Br. J. Surg. 1992, Vol. 79, December, 1300-1 302
J. Rosenberg, M. H. Pedersen, T. Ramsing and H. Kehlet Department of Surgical Gastroenterology 235, Hvidovre University Hospital, D K- 2650 Hvidovre, Denmark Correspondence to: Dr J. Rosenberg
Circadian variation in unexpected postoperative death Unexpected deaths still occur following major surgical procedures. The cause is often unknown but may he cardiac or thrornboembolic in nature. Postoperative ischaemia, infarction and sudden cardiac death may be triggered by episodic or constant arterial hypoxaemia, which increases during the night. This study examined the circadian variation of sudden unexpected death ,following abdominal surgery between I985 and 1989 inclusive. Deaths were divided into those occurring during the day (08.00-16.00 hours), euening ( 16.00-24.00 hours) and night (24.00-08.00 hours). Twenty-three deaths were considered to have been totally unexpected. Of 16 such patients undergoing autopsy, pulmonary embolism was the cause of death inJive. In the remaining 11 patients, death occurred at night in eight ( P < 0~00.5). Five oftlze seven patients without an autopsy died at night ( P < 0.04); overall, 13 of 18 unexpected deaths occurred at night-time. These results suggest a need f o r further studies of sleep- and respiration-related efects on postoperative nocturnal cardiac function. The eficacy of monitoring during this apparent high-risk period should be evaluated.
Major surgery is still associated with significant morbidity and mortality’,’. Periodicity occurs for time of death in the general population not undergoing ~ u r g e r y ~ time - ~ , of onset of chest pain in acute myocardial infarction’-*, time of transient myocardial ischaemia in patients with coronary artery time of onset of ischaemic stroke” and of attacks of asthma’’ 14. Nocturnal severe episodic or constant hypoxaemia is common late after operation, with episodic hypoxaemia being observed predominantly during sleep”- ’. The clinical implications of late postoperative low oxygen saturation are not known, but when episodic it may be associated with cardiac ischaemia and arrhythmiaI6, and may represent an important factor in the development of late postoperative myocardial infarction’ * and sudden death. This study examined the circadian variation in unexpected postoperative death.
Patients and methods All deaths within the first week after surgery during the 5-year period from 1985 to 1989 were reviewed. In this period all patients over 40 years of age received prophylactic low-dose heparin and elastic stockings to prevent pulmonary embolism. Deaths were divided into those during the day (08.00-16.00 hours), evening ( 16.00-24.00 hours) and night (24.00-08.00 hours). Because many deaths occurred in patients with potentially life-threatening complications, a selection procedure (Figure 1 ) was used to find postoperative deaths where no explanatory pathology could be identified. The exclusion of patients with potentially life-threatening complications after operation was performed from the records by a senior doctor with no knowledge of the time of death. For statistical analyses the non-parametric sign test’ was used. To overcome the problem of mass-significance and the time factor (night lasts 8 h, day and evening 16 h), tests were performed for day- and evening-deaths uersus twice the recorded number of night-deaths. P < 0.05 was considered significant.
’
from terminal cancer ( F i g u r e I ). Fifty-eight patients died > 7 days after surgery. Of the 59 deaths remaining for further analysis, I8 patients were considered to have suffered from early postoperative life-threatening complications such as shock, heart failure and the sepsis syndrome. Another 18 patients suffered late postoperative ( > 4 h after surgery) potentially life-threatening complications such as cardiac arrhythmia, stroke and severe pulmonary dysfunction requiring oxygen therapy or artificial ventilation.. Thus, 23 deaths were considered to have been sudden and totally unexpected ( F i g u r e I ); a typical statement in the patient’s file was ‘found dead in bed’. In these 23 patients the median age was 74 (range 48-86) years; ten were men. Eight had preoperative cardiac or pulmonary disease requiring regular medication and 15 underwent surgery for an acute problem. Of 16 patients undergoing autopsy, pulmonary embolism was the cause of death in five. In the remaining 11 patients, death occurred at night in eight ( P < 0.005). In the seven patients without autopsy, five deaths occurred during the night ( P < 0.04). Thirteen of 18 unexpected deaths occurred during one-third of the day (24.00-08.00 hours) ( P < 0.0003). The circadian variation in unexpected postoperative death and the distribution of unexpected deaths in the first week after surgery are shown in Figure 2.
