Intraoperative gut hypoperfusion may be a risk factor for postoperative nausea and vomiting

British Journal of Anaesthesia 1997; 78: 476–482 CORRESPONDENCE Intraoperative gut hypoperfusion may be a risk factor for postoperative nausea and vomiting Sir,—We read with interest the article on nausea and vomiting after cardiac surgery by Grebenik and Allman.1 There are two issues which we believe are important but were not discussed by the authors. Gut mucosal hypoperfusion is common after cardiac surgery and may be another factor that explains the high associated incidence of postoperative nausea and vomiting.2 Animal studies have shown that damage to the gut is accompanied by an increase in 5-HT3 in gut mucosa.3 Radioligand binding studies demonstrated the presence of large number of 5-HT3 receptors on vagal terminals that innervate the intestinal mucosa and on the same vagal afferent nerves located in the brain stem vomiting centre.3 Furthermore, it has been demonstrated recently that plasma volume expansion during cardiac surgery is associated with maintenance of gut perfusion and reduction in postoperative morbidity, including persistent nausea and vomiting.4 The authors mentioned that “if the patients did not meet the criteria for tracheal extubation and required sedation to tolerate intubation and ventilation, an infusion of propofol 1–4 mg kg91 h91 was commenced until another trial of spontaneous breathing was performed at the discretion of the nursing staff”. Propofol is an antiemetic in its own right. Numerous studies have shown that propofol as a primary agent for maintenance of anaesthesia was associated with a lower incidence of postoperative nausea and vomiting.5 6 When used in this manner, it is more efficacious than ondansetron 4 mg.7 We have defined plasma concentrations of propofol for effective treatment of PONV as 405 ng ml91 (95% confidence interval 280–530 ng ml91).8 Data from computer simulation using the pharmacokinetic variables of Gepts and colleagues9 demonstrated that infusion of propofol 1–4 mg kg91 h91 for sedation, as used by the authors, produces plasma propofol concentrations in the effective range for treatment of nausea and vomiting for up to 180 min after termination of infusion (fig. 1). However, they did not provide information on the number of patients treated with propofol and the duration and dose of propofol used in each group, which may affect the overall results.

Figure 1 Simulated plasma decay of propofol concentrations after 2-, 4-, 6- and 8-h infusions.

Hence it is important to specify the details of propofol administration in studies that investigate the incidence of PONV. T. J. GAN M. G. MYTHEN P. S. A. GLASS Department of Anesthesiology Duke University Medical Center Durham, NC, USA 1. Grebenik CR, Allman C. Nausea and vomiting after cardiac surgery. British Journal of Anaesthesia 1996; 77: 356–359. 2. Gan TJ, Mythen MG. Does peroperative gut-mucosa

3.

4. 5. 6. 7.

8.

9.

hypoperfusion cause postoperative nausea and vomiting? Lancet 1995; 345: 1123–1124. Stables R, Andrews PL, Bailey HE, Costall B, Gunning SJ, Hawthorn J, Naylor RJ, Tyers MB. Antiemetic properties of the 5HT3-receptor antagonist, GR38032F. Cancer Treatment and Research 1987; 14: 333–336. Mythen MG, Webb AR. Per-operative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during cardiac surgery. Archives of Surgery 1995; 130: 423–429. Doze VA, Shafer A, White PF. Propofol–nitrous oxide versus thiopental isoflurane–nitrous oxide for general anesthesia. Anesthesiology 1988; 69: 63–71. Van Hemelrijck J, Smith I, White PF. Use of desflurane for outpatient anesthesia. A comparison with propofol and nitrous oxide. Anesthesiology 1991; 75: 197–203. Gan TJ, Ginsberg B, Grant AP, Glass PSA. Double-blind, randomized comparison of ondansetron and intraoperative propofol to prevent postoperative nausea and vomiting. Anesthesiology 1996; 85: 1036–1042. Gan TJ, Glass PSA, Ginsberg B, Canada A, Grant AP, Goodman D, Smith LR. Determination of the plasma concentration of propofol to treat postoperative nausea and vomiting. Anesthesiology 1996; 85: A341. Gepts E, Camu F, Cockshott ID, Douglas EJ. Disposition of propofol administered as constant rate intravenous infusions in humans. Anesthesia and Analgesia 1987; 66: 1256–1263.

Sir,—Thank you for the opportunity to respond to the comments of Gan, Mythen and Glass. While gut mucosal hypoperfusion may be among the factors that are associated with postoperative nausea and vomiting (PONV), I would guess that this is not confined to cardiac operations and may well contribute to PONV after other types of major surgery. Propofol undoubtedly possesses antiemetic properties. All our patients received an infusion of propofol during cardiopulmonary bypass. Patients whose tracheas were extubated immediately after return to the recovery unit would still have had propofol concentrations in the effective range for treatment of PONV at the time of tracheal extubation. Similarly, those patients who required a period of propofol sedation after operation would also have had therapeutic concentrations at tracheal extubation. We did not measure the amounts of propofol given, but duration of intubation did not differ significantly between our two groups. Given the large number of patients in each group (n:200 and n:198) it seems likely that the amounts of propofol given were similar in each group. Thus while propofol used either during anaesthesia or in the early postoperative period may have exerted an antiemetic effect during the first 2 h after tracheal extubation, this does not alter our findings that addition of droperidol to a continuous infusion of morphine was effective in reducing nausea and vomiting after cardiac surgery. C. R. GREBENIK Department of Cardiothoracic Anaesthesia Oxford Heart Centre John Radcliffe Hospital Oxford

