Design: multicentre, open-label, randomised controlled trial with 2-by-2 factorial design, up to 28 days of treatment
Population: 1,956 critically ill mechanically ventilated patients aged 16 or over with acute respiratory failure and difficult-to-clear secretions
Primary outcome: duration of mechanical ventilation from randomisation to first successful unassisted breathing
Effect: carbocisteine 186.1 vs 172.7 hours (adjusted HR 0.96, 95% CI 0.87-1.05); hypertonic saline 184.5 vs 174.3 hours (adjusted HR 1.00, 0.91-1.10); GI bleeding RR 6.51 (1.47-28.76); nebulisation hypoxaemia RR 13.29 (4.12-42.83)

A multicentre open-label 2-by-2 factorial trial randomised 1,956 mechanically ventilated patients aged 16 or over with acute respiratory failure and difficult-to-clear secretions to enteral carbocisteine 750 mg three times daily, nebulised 6-7% hypertonic saline 4 ml four times daily, both, or usual care alone, for up to 28 days. There was no interaction between the two (HR 1.01, 95% CI 0.83-1.22).

Neither shortened ventilation. Median duration was 186.1 hours with carbocisteine against 172.7 without (adjusted HR 0.96, 0.87-1.05) and 184.5 hours with hypertonic saline against 174.3 without (adjusted HR 1.00, 0.91-1.10). Both directions of the point estimate favour not giving the drug, though not significantly.

The harms did reach significance. Clinically important upper gastrointestinal bleeding occurred in 13 of 965 on carbocisteine against 2 of 966 without (RR 6.51, 1.47-28.76). Hypertonic saline produced bronchoconstriction needing a bronchodilator in 2.4% against 0.4% (RR 5.73, 1.99-16.52) and hypoxaemia during nebulisation in 4.1% against 0.3% (RR 13.29, 4.12-42.83). These are drugs given for secretions that are already a problem, in patients whose oxygenation is already the reason they are ventilated — and the nebulisation itself is desaturating one in twenty-five of them.

Stop prescribing carbocisteine and nebulised hypertonic saline as routine secretion management in ventilated patients
Where nebulised hypertonic saline is used for another indication, monitor saturations through the nebulisation, not just before and after
Review any patient on enteral carbocisteine who develops a fall in haemoglobin or melaena
Direct the effort to what does clear secretions: positioning, humidification, suction technique and early mobilisation
These findings concern mucoactives in undifferentiated acute respiratory failure, not nebulised saline in cystic fibrosis or bronchiectasis

Why it matters

Two treatments given by habit for a problem everyone recognises turn out to do nothing for it and to cause something else.

The statistics, in plain English

Both efficacy hazard ratios sit essentially on 1.0 with tight intervals, so this is a definitive null rather than an underpowered one. The harm ratios look enormous (6.51, 13.29) because the comparator rates are tiny — 2 events and 3 events respectively. The absolute figures are the usable ones: about 1 extra gastrointestinal bleed per 85 patients treated with carbocisteine, and about 1 extra desaturation episode per 26 treated with hypertonic saline.