Hepatocellular integrity after inhalation anaesthesia

British Journal of Anaesthesia 1997; 78: 772–778 CORRESPONDENCE

Use of rocuronium in a pregnant patient receiving magnesium medication Sir,—The report by Gaiser and Seem1 highlighted some important aspects of the use of non-depolarizing neuromuscular blocking agents in the presence of magnesium infusions. In their patient, the use of rocuronium for neuromuscular block was followed by prolonged paralysis. However, there are several problems with this report. The dose of rocuronium used was well in excess of the usual recommended dose of 0.6 mg kg91. The prolongation of non-depolarizing neuromuscular block by magnesium is so well known that the decision to use such a high dose of a non-depolarizing blocker needed to be very carefully considered. If it was known at the outset that the case was to last 6 h, then the use of rocuronium in these circumstances is acceptable, but prolongation of its action should have been expected. The validity of the choice of blocker is also open to debate. The evidence against the use of suxamethonium in patients with open eye injuries is, at best, weak. There is little hard evidence that the small increase in intraocular pressure (IOP) seen after the use of suxamethonium poses a real threat to the injured eye and several strategies are available to limit this increase in IOP.2 In addition, magnesium decreases the fasciculations produced by suxamethonium,3 and is likely to limit the increase in IOP consequent on suxamethonium use. The aspiration risk in this patient was clear, and it seems inappropriate to consider the minute risk of ocular injury from suxamethonium ahead of the hazards associated with the use of a non-depolarizing neuromuscular blocking agent. The plasma concentrations of magnesium referred to in this case report must be wrong. I suspect that the units were erroneously quoted as mmol litre91 whereas the reported values are consistent with mEq litre91. It is extremely unlikely that, at the doses of magnesium given, the concentration in plasma would have been 5.8 mmol litre91. The correct therapeutic range is not 4–8 mmol litre91 as quoted in this article, but 2–4 mmol litre91. The quoted values tally with the normally quoted mEq litre91 range, those for molar concentrations being half the concentration reported when equivalents are used for divalent ions. The important lesson of this report is that the well known interactions of magnesium and neuromuscular blocking agents should always be borne in mind. Careful selection of appropriate drugs and dosage is vital if this type of complication is to be avoided. M. F. M. JAMES Department of Anaesthesia University of Cape Town South Africa 1. Gaiser RR, Seem EH. The use of rocuronium in a pregnant patient with an open eye injury, receiving magnesium medication, for preterm labour. British Journal of Anaesthesia 1996; 77: 669–671. 2. Zimmerman M, Funk KJ, Tidwell JL. Propofol and alfentanil prevent the increase in intraocular pressure caused by succinylcholine and endotracheal intubation during rapid sequence induction of anesthesia. Anesthesia and Analgesia 1996; 83: 814–817. 3. Stacey MR, Barclay K, Asai T, Vaughan RS. Effects of magnesium sulphate on suxamethonium-induced complications during rapid-sequence induction of anaesthesia. Anaesthesia 1995; 50: 933–936. Sir,—We appreciate the comments of James on our case report, especially those concerning the units for the plasma concentration of magnesium. In the original manuscript, it was mEq litre91. We apologize for the error in typesetting. However, we disagree with several of his other statements. After initiating the case, we knew the neuromuscular blocking effects of rocuronium would be prolonged by magnesium; it was the extent with which we were unsure. In determining the risk and benefit of using rocuronium, we felt the benefit of avoiding the increase in intraocular pressure (IOP) outweighed the risk of possible postoperative intubation and ventilation should surgery conclude before its action had terminated.

He also states that the dose of rocuronium, 0.9 mg kg91 (3⫻ED95), was excessive. As outlined in the manuscript, this dose, compared with 0.6 mg kg91, decreases the onset time to complete neuromuscular block by 50%.1 In the pregnant patient who is prone to oxygen desaturation on induction, we felt that rapid securing of the airway to be most important.2 As such, the higher dose was used. Suxamethonium increases IOP by approximately 8 mm Hg. James stated that magnesium may attenuate this increase as magnesium would decrease the fasciculations produced by suxamethonium. Contraction of the extraocular muscles does not completely account for the increase in IOP. Pretreatment with a non-depolarizing neuromuscular blocking agent to prevent fasciculations does not reliably prevent this increase.3 In fact, in patients scheduled for elective enucleation who had all of the extraocular muscles detached, suxamethonium still increased IOP.4 Therefore, one would have to question if magnesium would limit the increase in IOP by suxamethonium. Finally, James suggests the use of propofol and alfentanil to attenuate the increase in IOP. Although these agents may be helpful, it is important to remember that the patient was a 31-yr-old parturient with a 28-week gestation. Although propofol has been used successfully during pregnancy, it is not recommended.5 There are no adequate and well-controlled studies concerning the use of alfentanil during pregnancy. Its use as a continuous extradural infusion during labour has been investigated. Neonatal hypotonia was observed.6 R. GAISER Department of Anesthesia Hospital of the University of Pennsylvania Philadelphia, PA, USA 1. De Mey JC, Debrock M, Rolly G. Evaluation of the onset and intubation conditions of rocuronium bromide. European Journal of Anaesthesiology 1994; 11 (Suppl. 9): 37–40. 2. Archer GW jr, Marx GF. Arterial oxygen tension during apnoea in parturient women. British Journal of Anaesthesia 1974; 46: 358–360. 3. Myers EF, Krupin T, Johnson M, Zink H. Failure of nondepolarizing blockers to inhibit succinylcholine-induced increased intraocular pressure: A controlled study. Anesthesiology 1978; 48: 149–151. 4. Kelly RE, Dinner M, Turner LS, Haik B, Abramson DH, Daines P. Succinylcholine increases intraocular pressure in the human eye with the extraocular muscles detached. Anesthesiology 1993; 79: 948–952. 5. Bacon RC, Razis PA. The effect of propofol sedation in pregnancy on neonatal condition. Anaesthesia 1994; 49: 1058–1060. 6. Heytens L, Cammu H, Camu F. Extradural analgesia during labour using alfentanil. British Journal of Anaesthesia 1987; 59: 331–337.

