we can be an important resource for other medical departments utilizing anesthetic drugs. We thank Drs. Epstein and Schyve for their interest and comments. We agree that the directors of each department are ultimately responsible for the quality of patient care in their own departments; however, we should not avoid our responsibility in helping with the monitoring and evaluation of the quality of anesthesia outside the realm of our direct control . Who should know better the use of anesthetic drugs and safety concerns of administering an anesthetic than an anesthesiologist?

BACKGROUND:We determined whether the coupling between cerebral blood flow (CBF) and oxygen metabolism (CMRO2) is preserved during liver transplantation. Because of cerebrovascular dilatation, we hypothesized that cerebral metabolic autoregulation is impaired, because CBF becomes uncoupled from CMRO2 during the reperfusion phase of the operation.MATERIALS AND METHODS:In a prospective study, 13 patients (8 women, median age 46, range 21-6) with liver failure (10 with end-stage chronic liver disease and 3 with acute liver failure) were enrolled. Catheters were placed in a femoral artery and in the internal jugular vein for calculation of the cerebral arteriovenous oxygen content difference (AVDO2). CBF was recorded by the 133Xenon injection technique, and by transcranial Doppler sonography determined mean flow velocity (Vmean) in the middle cerebral artery. The CMRO2 was calculated as the AVDO2 times CBF and the cerebrovascular resistance (CVR) as the mean arterial pressure to CBF ratio. An index of large cerebral artery diameter was expressed by the CBF to Vmean ratio.RESULTS:From induction of anesthesia to the anhepatic period, CBF decreased from a median of 47 (interquartiles 31-55) to 41 (37-48) ml 100 g(-1) min(-1), whereas the CMRO2 remained unchanged (1.3 [0.9-2.5] vs. 1.7 [0.9-2.3] ml 100 g(-1) min(-1)). In the reperfusion phase, the CBF increased to 51 (45-54) ml 100 g(-1) min(-1), whereas the CMRO2 remained unchanged at 1.1 (1.0-2.5) ml 100 g(-1) min(-1). The CVR decreased from 2.0 mm Hg (1.4-2.1) to 1.4 (1.1-1.8) mm Hg(-1) min 100 g ml. In the anhepatic phase, mean arterial pressure decreased from 92 mm Hg (84-98) to 85 (80-92) mm Hg and at reperfusion it was 80 (71-105) mm Hg. From the anhepatic to the reperfusion phase, the CBF increased 7% (0 to 26) for each mm Hg concomitant increase in PaCO2. The CBF to Vmean ratio remained stable (1.0 [0.8-1.2] vs. 0.9 [0.7-1.1] ml 100 g(-1) min(-1) cm(-1) sec).CONCLUSION:During the reperfusion phase of liver transplantations, cerebrovascular dilatation uncouples cerebral oxidative metabolism from blood flow. The increase in CBF is beyond what can be explained by changes in arterial carbon dioxide tension and arterial pressure.

Objectives, To estimate the incidence of bleeding leading to death or hospital admission in out-patients treated with oral anticoagulants.Design, Population-based historical cohort study 1 January 1992 to 31 September 1994.Setting, The County of North Jutland, Denmark (488 000 inhabitants).Subjects, Six hundred and eighty-two consecutive patients commencing oral anticoagulant therapy.Main outcome measures. Fatal bleeding or bleeding demanding hospital admission.Results. In 756 treatment-years of follow-up, there were 45 major haemorrhagic events (6.0 per 100 treatment-years) in 42 patients, of which seven (0.9 per 100 treatment-years) were fatal. The risk of a first major haemorrhagic episode was highest during the first 90 days of treatment compared with duration above one year (incidence rate ratio, IRR, 1.9; 95% CI, 0.8-4.1). The rate was highest above the age of 60 years, 6.8 per 100 treatment-years, compared with 2.9 per 100 treatment-years below 60 years (IRR 2.3; 95% CI, 1.0-5.6). The rate for a bleeding event was slightly higher in females than in males (IRR 1.3; 95% CI, 0.7-2.3), but did not vary according to type of anticoagulant drug.Conclusions. The reported rates of major bleeding in this routine community setting implied a higher bleeding risk than was found in randomized trials or when patients are monitored in specialist anticoagulation clinics.