6 2 7 n o t operated on
or receiving highdose morphine
,
2 3 unexpected
58 died > 7 days after surgery
7 no autopsy
I
36 postoperative life-threatening ;mplications*
postoperative ,de;ths
5 pulmonary em bo Iism
11 no explanatory
pa thology
Results A total of 744 deaths occurred: 241 during the day, 236 in the evening and 267 at night ( not significant 1. Of the 744 deaths only 59 were both unexpected and Occurred within 7 days of surgery. The majority (627 patients) had either had no surgery or received high-dose morphine infusion for alleviation of pain
Figure 1 Selection procedure in the analysis of deaths during a 5-year period. * Life-threatening conzplications presumably influencing time of death. t P < 0.04: 2 P i0.005 (day- PIUS evening-deaths versus tuice number of night-deaths, sign test)
1300
0007 1121 92 12130&01
1 day
~
~
1 evening 5 n i g h t
- -~(
~
t
2 evening 8 night P
1 day
~
~~
~
1992 Butterworth-Heinemdnn Ltd
2/i,
Circadian variation in postoperative death: J. Rosenberg et al.
3
In
1
+ m
U 4-
0 0 Z
.. .. ...
1
... ... ...
0 08.00
..
a
Time o f day ( h o u r s )
n
In
5 ar
U
L 0
0
z
0
b
1
2
3
4
5
6
7
Time a f t e r operation ( d a y s )
Discussion This study demonstrates significant circadian variation in the frequency of sudden unexpected postoperative death. Compiled data for 437 51 1 non-surgical patients have previously shown that a major peak o f deaths occurs between 04.00 and 06.00 hours regardless of cause3. The subgroup of sudden cardiac death seems to occur between 07.00 and 11.00 hours4.'. Occurrence of acute myocardial infarction determined by the onset of chest pain is most likely to occur between 06.00 and 13.00 hours, with a peak incidence at 09.00 hours", or between 06.00 and 09.00 hours with an additional peak at 24.00 hours'. The morning peaks may be caused partly by an increase in platelet aggregabiiity'. The time of transient myocardial ischaernia in patierits with coronary artery disease shows the same trend as the onset of acute myocardial infarction, with peak incidences at 06.00- 10.00 hours" and 08.00- 13.00 hours"'. Onset of ischaemic stroke also shows a peak incidence in the morning between 10.00 and 12.00 hours". In some asthmatic subjects sleep may be a contributing factor in the worsening of bronchoconstriction' which leads to an increased number of asthmatic deaths at nightl3.I4. Several mechanisms of sudden unexpected death in the present study are possible. IJnexpected fatal pulmonary embolism was found in five patients; in the remaining deaths cardiac infarction or arrhythmia is most probably the pathogenetic factor. The majority of postoperative myocardial infarctions (50-70 per cent) and ischaemic episodes (94 per cent) arc silent"'. Repeated episodes of ischaemia may have a cumulative etkct and eventually cause myocardial infarction", and postoperative ischaemia may be associated with adverse cardiac outcome such as myocardial infarction and sudden death'2.'.3. A number of factors may be involved in the development of cardiac ischaemia. such as constant post-
',
Br. J . Surg., Vol. 79, No. 12, December 1992
