because acute URTIs usually last one week, we calculated the number of total events by multiplying the number of people by the number of weeks of follow‐up, and the target number of events by multiplying the number of people by the number of acute URTI episodes.

Unit of analysis issues

We planned to include cross‐over trials and individually randomised trials in the analyses. We would only include data from the first phase in cross‐over trials to reduce the impact of the sequelae effect. However, none of the included studies were cross‐over trials. We identified two cluster‐randomised trials, and we ensured that the analysis methods accounted for clustering in the data (Higgins 2022). Calculating the adjusted sample size of cluster‐randomised trials requires the intracluster correlation coefficient (ICC) value, but it was not reported by one of the cluster‐randomised trials we included (Rahmathullah 1991), and we did not find a suitable reference value. If multi‐arm studies were included, we planned to handle them according to the guidance in the Cochrane Handbook for Systematic Reviews of Interventions. However, none of the included studies were multi‐arm trials.

Dealing with missing data

If we identified studies published only in abstract form, we planned to contact the trial authors for full reports and data as needed. If data related to attrition rates were missing or unclear, and no reasons for withdrawal were reported, we planned to ask the trial authors to provide further information. We tried to contact the trial authors of one study to obtain missing numerical outcome data, but the email address was invalid (Coutsoudis 2000). If continuous outcome data such as standard deviations were missing, we planned to calculate them from other available statistics such as P values using the methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022). The included study Coutsoudis 2000 did not have SDs, but there were no statistics available in this study for us to calculate these.

Assessment of heterogeneity

We assessed heterogeneity between studies in two ways: first, at face value: heterogeneity in population, intervention, or outcome; and second, using the Chi² test, where a low P value (< 0.10) provided evidence of heterogeneity. However, the Chi² test has low power in the (common) situation of a meta‐analysis when studies have small sample sizes or are few. We used the I² statistic to quantify inconsistency (Higgins 2022). We reported heterogeneity and planned to explore possible causes through prespecified subgroup analysis. However, we were not able to perform subgroup analysis due to the small number of studies.

Assessment of reporting biases

We planned to assess reporting bias by constructing a funnel plot. However, we did not pool more than 10 studies for any given outcome. Therefore, reporting bias could not be assessed. For future updates, we plan to construct funnel plots to explore possible small‐study and publication biases for outcomes with more than 10 studies.

Data synthesis

We used RevMan Web to aggregate study data with clinical homogeneity (RevMan Web 2022). We included RCTs and cluster‐RCTs in the analyses. We used the random‐effects model for meta‐analysis due to the clinical and methodological heterogeneity between studies. We used the generic inverse‐variance approach in RevMan Web for effect estimates and their standard errors at the same allocation level (RevMan Web 2022). We presented the findings narratively where it was not possible to extract and pool quantitative data.

Subgroup analysis and investigation of heterogeneity

We planned to carry out the following subgroup analyses and use the Chi²test to test for subgroup interactions in RevMan Web (RevMan Web 2022). However, we were not able to perform subgroup analyses due to the small number of studies.

Possible differences in the vitamin A intervention.
1. Vitamin A dosage (< 100,000 IU, 100,000 IU to 200,000 IU, > 200,000 IU) (WHO 2019).
2. Duration of use of vitamin A (shorter than one year and one year or longer) (Chen 2008).
3. Vitamin A formulation (e.g. liquid, chewable tablet, injection).
Possible differences in the participants.
1. Gender.
2. Age (younger than five years old, and five to seven years old) (Chen 2008).
3. Underlying disease condition (e.g. dystrophy).
4. Baseline vitamin A deficiency (subclinical vitamin A deficiency: serum/plasma concentration of retinol < 0.70 μmol/L; severe vitamin A deficiency: serum/plasma concentration of retinol < 0.35 μmol/L) (WHO 2009).
With or without industry funding.

Sensitivity analysis

We planned to carry out the following sensitivity analyses.