Ethanol monitoring during hysteroscopy Sir,—The complications of endometrial resection were reviewed by Williamson and Mushambi in an editorial in the September issue of British Journal of Anaesthesia.1 The importance of fluid absorption in this context is emphasized by the fact that several fatal cases of the “female TUR syndrome” have occurred.2 3 In their article, the authors did not recommend the ethanol method for monitoring fluid absorption because they could not account for the sex-related differences in breath ethanol and serum sodium responses to fluid absorption that were described by me in a recent review article.4 I have stated that the ethanol response to fluid absorption is 10% greater in middle-aged women undergoing endometrial resection than in elderly men undergoing transurethral resection of the prostate. This difference is easy to account for as most women are of a smaller size than men. The authors’ distrust,

Correspondence however, was probably elicited by my suggestion that the serum sodium response to fluid absorption is 40% greater in females than in males. The magnitude of this difference can be illustrated by assuming that 1000 ml of irrigating fluid containing 1.5% glycine with 1% ethanol is absorbed by a male and a female. My nomograms suggest the reduction in serum sodium concentrations would amount to 8 mmol litre91 in males and 11 mmol litre91 in females.4 Is this difference reasonable? Yes, I think so. The amount of extracellular water in the extracellular space, which is the distribution volume for sodium, has been reported to be 200 ml/kg body weight in males and 150 ml/kg body weight in women.5 Based on these data, the difference that I have reported is what would be expected. I admit that in the original study of fluid absorption by myself and Olsson, 62 patients were too few in number to allow construction of a nomogram.6 However, as stated in my review article, the published nomogram was also based on a series of additional patients (up to a total of 70) who actually absorbed irrigating fluid. One of the additional cases has been described to show the elegance with which the ethanol method works.7 The “fluid deficit” is suggested in this editorial to be the best indicator of fluid absorption during hysteroscopy. This implies routine use of volumetric fluid balance. However, many factors need to be controlled to make routine use of this method accurate.4 To my knowledge, its sensitivity has never been tested. Poor correlations between key variables which should all reflect fluid absorption are usually reported in studies where the method seems to have been used on a routine basis. My group has had difficulty obtaining useful data with it during prostatectomy.8 The fluid deficit needs to be compared with more precise methods of measuring fluid absorption before I would recommend it as a standard method. As far as ethanol monitoring is concerned, several such studies have been conducted and its accuracy and precision are well known. R. G. HAHN Department of Anaesthesia Karolinska Institute Stockholm, Sweden 1. Williamson KM, Mushambi MC. Complications of hysteroscopic treatments of menorrhagia. British Journal of Anaesthesia 1996; 77: 305–308. 2. Arieff AI, Ayus JC. Endometrial ablation complicated by fatal hyponatremic encephalopathy. Journal of the American Medical Association 1993; 270: 1230–1232. 3. Baggish MS, Brill AIO, Rosenweig B, Barbot JE, Indman PD. Fatal acute glycine and sorbitol toxicity during operative hysteroscopy. Journal of Gynecological Surgery 1993; 9: 137–143. 4. Hahn RG. Ethanol monitoring of irrigating fluid absorption. European Journal of Anaesthesiology 1996; 13: 102–115. 5. Sendak MJ. Monitoring and management of perioperative fluid and electrolyte therapy. In: Rogers MC, Tinker JH, Covino BG, Longnecker DE, eds. Principles and Practice of Anesthesiology. St Louis: Mosby Year Book, 1993; 866. 6. Olsson J, Hahn RG. Ethanol monitoring of irrigating fluid absorption in transcervical resection of the endometrium. Acta Anaesthesiologica Scandinavica 1995; 39: 252–258. 7. Olsson J, Hahn RG. Early detection of the transcervical endometrial resection syndrome. Gynecologic and Obstetric Investigation 1996; 103: 558–561. 8. Olsson J, Rentzhog L, Hjertberg H, Hahn RG. Reliability of clinical assessment of fluid absorption in transurethral prostatic resection. European Urology 1993; 24: 262–266. Sir,—We thank Professor Hahn for his interest in our editorial on complications of hysteroscopic treatments of menorrhagia,1 and we are grateful for his comments. It was not the aim of our editorial to exclude the ethanol method for estimating fluid absorption during transcervical endometrial resection (TCER). As stated by Professor Hahn both in his letter and in his review,2 the nomograms for estimating fluid absorption from expired ethanol concentration during TCER were based on few patients. Incidentally, his review article stated that the TCER nomograms were based “on 62 patients with the addition of a few more patients with large scale absorption”. Therefore, while the method has been investigated and used extensively, during transurethral resection of the prostrate (TURP) we feel that more research and experience is necessary

477 before we can recommend it for routine use during TCER. There is no doubt that, despite its limitations, volumetric fluid balance is still the most Commonly used method of assessing fluid absorption during TCER. We agree with Professor Hahn that ethanol monitoring is a non-invasive, easy and accurate method of assessing fluid absorption during TURP. Hopefully in the near future this will also be applicable to its use during TCER. M. MUSHAMBI Department of Anaesthesia K. M. WILLIAMSON Department of Obstetrics and Gynaecology Leicester Royal Infirmary Leicester 1. Williamson KM, Mushambi MC. Complications of hysteroscopic treatments of menorrhagia. British Journal of Anaesthesia 1996; 77: 305–308. 2. Hahn RG. Ethanol monitoring of irrigating fluid absorption. European Journal of Anaesthesiology 1996; 13: 102–115.

