ATROPINE REVERSAL OF HYPERCARBIA DURING ENFLURANE ANAESTHESIA

Atropine was given i.v. to 35 spontaneously breathing healthyadults anaesthetized with enflurane to determine if it wouldreduce Paco₂ by improving gas exchange. Five minutes after administrationof atropine, Paco₂, decreased from mean 7.18 to 6.65 kPa ( —7%),while mean E increased from 4.8 to 6.1 litre min–¹ ( +27%). These changes were maintained 15 min later in the 20 patientsstudied at that time. Older and more obese patients showed amore significant change. Respiratory frequency and Pao₂ didnot alter.     

PULMONARY MECHANICS DURING GENERAL ANAESTHESIA: The influence of mechanical irritation on the airway

The changes in pulmonary resistance (R_L) and compliance (C_L),following airway irritation and surgical stimulation, were investigatedin 19 anaesthetized, paralysed patients. Thirteen were normal,while six had objective evidence of chronic obstructive pulmonarydisease (COPD). Broncho—carinal irritation with a suctioncatheter produced a 27% increase in R_L (0.58±0.32 kPalitre^–1 s to 0.74±0.40, P < 0.01) and a 10%decrease in C_L (0.87±0.19 litre kPa^–1 to 0.81±0.22,P < 0.01). These changes were associated with a significantincrease in systolic arterial pressure and heart rate duringclinical stages of anaesthesia (end-tidal enflurane, 1.3±0.4%,Pa_CO2 5.20±0.53 kPa). These changes did not correlatewith depth of anaesthesia, but all returned to the pre-irritationvalue within 5 min. There was no difference in subjects considerednormal and those with COPD. Surgical stimulation did not producesignificant changes in pulmonary mechanics; however, repeatedbroncho—carinal irritation during surgery was associatedwith a transient increase in R_L (P < 0.01). ^*Present address: Department of Anesthesiology, Sapporo MedicalCollege and Hospital, Sapporo, Japan.     

EFFICIENCY OF TWO VARIABLE PERFORMANCE TECHNIQUES OF OXYGEN THERAPY IN RELIEVING POSTOPERATIVE HYPOXAEMIA

The efficiency of two patient-dependent, variable performancetechniques of oxygen therapy in relieving hypoxaemia after upperabdornir,al surgery was compared. A high-flow system delivereda humidified mixture of oxygen 2 litre min^–1 and air 13litre min^–1 through a cannula inserted into the anteriornares. The low-flow system delivered 2 litre min^–1 ofdry oxygen into the nasopharynx through a catheter. Thirteenotherwise healthy patients received either high-flow oxygentherapy for 30 mm followed by low-flow oxygen therapy for afurther 30 mm after operation or the same therapy in reverseorder. With the patients breathing room air, arterial hypoxaemiacould be demonstrated with Pao₂ inversely related to age. Theincrease of Pao₂ during either oxygen treatment was significantlygreater with the low flow system.     

Effect of low-dose enflurane on the ventilatory response to hypoxia in humans

To investigate the effects of enflurane on the control of breathingwe have studied the ventilatory responses to isocapnic hypoxiain 12 adults with and without sedation with enflurane. Design1 consisted of three steps into hypoxia (PE'_O2=6.7 kPa), eachlasting 3 min, separated by periods of euoxia lasting 5 min(PE'_O2=13.3 kPa). Design 1 was repeated four times in each subjecton the same day in random order: with carrier gas (control)and with 0.04 MAC, 0.07 MAC and 0.13 MAC of endetidal enfluraneconcentrations. Design 2 consisted of 20-min exposures to hypoxiawith and without 0.07 MAC of enflurane. Each exposure was precededand followed by 5 min of euoxia. End-tidal PC0₂ was held constantat 0.13–0.27 kPa greater than the resting level throughoutboth designs. Mean (SEM) ventilatory responses to hypoxia fordesign 1 were: 8.2 (1.3) litre min^–1 (control), 6.6 (1.4)litre min^–1 (0.04 MAC), 5.7 (1.1) litre min^–1 (0.07MAC) and 3.7 (0.5) litre min^–1 (0.13 MAC) (P<0.001).For design 2, enflurane produced a 15% reduction in restingventilation (P<0.001), a 40% decrease in the acute ventilatoryresponse to hypoxia (P<0.001) and a 32% reduction in ventilatorydecline(ns) which occurred during sustained hypoxia. (Br. J.Anaesth. 1994; 72: 509–514)     

