Vitamin A for treating measles in children.

Background 

Measles is a leading cause of childhood morbidity and mortality in developing countries with fatality rates in hospitalised children often exceeding 10% (Morley 1969A). An estimated 36.5 million cases and 1 million deaths caused by measles still occur each year. About half of these deaths occur in Africa (MMWR 1998). The problem is by no means limited to developing countries. There were 1750 cases reported in the Netherlands in 1999 despite a 96% immunisation rate in children over 14 months (Sheldon 2000). 

Ellison (Ellison 1932) first documented the protective effect of vitamin A on measles mortality almost 70 years ago. Barclay's study (Barclay 1987) drew attention to the importance of vitamin A therapy in reducing measles mortality and led to the 1987 World Health Organization (WHO) recommendation. In 1987 the WHO and United Nations International Children's Fund (UNICEF) jointly recommended administration of a single oral dose of 200,000 IU (or 100,000 IU in infants) of vitamin A, at the time of initial measles diagnosis, to non-xerophthalmic children who live in areas where measles case fatality rates were greater than 1% (WHO 1988). In 1993, the WHO expanded its recommendation that vitamin A be given to all cases of severe measles but the dose remained the same (WHO 1993). There was sufficient evidence at that time to demonstrate that vitamin A supplementation reduced childhood mortality and morbidity (Sommer 1996) but there were only two studies demonstrating the effect of vitamin A in the treatment of children with measles. Hussey's work (Hussey 1990) confirmed that treatment with vitamin A reduced measles morbidity and mortality. In 1997, WHO and UNICEF recommended 200,000 IU of vitamin A to be given twice to children with measles over the age of 1 year in populations where vitamin A deficiency may be present (WHO 1997).

Vitamin A deficiency is a recognised risk factor for severe measles (Frieden 1992). It is biologically possible for vitamin A to be of benefit in measles (Lancet 1987). Vitamin A can be used for the treatment of measles and may be beneficial either by reducing the effects of measles infection (therapeutic effect), or by preventing the subsequent development of secondary infection (protective effect), or both (Coutsoudis 1991).

Measles can decrease serum concentrations of vitamin A in well nourished children to less than those observed in malnourished children who did not have measles (Inua 1983). There could be two mechanisms of how hyporetinemia occurs in measles. One has been postulated to be through depletion of hepatic stores. Another possible explanation could be that vitamin A is not mobilised fast enough even in the presence of adequate hepatic stores (Hussey 1990). This could be the reason for the hyporetinemia in children with severe measles where vitamin A deficiency is uncommon as in Zaire (Markowitz 1989) and Cape Town (Hussey 1990) and in Nairobi (Ogaro 1993). Retinol concentrations have been found to be depressed in children with measles even in industrialised countries like the US. The degree of retinol depression was associated with the severity of illness (Butler 1993).

Vitamin A is essential for the maintenance of normal epithelial tissues throughout the body (Wolbach 1925). Measles is a viral disease that infects and damages these tissues (Morley 1969A). Vitamin A deficiency is known to depress the immune function and destroy epithelial tissue and measles produces similar effects (Coutsoudis 1991). The combined effect of vitamin A deficiency and measles infection could be quite serious. Therefore, when a child who has marginal vitamin A stores gets measles this exhausts the already depleted vitamin A stores thereby reducing the ability to resist secondary infection or their consequences (Bhaskaram 1975). This would also further reduce the immunocompetence that is associated with measles infection (Whittle 1979).

In Asia, measles was found to be an important risk factor in severe vitamin A deficiency (Tielsch 1984). In a number of community studies in Asia, vitamin A deficiency has been linked to an increased risk of childhood morbidity (Sommer 1984Milton 1987Bloem 1990) and mortality (Sommer 1983). Reductions in mortality of 6 - 54% were reported in children who were given vitamin A (Sommer 1986Muhilal 1988Rahmathullah 1990West 1991Daulaire 1992Vijayaraghavan 1990). In four studies that reported large reductions in mortality, measles mortality fell but the acute respiratory infection (ARI) mortality did not change (Rahmathullah 1990West 1991Daulaire 1992Ghana VAST Study)

Some studies have found vitamin A to have little effect on morbidity, while causing a significant reduction in mortality (Sommer 1986Rahmathullah 1990Rahmathullah 1991) and yet another reported no significant effect on morbidity or mortality (Vijayaraghavan 1990) even though it was a sufficiently large study. In some clinical trials, vitamin A reduced the severity of illness and mortality in children with measles (Barclay 1987Hussey 1990Coutsoudis 1991) even in areas where eye signs of vitamin A deficiency were rare (Ghana VAST Study). 

A meta-analysis by Glasziou et al (Glasziou 1993) on the role of vitamin A supplementation for infectious diseases showed clear evidence that vitamin A reduced all cause mortality in children in developing countries by around one third. A similar but apparently stronger effect was seen in children hospitalised with measles, with a reduction of 66%, although this was not significantly different from the 30% seen in developing country community settings. The reduction in deaths from respiratory disease was seen only in the measles studies. The results of this meta-analysis confirmed the 1987 WHO recommendation to give vitamin A to children in countries where vitamin A deficiency is a recognised problem (WHO 1987). The authors say that "the sum of current evidence is now too strong to justify further trials in hospitalised measles patients in developing countries".