ACE inhibitors and renal protection Sir,—I read with interest the article by Licker and co-workers1 and agree that angiotensin converting enzyme (ACE) inhibitors, may be of some value in renal protection during surgery. In a previous study we found that aortic cross-clamping-induced haemodynamic changes were less marked in enalapril-treated patients compared with controls,2 in agreement with the study of Licker and colleagues. However, regarding renal perfusion, we established that effective renal plasma flow (ERPF) and glomerular filtration rate (GFR) were not maintained at baseline levels during aortic cross-clamping in enalapril-treated patients, although ERPF and GFR were good before clamping.2 Licker and colleagues established that ACE inhibitors were effective in improving renal perfusion during aortic abdominal surgery and this seems to refute our own conclusions. In fact, these differences may be explained easily. Their renal haemodynamic studies showed that aortic crossclamping induced a significant decrease in ERPF in controls only (approximately 921%). However, although this decrease in ERPF in enalapril-treated patients was not statistically significant (approximately 914%), the renal blood flow:cardiac output ratio decreased significantly only in enalapril-treated patients (925%); this strongly suggests that blood flow was redistributed away from the kidney during clamping. Moreover, although not statistically significant (but the number of enalapril-treated patients was small), the mean increase in renal vascular resistance and mean decrease in GFR during aortic cross-clamping (;43% and 920%, respectively) were not less relevant than the decrease in ERPF in controls. The data may have been confusing therefore because the threshold of statistical significance was not reached because of the small number of patients studied. Moreover, patients had different pathologies, notably both aortic aneurysm and atherosclerotic occlusive disease. In my opinion, Licker and colleagues cannot assert that, in contrast with calcium antagonists,2 ACE inhibitors prevent the renal vasoconstrictive effect of infrarenal aortic abdominal crossclamping. Nevertheless, I agree that ACE inhibitors may boost renal perfusion before and after aortic clamping. P. COLSON Department of Anaesthesiology and Intensive Care University of Montpellier Montpellier, France 1. Licker M, Bednarkiewicz M, Neidhart P, Pretre R, Montessuit M, Favre H, Morrel DR. Preoperative inhibition of angiotensin-converting enzyme improves systemic and renal haemodynamic changes during aortic abdominal surgery. British Journal of Anaesthesia 1996; 76: 632–639. 2. Colson P, Ribstein J, Seguin JR, Marty-Ane C, Roquefeuil B. Mechanisms of renal hemodynamic impairment during infrarenal aortic cross-clamping. Anesthesia and Analgesia 1992; 75: 18–23. Sir,—We thank Dr Colson for his comments. We acknowledge that the small sample size of the study precludes firm conclusions concerning the haemodynamic changes induced by renin–angiotensin system block. The numbers of patients whose

478 renal data were evaluated were nine and 11 for the saline and enalapril groups, respectively. At P
0.05 this number of patients was sufficient to detect a significant decrease in ERPF from a mean value of 472 to 371 ml min91 after aortic cross-clamping in the saline group whereas the (ns) changes observed in the enalapril group (from 582 to 499 ml min91) may represent a type II error. If we assume similar ERPF values in larger groups (n:69), “true differences” would have been reached for (within and between) group comparisons while maintaining a power of 0.90 (␤:0.10); such data would confirm lesser impairment of renal haemodynamic state after pretreatment with enalapril. During aortic cross-clamping, the renal blood flow:cardiac output ratio (RBF:CO) decreased by 25% in enalapril-treated patients and this was attributed entirely to a decrease in RBF as there was no change in CO. In contrast, proportional changes in RBF and CO were observed in saline-treated patients resulting in an unchanged RBF:CO ratio. Activation of unblocked vasopressor support systems (such as arginine vasopressin and the sympathetic nervous system) in response to aortic clamping could be implicated in the renal vasoconstriction in the two groups. In addition, the systemic haemodynamic pattern suggested that enalapril-induced vasodilatation was effective in renal and extrarenal vascular tissues; these data agree with experimental data supporting the notion that the splanchnic vascular bed is more sensitive to the vasoconstrictor effects of sympathetic stimulation and angiotensin II.1 2 The beneficial renal haemodynamic effects of enalapril were demonstrated clearly in the present study as: (1) the RBF:CO ratio increased before aortic cross-clamping as a result of a greater increase in RBF compared with CO, (2) during aortic clamping, ERPF was reduced to a lesser extent, (3) a greater glomerular filtration rate (GFR) was maintained; GFR and ERPF were strongly associated (r:0.65 in the enalapril group) after aortic clamping, implying that approximately 40% of the observed changes in GFR could be explained by haemodynamic variability. Taken together, our preliminary data suggest lesser impairment of renal perfusion and function during aortic abdominal surgery after enalapril pretreatment. M. LICKER Division of Anaesthesiology D. R. MOREL Division of Anaesthesiological Investigations University Hospital Geneva, Switzerland 1. Cserep G, Abraham G, Tolins JP, Simon G. Different vascular responses to suppressor angiotensin II administration in the mesenteric and renal circulation of rats. American Journal of Hypertension 1996; 9: 385–389. 2. Reilly PM, Bullcley GB. Vasoactive mediators and splanchnic perfusion. Critical Care Medicine 1993; 21 (Suppl.): S55–68.