ARTERIAL OXYGENATION AFTER MAJOR HEAD AND NECK SURGERY INVOLVING TRACHEOTOMY

Changes in vital capacity, and in arterial Po₂ and Pco₂ werestudies after major head and neck surgery involving tracheotomyand after surgery on the upper abdomen. After major head andneck surgery, changes in vital capacity were comparable to thoseseen after upper abdominal operations, with a decrease on day1 of 39%; recovering to 78% of the preoperative volume by day7. Arterial PO₂ in both groups decreased on day 1 but whereasPao₂ after upper abdominal surgery was recovering by day 7,Pao₂ in the head and neck group, at day 7, showed no sign ofrecovery. Arterial Po₂ 7 days after major head and neck surgerywas 2.3 (±0.4) kPa below the preoperative value (P <0.001). A preliminary report of this work was presented to the AnaestheticResearch Society in London, November 1985.     

DOMICILIARY OXYGEN CONCENTRATORS IN ANAESTHESIA: PREOXYGENATION TECHNIQUES AND INSPIRED OXYGEN CONCENTRATIONS

A conventional preoxygenation technique was compared in twovolunteers with two techniques based on a drawover breathingsystem supplied with oxygen enriched air from a Devilbiss MiniVO2 domiciliary oxygen concentrator. Breathing from a 20-litrebag, filled previously at 2 litre min^–1 from the concentratorwas as effective as the conventional technique (Pa_O2 73.2 kPa).Supplying 4 litre min_–1 to a drawover system was lesseffective (Pa_O2 35.5 kPa). We present a simple method for determiningthe inspired oxygen concentration obtained under conditionsof drawover anaesthesia using an oxygen concentrator. ^*Present address, for correspondence: Department of Anaesthesia,Bristol Royal Infirmary, Bristol BS2 8HW.     

HYPOXAEMIA AND ATELECTASIS PRODUCED BY FORCED EXPIRATION

Ventilation at maximal voluntary reduction of lung volume causedsignificant desaturation in some healthy subjects breathingair. Saturation rapidly returned to control levels when normallung volume was regained. These changes are probably due toreversible airway obstruction. During the inhalation of oxygen,ventilation at maximal voluntary reduction of lung volume caused,in one subject, a reduction of arterial Po₂ of 243 mm Hg. Normalarterial Po₂ was not immediately restored on regaining normallung volume. Chest radiographs showed extensive atelectasiswhich persisted for several hours in an ambulant subject. Thesechanges are probably due to absorption of oxygen from alveolibeyond obstructed airways. Reduction of lung volume may be harmfulfor patients who are breathing oxygen. Caution is thereforenecessary in the use of a sub atmospheric pressure phase duringartificial ventilation and during suction of the tracheobmchialtree. ^* Present address: Department of Anaesthesia, University ofLeeds, Leeds 2.     

Does the optimization of cardiac output by fluid loading increase splanchnic blood flow?

We studied the effects of increasing cardiac output by fluidloading on splanchnic blood flow in patients with haemodynamicallystabilized septic shock. Eight patients (five female, 39–86yr) were assessed using a transpulmonary thermo-dye-dilutiontechnique for the measurement of cardiac index (CI) intrathoracicblood volume (ITBV) as a marker of cardiac preload and totalblood volume (TBV). Splanchnic blood flow was measured by thesteady state indocyanine-green technique using a hepatic venouscatheter. Gastric mucosal blood flow was estimated by regionalcarbon dioxide tension (PR_CO2). Hydroxyethyl starch was infusedto increase cardiac output while mean arterial pressure waskept constant. In parallel, mean norepinephrine dosage couldbe reduced from 0.59 to 0.33 µg kg^–1 min^–1.Mean (SD) TBV index increased from 2549 (365) to 3125 (447)ml m^–2, as did ITBV index from 888 (167) to 1075 (266)ml m^–2 and CI from 3.6 (1.0) to 4.6 (1.0) litre min^–1m^–2. Despite marked individual differences, splanchnicblood flow did not change significantly neither absolutely (from1.09 (0.96) to 1.19 (0.91) litre min^–1 m^–2) norfractionally as part of CI (from 28.4 (19.5) to 24.9 (16.3)%).Gastric mucosal PR_CO2 increased from 7.7 (2.6) to 8.3 (3.1)kPa. The PCO₂-gap, the difference between regional and end-tidalPCO₂, increased slightly from 3.2 (2.7) to 3.4 (3.1) kPa. Thus,an increase in cardiac output as a result of fluid loading isnot necessarily associated with an increase in splanchnic bloodflow in patients with stabilized septic shock. Br J Anaesth 2001; 86: 657–62     