In another meta-analysis of 12 controlled trials on the role of vitamin A supplementation and child mortality including many community, preventative studies (Fawzi 1993), vitamin A supplementation to hospitalised measles patients (children) was found to be highly protective against mortality. The most recent review (Beaton 1993) concludes that " ... in the specific case of measles, there is evidence that improvement of vitamin A status even after the onset of infection can improve both the course of the episode and the case fatality rate."

The World Bank (World Bank 1993) has declared vitamin A supplementation to be one of the most cost effective of all health interventions. Programs to control vitamin A deficiency are now in place or in planning in more than 60 countries (Sommer 1997).

Despite all this the situation is far from satisfactory. According to Ogaro, "the WHO recommendation of vitamin A supplementation has not been implemented in developing countries because vitamin A deficiency is usually identified because of high rates of xerophthalmia, a problem that exists in only selected places in the developing world. More commonly, developing country populations have inadequate or marginal vitamin A body stores without a high incidence of eye disease. Secondly, not all settings even in Africa have high measles case fatality rates and the usefulness of vitamin A supplementation where mortality and severe complications are much less frequent has had limited study" (Ogaro 1993).

Objectives 

To determine whether vitamin A is beneficial in preventing mortality and pneumonia and other secondary infections in children with measles.

Criteria for considering studies for this review 

Types of studies 

All randomized controlled trials in which children with measles were given vitamin A or placebo along with standard treatment were considered.

Types of participants 

Children with measles under the age of 15 of either gender.

Types of intervention 

Vitamin A or placebo given orally.

Types of outcome measures 

As stated a priori (D'Souza 1999) outcomes were mortality, pneumonia specific mortality, development of pneumonia, diarrhoea, croup and otitis media, and duration of hospitalisation, fever, pneumonia and diarrhoea. The definition of pneumonia used was a clinical case definition or radiological confirmation.

Search strategy for identification of studies 

See: Cochrane Acute Respiratory Infections Group search strategy

The reviewers used the search strategy developed for the Acute Respiratory Infections Group (Cochrane 1999). A MEDLINE search was done in July 1999 on PubMed for the period 1994 - 1998. The Cochrane Library now includes searches of MEDLINE 1966-97 and EMBASE 1974-1997, and Issue 4, 1999 was searched. Keywords used were measles, vitamin A, randomized, controlled trial, respiratory disease, pneumonia, random allocation and clinical trial. Sixty six references were found using this search strategy. An additional trial (Kawasaki 1999) was found in a subsequent search of MEDLINE in December 1999. 

In addition, the references of the available primary studies, review articles, and editorials were checked to identify trials not found in the database searches. One additional trial was found (Ellison 1932). 

Experts in the field were also contacted and one additional, unpublished trial was found but the data was unavailable (Lucero 1993). Trialists were also contacted for missing data.

Finally, a permanent search has been registered with Current Contents to notify the reviewers by e-mail of any new trials published in journals indexed by Current Contents.

Methods of the review 

Sixty eight references were found using the search strategy described. From the abstracts 28 of these appeared to meet the inclusion criteria. Two reviewers rated these blinded, using the Jadad method (Jadad 1996) for assessing the quality of trials. 

This method assigns points as follows:

1. Was the study described as randomised? (0=no; 1=yes)
2. Was the study described as double-blind? (0=no; 1=yes)
3. Was there a description of withdrawals and drop-outs? (0=no; 1=yes)
4. Was the method of randomisation well described and appropriate? (0=no; 1=yes)
5. Was the method of double blinding well described and appropriate? (0=no; 1=yes)
6. Deduct 1 point if methods for randomisation or blinding were inappropriate.

From this assessment six trials, each with a score of three or more (out of a possible maximum of five) were included. See characteristics of included studies table for scores.

Although Ellison (Ellison 1932) is the largest study yet (600 children), it received a low quality score because it was not randomised. Secondly, the quality of health care, availability of antibiotics and immunisation have affected the incidence and case fatality of measles quite substantially in the last 30 years but the case fatality rate is quite similar to the studies done in Africa 60 years later. The other 21 studies were excluded not for their scores but because vitamin A was given to all children in communities and not just children with measles.

The scores from this assessment were used to do a sensitivity analysis (i.e. including studies of low quality to determine how robust the summary effect measures were). Only one study (Ellison 1932) received a score of less than three and was included in the sensitivity analysis.

To assess the strength of the evidence for giving vitamin A to all children with measles, a meta-analysis of the selected studies was done in which administration of vitamin A was compared with placebo. As the outcomes had small numbers, the analysis for dichotomous outcomes was done using the StatXact software package (StatXact; Cytel Software Corp. Version 3.1, 1997). Exact odds ratios and their 95% confidence intervals (CI) were used to calculate the relative risks and 95% CI. The data used for calculating the exact odds ratios and 95% CI are given in the RevMan graphs. Weighted mean difference with 95% CI using random effects model were calculated for continuous outcomes using the Review Manager4.0.4 software with MetaView 3.1.

The following sub-group analyses which were determined a priori (D'Souza 1999) were done:
age, dosage, formulation (oil or water based), setting (hospital or community) and geographic area (varying measles case fatality rates).

