Clinical application of forced oscillation.

This review summarizes current clinical use of the forced oscillation technique (FOT) for analysis of lung function. It presents an intuitive approach to FOT pattern recognition for interpretation of results in human subjects, and the view that FOT is now well established and, clinically, eminently useful in patients with airflow obstruction. The focus of this review is on findings that relate directly to clinical utility, with less emphasis on theoretical mechanisms. The major thrust for clinical application of FOT derives from a number of European clinical research centers. Farre and Navajas and their colleagues in Barcelona, Harf and the Lorinos and their coworkers in Paris, Peslin and Duvivier and their coworkers in Vandoeuvre-les-Nancy, Pride and coworkers in London, and Van de Woestijne, Clement, Demedts, Landser, Van Noord, and their colleagues in Leuven have essentially been responsible for clinical development of FOT over the past 25 years. Publishing space does not permit an exhaustive listing of the many contributions of these investigators, but it is intended that the present review will provide a useful infrastructure from which the reader may progress to other research citations as desired.

Breathhold cine MRI of left ventricular function in patients with obstructive sleep apnea: work-in-progress.

Obstructive sleep apnea (OSA) is a sleep-related breathing disorder that can cause left ventricular (LV) dysfunction. In patients with OSA, the LV dysfunction is usually evaluated by echocardiography. The purpose of this study was to evaluate whether the use of breathhold cine MRI for the study of LV dysfunction would be feasible and well tolerated by patients with OSA. Six volunteers and five patients underwent a breathhold cine MRI study of the LV using a 1.5 Tesla MR imager. Cine MRI was performed using a breathhold k-space segmented TurboFLASH technique during end-expiration. Systolic thickening of the LV septal wall was 49% +/- 16% in normals vs. 25% +/- 10.5% in patients (p < 0.05). Systolic thickening of the LV free wall was 42% +/- 12% in normals vs. 22% +/- 9% in patients (p < 0.05). There was a significant difference in end-diastolic wall thickness between the two groups. All patients tolerated the procedure well. The total duration of each study was relatively short (less than 11 min). Breathhold MRI techniques can be used to study LV dysfunction in patients with respiratory disability such as OSA.

Characteristics of nasal airflow and the effect of a nasal dilator in normal human subjects.

Nasal resistance contributes to negative airway pressure during breathing. We sought to define normal patterns of nasal flow and the effects of mechanical dilatation and splinting of the nares on flow during forced inspiration and expiration. Maximal inspiratory and expiratory flow volume loops (FVL) were determined in 17 normal subjects. Oral FVL were obtained with nares clamped and nasal FVL through a mask with and without dilatation of nares using a plastic splint (Nozovent). Oral FVL were normal in all. Two patterns of nasal FVL were observed: one indicating 'variable' extrathoracic obstruction, the other indicating 'fixed' extrathoracic obstruction. Maximal inspiratory flow at 50% of vital capacity (FiF50) was improved by the Nozovent only in those with a 'variable' pattern (FIF50 (L/sec): 1.54 +/- 0.3 to 2.86 +/- 0.5; P < 0.05, versus 1.92 +/- 0.3 to 2.21 +/- 0.3: P = 0.5). In subjects with a fixed pattern, failure of dilatation of the nares to increase flow suggests that the site of inspiratory flow limitation is within the bony nostril.

Efficacy of salmeterol xinafoate in the treatment of COPD.