EFFECT OF HALOTHANE ON HYPOXIC AND HYPERCAPNIC VENTILATORY RESPONSES OF GOATS

We have measured the ventilatory responses to increased inspiredcarbon dioxide and to hypoxia in four goats awake and at 0.5%,1.0% and 1.25% end-tidal halothane concentration. While maintainingPE'_CO2 constant at each of three values (means 5.86, 6.45 and7.2 kPa), PE'_O2 was reduced rapidly from more than 25 kPa to5.3–6 kPa for 3 min to record the increase in ventilation.Eleven sets of these 24 steady state points were obtained (2PO₂ x 3 P_CO2 x 4 anaes. = 24). The mean isocapnic hypoxic ventilatoryresponse (HVR) was 6.52 (SD 2.58) litre min^–1 (n = 33)when awake, 5.62 (3.48) litre min^–1 at 0.5% end-tidalhalothane (ns), 3.05 (2.02) litre min^–1 at 1 % and 2.91(2.12) litre min^–1 at 1.25%, the last two being reducedsignificantly from awake and 0.5 % halothane (P < 0.05).With 1.25 % halothane, HVR was reduced to 44.5 (18.6)% of theawake HVR. However, when HVR was expressed as % increase inventilation produced by isocapnic hypoxia, it was 71 (19) %awake but 124 (65) % with 1.25% halothane, a significant increasewith halothane (P < 0.05). With 1.25% halothane, the carbondioxide response slope decreased to 36.4 (26.4) % of control;hypoxia did not increase the slope significantly. Whereas previousstudies in man have shown that halothane preferentially depresseshypoxic chemosensitivity and has a significant effect at 0.1MAC, in the goat the hypoxic and carbon dioxide chemosensitivitieswere depressed equally. At 0.5% end-tidal concentration (about0.5 MAC), halothane did not significantly depress hypoxic response.     

CHANGES IN FUNCTIONAL RESIDUAL CAPACITY DURING CARDIAC SURGERY

A gas washout technique was used to measure the functional residualcapacity (FRC) in eight patients during anaesthesia for cardiacsurgery. The patients were anaesthetized with droperidol, fentanyland nitrous oxide, alcuronium was given and the lungs were ventilatedwith a volume controlled ventilator. FRC was measured at threestages: before skin incision, after sternotomy but before cardiopulmonarybypass, and after closure of the sternum. The pleural cavitieswere intact in all patients during the operation. FRC beforeskin incision was 1.7±0.5 litre (mean±1 SD). A55% mean increase in volume was noted after sternotomy and placementof the sternal retractor (P<0.001). Mean FRC after sternalclosure was 16% lower than the preincision value (P<0.05).Arterial Po₂ was measured in 22 other patients who underwentcoronary artery bypass surgery and in whom F1_o2 was 0.5. Pa_o2increased significantly when the sternum was opened, but decreasedafter cardiopulmonary bypass. There was a further significantdecrease on closure of the sternum.     

DEADSPACE AND THE SINGLE BREATH TEST FOR CARBON DIOXIDE DURING ANAESTHESIA AND ARTIFICIAL VENTILATION: Effects of tidal volume and frequency of respiration

Using the single breath test for carbon dioxide (SBT-CO₂) thecomponent of physiological deadspace were investigated duringanaesthesia with IPPV in 58 patients. A square-wave inspiratoryflow and an end-inspiratory pause (25% and 10% of cycle time,respectively) were used. At tidal volumes of 0.45 litre (f =17 b.p.m.),and 0.75 litre (f = 9 b.p.m.), median values forVD_phys/VT were 0.44 and 0.31. Increasing VT and decreasing fdid not change airway deadspace (VD^RW) so that the fractionVD^RW/VT was decreased (P<0.001). The alveolar deadspace fraction,VD^alv/VT^alv, was decreased in 93% of patients (P<0.001).These improvements with increasing VT can be attributed to beneficialeffects on gas distribution and diffusion time. Patients withlarge alveolar deadspaces had steeply sloping SBT-CO₂ phaseIII, and increased expiratory time constants of the respiratorysystem. The median arterial—end-tidal PCO₂ difference,(PaCO₂ – PE'CO₂), was 0.6 kPa at small and 0.3 kPa atlarge tidal volumes (P<0.001). Three patients had zero andfour had negative (PaCO₂ - PE'CO₂) values at large tidal volumes.When phase III slopes steeply, negative (PaCO₂ – PE'CO₂)values may be observed in the presence of alveolar deadspace.     