Possible mechanism to explain increased coagulability of blood after haemodilution Sir,—We read with interest the discussion ensuing from the article by Ruttman, James and Viljoen1 on hypercoagulability of blood after administration of i.v. crystalloids. Farmery and Kong2 suggested that the results “challenge intuitive assumptions”. We suggest that an increase in coagulability is reasonable for the following reasons. Blood coagulation is controlled by continuous interaction between pro- and anticoagulant factors. Haemodilution reduces the concentration of anticoagulant factors as much as procoagulant factors. Thus there is no intuitive reason to assume that anticoagulation would occur any more than procoagulation. A procoagulant effect could easily be brought about if the concentration–effect relationship of the anticoagulant factors were initially steeper than that of the procoagulant factors, that is dilution reduces the effect of anticoagulant factors more rapidly than procoagulant factors. This makes sense teleologically, where coagulability should increase after haemorrhage when blood volume is re-expanded with interstitial fluid. As pointed out by the above authors, regional anaesthesia reduces the incidence of deep vein thrombosis despite the likely need for increased crystalloid administration. A possible explanation is that regional anaesthesia reduces the surgical stress response and its associated procoagulant effects. Could it be that crystalloid “primes” the coagulation system by the above

British Journal of Anaesthesia proposed mechanism, and that the procoagulant tendency is then activated by surgical stress? We feel that these considerations would explain the observations reported, although we accept that evidence to confirm or refute them is not yet available. J. E. FLETCHER C. M. B. HEARD Department of Anesthesiology Children’s Hospital of Buffalo Buffalo, NY, USA 1. Ruttmann TG, James MFM, Viljoen JF. Haemodilution induces a hypercoagulable state. British Journal of Anaesthesia 1996; 76: 412–414. 2. Farmery AD, Kong A. Hypercoagulability induced by crystalloids. British Journal of Anaesthesia 1996; 77: 132. Sir,—We thank Fletcher and Heard for their comment on our article. The argument they make that anticoagulant factors are diluted to the same degree as procoagulant factors is valid, and has been suggested previously by Monkhouse,1 who suggested that haemodilution may disturb the ratio of thrombin to antithrombin III. An imbalance between the two may predispose to a procoagulant effect. Subsequent in vivo data, which are currently under consideration for publication, suggest that antithrombin III is decreased to a greater extent after haemodilution than could be explained by haemodilution alone, the inference being that the act of haemodilution per se, had induced the conversion of prothrombin to thrombin with subsequent thrombin–antithrombin III interaction. The suggestion by Fletcher and Heard that this effect may have a teleological basis is particularly elegant, as it could explain the reason for its occurrence. T. RUTTMAN M. F. M JAMES J. F. VILJOEN Department of Anaesthesia University of Cape Town Cape Town, South Africa 1. Monkhouse FC. Relationship between anti-thrombin and thrombin levels in plasma and serum. American Journal of Physiology 1959; 197: 984–988.

Regional anaesthesia in moyamoya disease Sir,—We have read the article of Ngan Kee and Gomersall on extradural anaesthesia for Caesarean section in a patient with moyamoya disease.1 We wish to describe our own experience with an elective Caesarean section managed with spinal anaesthesia in a patient with moyamoya disease. A 34-yr-old woman suffered intraventricular bleeding in her 24th week of pregnancy, requiring intensive care and mechanical ventilation for 8 days. During this period the only unusual problem was development of pneumonia. She recovered without any sequelae and moyamoya disease was diagnosed by arteriography. The patient remained healthy for the remainder of the pregnancy and in the 38th week an elective Caesarean section was planned. Preoperative investigations were normal, including anticardiolipin antibody concentrations. After a preload of 1000 ml of lactated Ringer’s solution, a left radial artery cannula and central venous catheter (via the antecubital fossa) were inserted under local anaesthesia. Analysis of arterial blood was carried out at 15-min intervals. Anaesthesia was perforated with 0.5% spinal bupivacaine 15 mg and fentanyl 50 ␮g at L2–3, using a 25-gauge Whitacre needle, and 28% oxygen was given by face mask. The level of sensory analgesia to pinprick was T5. Total intraoperative fluid given was 1500 ml of mild lactated Ringer’s solution. There were no changes in heart rate, direct arterial pressure or temperature, apart from transient hypotension (80% of preoperative value) treated with ephedrine 15 mg i.v. A healthy 3250-g male baby was delivered with Apgar scores of 8 (1 min) and 9 (5 min). After surgery the patient remained in the PACU for 6 h and was discharged uneventfully without any change in her neurological status. There were no other problems in the subsequent 3 months. Several cases have been reported of patients with moyamoya