STUDY OBJECTIVES: To examine and compare the efficacy and safety of salmeterol xinafoate, a long-acting inhaled beta2-adrenergic agonist, with inhaled ipratropium bromide and inhaled placebo in patients with COPD. DESIGN: A stratified, randomized, double-blind, double-dummy, placebo-controlled, parallel group clinical trial. SETTING: Multiple sites at clinics and university medical centers throughout the United States. PATIENTS: Four hundred eleven symptomatic patients with COPD with FEV1 < or = 65% predicted and no clinically significant concurrent disease. INTERVENTIONS: Comparison of inhaled salmeterol (42 microg twice daily), inhaled ipratropium bromide (36 microg four times a day), and inhaled placebo (2 puffs four times a day) over 12 weeks. RESULTS: Salmeterol xinafoate was significantly (p < 0.0001) better than placebo and ipratropium in improving lung function at the recommended doses over the 12-week trial. Both salmeterol and ipratropium reduced dyspnea related to activities of daily living compared with placebo; this improvement was associated with reduced use of supplemental albuterol. Analyses of time to first COPD exacerbation revealed salmeterol to be superior to placebo and ipratropium (p < 0.05). Adverse effects were similar among the three treatments. CONCLUSIONS: These collective data support the use of salmeterol as first-line bronchodilator therapy for the long-term treatment of airflow obstruction in patients with COPD.

Failure of physician documentation of sleep complaints in hospitalized patients.

Sleep disorders are acknowledged to be common but remain underrecognized by the medical community, often attributed to the failure to question patients about their sleep quality. We examined the prevalence of sleep complaints (insomnia or excessive daytime sleepiness) in a group of general medical patients by administering a questionnaire to hospitalized patients in a Veterans Affairs tertiary care medical center. A total of 222 consecutive adults (215 men, 60 +/- 14 years; body mass index, 24.8 +/- 5.6) completed the questionnaire. Of these, 105 patients (47%) had either insomnia, excessive daytime somnolence, or both; 63 (28%) had excessive daytime somnolence, which was severe in 27 (12%). Of 75 patients (34%) who had insomnia, a third were taking hypnotic medication. Forty patients (18%) had snoring, which was associated with excessive daytime somnolence in 36, whereas 46 patients (21%) had either restless legs or a combination of leg jerks and leg kicking or twitching during sleep, associated with a sleep complaint (insomnia in 32). The medical records were subsequently reviewed to assess the admitting physicians' recognition of these symptoms. No record included mention of any patient symptom related to sleep. We conclude that symptoms related to sleep, some of which may be clinically important, are common, and that none of these complaints appear to be recognized by the physicians of record.

Leptin indirectly affects estrous cycles by increasing metabolic fuel oxidation.

In previous experiments, lean Syrian hamsters fasted on days 1 and 2 of the estrous cycle failed to show sex behavior and ovulation normally expected to occur on the evening of day 4. The first goal of the present experiment was to determine whether systemic treatment with the ob (obese) protein leptin could reverse the effects of fasting on estrous cyclicity, social behaviors, and ovulation rate. Fasting-induced anestrus was reversed and normal sex and social behavior and ovulation rate were restored in hamsters injected intraperitoneally with 5 mg/kg leptin every 12 h during fasting on days 1 and 2 of the estrous cycle. A second goal was to test whether the effects of leptin could be prevented by treatment with pharmacological agents that block the oxidation of metabolic fuels. Glucose oxidation was blocked by treatment with 2-deoxy-d-glucose (2DG) and fatty acid oxidation was blocked by treatment with methyl palmoxirate (MP). 2DG (1000 mg/kg) or MP (20 mg/kg) was administered at doses that did not induce anestrus in hamsters fed ad libitum. As in the first experiment, fasting-induced anestrus was reversed by leptin treatment. However, when each injection of leptin was preceded by an injection of 2DG or MP, leptin treatment did not reverse fasting-induced anestrus. In summary, estrous cyclicity was not restored when oxidation of metabolic fuels was blocked, despite high endogenous levels of leptin. These results are consistent with the hypothesis that leptin acts indirectly on the reproductive system by increasing fuel oxidation.

Central vs. peripheral metabolic control of estrous cycles in Syrian hamsters. II. Glucoprivation.