Simultaneous measurement of cardiac output by thermodilution, thoracic electrical bioimpedance and Doppler ultrasound

To evaluate the accuracy of two non-invasive techniques forcardiac output (CO) measurement, we have measured CO simultaneouslyby thoracic electrical bioimpedance (TEB), pulsed Doppler ultrasound(DU) and standard thermodilution methods (TD) under differentclinical conditions. Measurements were made in 10 patients:(I) during steady state anaesthesia with controlled IPPV ventilation(n = 131), spread over the entire ventilatory cycle; (II) duringapnoea (n = 56); (III) during spontaneous breathing (n = 152)in the intensive care unit. Mean (SD) cardiac output valueswere: (I) CO_TD 3.5 (1.0) litre min^–1, CO_TEB 3.4 (0.7)litre min^–1 CO_DU 2.8 (0.7) litre min^–1; (II) CO_TD3.6 (0.6) litre min^–1, CO_TEB 3.5 (0.4) litre min^–1,CO_DU 2.9 (0.7) litre min^–1; (III) CO_TD 7.7 (1.5) litremin^–1, CO_TEB 7.6 (1.9) litre min^–1, CO_DU 5.2 (1.4)litre min^–1. The mean percentage deviation of TEB fromTD ranged from –2.2% to 1.4% and that of DU from TD wasfrom –16% to –32%. There were no statistically significantdifferences between TD and TEB, but TD and DU differed significantlyduring IPPV, apnoea and spontaneous ventilation (P < 0.0001).(Br. J. Anaesth. 1994; 72:133–138) ^*Department of Anaesthesiology, Caritas Krankenhaus, Werkstr.1, 66763 Dillingen/Saar, Germany      

CEREBROVASCULAR CARBON DIOXIDE REACTIVITY DURING EXPOSURE TO EQUIPOTENT ISOFLURANE AND ISOFLURANE IN NITROUS OXIDE ANAESTHESIA

We have studied the effects of hypocapnia on cerebrovascularchanges in two MAC-equivalent anaesthetic regimens, using thetranscranial Doppler technique as an index of cerebral bloodflow (CBF) in 24healthy ASA I patients undergoing spinal surgery.Eight of the patients were subjected to carbon dioxide reactivitychallenges in the awake state. Before surgery, the other 16patients received, in random order, either 1.15% isofluranein oxygen or 0.5% isoflurane with 70% nitrous oxide. Carbondioxide reactivity was calculated for each group as the increasein flow velocity per kPa change in PÉ_CO2 (cm s^–1kPa^–1). It was significantly greater for the isofluranegroup (14.09 (SD 2.44) cm s^–1 kPa^–1) and significantlyless for the isoflurane—nitrous oxide group (7.95 (1.32)cm s-^–1 kPa^–1) compared with the awake group (11.24(0.95) cm s^–1 kPa^–1). We conclude that cerebrovascularresponsiveness to changes in arterial carbon dioxide concentrationis influenced markedly by the anaesthetic procedure. Hyperventilationis more likely to affect CBF during isoflurane anaesthesia thanduring an MAC-equivalent isoflurane—nitrous oxide anaesthesia.     

CONTINUOUS FLOW VENTILATION WITHOUT RESPIRATORY MOVEMENT IN CAT, DOG AND HUMAN

The insufflation of oxygen at 1 litre kg^–1 min^–1via two endobronchial catheters (called continuous flow ventilation(CFV)) maintained a normal Pa_co2 and a constant Pa_o2 in anaesthetizedparalysed dogs and in five out of seven cats. In two cats witha high carbon dioxide production, CFV failed to maintain carbondioxide homeostasis since gas flows greater than 1 litre kg^–1min^–1 caused thoracic distension and a decrease in arterialpressure. In five patients, endobronchial insufflation of oxygen0.5 litre kg^–1 min^–1 caused approximately a 30%decrease in the increase in Pa_co2 compared with apnoeic oxygenation(P < 0.05) during a period of 6 min. CFV at 1 litre kg^–1min^–1 can be used for physiological measurement withoutrespiratory movement while maintaining blood-gas homeostasisin dogs and in cats with a normal carbon dioxide production.Ethical constraints have so far prevented the investigationof the effects of comparable gas flows in man. ^*Present address: The Edith Wolfson Hospital, Holon, 58000,P.O. Box 5, Israel.     

