BRITISH MEDICAL JOURNAL 14 MAY 1977 1241 to temper enthusiasm with caution. Abrams9 questions the disease, particularly those who have had episodes of ence- common description of colonoscopy as being "simple, rapid, phalopathy.4 This sensitivity is not associated with abnormally thorough, reliable, inexpensive and safe." Far from being high concentrations of in the blood, and simple, it is the most difficult of the alimentary endoscopies, increased cerebral sensitivity rather than impaired drug and it has been suggested that a doctor should perform at metabolism has been implicated.5 In recent years the benzo- least 200 examinations before attempting polypectomy. The diazepine drugs have been increasingly used and studied in use of the sideviewing telescope, however, does enable the liver disease. Murray-Lyon et a16 in a clinical and EEG study trainee to watch the expert at work and is an invaluable found that there was little tendency for sedation to be pro- teaching accessory. Abrams emphasises that the examination longed in patients who had received a 5-mg intravenous dose is often extremely time consuming; that it is not always possible and no increase in slow wave activity in the EEG such as is to examine the whole of the colon; and that the presence of found after administration of opiates or phenothiazines. fixed loops or bends (particularly as a result of previous Despite these findings many clinicians will have encountered surgery, diverticular disease, or irradiation) may make the occasional patient with cirrhosis who develops prolonged it impossible to pass the instrument the length of the colon. stupor or coma after a single dose of . Such an ob- Colonoscopic biopsy is limited technically by the size and servation prompted the recent study of Branch et al.7 These depth ofthe bite and the sample may not reflect the true histo- workers measured the EEG responses and drug kinetics after logical nature of the lesion. In many cases the polyp may be an intravenous infusion of diazepam (1 mg/min) sufficient to "lost"; this occurred in 500 of all tissues removed in Abrams's produce enough sedation for a procedure such as endoscopy. own practice. Abrams points out the considerable expense of Five normal subjects and 17 patients with histologically proved setting up a colonoscopic unit and estimates that the 3000 chronic liver disease were investigated. The dose of diazepam colonoscopes in the USA represent a total initial expenditure needed to produce the required response was significantly of about 75 million dollars. The average life of the instrument lower in the patients (mean 17-9 mg) than in the controls is about 200 examinations, so there is a continued need for (mean 27 mg), and the dose correlated well with plasma further expenditure on repairs and replacements. Incidences albumin levels. Plasma diazepam concentrations at the end of of perforation and haemorrhage of up to 5% have been re- infusion, however, were similar in the two groups despite the ported after polypectomy, and the true complication rate for lower dose requirement in those with liver disease, and the colonoscopy is unknown. authors suggest that this may be explained by a dose-dependent Nevertheless, an instrument is now available of high diag- increase in the volume of distribution of the drug. Although nostic accuracy which also allows open surgery to be avoided the plasma drug concentrations were similar, there was a in many cases. Hence the financing of expensive equipment greater slowing in the EEG in the patients with liver disease, and, particularly, the training of skilled endoscopists are particularly those who had had encephalopathy, and this problems that will have to be faced if our patients are to reap strongly suggests increased cerebral sensitivity. Plasma the benefits of this important advance. diazepam clearance was reduced in the patients with liver disease to a degree that correlated with albumin concentration. 1 Berci, G, Endoscopy. New York, Appleton Century Crofts, 1976. No patients in the study developed prolonged sedation. 2 Williams, C, and Teague, R, Gut, 1973, 14, 990. 3 Loose, H W C, and Williams, C B, Proceedings of the Royal Society of The major site of metabolism of diazepam is the liver, and Medicine, 1974, 67, 1033. drug half life is about two days. In normal subjects the drug 4 Teague, R H, Salmon, P R, and Read, A E, Gut, 1973, 14, 139. is 5 Williams, C B, et al, British Journal of Surgery, 1974, 61, 673. will accumulate if repeated doses are given.8 As this clearly Wolff, W I, and Shinya, H, New England Journal of Medicine, 1973, 288, more likely to occur in patients with liver disease, great care 329. must be taken with repeated doses. Other 7 Figiel, L S, Figiel, S J, and Wietersen, F K, Acta Radiologica, 1965, 3, 417. 