Estrous cycles in Syrian hamsters are inhibited by food deprivation or treatment with pharmacological inhibitors of intracellular glucose utilization (glucoprivic treatments). These same metabolic challenges increase neural stimulation in areas of the caudal brain stem thought to be involved in detection of metabolic signals. Experiment 1 was designed to examine whether vagally transmitted signals are important for glucoprivic effects on estrous cycles and on neural stimulation in the caudal brain stem. Vagotomized or sham-operated hamsters were treated with 2-deoxy-D-glucose (2-DG) at a dose known to decrease cellular glucose utilization and inhibit estrous cycles (1,750 mg/kg). Vagotomized and sham-operated hamsters did not differ significantly in incidence of 2-DG-induced anestrus or in neural stimulation in the caudal brain stem, but the effects of 2-DG on estrous cycles and neural stimulation appeared to have been attenuated in vagotomized hamsters. In experiment 2, hamsters were injected intracerebroventricularly with 2-DG or glucose at doses that did not induce anestrus when injected systemically (125 and 250 mg/kg). Groups treated with intracerebroventricular injections of 2-DG showed a significantly higher incidence of anestrus than those treated with glucose. In experiment 3, effects of systemic injections of 2-DG were prevented by prior injection of glucose or fructose at the same concentration, indicating that 2-DG acts via effects on glucose metabolism, rather than via a nonspecific pharmacological effect or generalized stress response. Results of these experiments and those reported elsewhere (J. E. Schneider, A. J. Hall, and G. N. Wade. Am. J. Physiol. 272 (Regulatory Integrative Comp. Physiol, 41) R400-R405, 1997] are consistent with the notion that central glucoprivation is sufficient, whereas peripheral lipoprivation is not critical, for metabolic effects on estrous cycles.

Acute effects of intravenous cocaine on pulmonary artery pressure and cardiac index in habitual crack smokers.

BACKGROUND: Some habitual crack cocaine smokers who deny IV drug abuse show decreased pulmonary transfer of carbon monoxide (DCO). We speculated that repeated elevations in pulmonary artery pressure (PAP) might cause pulmonary capillary damage and result in a lowered DCO, or that the reduction could be due to anoxic lung injury secondary to repeated episodes of cocaine-induced pulmonary vascular constriction. STUDY OBJECTIVE: Compare the acute effects of i.v. cocaine HCl and placebo on PAP, cardiac stroke volume, and cardiac output estimated indirectly by continuous Doppler echocardiography. DESIGN: A single-blind crossover study in which placebo always preceded the active drug. SUBJECTS: Ten current crack-smoking subjects, 32 to 47 years of age, with a history of limited previous i.v. cocaine use. METHODS: PAP, cardiac stroke volume, heart rate, and BP were measured continuously after injection of placebo followed by cocaine HCl (0.5 mg/kg). RESULTS: i.v. cocaine resulted in no significant change in PAP (-0.14 +/- 3.3[SD] mm Hg, 95% confidence interval [CI] for difference -2.48, +2.21). Stroke volume index showed no significant change after cocaine (-0.1 +/- 2.0 mL; 95% CI, -1.5, +1.3). Heart rate showed a significant increase (10.0 +/- 7.2 min-1; p = 0.0017, 95% CI, +4.9, +15.1). Cardiac index showed a significant increase (0.48 +/- 0.32 L/min; p = 0.0012, 95% CI, +0.25, +0.71). Pulmonary vascular resistance showed no significant change (-44 +/- 101 dyne.s.cm-5/m2, 95% CI, -116, +29). CONCLUSIONS: i.v. cocaine HCl does not cause short-term increases in PAP or stroke volume index, but causes an increase in cardiac index due to its chronotropic effect.

Acute effects of inhaled and i.v. cocaine on airway dynamics.

