Skip navigation

Vomiting after Squint Surgery

Nausea and vomiting are unpleasant but common postoperative complications. It is a particular problem for squint correction, because many of these procedures are carried out as day-cases and vomiting can occur as long as 40 hours after operation. In one study of more than 300 outpatients having squint correction, as many as 38% of overnight admissions resulted from vomiting.

Measures which reduced postoperative vomiting, both early and late after operation, would not only benefit the children and their parents, but would also reduce stress on the providers.


In these circumstances effective prophlaxis would seem to be a good idea. Is prophylaxis effective?

Some answers come from a systematic review of the prevention of vomiting after paediatric strabismus surgery [1]. Information was available on 2033 children in 27 randomised controlled studies in which drugs from eight different pharmacological classes were used. Because vomiting is an "all-or-nothing" phenomenon, information was available in dichotomous (yes/no) form, allowing calculations of numbers-needed-to-treat (NNT). Vomiting was examined early (up to six hours after operation) and late (up to 48 hours).


Because of the large number of drugs and doses used, there were few treatment groups where sufficiently large numbers of children had been studied to warrant any certainty about results. Ondansetron, for instance, was studied in one trial with 30 patients, though the data for these studies is given in the report [1].

Nevertheless, a number of solid conclusions could be drawn.

Vomiting without active prophylaxis

In the absence of any active prophylaxis (about 400 children in placebo or no-treatment control groups), just over half of the children vomited within six hours. However, the range was great; in one trial only 18% of children vomited while in another 88% of children vomited.

Late vomiting (studied in about 300 children) occurred on average in about 60% of children, but again the range in individual trials was wide (43 - 97%).

In these studies children were undergoing the same operation with very similar anaesthetic techniques. Despite this, huge variations in the rate of vomiting occurred. Clearly there are factors other than surgery and anaesthesia which are important (an unknown).


The best studied antiemetics were droperidol at 75 µg/kg and metoclopramide at 0.15 mg/kg.

Droperidol (in about 270 children) was better than placebo in preventing both early and late vomiting, with NNTs (with 95% confidence intervals) of 3.5 (2.8 - 4.8) and 4.4 (3.1 - 7.1) respectively. Metoclopramide (in 120 children) produced NNTs for early and late vomiting of 4.0 (2.7 - 7.6) and 4.1 (2.5 - 12.3) respectively.

Minor adverse effects (drowsiness, restlessness, agitation) occurred more frequently with droperidol than in controls with a NNT of 6.3 (4.6 - 10.2). The occurrence of a major adverse effect (extrapyramidal symptoms)was not statistically significant, but had a point NNT estimate of 123.


Propofol is an induction agent used intravenously, and it has been suggested that its use is associated with a lower incidence of postoperative nausea and vomiting. In a number of studies in which anaesthesia with propofol was compared with anaesthesia without propofol, the incidence of both early and late vomiting was somewhat less, but in the study with the largest numbers the effects barely achieved statistical significance and NNTs were large with wide confidence intervals.

The use of propofol was, however, associated with an incidence of oculocardiac reflex (OCR) of about 50%. This reflex is decrease in heart rate of more than 15-20% which occurs when the eye muscles are pulled during surgery. The bradycardia occurred despite measures to prevent it. The NNT for OCR was 3.6 (2.6 - 6.3); one in every four children given propofol rather than halogenated anaesthetic had a bradycardia.


There are a number of issues raised by these results. For droperidol 75 µg/kg, four children have to be given the drug to prevent one vomiting.; of the other three, one may vomit and two would not have vomited anyway. Bandolier 16 examined the issues involved in prophylaxis, including the severity of the medical condition being treated, the effectiveness of the treatment and any adverse effects which might ensue. It is an interesting judgement as to whether prophylaxis is worthwhile here.

This judgement is particularly important when the rate of vomiting without prophylaxis varies so widely. If the incidence of early vomiting in controls (on average 54%) could be brought down to the lowest figure (18%), that would be much more effective than using antiemetic prophylaxis.

The other major point is adverse effects. Perhaps the most important result from this systematic review was that concerning the link between propofol use and OCR. Given that OCR can occasionally become chaotic arrhythmia or sinoatrial arrest, the conclusion that this form of anaesthesia should not be used would seem about right.

That in turn gives purchasers an interesting problem. Just how far into the purchasing system should this be carried. One argument is that the professionals know what they are doing; the contrary is that purchasers are to some extent "in loco parentis" and should strive to avoid potentially harmful interventions when adequate and safe alternatives exist. Food for thought.


M Tramèr, A Moore, H McQuay. Prevention of vomiting after paediatric strabismus surgery: a systematic review using the numbers-needed-to-treat method. British Journal of Anaesthesia 1995 75: 556-61.

next story in this issue