Correspondence disease managed successfully with general anaesthesia for Caesarean section.2 In our opinion, regional anaesthesia is preferable for these patients in order to obviate intubationinduced hypertension and the risk of aspiration. Some cases have also been reported using extradural anaesthesia.3 However, the complication rate is less with spinal anaesthesia.4 We believe that avoidance of hypotension, or treating it carefully, allows the use of spinal anaesthesia in these cases. By avoiding hypotension, hypothermia and hyperventilation, there were no neurological signs in our awake patient whose conscious state and cerebral function could be observed continually. A case of convulsions after spinal anaesthesia in a child with moyamoya disease has been reported,5 probably caused by changes in cerebral blood flow (CBF). We believe that these convulsions occur more frequently in younger patients with moyamoya disease and may be avoided by strict management of CBF. The doses of bupivacaine used for extradural anaesthesia are higher than those used in spinal anaesthesia, with increased risk of neurological and cardiological problems. A. LLORENTE DE LA FUENTE Department of Anesthesia Hospital de León, León, Spain M. C. GIMENEZ GARCIA Obra Hospitalaria Nuestra Senora de Regla León, Valladoltd, Spain F. LOPEZ SANCHEZ Department of Obstetrics and Gynaecology Hospital de León León, Palencia, Spain 1. Ngan Kee WD, Gomersall CD. Extradural anaesthesia for Caesarean section in a patient with moyamoya disease. British Journal of Anaesthesia 1996; 77: 550–552. 2. Venkatesh El, Taggart PCM. Anaesthetic management of a patient with moyamoya disease for Caesarean section. Canadian Journal of Anaesthesia 1994; 41: 79–80. 3. Sharma SK, Wallace DH, Gajraj NM, Willis C, Sidawi E. Epidural anesthesia for a patient with moyamoya disease presenting for cesarean section. Anesthesia and Analgesia 1994; 79: 183–185. 4. Riley ET, Cohen SE, Macario A, Desai JB, Ratner EF. Spinal versus epidural anesthesia for cesarean section: A comparison of time efficiency, cost, charges, and complications. Anesthesia and Analgesia 1995; 80: 709–712. 5. Yasukawa K, Akagama S, Nagakawa Y, Migasaka K. Convulsions and temporary hemiparesis following spinal anesthesia in a child with moyamoya disease. Anesthesiology 1988; 69: 1023–1024. Sir,—Thank you for the opportunity to reply to Drs Llorente de la Fuente, Gimenez Garcia and Lopez Sanchez. Although they have described the successful use of spinal anaesthesia for Caesarean section in a patient with moyamoya disease, we believe this is not the optimum technique in this circumstance. While we concur that regional anaesthesia has advantages over general anaesthesia, we believe that extradural anaesthesia is preferable to spinal anaesthesia.1 Llorente de la Fuente, Gimenez Garcia and Lopez Sanchez have emphasized the importance of strict management of cerebral blood flow in these patients. We sought to achieve this by inducing central neural block gradually under close monitoring; we believe the slower onset of sympathetic block in extradural anaesthesia is more amenable to accurate titration of interventions compared with spinal anaesthesia.2 W. D. NGAN KEE C. D. GOMERSALL Department of Anaesthesia and Intensive Care Chinese University of Hong Kong Prince of Wales Hospital Hong Kong 1. Sharma SK, Wallace DH, Gajraj NM, Willis C, Sidawi E. Epidural anesthesia for a patient with moyamoya disease presenting for cesarean section. Anesthesia and Analgesia 1994; 79: 183–185. 2. Snider SM, Levinson G. Anesthesia for cesarean section. In: Snider SM, Levinson G, eds. Anesthesia for Obstetrics. Baltimore: Williams and Wilkins, 1993; 211–245.

479

Converting pH and H;: a “rule of thumb” Sir,—Many centres have changed to the use of hydrogen ion concentration (H;) rather than the traditional pH when dealing with acid–base problems. Converting H; to pH of vice versa is very simple when tables are available. We propose a straightforward method for conversion when they are not. We have observed that, between pH values of 7.10 and 7.60, the sum of the hydrogen ion concentration (in nmol litre91) and the numerical value of the two digits after the pH decimal point is relatively fixed around 83, the mean of these sums in this pH range (table 1). Table 1 Method for conversion of H; to pH pH 7.00 7.10 7.20 7.30 7.40 7.50 7.60 7.70 H+ 100 79.4 63.1 50.1 40 31.6 25.1 20 Sum 100 89.4 83.1 80.1 80 81.6 85.1 90 To convert pH from H;, the numerical value of the two digits after the pH decimal point is calculated by subtracting H; from 83. For example, if H; is 70.8 nmol litre91, the estimated value of pH would be (7.83–70.8), or 7.12. The actual pH value at this hydrogen ion concentration is 7.15. To convert H; from pH, the numerical value of the two digits after the pH decimal point is subtracted from 83. Therefore, if pH is 7.50, the estimated value of H; would be 83–50, or 33 nmol litre91. The actual hydrogen ion concentration at this pH is 31.6 nmol litre91. The errors are small and clinically inconsequential for conversions performed between pH values of 7.10 and 7.60. We feel that our simple “rule of thumb” will be useful to doctors dealing regularly with clinical situations in which acid–base disturbances are a feature. R. J. BURDEN P. J. MCQUILLAN Department of Anaesthesia and Intensive Care Queen Alexandra Hospital Portsmouth

Local anaesthesia to the airway reduces sedation requirements in patients undergoing artificial ventilation Sir,—We read the article of Mallick, Smith and Bodenham1 with interest and wish to make a few comments. In their study, although the use of local anaesthetic solution reduced the need for sedation in patients who underwent ventilation, it also suppressed the cough reflex. Patients appeared to be less bothered by pulmonary suctioning. The technique may be safe if used for short periods but our concern is that, if used over a longer period it may increase the risk of pulmonary infection. Mucociliary clearance is an important mechanism of pulmonary defence. Effective mucociliary clearance depends on the function of the cilia and their coordinated beating.2 Corssen and Allen3 studied ciliary function in tracheobroachial epithelium obtained with a punch biopsy and found that lignocaine caused a dose-related but reversible effect on ciliary activity. At low doses ciliary beating was stimulated and as the dose increased the cilia stopped beating. In contrast, amethocaine and cinchocaine caused irreversible stoppage of ciliary activity, It is clear that topical anaesthesia may cause areas of ciliary dysfunction and could predispose to airway colonization and pneumonia. The effects of long-term local anaesthesia on ciliary motility have not been studied. We believe that further studies are needed to evaluate the effect of local anaesthetics on respiratory epithelium. One suggestion is to compare the effect of local anaesthetics on mucociliary clearance using either a dye or a radio-opaque contrast as a marker. If the use of topical anaesthesia is found to be safe, then reduced requirements for sedatives may prove beneficial and may reduce the morbidity and mortality of patients undergoing artificial ventilation. P. BAPAT Department of Anaesthetics Charing Cross Hospital London

480

British Journal of Anaesthesia C. VERGHESE Department of Anaesthetics Royal Berkshire Hospital Reading

1. Mallick A, Smith SN, Bodenham AR. Local anaesthesia to the airway reduces sedation requirements in patients undergoing artificial ventilation. British Journal of Anaesthesia 1996; 77: 731–734. 2. Skerrett SJ, Niederman MS, Fein FM. Respiratory infections and acute lung injury in systemic illness. Clinics in Chest Medicine 1989; 10: 469–502 3. Corssen G, Allen CR. Cultured human respiratory epithelium: its use in the comparison of the cytotoxic properties of local anesthetics. Anesthesiology 1960; 21: 237–243.