Hepatocellular integrity after inhalation anaesthesia Sir,—In their study of hepatocellular integrity after inhalation anaesthesia, Tiainen and Rosenberg1 found that GSTA concentrations increased in patients who received isoflurane or halothane, but the increase appeared greater in those receiving halothane. As they point out, this contrasts with previously published studies,2 3 and they speculated on possible reasons for this. We believe the discrepancy between their findings and ours is explained by the use of different modes of ventilation in the two studies, and not from differences in the concentrations of agents administered. There is little doubt that GST concentration measured by specific radioimmunoassay does not increase in normotensive, spontaneously breathing patients who receive isoflurane. Of the 18 patients who received isoflurane, only two showed an increase in GST concentration (4.1 and 4.3 ␮g litre91) after anaesthesia.2 This confirmed results from a previous study in which no patient who received isoflurane showed an increase in GST concentration after anaesthesia.4 In contrast, we have shown that the mode of

Correspondence ventilation significantly influences GST concentration more profoundly than the use of different volatile anaesthetic agents.5 We studied 48 patients undergoing controlled hypotension for middle ear surgery who were allocated to one of four groups to receive either IPPV or SV, and halothane or isoflurane. GST concentrations increased at 1–3 h after the end of anaesthesia in all four groups but there was no significant difference in GST concentration between patients who had received halothane or isoflurane, irrespective of the mode of ventilation. Combining the results obtained using the two volatile agents allowed assessment of the effect of ventilatory mode and revealed that GST concentration was significantly greater at 1 h after the end of anaesthesia in patients who had received IPPV (pre-induction 2.3 ␮g litre91; 1 h 3.6 ␮g litre91) than in those who had breathed spontaneously (pre-induction 2.5 ␮g litre91; 1 h 2.6 ␮g litre91). We do not consider that differences in concentration of volatile agents administered account for the differences observed between the study of Tiainen and Rosenberg and ours. We have never found that the change in GST from preoperative values is related to dose of anaesthetic agent or to duration of anaesthesia. Murray, Rowlands and Trinick did not find any increase in GST concentration after isoflurane anaesthesia in patients who received IPPV.3 We agree with Tiainen and Rosenberg that the likely reasons for this include the relatively insensitive GST assay used and infrequent sampling after anaesthesia. It should also be noted that whereas Murray and Tiainen’s groups measured total alpha class GST (i.e. B1 and B2 subunits), we measured only GST B1 subunits in our studies; measurement of B1 subunits appears more sensitive than B2 at detecting hepatic damage. We believe that hepatocellular integrity is not impaired in spontaneously breathing patients who receive isoflurane, and that mode of ventilation has a greater influence on GST concentration after anaesthesia than the use of different inhalation anaesthetic agents. D. C. RAY Department of Anaesthetics G. J. BECKETT Department of Clinical Biochemistry Royal Infirmary Edinburgh L. G. ALLAN Department of Anaesthetics Northwick Park Hospital London 1. Tiainen P, Rosenberg PH. Hepatocellular integrity during and after isoflurane and halothane anaesthesia in surgical patients. British Journal of Anaesthesia 1996; 77: 744–747. 2. Hussey AJ, Aldridge LM, Paul D, Ray DC, Beckett GJ, Allan LG. Plasma glutathione S-transferase concentration as a measure of hepatocellular integrity following a single general anaesthetic with halothane, enflurane or isoflurane. British Journal of Anaesthesia 1988; 60: 130–135. 3. Murray JM, Rowlands BJ, Trinick TR. Indocyanine green clearance and hepatic function during and after prolonged anaesthesia: comparison of halothane with isoflurane. British Journal of Anaesthesia 1992; 68: 168–171. 4. Allan LG, Hussey AJ, Howie J, Beckett GJ, Smith AF, Hayes JD, Drummond GB. Hepatic glutathione S-transferase release after halothane anaesthesia: open randomised comparison with isoflurane. Lancet 1987; 1: 771–774. 5. Aldridge LM, Ray DC, Noble DW, Howie AF, Beckett GJ, Drummond GB. Glutathione S-transferase measurement can detect impaired hepatocellular integrity in anaesthetised, ventilated subjects. Clinical Chemistry and Enzymology Communications 1993; 5:195–203. Sir,—Thank your for the opportunity to respond to the comments by Ray, Beckett and Allan concerning our study on hepatocellular integrity after isoflurane and halothane anaesthesia.1 The two differences between our study1 and that by Hussey and co-workers2 were: (1) in our study there was no significant difference in GSTA between the isoflurane and halothane groups; and (2) in our study GSTA increased also in the isoflurane group. The latter fact may, indeed, have been caused by the influence of positive pressure ventilation in normotensive patients as seems to be the case in patients operated on under deliberate hypotension.3 However, the former requires other explanations. For about 9 yr it has been thought that halothane anaesthesia is usually associated with a disturbance in hepatocellular integrity