operative tachycardia induced by pain and surgical stress' h . 2 4 , disruption of fluid and electrolyte homoeostasisZ5,and arterial hypoXaemia1h.1 8 . 2 6 . 2 7 . Episodes of arterial hypoxaemia may be harmful for the heart'h.'s.26,since they are often accompanied by rises in pulse rate15.i6.2hand blood p r e ~ s u r e ~ ' . ' ~leading , to a marked increase in the rate-pressure product. Severe episodic changes in arterial oxygen saturation are especially seen in the sleep apnoea syndrome that may prove to be an important surgical risk f a c t ~ r ~ ~ The , ~ ~prevalence ,~'. of this syndrome in the surgical population is unknown, but may be as high as 10 per cent, especially in old age3'. Lying supine, as in the postoperative period because of pain, aggravates the ventilatory disturbance with pronounced apnoea and worsened episodic oxygen d e ~ a t u r a t i o n ~Snoring ~. is associated with considerable morbidity and r n ~ r t a l i t y ' ~ . ~and ~ -patients ~ ~ , with obstructive sleep apnoea are more likely to die during sleep". An undiagnosed sleep apnoea syndrome could thus have been a contributing pathogenetic factor in some of the present patients with sudden postoperative death. Patients with obvious or possible causes of death were excluded from analysis to allow an examination of deaths in the light of unopposed postoperative hypoxaemia and sleep-related cardiopulmonary problems. These were less likely to occur beyond the first week after surgery3' 40. The 36 patients so excluded in Figure 1 may have been affected by an unexpected cardiac event that could have been attributed to postoperative hypoxaemia and/or sleep-related cardiopulmonary problems, but because they had several competing severe organ dysfunctions it was impossible to determine the exact cause of death or time of occurrence of the initial event leading to complications and death. It may be argued that the events leading to unexpected death would have been discovered and successfully treated if the patients had been monitored more intensely, especially during the night. Routine nocturnal surveillance late after operation in the majority of surgical departments is by hourly checks by a nurse usually recording whether the patient is breathing or not. In the light of the present results and recent studies demonstrating severe nocturnal hypoxaemia and related cardiac eventsl"'6.2h , routine monitoring with pulse oximetry or continuous electrocardiography may be indicated for several days after surgery.
Acknowledgements This work was supported by the Danish Heart Foundation. the P. Carl Petersen Foundation, the Velux Foundation and the Foundation of 17 December 1981.
References I. 7 -.
3. 4.
5. 6.
7. 8.
Fielding LP, Phillips RKS, Hittinger R . Factors influencing mortality after curativc resection for large howcl cancer in elderlj patients. Ltitic~,r 1989: i : 595-7. Kind P. Outcome measurement using hospital activity data: deaths aftcr surgical procedures. BrJSurcg 1990;77: 1399-402. Mitler M M , Carskadon MA, Czeisler CA, Dement WC. Dinges DF. Graeber RC. Catastrophes. sleep and public policy: cmseiiws report. Slwp 1988; 11: 100-9. Willich S N . Levy D, Rocco MB, Tofler G H , Stone PH. Muller J E . Circadian variation in the incidence of sudden cardiac death in the Framingham heart study population. A i i i J Cttrcliol 1987; 60: 801-6. Muller JE, Ludmer PL. Willich SN t't ril. Circadian variation in the frequency of sudden cardiac death. Circ,ulrrtioti 1987: 75: 131 8. Muller JE, Stone PH. Turi ZG ( ' I t i / . Circadian variation in thc frequency of onset of acute myocardial infarction. N Eiiyl J M t ~ l 1985; 313: 1315 22. Thompson DR, Sutton TW, Jowett NI. Pohl JEF. Circadian variation in the frequency of onset of chest pain in acute myocardial infarction. Br H w r t J 1991: 65: 177 X. Totler GH, B r u i n s k i D, Schaefer A1 e / t i / . Concurrent morning increase in platelet ageregability and the rish of myocardial infarction and suddcn cardiac death. N En;]/ J M d 1987; 316: 1514 18.
1301
Circadian variation in postoperative death: J. Rosenberg et al.
9. 10.
11. 12. 13. 14. 15.
16. 17.
18. 19.
20.
21. 22.
23.
24.