Objective: To evaluate the haemodynamic changes during treatment with high-volume plasmapheresis in patients with chronic liver failure compared to patients with acute liver failure.Methods: Haemodynamic measurements were performed with a Swan-Ganz catheter and thermodilution technique. High-volume plasmapheresis (mean plasma exchange of 8.6 litres) was performed in 11 patients with chronic and 16 patients with acute liver failure.Results: In patients with chronic liver failure, systemic vascular resistance index was unaltered: 1193 +/- 494 dyn s cm(-5) m(2) before treatment versus 1180 +/- 399 dyn s cm(-5) m(2) after. Mean arterial pressure increased from 69 +/- 11 mmHg to 78 +/- 13 mmHg (P < 0.05) and cardiac output increased from 8.1 +/- 2.4 l/min to 8.9 +/- 2.4 l/min (P < 0.05) during high-volume plasmapheresis. In patients with acute liver failure, systemic vascular resistance index increased from 1154 +/- 628 dyn s cm(-5) m(2) to 1614 +/- 738 dyn s cm(-5) m(2) (P < 0.001). In this group mean arterial pressure increased from 78 +/- 6 mmHg to 95 +/- 10 mmHg (P < 0.001) and cardiac output decreased from 9.6 +/- 3.7 l/min to 8.2 +/- 2.9 l/min (P < 0.01).Conclusion: The hyperkinetic circulation in chronic and acute patients was differently affected by high-volume plasmapheresis. We suggest that in chronic liver failure both portosystemic shunting and chronic peripheral vasodilation may contribute to the hyperkinetic syndrome, whereas in acute liver failure a humoral factor which can be removed by high-volume plasmapheresis is a main contributor.

Inferior vena cava (IVC) clamping during liver transplantation causes venous congestion in the splanchnic and IVC beds. A venovenous bypass relieves congestion and improves cardiac output (CO), but the bypass flow required for adequate drainage of the vascular beds is controversial. In this study we evaluated the bypass flow necessary to compensate for the IVC clamping. Lower body impedance (BI) is inversely related to tissue fluid content and was used to reflect congestion. A venovenous bypass was successfully applied to 59 of 62 patients. BI was measured across the left buttock and related to bypass flow, CO, bypass flow ratio (bypass flow/CO before IVC clamping; n = 62), and right femoral venous pressure (n = 8). The bypass flow was 1.7 (0.0-3.0) L.min-1 (median and range). BI decreased (delta BI; -2.2 [-10.3-1.1)] omega) as the femoral venous pressure increased (29 [21-49] mm Hg; r = -0.81; P < .05), and the femoral venous pressure correlated inversely to bypass flow (r = -0.35; P < .01). The change in CO at IVC clamping (delta CO; -2.3 [-6.3-1.6] L.min-1) related to bypass flow ratio (0.25 [0-0.51]; r = 0.57, P < .01), whereas delta BI related only minimally to bypass flow or bypass flow ratio (r = 0.37; P < .05). In conclusion the median bypass flow of 1.7 L.min-1 was too small to prevent fluid accumulation in the lower caval region, and extrapolation of data suggests that bypass flow should have approached 3.5 L.min-1 or 50% of CO in order to prevent fluid accumulation in the lower caval region. However the minimal correlation between lower BI and bypass flow indicates that bypass flow per se is not the only determinant of lower body fluid accumulation.

Under normal conditions cerebral blood flow (CBF) is regulated to secure oxidative brain metabolism, but in patients with fulminant hepatic failure (FHF), insufficient CBF has been suggested to precede cerebral edema and intracranial hypertension. In order to determine if insufficient CBF and hypoxia are present in patients with FHF we increased the mean arterial pressure and measured cerebral metabolism. In six patients with FHF CBF determined by 133Xenon injection technique, transcranial Doppler mean flow velocity in the middle cerebral artery (Vmean) and cerebral metabolism were determined, before and after an increase in mean arterial pressure by norepinephrine infusion. Mean arterial pressure was measured in a radial artery, and blood samples from the radial artery and internal jugular vein allowed calculation of the cerebral arteriovenous oxygen (AVDO2), -glucose (AVDgl), and -lactate (AVDlac) differences. Cerebral metabolic rates (CMRO2,-gl,-lac) were calculated as AVDO2,-gl,-lac times CBF. Mean arterial pressure was raised from 70 (54-105) to 111 (93-128) mm Hg during intravenous infusion of norepinephrine. CBF increased from 34 (12-55) to 47 (27-81) mL . 100g-1. min-1 (p < 0.05) and Vmean from 53 (42-60) to 67 (61-79) cm.s-1 (p < 0.05), whereas CMRO2 (1.4 (0.9-2.4) mL . 100g-1 . min-1), CMRgl (11 (4.8-20) mumol 100g-1 . min-1), and CMRlac (3.2 (0-8.9) mumol . 100g-1 . min-1) remained unchanged. Our finding indicates that cerebral oxidative metabolism is preserved in patients with FHF. Cerebral autoregulation is absent, however, and neuroprotective critical care is suggested to be guided by internal jugular vein oxygen saturation to secure appropriate cerebral oxygenation.