773 while isoflurane anaesthesia is not.2 Our study challenged this “truth”1 We cannot totally reject our speculation that the former difference between the studies is caused by previous exposure to halothane or by differences in concentration of anaesthetics. Halothane was probably more commonly used in Scotland in the 1980s than in Finland in the 1990s. The change in GSTA seems not to be related to the dose of anaesthetic agent or to duration of anaesthesia, but we have not found any study on GSTA comparing different concentrations of volatile anaesthetics, or comparing different durations of anaesthesia to prove these concepts. Murray, Rowlands and Trinick4 used a relatively insensitive assay and found a difference in GSTA concentration between halothane and isoflurane at 2 MAC for 10 h. There was also an increase in aspartate aminotransferase.4 On the other hand, we found no significant difference between these anaesthetics at 1 MAC for 3 h.1 In the studies in which Ray, Beckett and Allan have been co-authors,2 3 the concentrations of volatile anaesthetics and the changes in GSTA concentration seemed to fall between these extremes. We agree that positive pressure ventilation may be an important cause for the increase in GSTA after isoflurane anaesthesia but we need other explanations for the lack of difference between isoflurane and halothane. The increase in GSTA concentration after sevoflurane anaesthesia is also interesting because it occurs after positive pressure ventilation5 but also after spontaneous ventilation.6 P. TIAINEN P. H. ROSENBERG Department of Anaesthesiology Helsinki University Central Hospital Helsinki, Finland 1. Tiainen P, Rosenberg PH. Hepatocellular integrity during and after isoflurane and halothane anaesthesia in surgical patients. British Journal of Anaesthesia 1996; 77: 744–747. 2. Hussey AJ, Aldridge LM, Paul D, Ray DC, Beckett GJ, Allan LG. Plasma glutathione S-transferase concentration as a measure of hepatocellular integrity following a single general anaesthetic with halothane, enflurane or isoflurane. British Journal of Anaesthesia 1988; 60: 130–135. 3. Aldridge LM, Ray DC, Noble DW, Howie AF, Beckett GJ, Drummond GB. Glutathione S-transferase measurement can detect impaired hepatocellular integrity in anaesthetised, ventilated subjects. Clinical Chemistry and Enzymology Communications 1993; 5: 195–203. 4. Murray JM, Rowlands BJ, Trinick TR. Indocyanine green clearance and hepatic function during and after prolonged anaesthesia: comparison of halothane with isoflurane. British Journal of Anaesthesia 1992; 68: 168–171. 5. Taivainen T, Tiainen P, Meretoja OA, Räihä L, Rosenberg PH. Comparison of sevoflurane and halothane on the quality of anaesthesia, serum glutathione transferase alpha and fluoride in paediatric patients. British Journal of Anaesthesia 1994; 73: 590–595. 6. Ray DC, Bomont R, Mizushima A, Kugimiya T, Forbes Howie A, Beckett GJ. Effect of sevoflurane anaesthesia on plasma concentrations of glutathione S-transferase. British Journal of Anaesthesia 1996; 77: 404–407.

Safety issues and volatile agent analysers Sir,—Drs Strachan and Richmond raised some interesting points in their case report concerning the presence of diethyl ether in an isoflurane vaporizer.1 Older Datex anaesthetic agent monitors are calibrated to measure levels of halothane, enflurane, isoflurane and methoxyflurane. It has been noticed in this hospital that when using sevoflurane the agent monitor seems to give readings corresponding to the vaporizer settings when set to measure methoxyflurane which is not seen when the agent monitor is set for any other volatile agent. This was purely a chance observation and has certainly not been used clinically as a monitoring method. It does however seem surprising, given the different physical and chemical properties of sevoflurane (molecular weight 200.1, blood:gas partition coefficient 0.69, boiling point 58.6 ⬚C) and methoxyflurane (molecular weight 165, blood:gas partition coefficient 13, boiling point 105 ⬚C) and the seemingly closer similarity of sevoflurane to the routinely used volatile agents.2 This must reflect similar infrared absorption bands of sevoflurane and methoxyflurane.

774 This would seem to indicate that in some circumstances two compounds with quite distinct properties can effect monitoring in similar ways given similar absorption spectra. This shows that advancing the scope of anaesthesia in one area must be paralleled by the ability to use appropriate monitoring techniques and, at best, monitoring is only an adjunct to clinical vigilance. P. J. SHIRLEY Department of Anaesthetics Raigmore Hospital Inverness 1. Strachan AN, Richmond MN. Ether in an isoflurane vaporizer and the use of vapour analysers in safe anaesthesia. British Journal of Anaesthesia 1997; 78: 107–108. 2. Smith I, Nathanson M, White PF. Sevofluorane—a long awaited volatile anaesthetic. British Journal of Anaesthesia 1996; 76: 435–445.

Ether in an isoflurane vaporizer and the use of vapour analysers in safe anaesthesia Sir,—We read with interest the case report by Strachan and Richmond1 and wish to make the following comments. The ether-filled temperature compensation unit used in the PPV Sigma has been used safely in Penlon vaporizers for more than 40 yr and was one of the world’s first automatic TC units, having been first introduced into the EMO vaporizer. The design of the now obsolete PPV Sigma vaporizer discussed in the article necessitated a significant TC movement that was not possible to obtain by bimetallic expansion alone. That movement could only be provided by that type of liquid-filled expansion system. The current range of Penlon Elite vaporizers makes use of significantly different control variables and as a result of these major design changes Penlon vaporizers no longer use liquid-filled TC units but do use a solid state bimetallic TC unit in common with some other manufacturers. A. C. GREEN Penlon Ltd Abingdon 1. Strachan AN, Richmond MN. Ether in an isoflurane vaporizer and the use of vapour analysers in safe anaesthesia. British Journal of Anaesthesia 1997; 78: 107–108.

Non-invasive measurement of cerebral blood flow Sir,—I read with interest the recent article by Gupta and coworkers.1 It was only when I set out to derive the equations presented within this article that I realized that one of the definitions is wrong. The authors define CLVHR as “the coefficient of the large to small vessel packed cell ratio” whereas it is the inverse, the ratio of cerebral (small vessel) to large vessel haematocrits, as derived by Lammertsma and colleagues2 and calculated as 0.69. It is the definition that is misleading, the equations are correct for the value quoted. Unfortunately, Gupta and co-workers have simply compounded the original error in definition given by Elwell and co-workers.3 My second comment is on the use of FE′CO2 as equivalent to PaCO2 when calculating carbon dioxide reactivity. First, the normal value for PaCO2 is 5.3 kPa but this does not equate to an FE′CO2 of 5.3 kPa. At rest, with quiet respiration, there is a small difference of 0.3–0.6 kPa between PaCO2 and FE′CO2 . Thus data should have been indexed to a normal value for FE′CO2 , namely 4.85 kPa. The difference between PaCO2 and FE′CO2 is important in the context of this study as respiration is altered to achieve different FE′CO2 values. We are not told how the volunteers hyperventilated to reach a PaCO2 of 3.5 kPa, but most people would increase ventilatory frequency and tidal volume may decrease; conversely, adding carbon dioxide to the gas mixture triggers hyperventilation largely by increasing depth of ventilation in addition to ventilatory frequency. As the difference between PaCO2 and FE′CO2 is dependent on tidal volume, the slope of the CBV vs FE′CO2 line is no longer identical to the slope of the CBV vs PaCO2 relationship. Physiologically, carbon dioxide reactivity is dependent on PaCO2 not FE′CO2 and so this source of error in determining carbon dioxide reactivity should be discussed as we are not told if the volunteers were asked to maintain a constant tidal volume for all parts of the study.