Repeating the analysis excluding each study with a high risk of bias.
Repeating the analysis excluding published studies with abstract only or unpublished studies.
Repeating the analysis excluding studies with missing data.
Conducting the analysis using the fixed‐effect model.

However, due to the lack of data, we only performed sensitivity analysis 4.

Summary of findings and assessment of the certainty of the evidence

We produced a summary of findings table using the following outcomes:

Incidence of acute URTIs during the study period (the number of acute URTIs over a period of time)
Incidence of acute URTIs during the study period (proportion of participants with an acute URTI)
Adverse events
Severity of subjective symptoms
Time to occurrence of acute URTIs
Number of cases with complications
Rate of hospitalisation

We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of evidence (Atkins 2004). We used the methods and recommendations outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2022), using GRADEpro GDT software (GRADEpro GDT). We used footnotes to justify all decisions to down‐ or upgrade the certainty of evidence and made comments to aid the reader's understanding of the review where necessary.

Results

Description of studies

For study details, see Characteristics of included studies and Characteristics of excluded studies tables.

Results of the search

The searches of the four databases (see Electronic searches) retrieved 1925 records. Our searches of the trial registers identified 16 further studies. Our screening of the reference lists of the included publications revealed 12 additional RCTs. We identified a total of 1953 records.

Once duplicates had been removed, we had a total of 1786 records. We excluded 1771 records based on titles and abstracts. We obtained the full text of the remaining 15 records. We excluded nine studies (Chen 2013; Donnen 1998; Donnen 2007; Kartasurya 2012; Long 2006; Shaker 2018; Swami 2007; Venkatarao 1996; Vijayaraghavan 1990).

We included six studies reported in six references. For a further description of our screening process, see the study flow diagram (Figure 1).

Included studies

We included six studies (27,351 participants) in this review (Biswas 1994; Coutsoudis 1995; Coutsoudis 2000; Rahmathullah 1991; Sempértegui 1999; Stansfield 1993).

Study design

All included studies were randomised, placebo‐controlled, double‐blind trials. Two of the studies were cluster‐RCTs.

Participants

The included studies involved participants in different age groups. Biswas 1994 included children aged 12 to 71 months of age. Sempértegui 1999 included children aged 6 to 36 months (mean age 20.1 months). Coutsoudis 1995 included neonates. Coutsoudis 2000 included infants of gestational age < 36 weeks. Rahmathullah 1991 included children aged 6 to 60 months. Stansfield 1993 included children aged 6 to 83 months. Three studies included healthy children who had no vitamin A deficiency (Biswas 1994; Rahmathullah 1991; Sempértegui 1999), one study included children born to HIV‐infected women (Coutsoudis 1995), one study included low‐birthweight neonates (Coutsoudis 2000), and one study included children in areas with a high local prevalence of malnutrition and xerophthalmia (Stansfield 1993). The sample sizes varied widely, from 116 (Coutsoudis 2000) to 15,419 (Rahmathullah 1991).

Settings

Two studies were conducted in hospital environments (Coutsoudis 1995; Coutsoudis 2000), and the remaining studies were conducted in community populations.

Interventions

Biswas 1994 gave children a single dose of vitamin A 200,000 IU (2 mL retinyl palmitate in oil). Coutsoudis 1995 gave children vitamin A six times in 15 months (at one and three months, 50,000 IU, at six and nine months, 100,000 IU, and at 12 and 15 months, 200,000 IU of retinyl palmitate). Infants in Coutsoudis 2000 received 25,000 IU of vitamin A (retinyl palmitate) on study days 1, 4, and 8. Sempértegui 1999 gave children vitamin A 10,000 IU once a week (3000 mg of retinol in 0.2 mL of syrup). Rahmathullah 1991 gave children 8250 IU of vitamin A (retinyl palmitate) and 46 μmol of vitamin E once a week for 52 weeks. Sempértegui 1999 gave children vitamin A 10,000 IU once a week (3000 mg of retinol in 0.2 mL of syrup). Stansfield 1993 gave children vitamin A megadose supplements (infants 6 to 11 months of age 100,000 IU and older children 200,000 IU) and 40.6 mg of vitamin E every four months, for three cycles.

