1. Objectives

To evaluate vitamin E supplementation for the prevention of morbidity and mortality in preterm infants

2. How studies were identified

The following databases were searched:

• CENTRAL (The Cochrane Library 2007, Issue 1)
• MEDLINE (March 2007)
• EMBASE (March 2002)

Reference lists and personal files were also searched. The updated search in 2007 identified one further study, which was not included in the review

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials

3.2 Study participants

Preterm (<37 weeks’ gestation) or low birth weight (<2500 g) infants, regardless of maternal vitamin E intakes

3.3 Interventions

Parenteral or enteral vitamin E supplementation at any dose compared to placebo, no treatment, or to another type, dose or route of administration of vitamin E

(Studies assessing co-interventions, such as erythropoietin, were excluded)

3.4 Primary outcomes

• Mortality until discharge
• Combined mortality and morbidity at 18 months (morbidity included bronchopulmonary dysplasia, blindness, mental retardation or cerebral palsy)
• Combined morbidity at 18 months (including bronchopulmonary dysplasia, blindness, mental retardation or cerebral palsy)

Secondary outcomes included E-Ferol-related “clinical syndrome” (direct tissue toxicity from vitamin E and/or from specific reagents which can increase the risk of necrotising enterocolitis), combined mortality and morbidity before discharge (morbidity included bronchopulmonary dysplasia, severe retinopathy of prematurity, severe cerebral abnormality, blindness, mental retardation or cerebral palsy), bronchopulmonary dysplasia, patent ductus arteriosus, sepsis, severe intraventricular haemorrhage, mild-to-moderate intracranial haemorrhage, severe cerebral abnormality (intraventricular haemorrhage, cystic periventricular leukomalacia, severe ventricular dilatation), severe retinopathy of prematurity, retinopathy of prematurity, necrotising enterocolitis with gastrointestinal perforation, necrotising enterocolitis not requiring surgery, renal failure (serum creatinine concentration >1.5 mg/dL), haemolysis (anaemia, reticulocyte count, haemoglobin concentration, need for transfusion, number of transfusions, bilirubin concentration), platelet count, coagulation tests, and local reaction at the injection site

4. Main results

4.1 Included studies

Twenty-six randomized controlled trials, enrolling 3248 infants, were included in this review

• Twenty-five trials assessed vitamin E versus placebo: in eight trials vitamin E was administered enterally, in 15 trials vitamin E was administered intramuscularly with or without enteral supplementation, and in two trials vitamin E was administered intravenously
• In 20 of the 25 vitamin E versus placebo trials, supplementation was initiated within the first 48 hours of life and after 48 hours in four trials; the total daily dose of vitamin E was ≤30 IU/kg in 15 trials and >30 IU/kg in nine trials; the control groups received up to 10 mg of vitamin E/100 kcal in 22 trials and >10 mg/ 100 kcal in three trials
• One trial compared two formulations of intramuscular vitamin E: 20 mg/kg of vitamin E from eight hours of life in either aqueous colloidal or olive oil solutions

4.2 Study settings

• Five trials were conducted in the 1970’s, 17 in the 1980’s, three in the 1990’s, and one was conducted in 2003

4.3 Study settings

How the data were analysed
Two comparisons were made: vitamin E versus placebo or no vitamin E and intramuscular vitamin E (aqueous colloidal) versus another form of intramuscular vitamin E (olive oil). Dichotomous data were summarized using risk ratios (RR), risk differences (RD), and the number needed to treat or harm. Continuous data were summarized using mean differences (MD). Results were presented with 95% confidence intervals (CI). Analyses were conducted for all infants, and for very low birth weight infants (defined as ≤1500 g). The following subgroup analyses were also conducted:

• By route of vitamin E administration: enteral, enteral hypertonic formulation at pharmacologic dose, parenteral with or without enteral, parenteral with hypertonic enteral formulation at pharmacologic dose, excluding intravenous vitamin E administration
• By total dose of vitamin E in the treatment group: ≤30 IU/kg/day versus >30 IU/kg/day
• By serum tocopherol level in the treatment group: ≤3.5 mg/dL versus >3.5 mg/dL
• By timing of initiation of supplementation: within the first 48 hours of life versus after the first 48 hours
• By duration of treatment: ≤1 week versus >1 week
• By total dose of vitamin E in the control group: ≤10 mg vitamin E/100 kcal versus >10 mg/100 kcal
• By iron supplementation: ≤2 mg/kg/day versus >2 mg/kg/day
• By polyunsaturated fatty acid (PUFA) content of formula: ≤400 mg/100 mL of milk versus >400 mg/100 mL of milk
• By birth weight: ≤1000 g, >1000 g, ≤1500 g
• By vitamin A supplementation: ≥1500 IU/day

Results
Vitamin E versus placebo or no treatment
Mortality until discharge
Overall, vitamin E supplementation had no statistically significant effect on the risk of mortality in comparison with placebo (RR 0.97, 95% CI [0.83 to 1.14]; RD -0.01, 95% CI [-0.04 to 0.03], 12 trials/994 individuals). Results were not altered meaningfully in subgroup analysis by birth weight, route of administration, total dose, serum tocopherol level, onset and duration of therapy in the treatment group, dose of vitamin E in the control group, or by intake of iron or PUFA. Similarly, in pooled analysis of seven trials including 628 very low birth weight infants, vitamin E had no significant effect on mortality until discharge (RR 0.97, 95% CI [0.80 to 1.16]; RD -0.01, 95% CI [-0.05 to 0.04]). Subgroup analysis by birth weight, route of administration, total dose, serum tocopherol level, onset and duration of therapy in the treatment group, dose of vitamin E in the control group, and by intake of iron or PUFA demonstrated no significant effect of vitamin E on mortality. No trials evaluated the other primary outcomes, combined mortality and morbidity at 18 months, and combined morbidity at 18 months.

Sepsis
The risk of sepsis after study entry was statistically significantly increased by 52% among infants receiving vitamin E treatment (RR 1.52, 95% CI [1.13 to 2.04], p=0.0053; 4 trials/1009 infants). The risk of sepsis after study entry among very low birth weight infants was also increased with vitamin E supplementation (RR 1.53, 95% CI [1.13 to 2.08], p=0.0057; 4 trials/807 infants), and was increased among very low birth weight infants treated for more than one week (RR 1.63, 95% CI [1.17 to 2.26], p=0.0039; 4 trials/726 infants). Subgroup analyses revealed that the risk of sepsis after study entry was associated with parenteral vitamin E supplementation, intravenous supplementation, dosage >30 IU vitamin E/kg/day, supplementation beginning within 48 hours of life, serum tocopherol levels >3.5 mg/dL, and with vitamin E intake in the control group≤10 mg/100 kcal.

Intracranial haemorrhage
The risk of germinal matrix/intraventricular haemorrhage was significantly reduced by 15% with vitamin E administration overall (RR 0.85, 95% CI [0.73 to 0.99], p=0.032; 7 trials/1755 infants), but not among very low birth weight infants (RR 0.95, 95% CI [0.75 to 1.18], 3 trials/777 infants). The risk of germinal matrix/intraventricular haemorrhage was also reduced in one trial among infants with negative initial ultrasonograms (RR 0.57, 95% CI [0.40 to 0.80], p=0.0011; 210 infants). In a single study involving 85 infants, intravenous high-dose vitamin E supplementation among very low birth weight infants statistically significantly increased the risk of parenchymal brain haemorrhage (RR 9.21, 95% CI [1.22 to 69.58], p=0.031). No clear effect of vitamin E supplementation on severe intraventricular hemorrhage or parenchymal hemorrhage was observed among all infants or among very low birth weight infants.

