Objective

Pulmonary embolism (PE) is a complication of long-term parenteral nutrition (PN) and has been associated with high mortality with incidence of 35%.

The aim was to review the incidence of PE in children receiving PN since 2003, compare with previous incidence and relate to content of PN, including use of new intravenous lipid emulsions.

Design

Surveillance ventilation-perfusion (V/Q) scintigraphies for all 48 children aged from 0.1 to 15 years on PN between 2003 and 2010 were retrospectively reviewed. A comparative analysis between the incidence of PE and risk factors pre- and post-2003 was performed. The underlying diagnosis was short bowel syndrome in 13 children or 27%, small intestinal enteropathy in 21 children or 44% and intestinalmotility disorder in 14 or 29%.

Results

Forty-eight children had a total of 170 V/Q-scans. Ten (6%) episodes of PE were detected accounting for mean 0.2/1000 CVC days in the before 2003 period to 0.02/1000 CVC days after 2003, p=0.01. Nine of the 48 children or 19% had at least one episode of PE (one child had two episodes), all without recognised symptoms. In seven scans positive for PE the perfusion defect was single and sub-segmental, one demonstrated two sub-segmental perfusion defects. In two scans there were large bilateral perfusion defects. Two patients with new small sub-segmental perfusion defects were already on treatment with warfarin. There were 22 patients who received PN before 2003 compared to 26 children after 2003.Demographic data including primary diagnosis that led to IF, age and duration of PN was similar between the two periods. The underlying diagnosis of the nine children with PE was short bowel syndrome in four (44%), intestinal motility disorder in three (33%) and small intestinal enteropathy in two (23%). The type of lipid infused at the time of the 10 PEs was pure soya lipid in six (including all 4 pre-2003), or 60%, a mixture of LCT/MCT in two cases or 20% and in the other two, 20%, no lipids were administered. There was no association between the age at onset of PN, duration of PN or the primary diagnosis and PE.

Conclusion

There was a 10-fold fall in the rate of PE during PN since 2003. No significant association with common risk factors could be found. Following diagnosis, anticoagulation is recommended at least for the duration of PN. Surveillance programmes with V/Q scans(or perfusion alone) are still important.