Rocco MB, Barry J, Campbell S et al. Circadian variation of transient myocardial ischemia in patients with coronary artery disease. Circulation 1987;75: 395-400. Mulcahy D, Keegan J, Cunningham D et al. Circadian variation of total ischaemia burden and its alteration with anti-anginal agents. Lancet 1988;ii: 755-8. Marler JR, Price TR, Clark G L et a[. Morning increase in onset of ischemic stroke. Stroke 1989;20: 473-6. Ballard RD, Saathoff MC, Pate1 DK, Kelly PL, Martin RJ. Effect of sleep on nocturnal bronchoconstriction and ventilatory patterns in asthmatics. J Appl Physiol 1989;67: 243-9. Hetzel MR, Clark TJH, Branthwaite MA. Asthma: analysis of sudden deaths and ventilatory arrests in hospital. BMJ 1977;i : 808- 11. McNicholas WT, Fitzgerald MX. Nocturnal deaths among patients with chronic bronchitis and emphysema. BMJ 1984; 289: 878. Rosenberg J, Dirkes WE, Kehlet H. Episodic arterial oxygen desaturation and heart rate variations followingmajor abdominal surgery. Br J Anaesth 1989;63: 651-4. Rosenberg J, Rasmussen V, von Jessen F, Ullstad T, Kehlet H. Late postoperative episodic and constant hypoxaemia and associated ECG abnormalities. Br J Anaesth 1990;65: 684-91. Catley DM, Thorton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced episodic oxygen desaturation in the postoperative period: its association with ventilatory pattern and analgesic regimen. Anesthesiology 1985;63: 20-8. Pateman JA, Hanning CD. Postoperative myocardial infarction and episodic hypoxaemia. Br J Anaesth 1989;63: 648-50. Seigel S, Castellan NJ. Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York: McGraw-Hill, 1988. Mangano DT, Hollenberg M, Fegert G et al. Perioperative myocardial ischemia in patients undergoing noncardiac surgery. I: Incidence and severity during the 4 day perioperative period. J Am Coll Cardiol 1991; 17: 843-50. Geft IL, Fishbein MC, Ninomiya K et al. Intermittent brief periods of ischemia have a cumulative effect and may cause myocardial necrosis. Circulation 1982;66: 1150-3. Mangano DT, Wong MG, London MJ, Tubau JF, Rapp JA. Perioperative myocardial ischemia in patients undergoing noncardiac surgery. 11: Incidence and severity during the first week after surgery. J Am Coll Cardiol 1991; 17: 851-7. Mangano DT, Browner WS, Hollenberg M, London MJ, Tubau JF, Tateo IM. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery. N Engl J Med 1990;323: 1781-8. Kehlet H. Modification of responses to surgery by neural blockade: clinical implications. In: Cousins MJ, Bridenbaugh
1302
PO, eds. Neural Blockade in Clinical Anesthesia and Management of Pain. 2nd ed. Philadelphia: JB Lippincott, 1988: 145-88.
25. Shires GT. Important factors in the management of homeostasis in the surgical patient. Acta Chir Scand Suppl1988;550: 29-35. 26. Reeder MK, Muir AD, Foex P, Goldman MD, Loh L, Smart D. Postoperative myocardial ischaemia: temporal association with nocturnal hypoxaemia. Br J Anaesth 1991;67: 626-31. 27. Tirlapur VG, Mir MA. Nocturnal hypoxemia and associated electrocardiographic changes in patients with chronic obstructive airways disease. N Engl J Med 1982;306: 125-30. 28. Reeder MK, Goldman MD, Loh L, Muir AD, Casey KR, Gitlin DA. Postoperative obstructive sleep apnoea: haemodynamic effects of treatment with nasal CPAP. Anaesthesia 1991; 46: 849-53. 29. Reeder MK, Goldman MD, Loh L, Muir AD, Casey KR. Haemodynamic effects of periodic ventilation: abolition with supplementary oxygen. Br J Anaesth 1991;67: 326-8. 30. Rosenberg J, Kehlet H. Postoperative episodic oxygen desaturation in the sleep apnoea syndrome. Acta Anaesthesiol Scand 1991;35: 368-9. 31. Hanning CD. Obstructive sleep apnoea. Br J Anaesth 1989;63: 477-88. 32. Guilleminault C. Obstructive sleep apnea syndrome: a review. Psychiatr Clin North Am 1987;10: 607-21. 33. Miki H, Hida W, Kikuchi Y, Takishima T. Effect of sleep position on obstructive sleep apnea. TohokuJExp Med 1988;156: 143-9. 34. Rees J. Snoring. BMJ 1991;302: 860-1. 35. Guilleminault C, Connolly SJ, Winkle RA. Cardiac arrhythmia and conduction disturbances during sleep in 400 patients with sleep apnea syndrome. Am J Cardiol 1983;52: 490-4. 