When readymade parenteral nutrition in multi-chamber bags is supplied without vitamins and minerals, these have to be added or given separately. Separate rapid infusion has logistical advantages but has been claimed to saturate tissue mechanisms, potentially increasing urine micronutrient losses. The present study compared urinary losses after fast (1 h) v. slow (10 h) trace elements infusion. The study enrolled thirty-nine consecutive patients who were starting parenteral nutrition postoperatively. One day’s dose of a complete intravenous micronutrient product was infused over 1 h and over 10 h, in random order, with a washout day after each infusion day. Urinary Zn, Mn, Se, Cr, Cu and Fe losses were measured by 24-h urine collection, recorded for each infusion day and subsequent washout day. For Zn, a dose of 100 μmol was given, and total urinary loss over the next 2 d was mean 40·6 (sd 23·8) μmol after the fast (1 h) infusion v. 33·4 (sd 25·4) μmol after the slow (10 h) infusion, that is, 7 % more of the 1-d dose was lost after fast infusion (P = 0·01). For Mn, after a dose of 1000 nmol, losses were 9·8 (sd 23·9) nmol after the fast infusion v. 22·1 (sd 47·2) nmol after the slow infusion, that is, 1 % more of the 1-d dose was lost after slow infusion (P = 0·04). There were no other significant differences: after 1 μmol Se, the losses were 1·5 (sd 0·6) μmol fast v. 1·3 (sd 0·5) μmol slow; after 200 nmol Cr, 257 (sd 92) μmol fast v. 246 (sd 107) nmol slow; after 8 μmol Cu, 1·6 (sd 1·4) μmol fast v. 1·5 (sd 1·3) μmol slow; and after 20 μmol Fe, 0·6 (sd 1·1) μmol fast v. 0·8 (sd 1·6) μmol slow (P > 0·05 for all). Overall, trace element retention appears to be minimally affected by infusion time.