THE HUMPHREY ADE SYSTEM: EVALUATION IN PAEDIATRIC USE

The Humphrey ADE single lever parallel form breathing systemwas evaluated in 17 children whose lungs were ventilated mechanically,mean age 28 (range 9–58) months, weight 13.3 (8.3–18.0)kg and in 10 spontaneously breathing children, age 50.7 (18–99)months and weight 17.5 (10.9–24.3) kg. During controlledventilation in the E mode, the ADE behaved in a manner similarto the Jackson-Rees modification of Ayre's T-piece with thefresh gas flows (FGF) set either according to weight or at astandard 3 litre min^–1. After 5 min hand ventilation inthe A mode with FGF 3 litre min^–1 the mean end-tidal partialpressure of carbon dioxide was 5.3 (SD 0.8) kPa (range 4.1–7.7kPa). Whilst breathing spontaneously in the A mode, no patientexperienced rebreathing at FGF 3 litre min^–1; rebreathingstarted at 124 (31.6) ml kg^–1 min^–1. The ADE circuitperformed satisfactorily in all three modes. During controlledventilation in the E mode, FGF should be adjusted accordingto the standard formulae recommended for the T-piece circuit.During hand ventilation in the A mode, an FGF greater than 3litre min^–1 should be used.     

NITROUS OXIDE SEDATION CAUSES POST-HYPERVENTILATION APNOEA

We have studied, in six normal subjects, the effect of nitrousoxide sedation on the ventilatory pattern and oxygen saturationusing pulse oximetry (Sp_O2) after hyperventilation to an endtidalcarbon dioxide partial pressure (P_CO2) of 3 kPa. This valueof P_CO2 was shown to be less than the apnoeic threshold of allthese subjects when their ventilation vs P_CO2 response curveswere plotted. All subjects became apnoeic when told to relaxfollowing hyperventilation while breathing 75% nitrous oxidefor 90 s. Apnoea was defined as cessation of breathing for 20s or more. The mean duration of apnoea was 78 s (range 29–130s). All subjects demonstrated arterial desaturation (mean Sp_O275%, range 44–87%). In contrast, following hyperventilationwith air, no apnoea was seen in any subject, although therewas some evidence of desaturation (mean Sp_O2 92.5%, range 88–98%).It was concluded that subjects who are sedated with nitrousoxide behave similarly to those who are anaesthetized ratherthan to those who were fully conscious, in that they becomeapnoeic below the apnoeic threshold point. The reduction inSp_O2 after hyperventilation was explained almost entirely byapnoea and may explain abnormalities of respiratory controland hypoxaemia in patients recovering from general anaesthesiaor sedation accompanied by hypocapnia. This mechanism may beof importance in obstetric patients after breathing Entonox,when apnoea and hypoxaemia may reduce oxygen delivery to thefetus. This work was presented to the Anaesthetic Research Societyat the Nottingham Meeting in July 1990. ^*Present address: Doncaster Royal Infirmary, Doncaster.     

PREDICTIVE VALUE OF FRC AND RESPIRATORY COMPLIANCE ON PULMONARY GAS EXCHANGE INDUCED BY HIGH FREQUENCY JET VENTILATION IN HUMANS

To determine if functional residual capacity (FRC), complianceof the respiratory system (C), or underlying pulmonary diseaseare predictive for the efficacy of high frequency jet ventilation(HFJV) on pulmonary gas exchange, we investigated six adultpatients within 4 h of abdominal surgery and six patients withsevere adult respiratory distress syndrome. Gas exchange duringintermittent positive pressure ventilation (IPPV) was comparedwith that during HFJV at frequencies of 100 b.p.m. (HFJV₁₀₀)and 200 b.p.m. (HFJV₂₀₀), resulting in a minute ventilationof about 400 ml kg^–1 with both ventilatory frequencies,and in both groups of patients. Baseline FRC and C were measuredduring IPPV with the multiple-breath nitrogen washout methodand from expiratory pressure-volume curves, respectively. Changesin the alveolar-arterial oxygen difference (PA_O2–Pa_O2):FlO2 ratio induced by HFJV correlated negatively with C (HFJV₁₀₀:r = –0.78, P <0.005; HFJV200: r = –0.84, P <0.005); that is, greater oxygenation was obtained in patientswith a better compliance. Similarly, changes in arterial partialpressure of carbon dioxide (Pa_co2) induced by HFJV correlatednegatively with C (HFJV₁₀₀: r = –0.77, P < 0.001; HFJV₂₀₀:r = —0.61, P < 0.05). In contrast, there was no significantcorrelation between FRC measured during IPPV and changes in(PA_O2–Pa_O2): Fl_O2 ratio or Pa_co2 induced by HFJV, as thesechanges were influenced more by the patient's pulmonary diseasethan by baseline FRC. These results should be interpreted inthe context of different underlying pathophysiological mechanismsreducing FRC in both groups of patients.     