36. D’Alessandro R, Magelli C, Gamberini G et al. Snoring every night as a risk factor for myocardial infarction: a case-control study. BMJ 1990;300: 1557-8. 37. Ancoli-Israel S, Klauber MR, Kripke DF, Parker L, Cobarrubias M. Sleep apnea in female patients in a nursing home: increased risk of mortality. Chest 1989;96: 1054-8. 38. Knudsen J. Duration of hypoxaemia after uncomplicated upper abdominal and thoraco-abdominal operations. Anaesthesia 1970; 25: 372-7. 39. Jones JG, Sapsford DJ, Wheatley RG. Postoperative hypoxaemia: mechanisms and time course. Anaesthesia 1990;45: 566-73. 40. Knill RL, Moote CA, Skinner MI, Rose EA. Anesthesia with abdominal surgery leads to intense REM sleep during the first postoperative week. Anesthesiology 1990;73: 52-61. Paper accepted 24 May 1992
Br. J. Surg., Vol. 79, No. 12, December 1992
J. Rosenberg, M. H. Pedersen, T. Ramsing and H. Kehlet Department of Surgical Gastroenterology 235, Hvidovre University Hospital, D K- 2650 Hvidovre, Denmark Correspondence to: Dr J. Rosenberg
Circadian variation in unexpected postoperative death Unexpected deaths still occur following major surgical procedures. The cause is often unknown but may he cardiac or thrornboembolic in nature. Postoperative ischaemia, infarction and sudden cardiac death may be triggered by episodic or constant arterial hypoxaemia, which increases during the night. This study examined the circadian variation of sudden unexpected death ,following abdominal surgery between I985 and 1989 inclusive. Deaths were divided into those occurring during the day (08.00-16.00 hours), euening ( 16.00-24.00 hours) and night (24.00-08.00 hours). Twenty-three deaths were considered to have been totally unexpected. Of 16 such patients undergoing autopsy, pulmonary embolism was the cause of death inJive. In the remaining 11 patients, death occurred at night in eight ( P < 0~00.5). Five oftlze seven patients without an autopsy died at night ( P < 0.04); overall, 13 of 18 unexpected deaths occurred at night-time. These results suggest a need f o r further studies of sleep- and respiration-related efects on postoperative nocturnal cardiac function. The eficacy of monitoring during this apparent high-risk period should be evaluated.
Major surgery is still associated with significant morbidity and mortality’,’. Periodicity occurs for time of death in the general population not undergoing ~ u r g e r y ~ time - ~ , of onset of chest pain in acute myocardial infarction’-*, time of transient myocardial ischaemia in patients with coronary artery time of onset of ischaemic stroke” and of attacks of asthma’’ 14. Nocturnal severe episodic or constant hypoxaemia is common late after operation, with episodic hypoxaemia being observed predominantly during sleep”- ’. The clinical implications of late postoperative low oxygen saturation are not known, but when episodic it may be associated with cardiac ischaemia and arrhythmiaI6, and may represent an important factor in the development of late postoperative myocardial infarction’ * and sudden death. This study examined the circadian variation in unexpected postoperative death.
Patients and methods All deaths within the first week after surgery during the 5-year period from 1985 to 1989 were reviewed. In this period all patients over 40 years of age received prophylactic low-dose heparin and elastic stockings to prevent pulmonary embolism. Deaths were divided into those during the day (08.00-16.00 hours), evening ( 16.00-24.00 hours) and night (24.00-08.00 hours). Because many deaths occurred in patients with potentially life-threatening complications, a selection procedure (Figure 1 ) was used to find postoperative deaths where no explanatory pathology could be identified. The exclusion of patients with potentially life-threatening complications after operation was performed from the records by a senior doctor with no knowledge of the time of death. For statistical analyses the non-parametric sign test’ was used. To overcome the problem of mass-significance and the time factor (night lasts 8 h, day and evening 16 h), tests were performed for day- and evening-deaths uersus twice the recorded number of night-deaths. P < 0.05 was considered significant.