Description of studies 

There was considerable variance particularly in the outcomes measured and reported in the studies. The only outcome reported by all seven studies was death. All included studies were published trials covering 1628 participants. Five of the studies (Barclay 1987Hussey 1990Coutsoudis 1991Ogaro 1993Rosales 1996) were done in Africa one in Japan (Kawasaki 1999) and one in England (Ellison 1932). Except for the community study by Rosales, all the rest were in hospitalised patients. One of the side effects of high doses of vitamin A is bulging fontanelles in some very young infants (WHO 1998). However, no side effects were reported in any of the studies.

Methodological quality 

Generally the quality of studies was high except for Ellison (Ellison 1932), which was not randomised and hence only included in a sensitivity analysis. Four studies (Hussey 1990Coutsoudis 1991Ogaro 1993Rosales 1996) were double blinded. In Barclay's study (Barclay 1987) the staff and patients were blinded but not the treating physician, who was also assessing the outcomes.

Results 

As mentioned in the description of studies, mortality was the only outcome reported by all trials. Nevertheless, all results have been reported even those measured and reported by a single study. The estimates for dichotomous outcomes have been derived from the StatXact software package.

1. Overall Mortality

There were deaths reported in all studies except Kawasaki. When the five studies reporting mortality (Barclay 1987Hussey 1990Coutsoudis 1991Ogaro 1993Rosales 1996) were pooled together, the summary estimate of the effect of vitamin A on the risk of mortality associated with measles was not significant even though there was a 39% reduction (RR=0.61, 95% CI 0.32 to 1.12) (StatXact estimate). The study done by Barclay showed a 48% reduction (RR=0.52; 95% CI 0.16 to 1.40) and Hussey showed a statistically significant 79% reduction in the risk of mortality (RR=0.21; 95% CI 0.02 to 0.95). The study by Rosales and Ogaro were associated with no effect on the risk of mortality in the supplemented group.

Five of the studies were hospital based and only Rosales' study was done in a community setting in a group of patients with mild disease (i.e. outpatients). Barclay, Hussey and Coutsoudis used 200,000 IU of vitamin A on the first day and a second dose the next day. Coutsoudis gave two additional doses on days 8 and 42. These three studies also had hospitalised patients (Barclay 1987Hussey 1990Coutsoudis 1991) and used at least two doses of vitamin A and were associated with a statistically significant 64% reduction in risk of mortality (RR=0.36; 95% CI 0.13 to 0.82). These three studies were also done in areas where the hospital case fatality rate was more than ten per cent. Coutsoudis study had only one death and dropping this study did not change the summary estimate. 

Two studies used water based vitamin A formulations (Hussey 1990,Coutsoudis 1991), while the others used an oil-based formulation. When the studies that used the two-dose regimen, were stratified by formulation (whether water or oil based); 81% reduction in the risk of mortality (RR=0.19, 95% CI 0.02 to 0.85) was seen in studies that used water based preparations. Barclay (Barclay 1987) used two doses of oil based vitamin A and was associated with a 48% reduction in risk of mortality (RR=0.52; 95% CI 0.16 to 1.40).

The two studies (Ogaro 1993Rosales 1996) that used an oil-based single dose of 200,000 IU of vitamin A done in areas where the case fatality was less than six per cent were not associated with any reduction in the risk of mortality (RR=1.25; 95% CI 0.48 to 3.12). Vitamin A status appeared to be better in the patients in both studies; at least 30% of Ogaro's patients had vitamin A levels greater than 20 microgram/dl. These factors in addition to the fact that a single dose of vitamin A was used in both studies, are probably the major reasons for perceived lack of efficacy of vitamin A treatment.

Ogaro's study which used a single oil-based dose when included with the other three hospitalised studies the effect of hospitalisation became non-significant.

As part of a sensitivity analysis, when the study by Ellison (Ellison 1932) which received a lower score is included with the five studies, vitamin A was associated with a 47% reduction in overall mortality (RR=0.53; 95% CI 0.33 to 0.83). The argument for including this study as part of the sensitivity analysis is that the mortality rate which was 8.66 and 3.66 in the placebo and vitamin A groups respectively were less than what was observed in studies done almost 60 years later in Africa suggesting that basic health care available to partients then was not dissimilar to that available in Africa in the 80's and 90's. The magnitude of mortality reduction in the Ellison study was remarkably similar to that of the other studies. 

Four of the six studies reported the age distribution of the participants and the ages of those who died. There was a 83% reduction in risk of mortality (RR=0.17; 95% CI 0.03 to 0.61) in the vitamin A supplemented group in children under two years of age in studies that used two doses of 200,000 IU of vitamin A (Barclay 1987Hussey 1990Coutsoudis 1991). The two dose oil-based vitamin A was associated with a statistically significant reduction in risk of mortality in the study by Barclay (RR=0.13; 95% CI 0.002 to 0.95) while the water based preparations almost reached statistical significance (RR=0.23; 95% CI 0.02 to 1.008). There was no evidence of reduction in the risk of mortality in the age groups older than two years (RR=0.94; 95% CI 0.23 to 3.1). Even when the study by Ellison was included as part of the sensitivity analysis, there was no reduction in the risk of mortality in the children older than two years (RR=0.64; 95% CI 0.20 to 1.76).