Anticholinesterases and subsequent duration of block of suxamethonium Sir,—I read with interest the article by Fleming and colleagues1 on the effects of anticholinesterases on subsequent duration of block with suxamethonium. The clinical observations are clear and in accord with conventional wisdom. The authors have however fallen into a common trap with regard to their interpretation of the cholinesterase activity values. Their pooled data are in accord with those of Viby-Mogensen,2 who they referenced, in that over the clinical range the correlation between measured activity and duration of suxamethonium block is poor. In this study, a correlation can be shown clearly only in the presence of extremely low levels of measured activity. Another part of the problem is that the substrate used is not that used clinically (suxamethonium) but a thiocholine derivative, and that the correlation between different assay techniques and suxamethonium sensitivity is less than ideal.3 In addition, although not stated by the authors, such measurements of cholinesterase activity involve a large dilution of the plasma sample; this varies between laboratories but is usually 30–300-fold dilution into the reaction mixture. For cholinesterase inhibitors with prolonged covalent binding such as neostigmine, any inhibition produced would be likely to persist over a short period of dilution. Neostigmine concentrations decrease rapidly in vivo to approximately 61097 mol litre91.4 This concentration is associated with approximately 50% inhibition of cholinesterase activity (unpublished personal data, see fig. 1) and accords with the authors observations. Other less tightly bound agents such as edrophonium5 would be expected to be associated after such dilution with a lower inhibition than might be predicted from plasma concentrations at the time of sampling. The assumption that enzyme inhibition measured after dilution ex vivo accurately reflects in vivo conditions is a common but erroneous one.

Figure 1 Effects of neostigmine on plasma cholinesterase catalysis. Data are means of repeated measures of plasma from a normal individual. The assay is a modification of that of Ellman and colleagues.6

S. G. GRAHAM Department of Anaesthesia University of Newcastle Royal Victoria Infirmary Newcastle upon Tyne 1. Fleming NW, Macres S, Antognini JF, Vengco J. Neuromuscular blocking action of suxamethonium after antagonism of vecuronium by edrophonium, pyridostigmine or neostigmine. British Journal of Anesthesia 1996; 77: 492–495. 2. Viby-Mogensen J. Correlation of succinylcholine duration of action with plasma cholinesterase activity in subjects with genotypically normal enzyme. Anesthesiology 1980; 53: 517–520. 3. Evans RT, Wroe J. Is serum cholinesteras activity a predictor of succinylcholine sensitivity? An assessment of four methods. Clinical Chemistry 1978; 24: 1762–1766. 4. Williams NE, Calvey TN, Chan K. Clearance of neostigmine from the circulation during the antagonism of neuromuscular block. British Journal of Anaesthesia 1978; 50: 1065–1067. 5. Taylor P. Anticholinesterase agents. In: Gilman AG, Goodman LS, Gilman A, eds. The Pharmacological Basis of Therapeutics, 6th Edn. New York: Macmillan Publishing 1980; 100–119. 6. Ellman GL, Courtney D, Andres V, Featherstone RM. A new and rapid colorimetric determination of acetlycholinesterase activity. Biochemical Pharmacology 1961; 7: 88–95. Sir,—We thank Dr Graham for his interest in our work and appreciate the opportunity to respond to his comments. Although we beg to differ as to whether the interpretation of our results constitutes “falling into a common trap”, we feel his letter highlights some limits of our studies and emphasizes aspects of this work which could benefit from further discussion. We agree that although Viby-Mogensen described the most commonly referenced relationship between measured activity of plasma cholinesterase and the duration of action of suxamethonium,1 when cholinesterase activity is in the normal clinical range, the slope of this graph is indeed shallow. The hyperbolic relationship he plots is largely determined by values outside the normal range. Although the sensitivity of the plasma cholinesterase activity assay for detection of less active variants of the enzyme may be improved slightly by the use of alternative substrates as described by Evans and Wroe,2 their study focuses on the detection of less active variants of the plasma cholinesterase enzyme, not on the characteristics of normal enzyme which has been exposed to a cholinesterase inhibitor. It is the effects of the mechanism of inhibition which we believe deserve emphasis. As non-competitive inhibitors which react to form a carbomoylated cholinesterase enzyme, neostigmine and pyridostigmine are unlikely to be affected by assay techniques which include a dilutional step. In contrast, edrophonium, as a competitive inhibitor, may be profoundly affected. This may explain the contrast between studies which demonstrate minimal inhibition of plasma cholinesterase activity by edrophonium3 4 while the metabolism of other substrates (e.g. mivacurium) is inhibited.4 For non-competitive inhibitors, dilutional effects on enzyme activity measurements should be at best, minimal, and at worst, comparable. There is no evidence that the substrate used for the assay should have a differential sensitivity for these chemically related inhibitors. In our studies, pyridostigmine produced profound inhibition of plasma cholinesterase (79% of baseline) compared with neostigmine (48% of baseline). In contrast, the duration of action of suxamethonium increased only 35% after pyridostigmine compared with 78% after neostigmine. Despite the limitations of the enzyme activity assay, we believe this contrast provides the strongest evidence that this is a much more complex drug interaction than previously realized. N. W. FLEMING S. MACRES J. F. ANTOGNINI J. VENGCO Department of Anesthesiology School of Medicine University of California, Davis CA, USA 1. Viby-Mogensen J. Correlation of succinylcholine duration of

