SummaryThirty patients who had undergone elective abdominal aortic surgery were studied in a prospective, randomised double-blind comparison of thoracic epidural 0.2% bupivacaine alone, thoracic epidural fentanyl alone and thoracic epidural 0.2% bupivacaine combined with ,fentanyl. Pain relief, pulmonary function. cardiovascular stability and side effects were assessed. Pain relief was e.ucellent in the combined bupivacaine-fentanyl series, being signiJicantly better than the other groups ( p < 0.05) during the entire study period and was not accompanied by hypotension. Forced expiratory parameters were reduced in all groups throughout the study to 50-60% of the pre-operative values, but there were no significant dixerences between groups. The incidence qf side effects attributable to either epidural bupivacaine or fentanyl was low. This study supports the increasing use of' epidurul infusion analgesia .for postoperative pain management after abdominal surgery.

Summary Midazolam, a new water-soluble benzodiazepine, was investigated as an intravenous anaesthetic agent in

In a previous study' a surprising finding was that patients premedicated with diazepam (Valium) 10 mg orally were better sedated during the first 90 minutes following administration than when the same dose was given by intramuscular injection into the buttock. This was confirmed in a 'double-blind, double-dummy' study involving 200 comparable patients in whom the drugs were given by the nursing staff.' Plasma diazepam levels were estimated to seek an explanation for these clinical findings, the study being carried out to show the influence of the route of administration on the plasma diazepam levels following a single 10 mg dose. Method Plasma diazepam estimationsThese were done by gas-liquid chromatography following benzene e~traction.~ The method of estimation, its reliability and reproducibility has been described in detail in another paper from this d e~a r t m e n t .~ SubjectsThese were fit young women scheduled for morning minor gynaecological operations who had fasted overnight. Patients who had had diazepam in the previous 4 weeks were excluded as were any patients who, despite the screening, were found to have diazepam or its metabolite in the control plasma sample. Table 1 shows the number of patients. The four groups studied were broadly comparable with respect to average age and weight. The preliminary results following the injection of diazepam into the buttocks by nurses, who normally used a 23 swg (3 cm) needle, led us to repeat part of the investigation when the drug was given into the upper outer quadrant of the buttock by one of the authors using a large 21 swg (4 cm)

Plasma diazepam concentrations were measured by gas-liquid chromatography in samples of blood from adult female patients follwing diazepam 10 mg orally, alone or in combination with metoclopramide, morphine, pethidine or atropine. Patients receiving metoclopramide had higher plasma diazepam concentrations than those in the control group, while the addition of morphine, pethidine or atropine resulted in lower plasma diazepam concentrations throughout the 90-min period of the study. In the control goup peak plasma concentrations were reached by 60 min. The addition of metoclopramide increased the rate of diazepam absorption and peak concentrations were reached by 30 min, while morphine, pethidine and atropine reduced the rate of absorption with no apparent peak being reached by 90 min.

It is generally agreed that absorption of drugs occurs more rapidly following their intramuscular injection than when given by mouth.'-3 This is the basis for the giving of premedicants by injection, which although challenged by Inglis & Barrow in 1965,4 is still widely practised.Recently, in a pilot study, McCaughey & Dundee' found better sedation when 10 mg diazepam was given by mouth than following its intramuscular injection, but their investigation was carried out on a small number of patients. In view of the importance of this finding a large-scale study was undertaken involving both the oral and intramuscular routes, in which neither the patient nor the observer knew whether the capsule or injection was the active drug. As a result of this, and studies of blood diazepam levels, a further smaller investigation was initiated when injections were all given by a doctor.The findings involve many aspects of the action of diazepam and this presentation is designed to show the relative merits and demerits of the two routes of administration. No attempt was made to compare the action of diazepam with that of other premedicants or a placebo. Method PatientsThe patients were fit adult women from one hospital unit, scheduled for morning minor gynaecological operations. They were first questioned as to their use of tranquillisers or sedatives, and if they had been on long-term therapy within the previous month or had been given a hypnotic on the night prior to operation they were excluded from the study. DrugsEach patient was given one capsule by mouth and a 2 ml injection about 90 minutes