Although the sub-group analyses were determined a priori, these factors are highly correlated which means that their effects cannot be seperately identified and also due to the small sample size of the studies. Three significant studies (Barclay 1987, Hussey 1990, Coutsoudis 1991) are represented most frequently in all the 5 sub-group analyses - dose, formulation, hospitalisation, age and case fatality in the study area. They tend to use two doses, water-based formulations, hospitalised patients, children under the age of two and done in areas where the case fatality is high. Only Barclay's study slightly deviates from the other two studies as he used an oil based preparation. Therefore the presence of collinearity between these factors cannot be ruled out as the data cannot be stratified due to the raw data not being available and secondly there are too few studies to stratify across the five sub-groups. There were no trials comparing mortality reductions in children with measles who were given a single dose compared to two doses of vitamin A. Although the precision of the estimates of trials that used a single dose were similar to the trials that used two doses.

  • Pneumonia specific mortality. 

Four studies specified the causes of death in the children. Most of the deaths in these studies were due to pneumonia. In Ogaro's study (Ogaro 1993) all children who died had pneumonia as did 10/18 in Barclay's (Barclay 1987), and 10/12 in Hussey's study (Hussey 1990). The pooled estimate of the three studies which used two doses (oil or water based) were associated with a 67% reduction in the risk of pneumonia specific mortality RR=0.33; 95% CI 0.08 to 0.92 while none of these studies showed statistically significant reductions on their own. Water based preparations were associated with a 77% reduction in the risk of pneumonia specific mortality RR=0.23; 95% CI 0.02 to 1.05. Ogaro's study that used a single dose did not show any benefit. 

2. Morbidity

  • 2.1 Respiratory outcomes:

2.11 Post-measles croup

Four studies (Barclay 1987Hussey 1990Coutsoudis 1991Ogaro 1993) reported post-measles croup. The summary estimate of the four studies were associated with a statistically significant 41% reduction in the risk of croup in the vitamin A treated group RR=0.59; 95% CI 0.36 to 0.,94. When these studies were stratified by dose, the reduction in the incidence of croup was greater for the three studies (Barclay 1987Hussey 1990Coutsoudis 1991) that used two doses of 200,000 IU of vitamin A (RR=0.53; 95% CI 0.29 to 0.89).

2.12 Development of pneumonia

Development of pneumonia was reported in only two studies (Ogaro 1993Kawasaki 1999). These two studies individually did not show any statistical reduction in the incidence of pneumonia. The estimate in Ogaro's study was RR 0.68, 95% CI 0.28 to 1.36 and for Kawasaki's it was RR 0.97 95% CI 0.66 to 1.18. These studies were not combined as they were done in completely different settings and used different doses.

2.13 Duration of pneumonia

Duration of pneumonia was reported in only two studies (Hussey 1990Coutsoudis 1991). The summary estimate of these studies shows a reduction in the duration of pneumonia by more than three days in the vitamin A treated group but this was not statistically significant WMD -3.69, 95% CI -7.53 to 0.16, although both studies were individually statistically significant. In Hussey's study there was almost 6 days reduction in duration of pneumonia in the vitamin A treated group; WMD -5.8, 95% CI -8.2 to -3.5 and two days in Coutsoudis' study WMD -1.9, 95% CI -2.2 to -1.6.

  • 2.2 Other outcomes:

2.21 Development of diarrhoea

Only Barclay and Ogaro (Barclay 1987Ogaro 1993) reported on development of diarrhoea. The summary estimate of these studies shows a slight reduction in diarrhoea in the vitamin A treated group but this was not statistically significant (RR 0.96, 95% CI 0.53 to 1.63). In Barclay's study which used two doses, there was a 65% reduction in risk of developing diarrhoea (RR 0.35, 95% CI 0.33 to 1.83) while there was no evidence of reduction in Ogaro's study which used a single dose (RR 1.13, 95% CI 0.69 to 1.62).

2.22 Duration of diarrhoea

Two studies (Hussey 1990Coutsoudis 1991) reported the duration of diarrhoea in days. The summary estimate of these studies shows a statistically significant reduction in duration of diarrhoea by almost two days in the vitamin A treated group WMD -1.92, 95% CI -3.40 to -0.44.

2.23 Duration of fever

Two studies (Coutsoudis 1991Kawasaki 1999) reported the duration of fever in days. Kawasaki showed a one and half day statistically significant reduction in the duration of fever WMD -1.5, 95% CI -2.04 to -0.96 while Coutsoudis showed a little over half a day WMD -0.60, 95% CI - 0.81 to -0.39. These studies were not combined because they were done in completely different settings and used different doses.

2.24 Days in hospital

Hussey (Hussey 1990) showed a statistically significant reduction in hospital stay by almost five days in the vitamin A treated group WMD -4.72, 95% CI -7.22 to 2.21 while Kawasaki (Kawasaki 1999) showed a reduction by almost half a day in the vitamin A treated group but this was not statistically significant WMD -0.40, 95% CI -1.08 to 0.28. These studies were not combined because they were done in completely different settings and used different doses.

2.25 Herpes stomatitis

Hussey (Hussey 1990) and Coutsoudis (Coutsoudis 1991) reported herpes stomatitis. None of the studies showed any statistically significant reductions in the risks of developing herpes stomatitis. The estimate by Hussey showed a reduction RR 0.24, 95% CI 0.02 to 1.09 while Coutsoudis showed an increased risk of developing Herpes Stomatitis RR 1.59, 95% CI 0.18 to 9.22. The summary estimate showed a decrease in herpes stomatitis in the vitamin A treated group but this was not statistically significant RR 0.48, 95% CI 1.13 to 1.43.