Correspondence

481

action with plasma cholinesterase activity in subjects with genotypically normal enzyme. Anesthesiology 1980; 53: 517–520. 2. Evans RT, Wroe J. Is serum cholinesterase activity a predictor of succinylcholine sensitivity? An assessment of four methods. Clinical Chemistry 1978; 24: 1762–1766. 3. Mirakhur RK. Edrophonium and plasma cholinesterase activity. Canadian Journal of Anaesthesiology 1986; 33: 588–590. 4. Hart PS, Wright PM, Brown R. Edrophonium increases mivacurium concentrations during constant mivacurium infusion, and large doses minimally antagonize paralysis. Anesthesiology 1995; 82: 912–918.

6. Harris B, Moody E, Skolnick P. Isoflurane anaesthesia is stereoselective. European Journal of Pharmacology 1992; 217: 215–216. 7. Moody EJ, Harris BD, Skolnick P. Stereospecific actions of the inhalation anaesthetic isoflurane at the GABAA receptor complex. Brain Research 1993; 615: 101–106. 8. Moody EJ, Harris B, Hoehner P, Skolnick P. Inhibition of [3H] israpidine binding to L-type calcium channels by the optical isomers of isoflurane. Anesthesiology 1994; 81: 124–128. 9. Graf BM, Boban M, Stowe DF, Kampine JP, Bosnjak ZJ. Lack of Stereospecific effects of isoflurane and desflurane isomers in isolated guinea pig hearts. Anesthesiology 1994; 81: 129–136.

I. v. anaesthetics and binding to calcium channels

Sir,—We thank Dr Wilkinson for this interest and comments on our article.1 Before we respond to his detailed comments We wish to state clearly our position. We are not totally convinced that voltagesensitive Ca2; channels (VSCC) have a role to play in anaesthesia but we are equally not convinced that they do not. We feel that it is premature to exclude this important class of ion channels as anaesthetic target sites and hoped that this is what our article and a recent editorial2 conveyed. We agree that the literature is bewildering but as Dr Wilkinson points out, inhibition of VSCC and hence neurotransmission is an attractive site for anaesthetic action. While our article was concerned specifically with L channels (probed using the radiolabelled dihydropyridine [3H]PN200–110) we do not exclude other types of VSCC, such as P and N channels, as target sites for anaesthetic agents and this was also noted in our recent editorial.2 Indeed, studies with radiolabelled conotoxin (specific for N channels) are currently at the planning stage. Dr Wilkinson cites several studies that show neurotransmitter release is not controlled by L channels and in general we agree with this. However, there are reports showing that L channels are involved (see Spedding and Lepagnol3). Of particular interest to us is the SH-SY5Y human neuroblastoma cell line that we have used in several studies of the mechanism of action of i.v. and volatile general anaesthetic agents. In this cell line L channelmediated noradrenaline release and K;-evoked increase in [Ca2;]i are inhibited by thiopentone, propofol, ketamine, etomidate and halothane.4–7 Moreover, we feel that modulation, by classic L channel agonists and antagonists, of ethanol, barbiturate and benzodiazepine potency8 9 indicates something important! In addition, L-VSCC blockers have antinociceptive actions which are known to contribute to anaesthesia.10913 There is strong support for the GABAA receptor as a target site for anaesthetic agents14 and we have never disputed this fact. Stereoselectivity is noted for the GABAA receptor for a range of resolved anaesthetic agents. However, as Dr Wilkinson points out “it is perhaps too early to say whether stereospecificity may be a marker for actions that are relevant to the mechanism of anaesthesia”. The point made regarding the “non-anaesthetic” effect of L channel antagonists is well taken but it still remains that in experimental animals L channel agonists and antagonists modify ethanol, barbiturate and benzodiazepine potency.8 9 In addition, Nacif-Coelho and colleagues reported that direct administration of L channel ligands into the locus coerulus modified the anaesthetic potency of dexmedetomidine in the rat.15 Collectively, these data suggest that provided passage of the blood–brain barrier is facilitated, L channel blockers can have “anaestheticlike” effects. In summary, our studies with L channels form a basis for further studies with N and P channels and we feel that it is premature to exclude the voltage-sensitive Ca2; channel as a target site for anaesthetic agents. K. HIROTA Department of Anaesthesiology Hirosaki University Hirosaki 036, Japan D. G. LAMBERT University Department of Anaesthesia Leicester Royal Infirmary Leicester