British Journal of Anaesthesia The lack of any significant change in packed cell volume in relation to FE′CO2 is indeed puzzling when the calculated values for CBV are clearly related to FE′CO2 . Although the possibility of following changes in CBV non-invasively may be useful clinically, particularly for head injuries, to do this by altering PaCO2 is not acceptable. It seems this point must be addressed before the methodology presented here is acceptable. S. A. HILL Shackleton Department of Anaesthesia Southampton General Hospital Southampton 1. Gupta AK, Menon DK, Czosnyka M, Smielewski P, Kirkpatrick PJ, Jones JG. Non-invasive measurement of cerebral blood volume in volunteers. British Journal of Anaesthesia 1997; 78: 39–43. 2. Lammertsma AA, Brooks DJ, Beaney RP, Turton DR, Kensett MJ, Heather JD, Marshall J, Jones T. In vivo measurement of regional cerebral haematocrit using positron emission tomography. Journal of Cerebral Blood Flow and Metabolism 1984; 4: 317–322. 3. Elwell CE, Cope M, Edwards AD, Wyatt JS, Delpy DT, Reynolds EOR. Quantification of adult cerebral hemodynamics by near-infrared spectroscopy. Journal of Applied Physiology 1994; 77: 2753–2760. Sir,—I read with interest the article by Gupta and colleague1, but have grave doubts on the validity of the conclusions regarding the ability to non-invasively assess cerebral blood volume, for several reasons. First, no currently available near-infrared (NIR) spectroscopic system has the ability to completely exclude the scalp signal, which may potentially cause a great deal of artefact. Second, the oxidized and reduced haemoglobin concentrations were derived from NIR absorption spectra by an algorithm, introducing a further potential inaccuracy. Third, calculation of [Hbcalc] involves examining the changes in the derived signals, thereby compounding the error of an algorithm with the influence of the scalp tissue signal. Possibly this helps explain the 25% difference in the calculated cerebral blood volume (CBV) reactivities of 1.25 and 1.06 ml 100 g91 kPa91 when HbO2 or Hbdiff concentrations were used to perform the calculation. Fourth, CBV calculated in ml 100 g91 assumed a large number of variables to be constant. Haemoglobin was assumed to be 15 g dl91, even though five of 13 subjects were female and of an age likely to be menstruating. The coefficient of large to small vessel packed red cell ratio is assumed to be constant throughout, even though the authors noted that flow, as measured by transcranial Doppler, changed with moderate differences in oxygen content indicating vasoreactivity, which must surely change the large to small ratio. Lastly, there is no “gold standard” with which the authors have compared their results, and although they may have measured changes in cerebral blood volume, one cannot be sure from this study. D. WILLIAMS University Department of Surgery Royal United Hospital Trust Bath 1. Gupta AK, Menon DK, Czosnyka M, Smielewski P, Kirkpatrick PJ, Jones JG. Non-invasive measurement of cerebral blood volume in volunteers. British Journal of Anaesthesia 1997; 78: 39–43. Sir,—We thank Drs Hill and Williams for their comments. In response to Dr Hill, we agree that the definition of CLVHR is the small to large vessel haematocrit ratio, as described by Lammertsma and colleagues.1 The study aimed to evaluate non-invasive methods of assessing cerebral blood volume. While we accept that there is a small difference in PaCO2 and end-tidal carbon dioxide ( FE′CO2 ), arterial blood-gas monitoring was not used as described in the article. The data were indexed to the calculated CBV at and F E′ CO2 5.3 kPa. Although this does not equate to a PaCO2 of 5.3 kPa, the regression equation of CBV reactivity should not be affected. Further, the expression of CBV reactivity as a percentage of baseline CBV compensates for individual variations in baseline CBV. End-tidal carbon dioxide manipulation is a recognized method of measuring carbon dioxide reactivity2 and is used clinically to assess carbon dioxide reactivity. However, we would not alter PaCO2 to measure CBV in head injured patients. This

Correspondence is not implied in the study; we have used PaCO2 variations in volunteers merely as a paradigm of CBV alterations in volunteers and these changes would result from spontaneous pathophysiological mechanisms in disease. In response to Williams’ comments, it is well recognized that there is extracranial contamination in the near infrared signal. It has been reported that there is approximately a 15% contribution of skin changes to the HbO2 signal during carbon dioxide challenge, and much of this change is related to changes in systemic arterial pressure.2 Our volunteers showed no significant change in MAP. It therefore seems unlikely that changes in skin blood flow contributed significantly to the carbon dioxide reactivity of the NIR signal. The use of an estimated Hb value and the approximations involved in the calculation may have biased the absolute level of CBV measured, but are unlikely to have affected reactivities. While there is as yet no direct comparison of NIR measurement of CBV and CBV reactivity against the gold standard of positron emission tomography (PET), we have discussed our results in the context of data obtained by other authors and different techniques, including PET.3 A. K. GUPTA D. K. MENON Department of Anaesthesia Addenbrooke’s NHS Trust Cambridge 1. Lammertsma AA, Brooks DJ, Beaney RP, Turton DR, Kensett MJ, Heather JD, Marshall J, Jones T. In vivo measurement of regional cerebral haematocrit using positron emission tomography. Journal of Cerebral Blood Flow and Metabolism 1984; 4: 317–322. 2. Smielewski P, Czosnyka M, Pickard JD, Kirkpatrick P. Clinical evaluation of near-infrared spectroscopy for testing cerebrovascular reactivity in patients with carotid artery disease. Stroke 1997; 28: 331–338. 3. Grubb RL, Raichle ME, Eichling JO, Ter-Pogossian MM. The effects of changes in PaCO2 on cerebral blood volume, blood flow, and vascular mean transit time. Stroke 1974; 5: 630–639.