Retinopathy of prematurity
Overall, no effect of vitamin E supplementation was found for retinopathy of prematurity (all grades) or severe retinopathy of prematurity in all infants or in very low birth weight infants. However, among examined very low birth weight infants, the risk for severe retinopathy of prematurity (RR 0.58, 95% CI [0.34 to 1.00], p=0.049; 7 trials/1062 participants) and the risk of blindness (RR 0.19, 95% CI [0.10 to 0.88], p=0.029; 4 trials/467 infants) were reduced.

Additional outcomes
Haemoglobin concentrations were increased with vitamin E treatment in pooled analysis of eight trials involving 416 infants (MD 0.46 g/dL, 95% CI [0.24 to 0.69]), and in pooled analysis of four trials involving 196 very low birth weight infants (MD 0.43 g/dL, 95% CI [0.09 to 0.77]). Reticulocyte counts were reduced with vitamin E supplementation when expressed in percent of total red blood cell count (MD -1.66, 95% CI [-2.24 to -1.07], 2 trials/201 infants) and as absolute counts in millions per litre (MD -34.0, 95% CI [-46.2 to -21.8], 2 trials/71 infants). Overall, no statistically significant effect of vitamin E supplementation was found for the outcomes bronchopulmonary dysplasia, patent ductus arteriosus, necrotising enterocolitis, serum bilirubin, blood transfusions, platelet counts, coagulation tests, retinal haemorrhage, retinal detachment, cicatricial retrolental fibroplasia, or reaction at the site of injection.

Subgroup analyses
Serum tocopherol concentrations >3.5 mg/dL in very low birth weight infants were associated with a reduced risk of severe retinopathy among those examined, an increased risk of sepsis, and an increased risk of necrotising enterocolitis among those treated for more than one week. Intravenous, high-dose vitamin E in very low birth weight infants was associated with an increased risk of sepsis and parenchymal haemorrhage, and an increased risk of necrotising enterocolitis among those treated for more than one week.

Intramuscular vitamin E (aqueous colloidal) versus another form of intramuscular vitamin E (olive oil)
In one trial involving 50 infants, intramuscular colloidal vitamin E solution had no effect on bronchopulmonary dysplasia, patent ductus arteriosus, sepsis, intracranial haemorrhage, parenchymal haemorrhage, necrotising enterocolitis, or exchange transfusion in comparison with intramuscular vitamin E in olive oil solution. However, in the olive oil group, plasma levels of free tocopherol were within the deficient range and were significantly lower than among infants receiving the colloidal solution.

5. Additional author observations*

Overall, 14 of the 26 included trials were at unclear risk of bias for allocation concealment, and 10 trials were double-blinded. Many subgroup analyses were conducted in this review, and thus caution should be taken in the interpretation of results. As the majority of trials were conducted prior to the 1990’s, very few surviving extremely preterm infants were included in analyses. Contemporary populations of surviving preterm and low birth weight infants are more premature and lower in birth weight than the majority of infants included in this review, and therefore the findings may not be applicable to current settings. This is particularly important in the interpretation of the retinopathy of prematurity findings, as currently the infants at greatest risk of severe retinopathy are extremely low birth weight infants.

Vitamin E supplementation reduced the risk of germinal/intraventricular haemorrhage, but increased the risk of sepsis. No clear effect was found on mortality, bronchopulmonary dysplasia, patent ductus arteriosus, severe intraventricular haemorrhage, parenchymal haemorrhage, or retinopathy of prematurity. Due to an increased risk of sepsis, current evidence does not support the use of higher doses of vitamin E or intravenous administration, and higher serum tocopherol concentrations should also be avoided.

Further trials to investigate the effects of non-high-dose, non-intravenous vitamin E on severe retinopathy of prematurity, blindness, severe intracranial haemorrhage, sepsis, and long-term neurodevelopmental outcomes in extremely low birth weight or extremely preterm infants are warranted.