This study was designed to investigate changes in the immune system of elite swimmers compared with well‐conditioned age‐ and sex‐matched controls in relation to a competition swim (field study). Furthermore, the aim was to reveal possible differences in immune system changes depending on the type of sport performed by comparing with an earlier study of similar design, from the same laboratory that tested elite runners in relation to a competition run. The swimmers were tested before, immediately after and 2 h and 24 h after a competition swim. Lymphocyte subsets (CD5, CD3, HLA‐DR, CD4, CD8, CD19, CD3/CD16+56, CD57, CD18, CD16/CD122) all increased after the run, decreased to normal or subnormal levels after 2h, and returned to normal after 24 h (absolute numbers). The findings were identical for the swimmers and the age‐ and sex‐matched control group. No change in polymorphonuclear granulocyte migration was found. The lymphocyte proliferative responses decreased 2 h after the exercise. No changes were seen in plasma cytokine levels (interleukin‐1β (IL‐1β), interleukin‐6 (IL‐6), and tumor necrosis factor α (TNF‐α) in relation to exercise, but significantly lower baseline values for IL‐6 were observed in the swimmers. An increase in total natural killer cell activity immediately after exercise, followed after 2 h by a decrease, was seen in both swimmers and controls. Finally, no complement activation was detected. Compared with an earlier study of elite runners, differences were seen in granulocyte chemotactic response, TNF‐α plasma activity and the lymphocyte proliferative response to mitogen. These differences might be explained by the degree of immune system activation following muscle damage during exercise, inducing an increase in cytokines, which are known to activate and modulate both lymphocytes and granulocyte function. Our findings demonstrate identical exercise‐induced, immune system changes in elite swimmers and well conditioned controls, and furthermore, the findings suggest that different types of sport performed at maximum intensity induce different immune system changes.

BACKGROUND/AIMS:Normally, cerebral blood flow responds to changes in the arterial carbon dioxide tension (PaCO2) but not to changes in mean arterial pressure, commonly referred to as the cerebral CO2-reactivity and autoregulation. In patients with fulminant hepatic failure and in the rat with thioacetamide-induced liver failure, autoregulation is absent, presumably due to cerebral vasoparalysis. Since also CO2-reactivity may then be compromised, it was studied in patients with fulminant hepatic failure and rats with thioacetamide-induced liver failure. METHODS:In ten patients (median age 32 (range 20-48) years)) and in ten age-matched volunteers, cerebral perfusion was elevated by transcranial Doppler assessed mean flow velocity (V(mean)) in the middle cerebral artery during hypo- and hyper-capnia. In six rats with liver failure and in six control rats, cerebral blood flow was measured repeatedly by the intracarotid 133 Xenon injection technique. RESULTS:In the patients and volunteers, PaCO2 was lowered from 33 (23-44) to 28 (23-39) mmHg by hypocapnia and raised to 40 (34-48) mmHg by hypercapnia or 5% CO2 inhalation. During hypocapnia, the CO2-reactivity did not differ significantly between patients and volunteers, 4.0 (1.1-7.4) vs. 3.0 (1.7-5.0)% mmHg(-1), while it was reduced during hypercapnia in the patients, 2.2 (1.8-5.2) vs. 4.6 (3.0-8.0)% mmHg(-1) (p < 0.05). In the rats, PaCO2 was reduced from 39 (37-40) to 30 (29-31) mmHg and then raised to 51 (41-55) mmHg. During hypocapnia, CO2-reactivity was similar in rats with liver failure and in control rats, 2.3 vs 2.7% mmhg(21), respectively. In all rats with liver failure CO2-reactivity was abolished during hypercapnia, while it was 1.5% mmHg(-1) in the control rats (p < 0.01). CONCLUSIONS:The finding that cerebral CO2 reactivity is reduced in hypercapnia, while it is preserved in hypocapnia, suggests that gradual dilation of the cerebral resistance vessels develops in fulminant hepatic failure and connects previous morphological studies with changes in the regulation of cerebral blood flow, i.e. impaired cerebral autoregulation and blunted CO2-reactivity.