Tolerance of laparoscopy for resection of phaeochromocytoma Sir,—In their report of two cases of intraperitoneal laparoscopic phaeochromocytoma resection, Mann and colleagues suggested that haemodynamic disturbance associated with pneumoperitoneum in one case may question the safety of this approach.1 Such a conclusion cannot be justified on their evidence in which adequate preoperative preparation was given such a low priority. They stated that 10 days preoperative treatment failed to provide satisfactory haemodynamic control in a patient, yet despite this, surgery was allowed to proceed. They then claimed that the hypertensive response to peritoneal insufflation in this patient reflected a risk associated with pneumoperitoneum rather than inadequate preoperative preparation. Notably, the second case (in which arterial pressure was stabilized before surgery with 20 days’ treatment) showed no vasopressor response to pneumoperitoneum. While no controlled, randomized study has been performed, it is generally agreed that preoperative preparation is the major factor in the reduction in perioperative morbidity and mortality and indeed preoperative therapy is mandatory in patients with sustained or paroxysmal hypertension.2 Their choice of antihypertensiye drugs is also worthy of comment. They used nicardipine and, in the first case, labetolol. Calcium channel blockers such as nicardipine have been used in the perioperative period but cannot be regarded as “agents of choice” on the references quoted.3 4 Labetolol is inadequate to control arterial pressure if large quantities of noradrenaline are released5 and has been reported to cause paradoxical hypertension in patients with phaeochromocytoma6 because of its potency at ␤ relative to ␣ receptors. Labetolol is therefore regarded as a poor choice of drug in the management of patients with phaeochromocytoma. ␣-Adrenergic blockade, restoration of plasma circulating volume and ␤-blockade if required to control tachyarrhythmias are widely accepted for preoperative preparation. Mann and colleagues refer to the retroperitoneal approach and

775 apparent cardiostability of such cases, as reported by Giebler and colleagues.7 It is worth noting that these patients were treated with phenoxybenzamine for 21 days before operation. C. JERWOOD B. WILLIS University Hospital of Wales Heath Park Cardiff 1. Mann C, Millat B, Boccara G, Atgar J, Colson P. Tolerance of laparoscopy for resection of phaeochromocytoma. British Journal of Anaesthesia 1996; 77: 795–797. 2. Desmonts JM, Marty J. Anaesthetic management of patients with phaeochromocytoma. British Journal of Anaesthesia 1984; 56: 781–789. 3. Arai T, Hatano Y, Mori K. Use of nicardipine in the anesthetic management of phaeochromocytoma. Anesthesia and Analgesia 1986; 65: 706–708. 4. Colson P, Ribstein J. Strategic simplifiee pour l’anesthesie du pheochromocytome. Annales Francoises d’Anesthésie et de Réanimation 1991; 10: 456–462. 5. Russel WJ, Kaines AH, Hooper MJ, Frewin DB. Labetolol in the preoperative management of phaeochromocytoma. Anaesthesia and Intensive Care 1982; 10: 160. 6. Navaratnarajah M, White DC. Labetolol and phaeochromocytoma. British Journal of Anaesthesia 1984; 56: 1179. 7. Giebler RM, Walz MK, Peitgen K, Scherer RU. Hemodynamic changes after retroperitoneal CO2 insufflation for posterior retroperitoneoscopic adrenalectomy. Anesthesia and Analgesia 1996; 82: 827–831. Sir,—We thank Drs Jerwood and Willis for their interest in our investigation1 but their letter deserves comment on two points: first, preoperative preparation, and second, effect of pneumoperitoneum. We agree that no controlled, randomized study has demonstrated that preoperative preparation for phaeochromocytoma reduces perioperative morbidity and mortality. It is also not established that preoperative treatment should last at least 21 days to improve perioperative haemodynamic stability. Moreover, no preoperative preparation can control all perioperative haemodynamic events.2 The issue of the need for any preoperative preparation has even been questioned elsewhere.3 ␣-Adrenergic blockade requires time to be fully effective (this sometimes “justifies” up to 21 days of preparation!) and has many drawbacks: (a) arterial pressure instability when inducing treatment may delay the time of completing vasodilator therapy, (b) decreased venous tone and cardiac preload, and (c) increased risk of severe postoperative hypotension, especially with long-acting drugs such as phenoxybenzamine. In our experience, dihydropyridines allow short-term preoperative preparation,4 do not induce arterial pressure instability, do not adversely affect venous return and cardiac preload, and do not cause severe postoperative hypotension. All of these features provide many reasons to choose dihydropyridines rather than ␣-adrenergic blocking drugs for preoperative preparation of phaeochromocytoma surgery. Labetolol was used as a ␤-blocking drug, and it was added to a preliminary vasodilator therapy; therefore the risk of paradoxical hypertension was avoided. In the first of the two case reports, peritoneal insufflation induced a severe increase in catecholamine concentration that mimicked adrenal tumour response to surgical manipulation. This is the evidence that insufflation of the pneumoperitoneum may induce direct stimulation of the adrenal tumour. Even 21 days of phenoxybenzamine preparation would not have avoided this catecholamine release from the adrenal tumour, as ␣-adrenergic blockade does not act on tumour secretion. Therefore, our conclusion is justified and we recommend caution with peritoneal insufflation for laparoscopic phaeochromocytoma resection. P. COLSON C. MANN Department of Anaesthesiology and Intensive Care University of Montpellier Montpellier, France 1. Mann C, Millat B, Boccara G, Atger J, Colson P. Tolerance of laparoscopy for resection of phaeochromocytoma. British Journal of Anaesthesia 1996; 77: 795–797. 2. Desmonts JM, Marty J. Anaesthetic management of patients with phaeochromocytoma. British Journal of Anaesthesia 1984; 56: 781–789.

776

British Journal of Anaesthesia

3. Angermeier KW, Montie JE. Perioperative complications of adrenal surgery. Urologic Clinics of North America 1989; 16: 597–606. 4. Colson P, Ribstein J. Stratégic simplifiée pour l’anesthésie du phéochromocytome. Annales Françoises d’Anesthésie et de Réxanimation 1991; 10: 456–462.