The outcomes below were reported in single studies:

  • 2.3 Respiratory outcomes:

2.31 Recovery from pneumonia in less than 8 days

Ogaro (Ogaro 1993) reported the number of patients that recovered from pneumonia in less than eight days. His study used a single dose of vitamin A and did not show any benefit RR=0.99, 95% CI 0.78 to 1.15. 

2.32 Pneumonia for more than 10 days.

Hussey (Hussey 1990) reported the number of patients who had pneumonia for more than ten days. This study used two doses and showed a 57% statistically significant reduction in the number of children in the vitamin A group who had pneumonia for more than 10 days, RR 0.43, 95% CI, 0.20 to 0.83.

2.33. Pneumonia for 14 days

The study by Rosales (Rosales 1996) used a single dose and did not show any benefit on pneumonia in the vitamin A group after two weeks RR=1.24, 95% CI 0.76 to 1.82. 

2.34 Days of cough

Kawasaki (Kawasaki 1999) alone reported the duration of cough in days. The study showed a statistically significant reduction by two days in the vitamin A treated group WMD -2.00, 95% CI -2.71 to -1.29.

2.35 Cough in two weeks

Rosales (Rosales 1996) reported on this outcome and there was an increased risk that the vitamin A group still had cough at the end of two weeks but this was not statistically significant RR 1.54, 95% CI 0.75 to 2.81.

2.36 Development of acute laryngitis

Kawasaki (Kawasaki 1999) reported on this outcome and there was an increased risk in the vitamin A group of developing acute laryngitis but this was not statistically significant RR 1.86, 95% CI 0.79 to 3.4.

2.37 Development of otitis media

Ogaro (Ogaro 1993) reported the development of otitis media with a 74% reduction in the incidence of otitis media in vitamin A treated patients and this was statistically significant RR 0.26, 95% CI 0.05 to 0.92.

2.4 Other outcomes

2.41 Recovery from diarrhoea in less than 5 days.

Ogaro reported on this outcome and it showed increased chances of recovery in the vitamin A supplemented group and was statistically significant RR 2.05, 95% CI 1.06 to 3.99.

2.42 Diarrhoea for more than ten days

Hussey (Hussey 1990) showed a statistically significant reduction of diarrhea at 10 days in the vitamin A treated group RR 0.41, 95% CI 0.14 to 0.89 

2.43 Diarrhoea for 14 days

Rosales (Rosales 1996) did not find any benefit in the vitamin A treated group RR 0.00, 95% CI 0.0 to 2.54.

2.44 Complete clinical recovery

Coutsoudis (Coutsoudis 1991) found that the vitamin A group had a 1.5 times higher chance of complete clinical recovery than the placebo group and was statistically significant RR 1.54, 95% CI 1.04 to 1.88.

2.45 Asymptomatic in week 2

Rosales (Rosales 1996) found that the vitamin A group did not show any benefit in 2 weeks in terms of complete clinical recovery than the placebo group RR 1.01, 95% CI 0.74 to 1.24.

2.46 Transferred to Intensive Care Unit

Hussey (Hussey 1990) reported that the vitamin A supplemented group had a lesser chance of being transferred to the intensive care unit but was not statistically significant RR 0.39, 95% CI 0.80 to 1.23. 

Discussion 

The quality of the trials included in this review is high. The factors included in the sub-group analysis of dose, formulation, setting and the age of children were highly correlated and represented in three studies (Barclay 1987Hussey 1990Coutsoudis 1991)

Dose and formulation

This review demonstrates that vitamin A administered to children with measles along with standard treatment was associated with a reduction in mortality when the 200,000 IU dose was repeated on the second day. The evidence supports the WHO's recommendation of two 200,000 IU doses.

Although the data do not allow us to examine the individual effects of dose and formulation, these are issues that need to be considered. Vitamin A preparations in oil and in water are different in terms of their action in the body over a period of time, including the processes of absorption, distribution, localisation in tissues, bio-transformation and excretion. Water based vitamin A preparations lead to greater absorption, which lead to higher serum retinol levels. The oil based preparation is more stable, readily available and costs less and for these reasons is recommended by the WHO.

Coutsoudis' study and others (Inua 1983Reddy 1986Markowitz 1989) support the finding that serum retinol concentrations are lowered during measles. In Coutsoudis' study (Coutsoudis 1991) the supplemented group had significantly higher concentrations than the placebo group which indicates that the liver stores were not depleted but there was temporary impairment of mobilisation of vitamin A and increased utilisation of vitamin A.

The children in Rosales's (Rosales 1996) and Ogaro's (Ogaro 1993) studies may not have benefited from receiving vitamin A oil based preparations in a single dose of 200,000 IU as it might not have been sufficient to reverse the hyporetinemia during measles due to it being stored mostly in the liver. Rosales reported a 70% increase in serum retinol after a single dose of oil based vitamin A as compared to a 215% increase observed by Coutsoudis (Coutsoudis 1991) who used two doses of water miscible vitamin A. Rosales' study was also a community based study and therefore the protective effect of vitamin A may not be as great as seen in the more severe hospital-based cases.