Controls/comparators

Four trials administered a placebo (Biswas 1994; Coutsoudis 1995; Coutsoudis 2000; Sempértegui 1999). In Rahmathullah 1991 and Stansfield 1993, the control group had the same dose of vitamin E as the intervention group.

Outcomes

Primary outcomes

Six of the studies measured and reported the primary outcome (incidence of acute URTIs during the study period, expressed as the number of acute URTIs over two weeks or a proportion of participants with an acute URTI) (Biswas 1994; Coutsoudis 1995; Coutsoudis 2000; Rahmathullah 1991; Sempértegui 1999; Stansfield 1993). One study reported adverse events (Coutsoudis 2000).

Secondary outcomes

Two of the studies measured and reported the secondary outcome (severity of subjective symptoms) (Biswas 1994; Coutsoudis 2000). None of the studies reported the time to occurrence of acute URTI, number of cases with complications, rate of hospitalisation, number of cases of antibiotics used for URTIs, time to return to school, rate of school absenteeism, or guardian satisfaction.

Funding

All studies reported details of support and sponsorship. Only one study received funding from industry/commercial funding. Biswas 1994 was supported by a research grant from the Indian Association of Preventive and Social Medicine. Coutsoudis 1995 was funded in part by grants from the Medical Research Council; the University of Natal, Faculty of Medicine Research Fund; the Fogarty International Center (grant TW00231); and the National Institute of Mental Health to the HIV Center for Clinical and Behavioral Studies (grant P50‐MH43520). Coutsoudis 2000 was supported by grants from Gerber‐Purity, the Medical Research Council, and the Faculty of Medicine, University of Natal. For Rahmathullah 1991, vitamin A, placebo, and dispenser bottles were provided by the Sight and Life Program, Hoffmann‐La Roche. Sempértegui 1999 was supported by a grant from the Pan American Health Organization. The US Agency for International Development (USAID) mission in Port‐au‐Prince, the USAID Office of Nutrition, Eye Care/Haiti’s PROVAX project, and Helen Keller International (through its Vitamin A Technical Assistance Project) offered financial and other support to Stansfield 1993.

Excluded studies

We excluded nine studies following a full‐text review. Three studies evaluated the effect of vitamin A supplementation on acute respiratory infection but did not distinguish acute LRTI and acute URTI (Swami 2007; Venkatarao 1996; Vijayaraghavan 1990). One study evaluated the effect of vitamin A supplementation on respiratory‐related illnesses but did not mention acute URTI (Chen 2013). One study included paediatric participants aged between two and 12 years (Shaker 2018). One study reported the morbidity of cough with fever but did not define acute URTI (Long 2006). One study reported on the effect of vitamin A supplementation on treatment, not prevention (Donnen 1998). One study assessed different dosages of vitamin A without a placebo or no intervention group (Donnen 2007). One study did not meet the criteria for randomisation because participants were not randomly assigned to the vitamin A or control group (Kartasurya 2012).

Risk of bias in included studies

See Figure 2 and Figure 3 for a summary of the risk of bias assessments for all included trials and individual ratings for each trial. Full descriptions and the review authors' justifications for assigned ratings are included in the risk of bias tables within the Characteristics of included studies table.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

Sequence generation

We judged Coutsoudis 2000 at unclear risk of bias because the study only mentioned that the infants were randomly allocated to a vitamin A or placebo group, and did not provide enough details. We rated Stansfield 1993 at high risk of bias due to 55% of children being assigned to the vitamin A group. It was reasonable to suspect that the probability of assignment to the intervention and control groups was not equal. We judged other studies to be at low risk of bias for sequence generation (Biswas 1994; Coutsoudis 1995; Rahmathullah 1991; Sempértegui 1999).

