Two methods of predicting difficult laryngoscopy were compared prospectively. Mallampati class and Wilson risk-sum were determined before operation and laryngeal view graded in 675 patients. Both tests identified five of 12 difficult laryngoscopies; twice as many patients were predicted to be difficult by Mallampati classification than by Wilson risk-sum. Inter-observer variation was minimal using Wilson risk-sum, but considerable for Mallampati classification. We prefer the Wilson risk-sum for assessment of the airway, while noting that both tests have poor sensitivities.

No abstract

In a placebo-controlled, double-blind study, we have compared the efficacy of ondansetron 16 mg, 8 mg and 1 mg administered 8-hourly for prevention of postoperative nausea and vomiting. We studied 995 patients undergoing major gynaecological surgery; 982 were included in the analysis. Study medication was administered 1 h before induction of anaesthesia and second and third doses were given 8 and 16 h after the first. The treatment groups were similar for patient characteristics, surgical procedures, anaesthetics administered and opioids given. The frequency of nausea was 75%, 70%, 56% and 55% after placebo and ondansetron 1 mg, 8 mg and 16 mg, respectively; the corresponding frequencies of vomiting were 60%, 55%, 37% and 37%. Ondansetron 8 mg was as effective as 16 mg and both resulted in significant reductions in nausea and vomiting compared with placebo and ondansetron 1 mg (P less than 0.001).

Aspirates were obtained from 142 masses in 120 patients who presented with palpable swellings in the head and neck region. 120 specimens (84.51 per cent) were adequate for diagnostic purposes and the remaining 22 (15.49 per cent) were unsatisfactory. Final diagnosis was based on resection histology in 87 cases and close clinical follow-up in 55 patients. The overall sensitivity and specificity including unsatisfactory aspirates was 81.37 per cent and 93 per cent respectively (if the technically inadequate specimens were deleted 98.81 per cent and 94.44 per cent). In comparison with the final diagnosis typing of malignant tumours was possible in 58.33 per cent. The correct diagnosis was made in 63.89 per cent of benign lesions.

No abstract