Airway problems after carotid endarterectomy Sir,—I read with interest the article by Munro, Makin and Read1 describing four cases of acute respiratory obstruction after development of postoperative wound haematomas following carotid endarterectomy. They make the point that such patients require continuous respiratory monitoring in order to detect developing respiratory obstruction at an early stage. It is worth emphasizing that the neurological complications they refer to after carotid endarterectomy may include changes in respiratory function and control.2 The presence of hypoventilation or even apnoea3 would make the obstructive effect of postoperative haematoma even more rapid in onset. Gradual onset of respiratory obstruction may also lead to hypercapnia, which by means of the development of cerebral steal syndromes might lead to a decrease in blood flow in already ischaemic areas of the brain, thereby exacerbating the already compromised control of respiration. Inadequate control of arterial pressure in the postoperative period increases this effect in addition to increasing the size of postoperative wound haematomas. D. BEAMISH Department of Anaesthetics Eastern General Hospital Edinburgh 1. Munro FJ, Makin AP, Reid J. Airway problems after carotid endarterectomy. British Journal of Anaesthesia 1996; 76: 156–159. 2. Bjoraker DG. Abdominal and major vascular surgery. In: Nimmo WS, Smith G, eds. Anaesthesia. Oxford: Blackwell Scientific Publications, 1989; 740–743. 3. Beamish D, Wildsmith JAW. Ondines curse after carotid endarterectomy. British Medical Journal 1978; 2: 1607–1608.

In vivo and in vitro dose–response curves Sir,—Our Japanese colleagues1 have made a good attempt in interpreting human anaesthetic dose–response curves at a molecular level. But they admit that not all the values they have derived are plausible. May I offer two suggestions? They used the method of least squares, however, this has underlying assumptions that the data are Gaussian and homoscedastic, neither of which is true for quantal data. One method of analysing quantal data was described by Waud2: the appendix provides a simple BASIC program, which will solve equation (2) of the Osaka study. Waud’s programme is translated easily into Microsoft Professional Basic, a language which overcomes most of the objections to old BASIC raised by computer professionals. Second, the Osaka team suggested that the in vivo curve is the outcome of a normal, or log-normal, array of events at the molecular level. If so then their equation (2) would match reality better as a series, not a single unit: y=

x n1 K1n1

+x

n1

+

x n2 K1n2

+x

n2

R. S. CORMACK Northwick Park Hospital, Harrow 1. Iwai T, Kihara H, Imaiand K, Uchida N. Dose–response curve for anaesthetics based on the Monod–Wyman–Changeux model. British Journal of Anaesthesia 1996; 77: 517–521. 2. Waud DR. On biological assays involving quantal responses. Journal of Pharmacology and Experimental Therapeutics 1972; 183: 577–607. Sir,—We appreciate the letter of Dr Cormack which relates to our article on dose–response curves.1 He suggests that assays of in vivo dose–response data have to be treated by the method of analysing quantal data, instead of the method of least squares, as described by Waud.2 According to Waud’s programme, we have reexamined the in vivo data and obtained revised results as follows: ED50:1.01%atm (SEM:0.08%atm), n:3.59 (SEM:0.89) for isoflurane, and ED50:0.76%atm (SEM:0.008%atm), n:29.2 (SEM:7.6) for halothane. These results are in agreement with those in our articles (ED50:1.03%atm, n:3.6 for isoflurane; ED50:0.75%atm, n:27 for halothane). Therefore, our least square fittings did not differ markedly from quantal assays. Second, he suggests better fitting of data to the Hill equation (2) in our text by adding another unit. This may be effective for reducing residual variance, but the biological or molecular meaning of the adding term is not clear. Such a modification should be plausible when two or more doses are applied simultaneously,2 or when some integrated effects take place on an organism, such as linkage effects between two or more receptor molecules each of which can bind with a drug molecule. In both cases, however, we have another problem, whether or not simple addition of one unit is suitable, because it depends on an adopted model. Lastly, he mentioned an effect of a series of synapses. This is an interesting suggestion, but if there were a long series of synapses, only a small amount of dose applied would anaesthetize any patient. In other words, it cannot show the existence of a critical dose value above which a patient does not respond; the critical value is different between patients. In a complicated system of neurone networks, information propagates backwards in addition to forwards, and then some integrated response emerges. We suggest that this must be responsible for the existence of the critical dose value mentioned above. Although our MWC model oversimplifies the system, the model inherently has an integrated nature. We thank Dr Cormack for his valuable suggestions and his interest in our article. T. IWAI H. KIHARA Physics Laboratory Kansai Medical University, Hirakata K. IMAI Department of Physiology, School of Medicine, Osaka University, Suita M. UCHIDA Depertment of Anaesthesia, Kitano Hospital, Osaka Japan 1. Iwai T, Kihara H, Imai K, Uchida M. Dose–response curve for anaesthetics based on the Monod–Wyman–Changeux model. British Journal of Anaesthesia 1996; 77: 517–521. 2. Waud DR. On biological assays involving quantal responses. Journal of Pharmacology and Experimental Therapeutics 1972; 183: 577–607.

+ ...

Adding one unit should reduce residual variance and the experiment repeated until there is no further improvement. With three units the solution would require inverting a 6⫻6 matrix, but that is no problem in Professional Basic, which has an efficient algorithm for any size of matrix. Lastly, anaesthetics act on the synapse; if 1% halothane reduces transmission across each synapse to 70% of normal, then over a relay of five synapses it is decreased to 17% (0.75). Does it not follow that the in vivo curve should be steeper than effects at single synapses? This is why a simple reflex, such as withdrawing the foot, can be present when consciousness has been extinguished, as John Snow noticed 150 yr ago. May I wish them good luck with their endeavours.