8 Beahrs, 0 H, and Sanfelippo, P M, Cancer, 1971, 28, 213. have not been much studied in liver disease, but the 9 Abrams, J S, American Journal of Suirgery, 1977, 133, 111. 10 mg (Mogadon) and 30 mg (Dalmane) impair visuomotor co-ordination performance for at least 16 hours after ingestion even in normal subjects.9 Diazepam is better in this respect, and ifnight sedation is needed in patients with liver disease then diazepam 5-10 mg seems appropriate. For alcoholic patients with chlormethiazole Sedation in liver disease (Heminevrin) is widely used in small doses, although pro- longed sedation may result and care must be taken. Pharma- Patients with liver disease are thought to be unduly sensitive to cological studies in patients with liver disease have not yet been many drugs, and doctors have been warned to be cautious in published. Patients with cirrhosis may sometimes need prescribing for any patient whose liver function is impaired.1 2 , but they are extremely sensitive to monoamine Sedation presents a particular problem, especially since oxidase inhibitors and these drugs are best avoided.'0 A prolonged hepatic coma may follow the injudicious use of tricyclic such as is safer but doses and paraldehyde.3 While it is sound clinical practice should be kept low. to avoid all unnecessary sedation in patients with liver disease, Branch et a17 point out that the serum albumin concentration it is occasionally essential to control anxiety and agitation, provides a useful index of liver function for predicting the which may be early features of hepatic encephalopathy, and response to intravenous diazepam-not unexpectedly, as to gain co-operation for a procedure such as upper gastro- protein binding of diazepam is thought to be due to albumin. intestinal endoscopy. Furthermore, alcoholic patients who Thus if the serum albumin concentration is low more of a may have cirrhosis are apt to develop delirium tremens. given dose may be available as free diazepam for uptake into What drugs may be recommended? Phenothiazines such as the tissue. Serum albumin concentrations have correlated chlorpromazine have been popular in the past, but chlorproma- with the clearance of other drugs,11 12 and Branch et a17 found zine causes stupor and considerable slowing on the electro- correlations between the clearance rate of diazepam and those encephalogram (EEG) in some patients with chronic liver of antipyrine, propranolol, and indocyanine green-drugs with 1242 BRITISH MEDICAL JOURNAL 14 MAY 1977 different clearance rates and routes of metabolism. This ribs broken, and the number and nature of associated injuries. suggests that impaired hepatic function can be expected to But the inappropriately ventilated group suffered more reduce the clearance of most drugs that are eliminated by the complications than the unventilated patients. These complica- liver irrespective of the mechanism of drug clearance, and tions included pneumonia, damage to the trachea, and induced furthermore the degree of impairment seems to be quantita- pneumothorax in nine out of 15 patients as against some tively similar. collapse of lung in four of the 11 who were not ventilated. It seems, then, that there is no cause for alarm, and the These findings were consistent with the authors' conclusion results of the Bristol study7 suggest that diazepam remains the from a review ofpublished work that the prospects ofsuiviving drug of choice for sedation in liver disease. Obviously it should paradoxical respiratory movement resulting from injury had be given with care in small doses, only when absolutely es- not shown any definite improvement in 10 years.7 sential, and preferably by slow intravenous infusion, since the We may justifiably ask whether our far better understanding maximum clinical effect is seen within minutes of injection of respiratory physiology can be used better than it is. On the and the correct dose is easily found. Particular caution is one hand, can we expect to improve the prospects of dealing needed in patients with low serum albumin concentrations or a successfully with the severe injury? On the- other, can we history of hepatic encephalopathy. reduce the need for artificial ventilation by using internal fixation ? We might expect that prompt restoration of stability IBritish Medical_Journal, 1973, 2, 193. to the chest wall and relief of pain would enable the patient to 2 James, I, British Journal of Hospital Medicine, 1975, 13, 67. ventilate his lungs with no more help than that of the physio- 3Laidlaw, J, Read, A E, and Sherlock S, Gastroenterology, 1961, 40, 389. 