BACKGROUND: Wheezing has been reported by 32% of habitual smokers of crack cocaine, and several cases of crack-related acute exacerbations of asthma have been reported. STUDY OBJECTIVE: To compare the acute effects of physiologically active doses of smoked cocaine base and, i.v. cocaine hydrochloride (HCl), a subphysiologic dose of cocaine base (smoked "placebo"), and i.v. saline solution placebo on bronchomotor tone, subjective level of intoxication, and cardiovascular responses in healthy habitual crack users. DESIGN: A single-blind crossover study in which the order of route of administration (inhaled vs i.v.) was random but placebo always preceded the active drug. SUBJECTS: Fourteen healthy, nonasthmatic current crack-smoking subjects, 34 to 48 years of age, with a history of previous i.v. cocaine use (1 to 12 times per lifetime). METHODS: Heart rate, BP, self-rated level of intoxication (scale of 0 to 10), and measurements of airway resistance (Raw) and specific airway conductance (SGaw) were recorded during separate sessions before and 3 to 5, 10, 15, and 30 min after administration of smoked cocaine base (38.5 +/- 2.3 [SEM] mg), smoked placebo (2.3 +/- 0.9 mg cocaine base), i.v. cocaine HCl (30.0 +/- 2.0 mg), and i.v placebo (saline solution). RESULTS: Both smoked active cocaine and i.v. cocaine HCl caused comparable, significant (p < 0.05) peak levels of acute intoxication (6.7 +/- 0.7 and 7.3 +/- 0.8, respectively) and increases in heart rate from baseline (29.6 +/- 2.9% and 21.4 +/- 3.7%, respectively, at 5 min). However, only smoked active cocaine caused significant decreases from baseline in SGaw (25.4 +/- 6.3% at 5 min), in contrast to nonsignificant changes after i.v. cocaine HCl (5.6 +/- 7.0% increase) and smoked placebo (10.2 +/- 6.0% decrease). CONCLUSIONS: Smoked cocaine base, but not systemically administered cocaine HCl, causes acute bronchoconstriction that is probably mediated by local airway irritation and could account for reports of crack-induced wheezing and asthma attacks in nonasthmatic and asthmatic individuals, respectively.

Monitoring the introduction of laparoscopic cholecystectomy in a district general hospital.

There has been recent concern regarding the widespread introduction of laparoscopic procedures without adequate training. We report a 2 year prospective series of 229 consecutive open and laparoscopic cholecystectomies, following the introduction of laparoscopic surgery at a district general hospital. All laparoscopic cholecystectomies were performed by surgeons without formal training in this procedure. The proportion of laparoscopic cholecystectomies rose from 27% in 1992 to 62% in 1993, with a conversion rate of 14%. An overall complication rate of 23% was recorded, 29% for open cholecystectomy and 16% for laparoscopic cholecystectomy. Procedure specific complication rates were 6% and 3% respectively. No major bile duct injuries occurred and the 30-day mortality was 0.9%. We conclude that laparoscopic cholecystectomy has been introduced as a safe procedure in this hospital, as compared to open cholecystectomy. It is recommended that new techniques should be introduced carefully and monitored by means of prospective audit.

Inspiratory fall in systolic pressure in normal and asthmatic subjects.