Ambulatory extradural analgesia in labour and mode of delivery Sir,—Drs May and Elton1 recently reported the results of a small prospective study comparing standard non-ambulatory extradural analgesia (boluses of 0.25% bupivacaine 10 ml) with ambulatory combined spinal–extradural analgesia (spinal bupivacaine 2.5 mg and fentanyl 2 ␮g followed by extradural top-ups of 0.1% bupivacaine 10 ml with fentanyl 2 ␮g ml91). They found that the rate of emergency Caesarean section was significantly lower in the ambulatory group (three of 30 vs 11 of 37; chi-square:3.89, P:0.05). If valid, this finding would be a major step forward for obstetric anaesthesia. The first ambulatory, or mobile, extradural at the Royal Surrey

Correspondence County Hospital was performed in late 1993. Because of the high rate (2.3%) of post-dural puncture headache after combined spinal–extradural analgesia reported by Collis and colleagues,2 we adopted successfully a technique using extradural boluses of 0.1% bupivacaine 20 ml with fentanyl 2 ␮g ml91, without a spinal and without a conventional test dose. In 1996, 1100 (95.8%) of the 1148 extradurals (42.4% of labouring mothers) given for pain relief used this mobile technique, compared with one (0.11%) of 907 extradurals (35.2% of labouring mothers) in 1993. Unfortunately, there has not been a reduction in the rate of emergency Caesarean section in mothers using extradural analgesia. In 1993, 78 (8.6%) of 907 mothers with extradurals required emergency Caesarean section compared with 146 (12.7%) of 1148 in 1996. This is a statistically significant increase (chi-square:8.8, P:0.01). However, there has also been a significant increase in the number of mothers having a spontaneous vaginal delivery, from 432 (47.6%) in 1993 to 646 (56.3%) in 1996 (chi-square:15.9, P:0.001) and a significant decrease in the number of instrumental deliveries, from 397 (43.8%) in 1993 to 356 (31.0%) in 1996 (chi-square:35.5, P:0.001). In those mothers who did not receive an extradural, the rates of spontaneous vaginal delivery, instrumental delivery and emergency Caesarean section have not changed significantly between 1993 (85.4%, 5.3% and 9.3%, respectively) and 1996 (87.8%, 4.2% and 8.0%, respectively). A randomized study by Collis, Davies and Aveling3 comparing ambulatory combined spinal–extradural analgesia with standard non-ambulatory extradural analgesia found no difference in the outcome of labour. However, Asselineau4 observed a significantly lower instrumental delivery rate but similar Caesarean section rates in the ambulant group when comparing ambulant and non-ambulant extradurals. Experience at this hospital suggests that the use of mobile, or ambulatory, extradurals may lead to a higher rate of spontaneous vaginal delivery and a lower rate of instrumental delivery. J. G. JENKINS Department of Anaesthesia Royal Surrey County Hospital Guildford 1. May AE, Elton CD. Ambulatory extradural analgesia in labour reduces risk of Caesarean section. British Journal of Anaesthesia 1996; 77: 692–693P. 2. Collis RE, Baxandall ML, Srikantharajah ID, Edge G, Kadim MY, Morgan BM. Combined spinal epidural analgesia with the ability to walk throughout labour. Lancet 1993; 341: 767–768. 3. Collis RE, Davies DWL, Aveling W. Randomised comparison of combined spinal–epidural and standard epidural analgesia in labour. Lancet 1995; 345: 1413–1416. 4. Asselineau D. La deambulation sous peridurale lors du travail modifie-t-elle les conditions d’extraction foetale? (Does ambulation under epidural analgesia modify the conditions of fetal delivery?). Contraception, Fertilite, Sexualite 1996; 24: 505–508. Sir,—We thank Dr Jenkins for his interest in our small randomized study1 comparing extradural analgesia using 0.25% bupivacaine with a previously described combined spinal– extradural (CSE).2 His large retrospective audit of extradural technique and mode of delivery confimed the findings of some,3 4 although not all, studies.5 6 We appreciate his anxiety about the use of the subarachnoid component of this technique with regard to reports of meningitis after CSE in labour.7 The evidence for an increased incidence of post-dural puncture headache is however not so clear.8 It is possible that the ability to provide pain-free ambulation in labour may be beneficial if it is produced using an extradural only technique or a CSE. We have recently commenced a much larger collaborative study with colleagues at Birmingham Women’s Hospital comparing these techniques and we hope this will demonstrate if the spinal component of a combined spinal–extradural is essential for the beneficial effect suggested in our study. C. D. ELTON A. E. MAY Department of Anaesthesia Leicester Royal Infirmary Leicester 1. May AE, Elton CD. Ambulatory extradural analgesia in labour reduces the risk of Caesarean section. British Journal of Anaesthesia 1996; 77: 692–693.

777 2. Collis RE, Baxandall ML, Srikantharajah ID, Edge G, Kadim MY, Morgan BM. Combined spinal epidural (CSE) analgesia: technique, management and outcome of 300 mothers. International Journal of Obstetric Anaesthesia 1994; 3: 75–81. 3. Vertommen JD, Vandermeulen E, Van Aken H, Vaes L, Soetens M, Van Steenberge A, Mourisse P, Willaert J, Noordvin H, Devlieger H, Van Assche A. The effects of the addition of sufentanil to 0.125% bupivacaine on the quality of analgesia during labor and the incidence of instrumental deliveries. Anesthesiology 1994; 74: 809–814. 4. Murphy TD, Henderson K, Bowden MI, Lewis M, Cooper GM. Bupivacaine versus bupivacaine plus fentanyl for epidural analgesia: effect on maternal satisfaction. British Medical Journal 1991; 302: 564–567. 5. Collis RE, Davies DWL, Aveling W. Randomised comparison of combined spinal epidural and standard epidural analgesia in labour. Lancet 1995; 345: 1413–1416. 6. Russell R, Reynolds F. Epidural infusion of low-dose bupivacaine and opioid in labour. Anaesthesia 1996; 51: 266–273. 7. Harding SA, Collis RE, Morgan BM. Meningitis after combined spinal–extradural anaesthesia in obstetrics. British Journal of Anaesthesia 1994; 73: 545–554. 8. Norris MC, Grieco WM, Borkowski M, Leigton BL, Akroosh VA, Huffnagle J, Huffnagle S. Complications of labor analgesia: Epidural versus combined spinal epidural techniques. Anesthesia and Analgesia 1994; 79: 529–537.