The results of this review have confirmed that two doses of 200,000 IU of vitamin A was associated with reduction in the risks of overall mortality and pneumonia specific mortality. In 1991, Rosales (Rosales 1996) came to the same conclusion at the end of his trial as did Sommer who suggested that it was prudent to follow the double dose schedule already proven in the Barclay, Hussey and Coutsoudis trials, rather than the single dose recommended by WHO at that time. Double the WHO dose was also advocated by Chan (Chan 1990) and Hussey (Hussey 1997). Although the two dose, water based product was associated with a greater reduction in risk of mortality, no recommendation can be made as to whether a single dose of water based preparation would have a similar benefit as no such studies have been conducted to see the effect of a single dose of water based vitamin A as compared to two doses. Therefore, given the limited number of trials our conclusion would be that the single-dose, water based and oil based preparations need to be compared to a two dose schedule. The trade off of using high dose oil based vitamin A versus a water-based formula has to be viewed in terms of the advantages of each product. As the water based product may be associated with greater mortality reductions, the advantage may be offset by its lower stability, higher cost, and non-availability. 

Only one study used the two-doses of oil-based vitamin A and was not significant on its own for overall mortality except for children under the age of two years. The evidence of oil based vitamin A having a protective effect on mortality is demonstrated when a study of a lesser score by Ellison was included as part of the sensitivity analysis. Although this study used very small doses of vitamin A i.e. 3000 for 7 days, the supplemented group had statistically significant reductions in risks of mortality even in the absence of antibiotics and immunisation 60 years ago. This is the largest study done to date but it was not randomised as two separate wards were allocated to receive the placebo or vitamin A supplementation. The study participants in this study could be comparable to the African children enrolled in the other 5 studies almost 60 years later as the case fatality rates in the Ellison study are very similar and in some cases lower than the case fatality in the placebo and supplemented groups in some studies.

Additional evidence of the effect of oil based vitamin A can be drawn from the community trials i.e. preventive trials. When community based trials were pooled; periodic vitamin A supplementation of children who were apparently healthy at baseline resulted in a 39% reduction in measles-related mortality in addition to decrease in overall mortality (Fawzi 1993).

The effect of vitamin A was more pronounced in children under the age of 2 years as greater reduction in the risk of mortality was observed in this age group. This was across all studies but more so in the studies that used the two-dose regimen (Barclay 1987; Hussey 1990; Coutsoudis 1991). In children under the age of two years formulation did not make any difference as the oil based product was associated with a statistically significant reduction in the risk of mortality and the water based vitamin A almost reached statistical significance. The study by Markowitz et al (Markowitz 1989) highlighted the fact that children aged less than two years of age and who had low vitamin A levels had a higher risk of dying than those with higher levels. The number of children in the age group older than two years were too few to observe any statistically significant difference.

Case fatality rate in country of study

As the studies (Hussey 1990;  Barclay 1987;  Coutsoudis 1991) using two doses were also done in areas where case fatality was more than ten percent, it is important to be careful in interpreting the results. It raises the issue of whether the decrease in mortality was a result of the higher dose, or whether the vitamin A supplementation in higher case fatality areas had a greater effect as there was greater potential for mortality decline in those populations. It may be possible that there would be a decline in mortality even with a single dose of vitamin A in high case fatality areas and this needs to be further explored. Although South Africa had a measles case fatality greater than 10% in hospitals, Coutsoudis had a very low case fatality rate in the vitamin A group and the control group and she remarked that this could be attributed to the absence of emergency and malnourished cases. 

Hospital versus community studies

The protective effect of vitamin A supplementation was seen only in hospitalised children. Hospitalisation may be a measure of severity of illness. There is the possibility that more severe clinical cases of measles are more likely to benefit from vitamin A treatment. Three (Barclay 1987Hussey 1990Coutsoudis 1991) of the four hospital based studies which used the two dose regimen demonstrated a protective effect on mortality. These studies were done under controlled conditions and their follow-up was relatively brief. Only Coutsoudis's study indicated some long term benefit of vitamin A as children were followed for 6 months but the outcomes used for this review were at the time of discharge from hospital. Four of the five RCT's were done in hospitals and this affects the generalisability of the results to the general population of patients with measles. 

An absence of a vitamin A effect or a smaller effect in the community study than that found in other studies may be due to the study population being healthier than previous studies (Rosales 1996). This study did not include children who were very sick as they were referred for hospitalisation. Rosales' study differed from the above four in patient setting, follow-up, disease severity, patient age, vitamin A preparation and analytical approach. It looked at ambulatory patients who were followed up closely for one month with daily and weekly visits to urban health centres. This study reflects the patient care conditions under which the majority of measles cases are diagnosed and treated in developing countries (Rosales 1996).

Baseline differences and the presence of complications on admission

The demographic, nutritional, immunological and clinical status at baseline all affect the comparability between the vitamin A treated and control groups (Coutsoudis 1991). Although all the studies did report the baseline nutritional status of the vitamin A supplemented and placebo groups only Barclay specified the nutritional status of the children who died and vitamin A recipients suffered lower mortality at every nutritional level. 

In Ogaro's study (Ogaro 1993), there were ten children severely malnourished in the vitamin A supplemented group and five children in the placebo group. This raises an issue about whether randomisation balanced this important confounder. This could have been an important difference possibly resulting in an inability to demonstrate a protective effect of vitamin A in the supplemented group. All the deaths were due to pneumonia (five in the vitamin A group and three in the placebo group). 

Five of the studies were done in Africa but the baseline prevalence of vitamin A deficiency and other baseline characteristics vary across the four countries and even within the same country as in South Africa. The health services in the five areas could be different and this could be one of the reasons in addition to dose that the studies showed different results.