We used a noninvasive monitor of arterial pressure to determine whether respiratory changes in arterial pressure were closely correlated with airflow obstruction in asthmatic patients during bronchial challenge with methacholine. To validate the noninvasive measurement of respiratory changes in arterial pressure, a preliminary study in 6 subjects with normal cardiovascular and respiratory systems was done during cardiac catheterization for suspected coronary artery disease. There were no significant differences between inspiratory falls in systolic pressure measured noninvasively and those measured from intraaortic pressure. In 11 otherwise healthy asthmatic patients we measured finger arterial pressure, end-expiratory lung volume (FRC), and forced expired volume (FEV1) during baseline and bronchial challenge in the supine posture. Finger arterial pressure was also measured in 11 normal control subjects seated and supine. Normal subjects had an inspiratory fall in systolic pressure (IFSP) of 3.2 mm Hg supine and 5.1 mm Hg seated (p < 0.01). Asthmatic patients when bronchodilated (seated FEV1 = 83 +/- 7% of predicted) had an IFSP of 5.9 mm Hg supine (p < 0.01 compared with supine normal subjects). During bronchial challenge (average fall in FEV1 = 22%), IFSP increased to 16.1 mm Hg (p < 0.001 compared with baseline). In asthmatic subjects, there was a significant correlation between IFSP and FEV1 (mean r = -0.92 +/- 0.05, p < 0.01), and the average change in IFSP/change in FEV1 was -0.38 mm Hg per percentage change in FEV1. During subsequent bronchodilation, IFSP decreased with a similar time course as relaxation of airway smooth muscle, assessed by the breath-to-breath fall in FRC.(ABSTRACT TRUNCATED AT 250 WORDS)

Time course of postoperative hypoxaemia.

OBJECTIVE: To study the time course of nocturnal episodic and constant hypoxaemia during the first five nights after a major abdominal operation in patients not given supplementary oxygen. DESIGN: Open study. SUBJECTS: 17 patients undergoing major elective operations and with no recognised risk factors. MAIN OUTCOME MEASURES: Arterial oxygen saturation measured by pulse oximetry. RESULTS: The level of constant hypoxaemia was lowest during night 2 (p < 0.01) as was the time spent below 90% saturation (p < 0.01) compared with the other four postoperative nights. Episodic hypoxaemia was most common during night 3 (p < 0.05). It was not possible to predict the postoperative occurrence of hypoxaemia from the preoperative spirometric measurements. Preoperative awake arterial oxygen saturation correlated significantly with mean constant hypoxaemia on postoperative nights 1, 3, 4, and 5, but not with episodes of sudden desaturation after operation. Preoperative overnight oximetry correlated significantly with constant hypoxaemia on all five postoperative nights and with episodic hypoxaemia on nights 3 and 4. CONCLUSION: In the light of these results which show the natural history of postoperative hypoxaemia without supplementary oxygen, and because postoperative hypoxaemia may be associated with dysfunction of organ systems after major operations, controlled studies of supplementary oxygen in the late postoperative period are warranted.

Surveillance policy for early detection of failing arteriovenous fistulae for haemodialysis.

We have developed a programme of surveillance for arteriovenous fistulae. Indications for investigation were a reduction of thrill or pulsation on the fistula, decreasing flow (< 200 ml/min) and/or increased venous return pressure (> 150 mmHg.) when on haemodialysis. Between March 1992 and February 1993 we performed intravenous digital subtraction fistulograms in 17 patients. These investigations demonstrated vein stenosis in 11 patients, nine with primary arteriovenous fistulae and in two with secondary access. There was disease of the arterial inflow in one, and no evidence of anatomical problems in the remaining five. As a result revision surgery was performed in 10 cases and percutaneous transluminal angioplasty in the remaining two cases. Five patients underwent dialysis the following day on the same site, avoiding temporary access. Sixteen patients (94%) are still using the same site for haemodialysis at mean follow-up time of 6.1 months (range 2-12 months) and one failed subsequently. Close surveillance of arteriovenous fistulae leads to detection of stenosis prior to occlusion and intervention increases patency, preserves alternative access sites, and prevents central venous cannulation for temporary access.

Comparison of thoracoabdominal calibration methods in normal human subjects.

