The great American epidemiologist Bill Silverman taught us t
The great American epidemiologist Bill Silverman taught us that the road to hell is paved with good intentions. From 1942, during a 12-year period, more than 10 000 infants were blinded through retinopathy of prematurity because paediatricians recommended a minor increase in the concentration of supplemental oxygen for preterm infants in incubators. Later, “untold thousands of premature infants succumbed in the first few days of life because incubator temperatures were set slightly too low (to avoid overheating)”. Indeed, Silverman\'s trial of adrenocorticotrophic hormone to treat retinopathy of prematurity showed no benefit and an increased purchase Regorafenib in the intervention group. He warned that we need trial evidence for any new preterm intervention. No shortcuts can be used—care of the preterm infant is a delicate and integrated process. Recently, antenatal corticosteroid treatment has been widely promoted to reduce preterm deaths in developing countries. On the face of it, the argument seems to be persuasive. Worldwide, 15 million infants are born prematurely every year, and south Asia accounts for two-thirds and Africa for three-quarters of deaths in these infants. The 2013 State of the World\'s Mothers report says: “Prenatal corticosteroids cost as little as 51 cents per treatment… are ready for rapid scale-up now…using skilled birth attendants…and could save 340 000 newborn lives each year”. This estimate does not come from trials but rather from a Lives Saved Analysis tool. The evidence is far more nuanced. A systematic review of 21 studies in high-income and middle-income country hospitals, which included a total of 3885 women and 4269 infants, showed that one course of corticosteroids to accelerate fetal lung maturation in women at risk of preterm birth reduced neonatal deaths by 31% (relative risk [RR] 0·69, 95% CI 0·58–0·81; in 18 studies of 3956 infants) and respiratory distress syndrome by 34% (0·66, 95% CI 0·59–0·73; in 21 studies of 4038 infants). Nonetheless, the statement made by the that “administering antenatal steroids to a mom in preterm labour, helps her baby speed up lung development, reducing the risk of newborn death by more than 50% in low-resource facilities”—is speculative. The clamour to roll out a so-called magic bullet treatment must be resisted until three crucial questions have been answered. The first of these questions is whether antenatal corticosteroids actually work for mothers and infants at 33 weeks\' gestation or less in poor populations? All studies in the systematic review were from hospitals in which infants had access to “level 2” special care: 24-h availability of skilled nursing; thermal stability; monitoring of blood gases, glucose, electrolytes, infection indicators, and bilirubin; apnoea alarms; oxygen and ventilatory support; and antibiotics and other essential drugs for infection and shock. Similar reductions in case-fatality rate are highly unlikely in settings where level 2 care is not available. Coverage of good-quality level 2 care in low-income settings is tiny—few facilities exist outside capital cities. One study of 1000 low-birthweight infants in Dhaka, Bangladesh, showed that even in a teaching hospital that provided level 2 care, three-quarters of neonates born at less than 33 weeks\' gestation died during the neonatal period. The causes of death were manifold and not just attributable to respiratory distress. Indeed, respiratory distress syndrome might be less prevalent in poor countries than in wealthy nations because intrauterine growth retardation causes fetal cortisol concentrations to rise, which might accelerate the production of lung surfactant. The second crucial question asks whether safety issues exist in poor settings. If steroids are given to millions of women in preterm labour in poor populations, can we be sure that there will be no significantly increased risk of maternal sepsis, perinatal death, or childhood disability? The 2006 systematic review showed that, from eight trials and 1003 women with data for puerperal sepsis, 57 of 496 treated mothers had sepsis, compared with 44 of 507 controls (RR 1·35, 95% CI 0·93–1·95). More worrying is the evidence from trials that used dexamethasone (the low-cost treatment recommended for scale-up), which significantly increased both puerperal sepsis (RR 1·74, 95% CI 1·04–2·89; in four trials of 536 women) and fever that needed antibiotics (2·05, 1·14–3·69; in one trial of 118 women), in comparatively wealthy populations. These findings ring alarm bells for scale-up in populations in which the prevalence of malnutrition and the risk of sepsis are much higher and access to antibiotics is low. In infants born at a gestation of at least 36 weeks, an almost significant trend was recorded towards an increase in combined fetal and neonatal death (RR 3·25, 95% CI 0·99–10·66; in two studies of 498 infants).