Rosales suggested that as the population in his study was a healthier population than in previous studies, it may explain an absence of a vitamin A effect in this setting or a vitamin A effect smaller than that found in other studies (Rosales 1996).

Morbidity

In Hussey's study, 64% of children had diarrhoea and pneumonia on admission, while in Barclay's study, pneumonia was the most frequent complication affecting 85 children, 43% in the vitamin A group and 51% in the controls.

Most of the morbidity outcomes are either based on single or two studies except for croup. As all studies did not report on all possible morbidity outcomes, the conclusions we were able to draw about the effect of vitamin A on measles related morbidity are limited.

Kawasaki's (Kawasaki 1999) study had no mortality and the morbidity outcomes were not pooled with the other studies as this study was done in a developed country i.e. Japan and had used only a single dose of 100,000 IU of vitamin A. The improvement in morbidity outcomes were mainly seen in the studies that used two doses. There was a significant decrease in the incidence of croup while there was no significant reduction in the incidence of pneumonia, although there was a reduction observed in the duration of diarrhoea, pneumonia, fever hospital stay and cough. Treatment of measles cases with vitamin A also have relevance to developed countries as reduction is seen in morbidity outcomes as seen in Kawasaki's study.

Limitations of this review

One of the main limitations of this review is that we were unable to control for confounding by nutritional status. Nutritional status is an important predictor of mortality and vitamin A deficiency. Small number of studies and sample sizes have made it difficult to stratify or do a meta-regression and therefore the sub-group analyses is very restricted as the same studies are represented in all of them. The apparent differences between trials may be related to the subgroup, but could equally be confounded by some other aspect of trial design.

In these trials it was not always apparent as to which day vitamin A was administered after the onset of measles. Another limitation is that the follow-up period is not the same in all studies. It is assumed that all have been followed up till the time they were discharged from hospital. For the purposes of this review the outcomes were taken at the time of discharge from hospital, hence it is difficult to make comparisons for delayed mortality across these studies .

It would have been useful to have baseline incidence of measles in the study populations reported and if there were epidemics during the study period. The cases enrolled during a measles epidemic could vary in severity from measles cases at other times.

There was also a lack of reporting on the immunisation status of these children and the immunisation coverage in the study population which was reported by only one study. The level of immunisation would have an impact on the severity of measles in the children as it could reduce the intensity of exposure and hence the dose of the infecting virus (Hussey 1997). This would have an impact on the severity of the disease in the children as well as the severity of the epidemics. The severity of measles would be less in already vaccinated children (vaccine failure) and in areas where the immunisation coverage was high.

Each study collected information on recovery from morbidity. There is some concern about whether some trialists collected data but later chose not to report these. As there were many outcomes reported by single studies, there is the possibility that some would appear to be significant by chance alone.

Cost

Vitamin A is not only effective but also cost saving as Hussey (Hussey 1990) demonstrated that the hospital stay in children given vitamin A decreased by an average of 4.7 days while half a day was reported by Kawasaki (Kawasaki 1999). The cost of a dose of vitamin A is around US$ 0.02 (WHO 1998). At this cost ... "to achieve significant reductions in hospitalisation and costs in terms of mortality and long-term morbidity, vitamin A therapy for the management of measles is highly cost-effective" (Cervinskas 1996). 

Side Effects

Until 1993, there were no reports of acute vitamin A toxicity in children with measles who took the WHO recommended dose, according to the Committee on Infectious Diseases of the American Academy of Pediatrics (Pediatrics 1993). Even doses up to 400,000 IU have been reported to be relatively safe (Frieden 1992). None of the studies included in this review reported any adverse effects. 

Headaches, loss of appetite, vomiting and bulging fontanelles (in infants) are some of the known adverse effects occasionally occurring with the administration of high doses of vitamin A. However, these symptoms are minor and transitory, with no known long term effects and require no special treatment (WHO 1998). Under these circumstances it would appear that two doses of vitamin A are not too expensive, are not likely to produce adverse effects and still have the capacity to reduce morbidity and mortality.

Comparison with other reviews

The conclusions of this review are in keeping with the previous three reviews (Glasziou 1993Fawzi 1993Beaton 1993) which were done at a time when only three trials (Barclay 1987Hussey 1990Coutsoudis 1991) were available. These are also the studies using double the dose and show a protective effect on measles mortality in the children treated with vitamin A. Later studies by Rosales (Rosales 1996), Ogaro (Ogaro 1993) and Kawasaki (Kawasaki 1999) used a single dose of oil based vitamin A and did not show reduced measles mortality. Hence authors of earlier reviews were not able to compare dosages in sub-group analyses. In addition, Fawzi's meta-analysis (Fawzi 1993) included Ellison's study of 1932 (Ellison 1932). Although it is the largest study it has been included in this review as part of the sensitivity analysis as it received a low quality score. It may also be worth mentioning that the objectives of those reviews were different from the objective of this review.

The findings of this review are consistent with one of the largest observational studies that have compared mortality as an outcome (Hussey 1997). A retrospective hospital record review of 1720 cases of measles, during 1985-86 and 1989-90 was carried out. There were 651 children in the latter time period who received two doses of 200,000 IU of vitamin A and had a shorter hospital stay, lower requirement for intensive care and lower death rate as compared to 1069 children during 1985-86 who received a single dose of 3000 IU. 