In 19 normal subjects in the supine posture, we compared accuracy and precision of calibration methods that utilized different ranges of tidal volumes and thoracoabdominal partitioning: spontaneous quiet breathing (QB), isovolume maneuvers, and voluntary efforts to breathe with variable tidal volume and thoracoabdominal partitioning. Thoracic and abdominal movements were measured with the respiratory area fluxometer. Calibration methods utilizing one or more types of respiratory efforts were applied to three measurement situations: QB, variable breathing (volume and thoracoabdominal partitioning), and simulated obstructive apnea (isovolume efforts). Qualitative diagnostic calibration (QDC) included QB data only. The isovolume method (ISOCAL) included isovolumetric efforts at end expiration (functional residual capacity) and QB. Multilinear regression analyses were performed on data sets that included 1) voluntary efforts to breathe with variable volume and thoracoabdominal partitioning (CAL 1), 2) QB in addition to variable volume and partitioning (CAL 2), and 3) isovolume maneuvers in addition to QB and variable volume and partitioning efforts (CAL 3). When calibration data included a wide range of tidal volume, variable thoracoabdominal partitioning, and isovolume efforts (CAL 3), a stable calibration with small bias and scatter during all respiratory patterns was obtained. Excluding isovolume maneuvers (CAL 2) and QB (CAL 1) did not diminish accuracy. Limiting data to isovolume efforts at functional residual capacity plus QB (ISO-CAL) caused a significant increase in scatter during variable breathing patterns. Limiting calibration data to that portion of QB with small variation in the uncalibrated sum of thoracic and abdominal movements (QDC) caused significant increases in scatter in both isovolume efforts and variable breathing.

Repetitive nocturnal arterial oxygen desaturation and silent myocardial ischemia in patients presenting for vascular surgery.

OBJECTIVE: To determine whether nocturnal respiratory abnormality (cyclic oxygen desaturation and tachycardia) is associated with nocturnal myocardial ischemia in older individuals with ischemic heart disease. DESIGN: Non-invasive monitoring on a single occasion. SETTING: Tertiary care referral hospital. PATIENTS: Thirty four consecutive older (68.5 +/- 6 yrs) patients referred for elective abdominal or carotid reconstructive vascular surgery. RESULTS: Seven patients (21%) had moderately severe nocturnal respiratory abnormality, defined by more than 50 dips in arterial oxygen saturation and increases in heart rate during the night. Two of these seven had clinical risk factors for ischemic heart disease and had nocturnal myocardial ischemia. Ten patients (29%) developed ischemia at some time during the study, of whom seven hand known ischemic heart disease, hypertension, and/or angina. Those with increased nocturnal ischemia showed very low frequency (1-2 cycles per minute) cyclic heart rate oscillations and repetitive nocturnal episodes of arterial oxygen desaturation, similar to patients with sleep apnea. CONCLUSION: Repetitive nocturnal cyclic arterial desaturation and cyclic increases in heart rate are associated with nocturnal myocardial ischemia in individuals with clinical risk factors for ischemic heart disease. Further investigation in a large patient sample utilizing non-invasive monitoring of saturation, heart rate, and blood pressure may provide definitive evidence regarding causation of some of the nocturnal myocardial ischemia occurring in older individuals with vascular disease.

Asynchronous chest wall movements during non-rapid eye movement and rapid eye movement sleep in children with bronchopulmonary dysplasia.

The aim of this study was to assess whether age-related changes in thoracic shape modify patterns of thoracoabdominal asynchrony (TAA) or applicability of phase angle analysis during sleep in young children with increased respiratory loads. We assessed TAA during polysomnographic monitoring in 14 young children (mean age 32 months, range 19 to 46; mean weight 12.5 kg, range 9.3 to 17) with severe bronchopulmonary dysplasia (BPD). Of the patients 10 were severely enough affected to require tracheostomy. We measured asynchrony of rib cage (RC) and abdominal (AB) movements at midinspiration and the corresponding phase angle from oscillographic recordings during both non-REM and REM sleep. We measured the amplitude of "paradoxical" displacement of either RC or AB during inspiration and expressed this as a percentage of the total displacement of the compartment. Of 9 children who manifested early inspiratory AB paradox during non-REM sleep, 7 showed a figure eight on the Konno-Mead diagram. The magnitude of abdominal paradox during non-REM sleep was significantly positively correlated with age (n = 14, r = 0.68; p < 0.01). Phase angle was significantly negatively correlated with dynamic lung compliance (n = 14, r = -0.66; p < 0.01). During REM sleep, expiratory abdominal muscle activity was abolished and all patients with abdominal paradox "converted" to an open loop with RC paradox during inspiration. Graphic assessment of the Lissajous figure on the Konno-Mead diagram indicated when midinspiratory phase angle analysis did not reflect the severity of TAA and can be used to infer patterns of respiratory muscle recruitment. We conclude that young children manifest patterns of TAA that differ from the early inspiratory RC paradox commonly observed in infants. Comparison of RC-AB loops between non-REM and REM sleep in the same child can assess increased thoracic inspiratory efforts and expiratory muscle activity as potential mechanisms for abdominal paradox, as distinct from diaphragm ineffectiveness.

