Results of the search
The original search strategy retrieved 77 unique citations of which we excluded 60 based upon title and abstract screening. We identified three randomised controlled trials of D-penicillamine in preterm infants for inclusion that reported on ROP outcomes involving 369 infants, along with an additional publication reporting one-year outcomes of one of the included trials.
A brief description of the included studies is as follows:
1. Lakatos 1986: randomised controlled trial of intravenous D-penicillamine given to 204 outborn infants of 750 to 2000 g birth weight between 1 January 1983 and 6 March 1984. Randomization after informed consent occurred in birth weight strata using sealed envelopes. D-penicillamine was started within 12 hours after birth, and given intravenously at 300 mg/kg/day (divided into three doses) for three days, then at 50 mg/kg/day (in a single dose) up to two weeks if the birth weight was less than 1500 g. Between 1500 g and 2000 g, drug was continued beyond three days only if the infant continued on oxygen. No placebo was used because of the characteristic odour of the drug. Starting at six weeks, infants were examined for ROP with an indirect ophthalmoscope without scleral depression. The ophthalmologist did not know which infants had earlier received D-penicillamine. Because scleral depression was not used, mild stages of ROP in zone 3 may have been missed in this study. However, since both groups were examined with the same technique, the rates of diagnoses in the two groups could be compared. The international classification of ROP was used to record the findings.
If acute ROP was diagnosed at any stage, children in the control group were then given D-penicillamine 50 mg/kg daily for three weeks. Therefore, the diagnosis of progression to severe ROP in this analysis, as well as other longer-term morbidities occurred following administration of active drug to some of the infants in the control group after six weeks of age. Most neonatal deaths occurred in the first week following birth, thus the outcome of death is unlikely to have been affected by administration of active drug to control infants who developed ROP. However, this study design feature contaminates to some degree other long-term outcomes such as growth and neurodevelopmental follow-up.
Deaths before discharge occurred at similar rates (29/100 in the treated infants, 34/104 in the controls), and ROP was diagnosed in none of the 71 treated survivors, and in six of the 70 control survivors. Two of the infants with ROP progressed to cicatricial stages; but neither of those cases progressed to blindness. No acute toxicity was reported in either group during hospitalisation, but there was no prospective, systematic collection of potential side effects either.
1a. Vekerdy-Lakatos 1987 (follow-up of infants in Lakatos 1986): at one year of age, the infants from the study reported above were evaluated in follow-up. There were three deaths after discharge so that 69 treated (69%) and 69 control infants (66%) survived to one year of age. Of these, 87% in each group returned for evaluation. Spasticity or seizures occurred in three (5%) of the treated infants and five (8%) of the children originally randomised to the control group. There were no significant differences in developmental quotients or growth parameters, but re hospitalizations occurred more frequently in control infants (59 times in 28 children) than in the treated infants (23 times in 15 children).
2. Lakatos 1987 (2nd trial data): the results of this trial were reported by the authors in combination with the results of Lakatos 1986 trial. Therefore, the results from this separate randomised controlled trial have been extracted from the data provided in these two publications. Seventy-seven infants of birth weights 751 to 1500 g were randomised to receive D-penicillamine or be controls between 1 July 1984 and 1 March 1985. Randomization was designed to be weighted 2:1 (treated:control) and the study was to end if three or more infants in the control group developed ROP. The dose was 300 mg/kg/day (divided into three doses) for three days, then 50 mg/kg/day (single dose) through to two weeks of age. Infants in the control group who later developed acute ROP were then given active drug, D-penicillamine, for three weeks (50 mg/kg/day intravenously) starting at the time of diagnosis of the ROP.
While a 2:1 randomisation ratio was intended through sealed envelopes, when the study was terminated after three infants developed ROP in the control group, the actual ratio of treated versus no treated infants was 1.4:1. Eighteen of 45 infants treated with D-penicillamine died compared to 10/32 control infants. Acute ROP occurred in 0/27 D-penicillamine survivors and in 3/22 control survivors.
3. Tandon 2010: randomised, double-blind, single-centre, placebo-controlled trial in which oral D-penicillamine was given to 88 inborn neonates with birth weight 750 to 1500 g, gestation 32 weeks or less, age 5 days or less. Subjects were stratified for birth weight (less than 1250 g and 1250 g or greater). Block randomisation with randomly varying block sizes was carried out. The random sequence was generated online from a website by an investigator who was not involved in recruiting subjects. Between 1 January 2007 and 31 December 2007, enrolled subjects were randomly allocated to receive oral D-penicillamine suspension at 100 mg/ kg/dose every eight hours for three days, followed by 50 mg⁄/kg/day for another 11 days or placebo. The primary outcome was 'any ROP or death'. Secondary outcomes included any ROP, treatable ROP, adverse effects and feeding intolerance. Retinal examination was performed by an ophthalmologist with expertise in ROP screening at four to six weeks after birth or at a postmenstrual age of 31 to 33 weeks, whichever was earlier. ROP screening was repeated depending on the status of the retina.
Subjects were followed up until death, treatment of ROP or until the last retinal examination was performed and the baby was declared to have no ROP. A total of 88 infants were enrolled. The D-penicillamine and placebo groups comprised 44 infants each. No subject was lost to follow-up. Ten of 44 infants (22.7%) in each group either died or had ROP. Seven of 44 infants (15.9%) from the D-penicillamine group and 4/44 from the control group developed ROP. One baby in the D-penicillamine group had ROP and subsequently died. The other deaths took place before the first ROP screen. Three subjects from the D-penicillamine group and one from the placebo group required treatment.