This review confirms that two doses of vitamin A is associated with a statistically significant reduction in the risk of overall mortality. Although the only conclusion that can be drawn with any degree of certainty is that high doses of oil or water based vitamin A were associated with greater reductions in mortality in children under the age of two years. It is possible that in high doses oil-based and water-based vitamin A have similar effects in children under the age of two years. Therefore in this age group formulation did not make any difference. On the other hand, as studies that used two doses were also done in high case fatality areas, there was no evidence to show that a single dose would not be effective as there were no studies using a single dose of oil based vitamin A in these areas. Similarly, sub-group analyses by causes of morbidity and mortality by age group and formulation could not be done as the information was not available.

Authors' conclusions 

 

Implications for practice

We support the WHO recommendation that two doses of 200,000 IU of vitamin A be given to all cases of severe measles in addition to the standard treatment.

The evidence from these studies can only be generalised to developing countries as there is limited information to generalise to developed countries. The only study done in a developed country ( Japan ) used one-fourth the recommended dose (100,000 IU) and showed reduced morbidity and no toxicity was reported.

 

Implications for research

This review has shown that mortality reductions were observed in children that were given two doses of vitamin A in children under the age of two years, in hospitalised children and in areas where the case fatality was greater than 10%. This review was unable to separate out which of these factors contributed greater benefit of vitamin A in preventing mortality. Therefore, RCTs need to be conducted that would compare single doses of 200,000 IU (oil or water based) vitamin A with two doses and have sufficiently large sample sizes that the results could be stratified across sub-groups like age, areas with low and high case fatality and hospitalised and non-hospitalised children.

To study the benefit in children older than two years of age, more children in this age group need to be enrolled. If trials are conducted, trialists should report on all outcomes and baseline findings like age, nutritional status, immunisation status, immunisation coverage, complications on enrolment, and vitamin A levels in the patients. In addition, the number of deaths and morbidity conditions should be reported in each of these groups.

 

Potential conflict of interest 

None known.

Acknowledgements 

We would like to thank the National Centre for Epidemiology and Population Health, The Australian National University for the use of its facilities in preparing this review. We would also like to thank Drs Owen Tinnion, Wafaie Fawzi, Mahomed Patel, Phil Alderson and Keith Dear, Prof Amy Zelmer and Ms Heather McIntosh for comments on the draft review, Prof Greg Hussey for providing unpublished data, and Dr Akira Shimouchi, Ms Kasumi Nishigaya and Dr Jianping Liu for translation.

Notes 

The Cochrane Acute Respiratory Infections Group would like to acknowledge funding for its editorial base from Medical Benefits Fund of Australia.

Characteristics of included studies 

Study
Barclay 1987

Methods
randomised clinical trial using random number table

Participants
180 children with measles in hospital

Interventions
200,000 Iu vitamin A orally for two days or routine treatment without vitamin A

Outcomes
death

Notes
Quality score=3

Allocation concealment
A

Study
Coutsoudis 1991

Methods
randomized, placebo controlled, double blind trial

Participants
60 children aged 4 - 24 months hospitalized with complicated measles

Interventions
WHO recommended dose (54.5 mg <12 mo or 109 mg >12 mo) of retinyl palmitate drops or a placebo syrup

Outcomes
death
recovery in <8 days
duration of pneumonia in days
duration of diarrhea in days
duration of fever in days
herpes stomatitis, laryngo-tracheobronchitis, integrated morbidity score

Notes
Quality score=5

Allocation concealment
A

Study
Ellison 1932

Methods
A controlled trial

Participants
600 children in two hospital wards

Interventions
300 Carr and Price units for 7-12 days

Outcomes
death

Notes
Quality score=1

Allocation concealment
C

 

Study
Hussey 1990

Methods
Randomised double-blind trial

Participants
189 children <13 years of age, hospatialised with measles complicated with pneumonia, diarrhea or croup

Interventions
Either 200,000 IU retinyl palmitate given orally for two days or a placebo, within 5 days of the onset of the rash

Outcomes
>10 days with pneumonia
>10 days of diarrhea
croup, duration of diarrhoea and pneumonia
herpes stomatitis
transferred to intensive care
hospital stay in days
death

Notes
Quality score=4

Allocation concealment
A

Study
Kawasaki 1999

Methods
A randomised controlled trial

Participants
105 children with measles age 5 months to 4 years in hospital

Interventions
Oral vitamin A (100,000 IU) supplementation

Outcomes
pneumonia, laryngitis, duration of cough, fever and hospitalisation.

Notes
Quality score=2

Allocation concealment
A

Study
Ogaro 1993

Methods
Randomised, double blind trial

Participants
294 children under five years admitted to hospital with measles in Kenya

Interventions
50,000 IU of vitamin A (retinyl palmitate) to children < 6 months, 100,000 IU to children between 6 - 12 months, and 200,000 IU to children > 12 months in a single dose on admission.

Outcomes
Croup, pneumonia, diarrhea, otitis media, death

Notes
Quality score=3

Allocation concealment
A

Study
Rosales 1996

Methods
randomised, double masked, placebo controlled clinical trial

Participants
200 children with acute measles not requring hospitalization

Interventions
Single dose of 200,000 IU vitamin A in oil (100,000 IU for infants) or placebo

Outcomes
measles-associate cough or pneumonia, croup fever, diarrhoea,
failure to improve from pneumonia

Notes
Quality score=5

Allocation concealment
A

 

http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD001479/full