Seat belts and the safe car

that reported in such patients at the Whittington Hospital. Possibly this higher incidence results from our retrospective study in which "simultaneous blood and urine samples" were not collected. I doubt, however, that this can be the explanation as Dr Kennedy and his colleagues have not excluded the 590' of their patients in whom urine analysis is not reported. Drs S J Iqbal and P J Ojwang (9 December, p 1640) express concern about the conclusion that biochemical analysis of the urine is of little value in such patients. I would go further and ask what conclusion one is entitled to make on data from 18 isolated estimations in 44 patients. Surely the value of this estimation lies in daily evaluation of electrolyte homoeostasis. To describe hyponatraemia as "dilutional" on the basis that postoperative patients have been infused with undisclosed volumes of 5%o dextrose is dangerously misleading. Most of our patients received more than 100 mmol sodium/24 h; 5° dextrose was only occasionally used. Is surgical practice at the Whittington Hospital so very different ? Surely the widely recognised views subsequently expressed by Dr C T G Flear and his colleagues (9 December, p 1640) are in no way challenged by the information presented. Finally, the observation that hyponatraemia cleared rapidly after intravenous dextrose infusion had been stopped and that "water restriction or hypertonic saline was not needed" leaves the surgical houseman confused. Should he then infuse saline to his fasting patients or merely deprive water by omission? Saline, and particularly hyperosmolar saline, should be used with caution as many such patients easily become overloaded with sodium. Fortunately, hyponatraemia tends to recover with improvement in general condition, but two of our postoperative and jaundiced patients required prolonged treatment with dextrose, potassium, and insulin. These patients survived, whereas the mortality in our hyponatraemic group was worse than the 270o in the Whittington patients. It is hard to imagine what evidence can justify a statement that "although 12 deaths occurred among the 44 patients hyponatraemia did not play a part in any." Dr Kennedy and his colleagues planned to comment on inappropriate antidiuretic hormone secretion. They conclude that it is not a common cause of hyponatraemia. It had not occurred to us that it was. I hope house officers and students will be advised to treat hyponatraemia by established methods. They should not be discouraged from the selective postoperative infusion of 500 dextrose as a valuable source of the water upon which life depends. C N MCCOLLUM