Objective: The effect of high-volume plasmapheresis on hepatic encephalopathy, cerebral blood flow (CBF) and cerebral metabolic rate for oxygen (CMRO2) was investigated in patients with fulminant hepatic failure (FHF). Methods: Twelve consecutive patients (8 women, 4 men, median age 34 years (range 19–51)) were studied before and after high-volume plasmapheresis with 10–16 litres fresh frozen plasma, while PaCO2 and body temperature were maintained at 30 (23–34) mmHg and 37.6°C (36.6–38.4), respectively. Blood samples from the internal jugular vein and a radial artery allowed calculation of the cerebral arteriovenous oxygen difference (AVDO2) and oxygen extraction (AVDO2 divided by arterial oxygen content). CBF was determined by a xenon-133 clearance method in eight patients and CMRO2 calculated as AVDO2 times CBF. Cerebral perfusion pressure (CPP) was determined as the difference between mean arterial and subdural pressures in eight patients. Results: High-volume plasmapheresis was initiated 22 (6–168) h after the development of hepatic encephalopathy and 11 patients had grade 4 encephalopathy. Following high-volume plasmapheresis the grade of encephalopathy improved in four patients. The CBF increased from a median of 31 (16–86) to 45 (18–97) ml/100 g/min and as oxygen extraction remained unchanged (32 (9–41) vs. 29 (7–39)%), CMRO2 increased from 1.24 (0.96–1.82) to 1.86 (1.00–2.07) ml/100 g/min (P<0.05). The CPP increased from 62 (19–76) to 92 (50–105) mmHg (P<0.01), whereas the intracranial pressure remained unchanged (19 (3–45) vs. 11 (5–33) mmHg). No statistical difference was found between the relative changes in the above parameters in survivors compared to non-survivors. Conclusion: Although the clinical status did not improve in all patients, both CBF and CMRO2 increased after high-volume plasmapheresis. The alleviation of brain oxygen metabolism by high-volume plasmapheresis may reflect partial removal of neuroinhibitory plasma factors.

Patients with severe liver disease and accompanying malnutrition may exhibit electrolyte disturbances including the magnesium balance. In 18 patients plasma magnesium (p-Mg) was determined at the start of the liver transplantation and during the anhepatic and reperfusion phases of the operation. The blood loss was 6.9 (2.5-8.8) 1 (median and range) and the cumulative transfusion volume was 10.2 (5.0-17.2) 1 of which 5.9 (2.5-14.2) 1 was with fresh frozen plasma. p-Mg was 0.72 (0.58-0.88) mmol.l-1 and it did not change significantly during the operation. Thus, in 4 patients it was at or below the lower reference value of 0.67 mmol.l-1. In 11 patients it changed less than 0.05 mmol.l-1, while in 4 patients the concentration was rose, and in 3 patients we noted a fall in each of 0.08 mmol.l-1. There was no correlation between p-Mg and the blood loss or the administered volume of fresh frozen plasma. In 10 randomly chosen fresh frozen plasma units, the p-Mg was 0.64 (0.61-0.71) mmol.l-1. These observations do not support a need for close monitoring or substitution of magnesium during human liver transplantation. On the other hand, the finding of a low value in 4 of 18 patients suggests that plasma magnesium should be monitored and eventually corrected while the patient is on the waiting list.

We conducted a randomized, double-blind trial to evaluate the early and late analgesic effect of preoperative wound infiltration with bupivacaine 0.25% (40 mL) compared to placebo (NaCl 0.9%, 40 mL) in patients undergoing major surgery. Forty-one patients scheduled for elective hysterectomy during general anesthesia were included. The pain management focused on pain prevention, including preoperative administration of nonsteroidal antiinflammatory drugs (NSAIDs), and peroperative administration of opioids. Postoperatively patients received buprenorphine and/or acetaminophen on demand. A significant difference between treatments was evident in the 3-day postoperative trial period. With identical pain scores in the two groups, the requested total amount of buprenorphine was greater in the placebo group (2.0 [0-5.1] mg) (median and [range]) than in the bupivacaine group (0.8 [0-2.8] mg) (P < 0.05). The demand for analgesics occurred earlier in those who received placebo (225 min) than in those who received bupivacaine (345 min), but did not reach the level of significance. In conclusion, preoperative wound infiltration with bupivacaine improved immediate and late postoperative pain management after hysterectomy compared to placebo.