Ventilation and thoracoabdominal asynchrony during halothane anesthesia in infants.

To evaluate the ventilatory consequences of high chest wall compliance during anesthesia in infants, we assessed the effects of halothane at different fractions of minimal alveolar concentration (0.75, 1.0, and 1.5 MAC) on ventilation and movements of the rib cage and abdomen in infants < or = 12 mo of age (group I) and children (group II) > or = 12 mo of age. Minute ventilation decreased in group I, (20.6%, 0.75 to 1.5 MAC), but the change in group II did not reach the level of statistical significance. Tidal volume decreased with halothane level between 0.75 and 1.5 MAC, and its fall was greater in group I (32.7 +/- 11.2 vs. 22.6 +/- 9.3% in group II, P < 0.05). Duty cycle, or ratio of inspiratory to total time (TI/TT), increased in group II with halothane level but did not change in group I, resulting in a decreased TI in group I at higher halothane levels. Thoracic paradox increased with halothane level in group I but not group II. The increase in thoracic paradox in association with the fall in tidal volume between 0.75 and 1.5 MAC was greater in group I than group II (P < 0.05). We conclude that smaller infants depend more on inspiratory intercostal muscle activity to stabilize the thorax, leading to a greater degree of depression of ventilation during halothane depression of inspiratory intercostal activity.

Surgical audit without autopsy: tales of the unexpected.

To establish our current practice and the potential value of the autopsy in general surgery, a retrospective review of general surgical autopsies was performed at one district general hospital from January 1989 to August 1991. There was considerable interconsultant variation in autopsy practice with a low 25% overall autopsy rate reflecting a low autopsy request rate. There were discrepancies between the clinical and pathological cause of death in 40 (63%) cases. There were important discrepancies which may have changed management in life in 18 (28%) autopsies, 7 (39%) of which were untreated visceral perforations. Autopsy is an important part of the surgical audit and will disclose considerable unsuspected pathology. Present autopsy rates are low and need to be improved. The unexpected finding of seven untreated visceral perforations requires further study.

Late postoperative nocturnal dips in oxygen saturation in patients undergoing major abdominal vascular surgery. Predictive value of pre-operative overnight pulse oximetry.

Twenty-four patients (23 male) who presented for aortic reconstructive surgery were studied with pulse oximetry on a pre-operative night and during the first five postoperative nights. Patients with five or more dips in oxygen saturation of greater than 4% (with a prompt recovery back toward baseline of 3% or more) per hour of monitoring were classified as having a significant abnormality of respiration. Pre-operatively, four of 24 patients (17%) demonstrated such an abnormality. Postoperatively, 12 patients (50%) met these criteria on at least one of the first five postoperative nights and six of these had two or more nights with severe episodic hypoxaemia. Frequent severe episodic dips in arterial oxygen saturation (to less than 85% saturation) occurred in the late postoperative period at a time when oxygen therapy would usually have been discontinued. Pre-operative overnight pulse oximetry studies fail to predict the development of abnormal respiratory patterns in the postoperative period in the majority of patients.