Effect of total flow rate on the concentration of degradation products generated by reaction between sevoflurane and soda lime

We have compared concentrations of degradation products in thecircle system during sevoflurane anaesthesia at different freshgas flows. Twenty-four patients underwent sevoflurane anaesthesiawith fresh gas flows of 1 litre min^–1 (1L group), 3 litremin^–1 (3L group), or 6 litre min^–1 (6L group) (n= 8 in each group). During anaesthesia, the concentrations ofdegradation products were measured every hour, and the temperatureof soda lime, end-tidal carbon dioxide concentration, inspiredand end-tidal sevoflurane concentrations, and carbon dioxideelimination were measured. CF₂=C(CF₃)—0—CH₂F (compoundA) was the only degradation product detected. The mean maximumconcentration of compound A was 19.7 (SD 4.3) ppm in the 1 Lgroup, 8.1 (2.7) ppm in the 3L group and 2.1 (1.0) ppm in the6L group (P > 0.05). The maximum temperature of soda limewas 44.6 (1.5) °C in the 1 L group, 37.0 (4.4) °C inthe 3L group and 29.1 (5.1) °C in the 6L group (P > 0.05).There were no significant differences in end-tidal sevofluraneconcentration or mean carbon dioxide elimination between thegroups. Only compound A was detected following anaesthesia,with higher concentrations observed at lower flow rates.     

ANALYSIS OF POSSIBLE FACTORS INFLUENCING Pao2 AND (PAo2--Pao2) IN PATIENTS AWAITING OPERATION

Data on arterial oxygen tension (Pa_o²) and alveolar—arterialPo₂ difference (PA_o²—Pa_o²) have been obtained from 337patients awaiting elective surgery. Statistical analysis ofthese data has assessed both the individual and the combinedinfluence of various factors on Pa_o² and (PA_o²—Pa_o²).The factors of importance in relation to Pa_o² include age, smokinghabits, body build and PA_o². In relation to (PA_o²—Pa_o²)the significant factors include age, Pa_CO2, weight and smokinghabits.     

Halothane affects ventilatory afterdischarge in humans

In awake humans, when ventilatory stimulation is suddenly removed,the subsequent change in minute ventilation (which remains athigher levels for longer times than expected from the dynamicsof the chemoreceptors) is termed ventilatory after discharge.In this study we investigated the effects of subanaestheticconcentrations of halothane on afterdischarge. The ventilatorypattern after sudden termination of brief periods (90–180s) of isocapnic hypoxia (PE'_co2 approximately 0.1 kPa aboveinitial resting values; PE'_o2 6.5 kPa) by normoxia (PE'_o2 14kPa) was determined in healthy volunteers. Six subjects underwent13 studies without halothane (control) and six others 10 studiesduring inhalation of 0.22% halothane. Isocapnic hypoxia causeda mean increase in ventilation of 10.8 (SD 2.4) litre min^–1in the control and 4.2 (2.4) litre min^–1 in the halothanestudies (P < 0.01). The transition to normoxia caused a slowventilatory decay in the control and a fast decay in the halothanegroups: the interval that occurred between the "last hypoxic"breath and the time required for ventilation to return to 110%of baseline was 60.7 (23) s for the control and 12.3 (6.0) sfor the halothane studies (P < 0.05). Taking into considerationthe different factors that determine the pattern of breathingimmediately after termination of a brief period of hypoxia bynormoxia (PE'_o2 waveform, transport delay time between lungsand carotid bodies, time constant of the peripheral chemoreflexloop and afterdischarge), the faster ventilatory decay observedwith halothane is probably related to suppression of afterdischarge.We conclude that afterdischarge was activated by brief periodsof isocapnic hypoxia, but not when combined with inhalationof a subanaesthetic concentration of halothane. (Br. J. Anaesth.1995; 74: 544–548)