The “Dumfries claim” Sir,—The reference to the first trial of ether anaesthesia at Dumfries and Galloway Royal Infirmary on 19 December 1846 (in your editorial1) is dismissive and, as such, deserving of comment. In an allusion to William Scott’s assertion of priority over Robert Listen with regard to the date, Professor Spence refers to the “Dumfries claim”, quoting, I presume, the title of my original paper in your journal.2 The “claim”, as Professor Spence correctly stated, was contained in William Scott’s letter to The Lancet3 on October 19, 1872, one of the no fewer than three reliable references to the validity of Scott’s assertion. While the question of precedence was an important personal issue for William Scott, at no time did he infer that his early application of anaesthesia at Dumfries had any influence on the rapid acceptance of ether. Nevertheless, had James Y. Simpson been in Edinburgh at the time, in which case the news of Scott’s operation would have reached him almost immediately, the Old World would doubtless have been informed of the anaesthetic properties of ether through alternative channels! In that case, Robert Listen might well have had to share the honours with his colleagues in south-west Scotland. The suggestion by Professor Spence that James Y. Simpson travelled to London “for the specific purpose of learning about anaesthesia” would appear to be incorrect. Simpson’s journey to London in December 1846 was the first of two visits in connection with his appointment as surgeon to Her Majesty Queen Victoria’s household in Scotland and, on the date of the Squire:Liston success, he was either on his way to, or had already arrived in, London. He received the news of ether, therefore, not from London but in London. The narrative surrounding the Dumfries operation is a fascinating one and has been the subject of several publications by this writer over the past 30 yr.4 5 The gentlemen at Dumfries provided the chronicles of anaesthesia with a remarkable story and the fact that particulars regarding the patient are, as yet, incomplete does not lend the right to dismiss the incident with, “the outcome, whatever it was, adds nothing to science or medicine”. T.W. BAILLIE Hengelo Gld The Netherlands 1. Spence AA. Ether anaesthesia comes to London. December 1846. British Journal of Anaesthesia 1996; 77: 705–706. 2. Baillie TW. The first European trial of anaesthetic ether: the Dumfries claim. British Journal of Anaesthesia 1965; 37: 952. 3. Scott W. The exhibition of ether as an anaesthetic. Lancet 1872; 585. 4. Baillie TW. From Boston to Dumfries. Dumfries: Dinwiddie, 1966 and 1969. 5. Baillie TW. The Dumfries Ether Diary. Dumfries: Creedon, 1996 (ISBN 1 899 3316 353).

778 Sir,—Writing from this address about the arrival of ether anaesthesia in Britain, what is one to do? Having mentioned Dumfries, thus hoping to avoid the letters that would have been sent had I failed to do so, it was necessary to give a brief evaluation of the significance of the event. I am grateful to Dr Baillie for drawing attention to the revised account of Simpson’s movements around Christmas, 1846. Dr Baillie will understand that I did not have the benefit of his interesting 1996 book1 until after the editorial had been published. As it happens I had been guided by the sentence “no sooner had word of the first successful demonstration of surgical anaesthesia in London reached the Scottish capital than Simpson was on his way south.” This is taken from Dr Baillie’s 1966 book.2 Even if Simpson had remained in Edinburgh in the relevant days, it is difficult to know how he might have been the vehicle through whom the Old World would have heard the news. The problem of the “Dumfries claim” is one of non-communication. Simpson was just as much in the dark as everyone else. A. A. SPENCE University Department of Anaesthetics Royal Infirmary of Edinburgh Edinburgh 1. Baillie TW. The Dumfries Ether Diary. Dumfries; Creedon, 1996 (ISBN 1 899 3316 353). 2. Baillie TW. From Boston to Dumfries. Dumfries: Dinwiddie, 1966 and 1969.

Allergies and anaesthesia In their case report,1 Rae, Milne and Wildsmith highlighted the importance of obtaining a detailed history, particularly in the area of allergies. I recall a similar case also during a Caesarean section under regional anaesthesia which reinforces this point. A 35-yr-old ex-nurse, with atopy but no known allergies, underwent a routine combined extradural–spinal anaesthetic for elective Caesarean section. Catheterization with a latex urinary catheter was uneventful. Just after delivery and a bolus of i.v. oxytocin she complained of “feeling unwell” and said her tongue “felt swollen”, she developed an erythematous rash on her upper body with visible facial oedema and her arterial pressure rapidly became unrecordable. A diagnosis of anaphylaxis was made; amniotic fluid embolus was part of the differential diagnosis. Treatment with oxygen, i.v. crystalloid and adrenaline rapidly restored her to her previous state. A similar, but less severe,

British Journal of Anaesthesia episode occurred before surgery was completed. Recovery was uneventful, the urinary catheter was removed 24 h after operation. It was only afterwards on direct questioning that the patient described past episodes of facial swelling on blowing up balloons, contact dermatitis to rubber gloves and vaginal irritation to condoms. Skin prick testing proved highly positive for latex, and negative for oxytocin, bupivacaine and Gelofusine. This case also demonstrates that the risk of latex allergy in medical staff (2.8%) is greater than that found in patients presenting for skin prick tests (0.8%) although theatre staff are at greatest risk (7.8%).2 Repeated latex exposure is believed to be responsible for these observations. Latex-containing products release variable amounts of antigen (water-soluble proteins), examination gloves and urinary catheters releasing far more than surgical gloves. This may explain why anaphylactic reactions associated with the former commonly occur within 2–3 min, as in Rae, Milne and Wildsmith’s case, while those associated with surgical gloves contacting the large absorptive surface of the abdomen (the commonest reported route) often take longer to develop, typically 40–300 min,3 as in the above case. Why the two patients reacted so differently I am at a loss to explain. Since the original case report4 latex sensitivity has become recognized as a relatively common cause of allergic dermatitis and anaphylaxis. This may be caused by a combination of change in the manufacturing process, greater antigen exposure, heightened clinical awareness and more frequent use of latex as an antigen during investigations of drug reactions. Whatever the cause it is necessary to inquire about latex sensitivity from all of our patients to avoid similar scenarios. R. SEIGNE Anaesthetic Department Addenbrooks Hospital Cambridge 1. Rae SM, Milne MK, Wildsmith JAW. Anaphylaxis associated with, but not caused by, extradural bupivacaine. British Journal of Anaesthesia 1997; 78: 224–226. 2. Turjanmaa K. Incidence of immediate allergy to latex gloves in hospital personnel. Contact Dermatitis 1987; 17: 270–275. 3. Gold M, Swatz JS, Braude BM, Dolovich J, Shandling B, Gilmour RF. Intraoperative anaphylaxis: An association with latex sensitivity. Journal of Allergy and Clinical Immunology 1991; 87: 626–666. 4. Nutter AF. Contact urticaria to rubber. British Journal of Dermatology 1979; 101: 597–598.