OBJECTIVES

This simulation-based economic evaluation aimed to assess the cost-effectiveness of lactoferrin in preventing late-onset infections (LOIs).

METHODS

2,203 very preterm infants (<32 weeks’ gestation) from the United Kingdom (UK) were randomised to receive either lactoferrin or placebo. Generalised linear models were used to analyse data on LOI, mortality and cost of primary hospitalisation. Non-linear continuous variables (e.g. gestation age) were modelled using restricted cubic splines. Assuming that LOIs increase the risk of neurodevelopmental impairments (NDI), the cohort model calculated the proportion of infants developing NDI and, based on NDI severity, extrapolated the infants’ life-expectancy. These survival estimates were adjusted for quality of life utilities to obtain quality-adjusted life years (QALYs) and were used to estimate lifetime costs. Between groups differences in costs and QALYs were used to calculate Incremental cost-effectiveness ratios (ICERs) and a probabilistic sensitivity analysis determined the ICERs uncertainty. This uncertainty was presented as Value of Information (VOI) at UK-population level. A 1.5% annual discount rate and lifetime time horizon were used to capture long-term costs and consequences.

The analyses were repeated for gestation age and feed-type subgroups, and under alternative assumptions on the discount rate and treatment effect. Missing data were multiple imputed assuming they were missing at random (MAR) or not MAR.

RESULTS

On average, lactoferrin would provide an additional 0.05 QALYs at the cost of £1787. Apart from infants with 24 weeks’ gestation (ICER: £6,665 [1.5% discount] and £11,187 [3.5% discount]), lactoferrin was not cost-effective in any of the subgroups. The VOI for UK infants with 24 weeks’ gestation is approximately £15 million. Our results were robust to alternative assumptions.

CONCLUSIONS

From the UK-NHS perspective, lactoferrin is not cost-effective based on conventional decision rules. More research in infants with 24 weeks’ gestation might be justifiable if the undergoing studies (e.g. LIFT trial) do not resolve these uncertainties.