In management of patients with fulminant hepatic failure, it is recommended that mean arterial pressure should be raised if cerebral perfusion pressure is lower than 50 mmHg, but the influence of such therapy on cerebral blood flow is unknown. We examined cerebral blood flow autoregulation in seven consecutive patients with fulminant hepatic failure during treatment of imminent insufficient cerebral perfusion pressure. Cerebral perfusion was evaluated by transcranial Doppler assessed mean flow velocity in the middle cerebral artery and by the arterio-venous difference for oxygen. Intracranial pressure was recorded by a subdural transducer and cerebral perfusion pressure calculated as the different between mean arterial pressure and intracranial pressure. After 20 (range 10 to 43) min, mena arterial pressure was raised from 74 (43–80) to 94 (76–114) mmHg by i.v. noradrenaline, cerebral perfusion pressure increased from 49 (26–75) to 82 (50–108) mmHg (p<0.01) as the intracranial pressure remained unchanged at 26 (3–35) mmHg. The mean flow velocity increased from 68 (30–134) to 108 (48–168) cm s−1 and the arterio-venous difference from oxygen by 46 (10–82) % (p<0.05). Both mean flow velocity (r=0.63) and arterio-venous difference for oxygen (r=0.71) were correlated to mean arterial pressure (p<0.001), and a lower blood pressure limit of autoregulation could not be identified in any of the patients. These data suggest that the cerebral blood flow is not autoregulated in patients with fulminant hepatic failure and therefore cerebral blood flow should be “clamped” within the normal physiologic range by manipulation of arterial blood pressure in order to avoid cerebral hypoxia and/or hypertensive induced cerebral oedema.

The mortality of critically ill patients who develop acute renal failure (ARF) is persistingly high. We reviewed all patients who developed ARF that required dialysis in a single intensive care unit (n = 167) during the period 1977 to 1989, in order to identify variables with possible influence on outcome. Overall mortality within hospital was 75%. Age above 60 (p < 0.02), requirement for mechanical ventilation (p < 0.0005), requirement for inotropic drugs (p < 0.0005) and increased levels of P-bilirubin (p < 0.005) had negative impacts on survival. Mortality increased significantly from 63% in the early period (1977-1985) to 84% in the late period (1986-1989) (p < 0.001), a rise that could not be explained by a single variable. The patients in the later period were, though, characterized by a higher morbidity with a higher frequency of additional organ failure.

We studied the utilization of oral anticoagulants and the degree of completeness and validity of some key research variables from the Danish health service in the county of North Jutland (482 000 inhabitants), studying the records of 12 855 prescriptions of oral anticoagulants during 1991 and 1992. The utilization of oral anticoagulants showed an increase of 16% from 1991 to 1992, when 0.3% of all women and 0.4% of all men were treated during one year. The increase was less than expected based on newly introduced indications for treatment of atrial fibrillation. The following variables were studied: the name (ATC commodity number) and amount of the prescribed drug, the defined daily dose, the personal registration number of the patient, type of prescription (e.g. in writing or by telephone), number of repeat prescriptions, date of prescription issued by the doctor, the registration number of the GP or hospital department (if the prescription is made by a hospital doctor) and date of expedition from the pharmacy. The number of wrong and missing data was less than 1%. The study suggests that Danish prescription information is of such quality that it can be used for pharmacoepidemiological studies.

Before induction of anaesthesia, 23 patients scheduled for major abdominal surgery had blood samples drawn from a central venous catheter for oxygen saturation (SvO 2 ) after graded infusion of isotonic saline. The infusion of saline was continued until further administration resulted in a stable SvO 2 (SvO 2 max). The SvO 2 increased from 69 (53–83) to 72 (66–83) % (median and range; P < 0. 0001), when the patients received 10 (0–26) mL‐ kg ‐1 , average 500 mL, of saline. At the same time central venous haematocrit decreased from 38 (32–47) to 36 (23–47) % which suggests that the intravascular volume was expanded by 420 (180–3070) mL or by 72 (18–174) % of the administered volume. The results demonstrate that volume expansion by saline can establish a maximal venous oxygen saturation in the surgical patient.

The use of strong analgesics was continuously registered in 90 patients throughout 12 months via the public health authorities, who are responsible for the control of prescription of strong analgesics. After the 12 months, questionnaires were sent to the prescribing doctors about the treatment during that period. Twenty-five patients were excluded mainly due to incomplete data and non-responding GPs. Analysis of validity of the GPs' registration, with the public health authorities' registration used as the reference standard, showed 90% (95% confidence limits (CL):68-99%) agreement concerning continuing treatment with strong analgesics and 98% (CL:88-100%) regarding GPs' registration of patients not being treated with strong analgesics. In all, misclassification occurred in 5% (CL:1-13%) of the patients. Our study suggests that the GPs' information about prescriptions of strong analgesics is valid, and that it can be used in research.