The DART Study of Low-Dose Dexamethasone Therapy

DART · Journal of AAP · 2006

The DART Study (Dexamethasone: A Randomised Trial, Doyle et al., Pediatrics 2006/2007) investigated whether a low-dose, tapering course of late postnatal dexamethasone could facilitate extubation in chronically ventilator-dependent preterm infants without causing the severe neurodevelopmental harm (such as cerebral palsy) associated with older, high-dose regimens. The Studied Regimen (The "DART Protocol")A 10-day tapering course with a cumulative dose of 0.89 mg/kg (roughly one-tenth to one-fifth of traditional high-dose regimens like the 42-day taper): • Days 1–3: 0.15 mg/kg/day (given as 0.075 mg/kg IV/oral every 12 hours) • Days 4–6: 0.10 mg/kg/day (given as 0.05 mg/kg every 12 hours) • Days 7–8: 0.05 mg/kg/day (given as 0.025 mg/kg every 12 hours) • Days 9–10: 0.02 mg/kg/day (given as 0.01 mg/kg every 12 hours)

• Extubation Success: Significantly higher in the dexamethasone group compared to placebo by day 10 (60% vs. 12%; OR: 11.2, 95% CI: 2.5–50.7; p < 0.001).

• Ventilator & Oxygen Weaning: Duration of mechanical ventilation was significantly reduced in the dexamethasone group (median: 3.5 days to extubation vs. non-extubation/prolonged ventilation; p < 0.001), alongside faster reduction in supplemental oxygen requirements.

• Short-Term Safety: No statistically significant differences in acute adverse events between dexamethasone and placebo groups:

* Hypertension: 6% vs. 0% (p = 0.51)

* Hyperglycemia requiring insulin: 17% vs. 9% (p = 0.44)

* Proven sepsis: 29% vs. 24% (p = 0.77)

* Gastrointestinal perforation: 0% vs. 0%

• Long-Term Neurodevelopment (2-Year Follow-up):

* Mortality: 11% vs. 21% (OR: 0.47, 95% CI: 0.12–1.77)

* Cerebral Palsy: 9% vs. 18% (OR: 0.44, 95% CI: 0.09–2.19)

* Major Neurosensory Disability: 16% vs. 30% (OR: 0.45, 95% CI: 0.13–1.59)

* Combined Death or Major Disability: 26% vs. 44% (OR: 0.44, 95% CI: 0.16–1.21; p = 0.11)

1. The trial was stopped early after recruiting only 70 infants across 11 centers — a fraction of the target sample — because clinicians grew reluctant to randomize amid rising concern about steroid neurotoxicity.

2. Substantial "contamination" occurred — a meaningful number of placebo-arm infants received open-label corticosteroids outside the protocol later in their course, which dilutes any true between-group difference. T

(Rademaker et al.'s letter: they argued this contamination made the 2-year comparison difficult to interpret cleanly)

Rademaker et al. went as far as to say that, given the underpowering and contamination, "the only proper conclusion... is that no conclusion" should be drawn about long-term safety from this trial alone

DART is good evidence that low-dose dexamethasone helps get chronically ventilated preterm infants off the vent faster — that short-term finding held up

It is not good evidence, on its own, about long-term neurodevelopmental safety, because the trial never reached the size it needed to answer that question

Although Recent Metaanalysis (2022) showed Late (8-14 Day) Dexamethasone at dose 2-4 mg/kg - Best for reducing Death or BPD. ( low quality evidence)