Sir,—The article by Hirota and Lambert1 demonstrating an interaction between a variety of i.v. agents and the dihydropyridine binding site on L-type voltage-sensitive Ca2; channels is an interesting addition to the bewildering literature on mechanisms of anaesthesia. The concept that i.v. anaesthetic agents may reduce neurotransmitter release via diminished influx of calcium ions is attractive as calcium influx is the common trigger for neurotransmitter release in invertebrates and vertebrates. The authors hypothesize that i.v. anaesthetics may act via inhibition of calcium influx through L-type channels. Although their discussion may appear convincing, there are reservations with this hypothesis. First, L-type channels do not seem to be crucial to stimulus–secretion coupling in the majority of vertebrate preparation studied2–5 as this coupling is generally insensitive to dihydropyridine antagonists. If the likely target for anaesthetics is transmission across central synapses, then the significance of L-type calcium channel inhibition for anaesthesia may be minimal. Interestingly, other calcium channels, for example P and N channels, have been implicated in neurotransmitter release in vertebrate neurones.3 4 Second, are the authors, therefore, observing non-specific actions of anaesthetic agents rather than actions which underlie anaesthesia? Stereospecific differences in anaesthetic effects are seen at the receptor level for isoflurane on the GABAA receptor.6 Hypnosis may also be Stereospecific.7 Although it is perhaps too early to say if stereospecificity may be a marker for actions that are relevant to the mechanism of anaesthesia, it is worthy of note that the anaesthetic effects of isoflurane on L-type calcium channels do not demonstrate stereoselectivity.8 9 Third, the overwhelming fact remains that traditional dihydropyridine antagonists have minimal action as anaesthetics in the conventional sense, even if the blood–brain barrier is crossed. I believe that these considerations significantly weaken the case for inhibition of calcium influx in L-type channels being an important mechanism of action of i.v. anaesthetic agents. P. R. WILKINSON Department of Anaesthesia Mater Misericordiae Public Hospitals South Brisbane, Queensland, Australia 1. Hirota K, Lambert DG. I.v. anaesthetic agents inhibit dihydropyridine binding to L-type voltage-sensitive Ca2; channels in rat cerebrocortical membranes. Britith Journal of Anaesthesia 1996; 77: 248–253. 2. Hirning LD, Fox AP, McCleskey EW, Olivera BM, Thayer SA, Miller RJ, Tsien RW. Dominant role of N-type Ca2; channels in evoked release of norepinephrine from sympathetic neurones. Science 1988; 239: 57–61. 3. Takahashi T, Momiyama A. Different types of calcium channels mediate central synaptic transmission. Nature (London) 1993; 366: 156–159. 4. Turner TJ, Adams ME, Dunlap K. Calcium channels coupled to glutamate release identified by omega-Aga-IVA. Science 1992; 258: 310–313. 5. Stanley EF, Atrakchi AH. Calcium currents recorded from a vertebrate presynaptic nerve terminal are resistant to the dihydropyridine nifedipine. Proceedings of the National Academy of Sciences USA 1990; 87: 9683–9687.

1. Hirota K, Lambert DG. I.v. anaesthetic agents inhibit dihydropyridine binding to L-type voltage-sensitive Ca2;

482

2. 3.

4.

5.

6. 7.

8.

channels in rat cerebrocortical membranes. British Journal of Anaesthesia 1996; 77: 248–253. Hirota K, Lambert DG. Voltage-sensitive Ca2; channels and anaesthesia. British Journal of Anaesthesia 1996; 76: 344–346. Spedding M, Lepagnol J. Pharmacology of sodium and calcium channel modulation in neurones: implications for neuroprotection. Biochemical Society Transactions 1995; 23: 633–636. Lambert DG, Willets JM, Atcheson R, Frost CL, Smart D, Rowbotham DJ, Smith G. Effects of propofol and thiopentone on potassium and carbachol evoked [3H]noradrenaline release and increased [Ca2;]i from SH-SY5Y human neuroblastoma cells. Biochemical Pharmacology 1996; 51: 1613–1621. Sikand KS, Smith G, Lambert DG. Ketamine inhibits K; evoked [3H]noradrenaline release from SH-SY5Y cells by reducing calcium influx. Biochemical Society Transactions 1995; 23: 417S. Lambert DG, Sikand KS, Hirota K, Lambert DG. Studies on the mechanism of action of etomidate. European Journal of Anaesthesiology 1996 (in press). Atcheson R, Hirst RA, Rowbotham DJ, Lambert DG. The effects of halothane on [3H]noradrenaline release from SH-SY5Y human neuroblastoma cells. British Journal of Pharmacology 1994; 112: 511P. Dolin SJ, Little HJ. Augmentation by calcium channel antagonists of general anaesthetic potency in mice. British

British Journal of Anaesthesia Journal of Pharmacology 1988; 88: 909–914. 9. Dolin SJ, Patch TL, Rabbani M, Taberner PV, Little HJ. Differential interactions between benzodiazepines and the dihydropyridines, nitrendipine and Bay K 8644. Neuropharmacology 1991; 30: 217–224. 10. Carta F, Bianchi M, Argenton S, Cervi D, Marolla G, Tamburini M, Breda M, Fantoni A, Panerai AK. Effect of nifedipine on morphine-induced analgesia. Anesthesia and Analgesia 1990; 70: 493–498. 11. Miranda HF, Bustamante D, Kramer V, Pelissier T, Saavedra H, Paeile C, Fernandez E, Pinardi G. Antinociceptive effects of Ca2; channel blockers. European Journal of Pharmacology 1992; 217: 137–141. 12. Wong CH, Dey P, Yarmush J, Wu WH, Zbuzek VK. Nifedipine-induced analgesia after epidural injection in rats. Anesthesia and Analgesia 1994; 79: 298–302. 13. Santillan R, Maestre JM, Hurle MA, Florez J. Enhancement of opiate analgesia by nimodipine in cancer patients chronically treated with morphine: a preliminary report. Pain 1994; 58: 129–132. 14. Franks NP, Lieb WR. Molecular and cellular mechanisms of general anaesthesia. Nature (London) 1994; 367: 607–614. 15. Nacif-Coelho C, Correa-Sales C, Chang LL, Maze M. Pertubation of ion channel conductance alters the hypnotic response to the ␣2-adrenergic agonist dexmedetomidine in the locus coerulus of the rat. Anesthesiology 1994; 81: 1527–1534.