Allocation concealment

We rated all studies except for Coutsoudis 2000 and Rahmathullah 1991 at low risk of bias, because all studies provided enough detail to fully describe the procedures used to hide the assignment group from the researchers. We judged Coutsoudis 2000 and Rahmathullah 1991 at unclear risk of bias because the studies gave insufficient detail about allocation concealment.

Blinding of participants and personnel

All studies were at low risk of bias and provided sufficient detail to adequately assess the procedures used to blind participants and personnel.

Blinding of outcome assessment

All studies were at low risk of bias and provided sufficient detail to adequately assess the procedures used to blind outcome assessors.

Incomplete outcome data

We rated Coutsoudis 1995 and Coutsoudis 2000 at high risk of bias due to per‐protocol analysis or loss to follow‐up. Other studies were at low risk of bias: reasons for withdrawals were explained and intention‐to‐treat analyses were used.

Selective reporting

We judged Sempértegui 1999 at high risk of bias because "Analysis of AURI in children who took vitamin A or placebo by baseline nutritional status", as planned in the methods, should have been done. We considered Rahmathullah 1991 to be at high risk of bias because cluster‐based analysis results were not reported. Since there was no published protocol available for reference, it was not possible to judge whether the other studies selectively reported results.

Other potential sources of bias

We rated Biswas 1994 at high risk of bias because the follow‐up interval was two weeks, which could lead to recall bias. Four trials were at low risk of other bias (Coutsoudis 1995; Coutsoudis 2000; Rahmathullah 1991; Sempértegui 1999). Stansfield 1993 was at unclear risk of other bias due to some minor flaws in the study design.

Effects of interventions

See: Table 1

See: Table 1, Table 2, and Table 3.

1. Main outcomes reported in the included studies.

Study	Incidence of acute URTIs during the study period	Adverse events	Time to occurrence of acute URTI	Severity of subjective symptoms	Number of cases with complications	Rate of hospitalisation	Number of cases of antibiotics used for URTIs	Time to return to school	Rate of school absenteeism	Guardian satisfaction
Biswas 1994	Yes^a	NR	NR	Yes^c	NR	NR	NR	NR	NR	NR
Coutsoudis 1995	Yes^a	NR	NR	NR	NR	NR	NR	NR	NR	NR
Coutsoudis 2000	Yes^b	Yes	NR	Yes^c	NR	NR	NR	NR	NR	NR
Rahmathullah 1991	Yes^b	NR	NR	NR	NR	NR	NR	NR	NR	NR
Sempértegui 1999	Yes^a	NR	NR	NR	NR	NR	NR	NR	NR	NR
Stansfield 1993	Yes^a	NR	NR	NR	NR	NR	NR	NR	NR	NR

Open in a new tab

NR: not reported URTIs: upper respiratory tract infections

^aThe number of acute URTIs over a period of time. ^bProportion of participants with an acute URTI. ^cMean duration of acute URTI.

2. Type and dosage of supplementation.

Author	Total duration of the study	Vitamin A dosage	Vitamin A supplementation period	Concomitant medications	Comparison	Supplementation period
Biswas 1994	6 months	200,000 IU	A single dose	‐	Placebo	A single dose
Coutsoudis 1995	2 years and 7 months	At 1 and 3 months, 50,000 IU, at 6 and 9 months, 100,000 IU, at 12 and 15 months, 200,000 IU of retinyl palmitate	6 times in 15 months	‐	Placebo	An equal volume of placebo at 1, 3, 6, 9, 12, and 15 months
Coutsoudis 2000	16 months	25,000 IU	On study days 1, 4, and 8	‐	Placebo	On study days 1, 4, and 8
Rahmathullah 1991	52 weeks (13 months）	8.7 μmol (2500 μg), equal to 8250 IU	52 weeks	46 μmol (20 mg) vitamin E	46 μmol (20 mg) vitamin E	52 weeks
Sempértegui 1999	10 months	10,000 IU	40 weeks	‐	Placebo	40 weeks
Stansfield 1993	12 months	Infants 6 to 11 months of age received half a capsule (100,000 IU) and older children the entire capsule (200,000 IU)	Every 4 months, 3 cycles	40.6 mg of vitamin E	40.6 mg of vitamin E	Every 4 months, 3 cycles