4 Read, A E, Laidlaw, J, and McCarthy, C F, British Medical3Journal, 1969, therapist. This is indeed the point of view of Paris and his 3, 497. colleagues8 and also of Moore,9 but they have not carried out 5 Maxwell, J D, et al, Clinical Science, 1972, 43, 143. controlled trials nor have they published the biochemical data 6 Murray-Lyon, I M, et al, British Medical3Journal, 1971, 4, 265. 7Branch, R A, et al, Gut, 1976, 17, 975. on which their case should rest. Except for a broken sternum,10 8 De Silva, J A F, Koechlin, B A, and Bader, G, Jfournal of Pharmaceutical for rupture of the diaphragm, and for stabilising broken ribs Sciences, 1966, 55, 692. 9 Nicholson, A N, in Twelfth Symposium on Advanced Medicine, ed D.K in the course of an otherwise necessary thoracotomy, the case Peters, p 163. Tunbridge Wells, Pitman Medical, 1976. for surgical repair of the chest wall has yet to be established. "' Morgan, M H, and Read, A E, Gut, 1972, 13, 697. In those cases in which assisted ventilation is essential for '1 Levi, A J, Sherlock, S, and Walker, D, Lancet, 1968, 1, 1275. Mawer, G E, Miller, N E, and Turnberg, L A, British Journal of Pharma- even immediate survival it has to be admitted that the outlook cology, 1972, 44, 549. is not encouraging, because severe damage to the chest wall is likely to be accompanied by severe damage to the lung,"I in the presence of which it becomes necessary to raise both the inflating pressure and the concentration of oxygen in the inspired air if dangerous hypoxaemia is to be avoided. Add the Management of the stove-in pulmonary complications of other injuries,'2-'4 and the resuscitating team finds itself between the devil of hypoxaemia chest with paradoxical and the deep blue sea of oxygen poisoning.'5 16 For these cases the only remaining hope is a membrane oxygenator'7 if one is movement accessible and can be spared from more certainly successful use. Nobody who has had any experience of the stove-in chest will Taxing as these cases are, they are few and far between. It deny that it may be a difficult clinical problem. Once it became is both comforting and salutary to reflect that the best prospect recognised that the aim should be to restore not only the of promoting recovery from most stove-in chest injuries lies in stability of the chest wall but also its shape, treatment moved prompt diagnosis and prompt attention to the basic and from compressive support to skeletal traction' and to internal essentially simple techniques of clearing the air passages, fixation.2' These methods had their successes, but the current expanding the lungs, relieving pain, and providing good practice of internal pneumatic splintage was ushered in by physiotherapy. Avery et a14 in 1956 and was reinforced in Britain by Griffiths5 in 1960. With time, however, the disadvantages and also the Proctor, H, and London, P S, British Joturnal of Surgery, 1955, 42, 622. 2 Coleman, F P, and Coleman, C L, Stur-gery, Gynecology and Obstetrics, dangers of Gram-negative infection of the bronchopulmonary 1950, 90, 129. tree have become increasingly clear. 3 Crutcher, R R, and Nolen, T M,Journal of Thoracic Surgery, 1956, 32, 15. ' Avery, E E, Morch, E T, and Benson, D W, J7ournal of Thoracic Surgery, Impressed by these dangers, Shackford and his colleagues6 1956, 32, 291. reviewed the patients with stove-in chest they had treated and Griffiths, H W C, Jouirtnal of the Royal Collcge of Surgeons of Edinburgh, realised that although some given artificial ventilation had need- 1960, 6, 13. Shackford, S R, et al, American Jounal of Surgcry, 1976, 132, 759. ed it because of much reduced ventilatory capacity, for others Relihan, M, and Litwin, M S,Journal of Trautma, 1973, 13, 663. the need was so much open to question that they regarded them 8 Paris, F, et al, Thorax, 1975, 30, 521. as being inappropriately ventilated. From then on, they resorted 9 Moore, B P,J7ournal of Thoracic and Cardiovascuilar Suirgery, 1975, 70, 619. 0 Henry, L, British Medical_Journal, 1957, 2, 339. to artificial ventilation only when there was manifest clinical "Garzon, A A, et al, Annals of Thoracic Surgery, 1966, 2, 629. distress and the Pao2 was less than 8-0 kPa (60 mm Hg) when 12 Martin, A M, Simmons, R L, and Heisterkamp, C A, Annals of Suirgery, air or the Pco2 was more than 8-0 kPa. They re- 1969, 170, 30. breathing ; Hadley, R N, Lulloff, R, and Birnbaum, M L, J7ournal of Bone and joint frained from ventilating those who would previously have Surgery, 1974, 56A, 396. fallen into the inappropriately ventilated group. 14 Prys-Roberts, C, Injuery, 1973-4, 5, 67. 5 Winter, P M, and Smith, G, Anaesthesiology, 1972, 37, 210. Patients in these two groups, unventilated and inappro- " Sevitt, S, Journal of Clinical Pathology, 194, 27, 21. priately ventilated, were closely similar in age, the number of 17 Hill, J D, et al, Nezw, England Journal of Medicine, 1972, 286, 629.