Open in a new tab

IU: international units

Vitamin A versus placebo or no intervention

Six studies compared vitamin A versus placebo or no intervention, which included 27,351 children (Biswas 1994; Coutsoudis 1995; Coutsoudis 2000; Rahmathullah 1991; Sempértegui 1999; Stansfield 1993; see Table 3 for further details on the intervention).

We could not conduct any planned subgroup analyses for this comparison due to the limited number of studies.

Primary outcomes

1. Incidence of acute upper respiratory tract infections (URTIs) during the study period (the number of acute URTIs over a period of time)

Five studies (11,938 participants) reported this outcome, although it was not possible to combine all the data in a meta‐analysis since the units were different (Biswas 1994; Coutsoudis 1995; Coutsoudis 2000; Sempértegui 1999; Stansfield 1993). We conducted a narrative analysis of two studies and a meta‐analysis of three additional studies.

There were 91 participants in the treatment group and 83 in the placebo group (Biswas 1994). The number of events was similar between the two groups: 249 in the treatment group and 235 in the placebo group (P > 0.05) (Biswas 1994). Coutsoudis 2000 reported 32 episodes of acute URTIs (incidence rate of 21.6 episodes per 100 visits) in the vitamin A group (56 participants), while 42 episodes (incidence rate of 24.3 per 100 visits) were reported in the placebo group (60 participants). The results were similar between the two groups.

We meta‐analysed three studies (Coutsoudis 1995; Sempértegui 1999; Stansfield 1993). We are uncertain of the effect of vitamin A supplementation on the number of acute URTIs measured over a two‐week period (risk ratio (RR) 1.00, 95% confidence interval (CI) 0.92 to 1.09; I² = 44%; 3 studies, 22,668 participants; low‐certainty evidence; Analysis 1.1; Figure 4). We downgraded the evidence due to study heterogeneity and risk of bias. We also performed a sensitivity analysis using a fixed‐effect model. The RR changed to 1.03 (95% CI 1.01 to 1.05). Although there was a statistical difference, the effect size was small and close to the null line.

1.1 — Comparison 1: Vitamin A versus placebo, Outcome 1: The number of acute URTIs over 2 weeks

2. Incidence of acute URTIs during the study period (proportion of participants with acute URTI)

Two studies (15,535 participants) reported this outcome (Coutsoudis 2000; Rahmathullah 1991). We could not conduct any planned subgroup analyses for this comparison due to the small number of studies.

There were 56 and 60 children in the treatment group and the control group, respectively. At the three‐month follow‐up, the cumulative percentage of children developing acute URTl was 30.2% and 41.6%, respectively; at the six‐month follow‐up, this was 55.1% and 53.5%, respectively, and at the nine‐month follow‐up this was 70.1% and 73.8%, respectively (Coutsoudis 2000). There were 7764 and 7655 children in the treatment group and the control group, respectively, and the number of children suffering from acute URTl was 4196 (54.0%) and 4026 (52.6%), respectively (Rahmathullah 1991). We are uncertain of the effect of vitamin A supplementation on the proportion of participants with an acute URTI (two studies, 15,535 participants; low‐certainty evidence), downgrading due to inconsistency.

3. Adverse events

Only one study (116 participants) reported this outcome (Coutsoudis 2000). We could not conduct any planned subgroup and sensitivity analyses for this comparison due to the limited number of studies.

No infant in either the placebo or vitamin A group was found to have feeding difficulties (failure to feed or vomiting), a bulging fontanelle, or neurological signs either before or after administration of vitamin A (Coutsoudis 2000). We judged the certainty of the evidence to be very low, downgrading due to imprecision and bias.