’
from terminal cancer ( F i g u r e I ). Fifty-eight patients died > 7 days after surgery. Of the 59 deaths remaining for further analysis, I8 patients were considered to have suffered from early postoperative life-threatening complications such as shock, heart failure and the sepsis syndrome. Another 18 patients suffered late postoperative ( > 4 h after surgery) potentially life-threatening complications such as cardiac arrhythmia, stroke and severe pulmonary dysfunction requiring oxygen therapy or artificial ventilation.. Thus, 23 deaths were considered to have been sudden and totally unexpected ( F i g u r e I ); a typical statement in the patient’s file was ‘found dead in bed’. In these 23 patients the median age was 74 (range 48-86) years; ten were men. Eight had preoperative cardiac or pulmonary disease requiring regular medication and 15 underwent surgery for an acute problem. Of 16 patients undergoing autopsy, pulmonary embolism was the cause of death in five. In the remaining 11 patients, death occurred at night in eight ( P < 0.005). In the seven patients without autopsy, five deaths occurred during the night ( P < 0.04). Thirteen of 18 unexpected deaths occurred during one-third of the day (24.00-08.00 hours) ( P < 0.0003). The circadian variation in unexpected postoperative death and the distribution of unexpected deaths in the first week after surgery are shown in Figure 2.
6 2 7 n o t operated on
or receiving highdose morphine
,
2 3 unexpected
58 died > 7 days after surgery
7 no autopsy
I
36 postoperative life-threatening ;mplications*
postoperative ,de;ths
5 pulmonary em bo Iism
11 no explanatory
pa thology
Results A total of 744 deaths occurred: 241 during the day, 236 in the evening and 267 at night ( not significant 1. Of the 744 deaths only 59 were both unexpected and Occurred within 7 days of surgery. The majority (627 patients) had either had no surgery or received high-dose morphine infusion for alleviation of pain
Figure 1 Selection procedure in the analysis of deaths during a 5-year period. * Life-threatening conzplications presumably influencing time of death. t P < 0.04: 2 P i0.005 (day- PIUS evening-deaths versus tuice number of night-deaths, sign test)
1300
0007 1121 92 12130&01
1 day
~
~
1 evening 5 n i g h t
- -~(
~
t
2 evening 8 night P
1 day
~
~~
~
1992 Butterworth-Heinemdnn Ltd
2/i,
Circadian variation in postoperative death: J. Rosenberg et al.
3
In
1
+ m
U 4-
0 0 Z
.. .. ...
1
... ... ...
0 08.00
..
a
Time o f day ( h o u r s )
n
In
5 ar
U
L 0
0
z
0
b
1
2
3
4
5
6
7
Time a f t e r operation ( d a y s )
Discussion This study demonstrates significant circadian variation in the frequency of sudden unexpected postoperative death. Compiled data for 437 51 1 non-surgical patients have previously shown that a major peak o f deaths occurs between 04.00 and 06.00 hours regardless of cause3. The subgroup of sudden cardiac death seems to occur between 07.00 and 11.00 hours4.'. Occurrence of acute myocardial infarction determined by the onset of chest pain is most likely to occur between 06.00 and 13.00 hours, with a peak incidence at 09.00 hours", or between 06.00 and 09.00 hours with an additional peak at 24.00 hours'. The morning peaks may be caused partly by an increase in platelet aggregabiiity'. The time of transient myocardial ischaernia in patierits with coronary artery disease shows the same trend as the onset of acute myocardial infarction, with peak incidences at 06.00- 10.00 hours" and 08.00- 13.00 hours"'. Onset of ischaemic stroke also shows a peak incidence in the morning between 10.00 and 12.00 hours". In some asthmatic subjects sleep may be a contributing factor in the worsening of bronchoconstriction' which leads to an increased number of asthmatic deaths at nightl3.I4. Several mechanisms of sudden unexpected death in the present study are possible. IJnexpected fatal pulmonary embolism was found in five patients; in the remaining deaths cardiac infarction or arrhythmia is most probably the pathogenetic factor. The majority of postoperative myocardial infarctions (50-70 per cent) and ischaemic episodes (94 per cent) arc silent"'. Repeated episodes of ischaemia may have a cumulative etkct and eventually cause myocardial infarction", and postoperative ischaemia may be associated with adverse cardiac outcome such as myocardial infarction and sudden death'2.'.3. A number of factors may be involved in the development of cardiac ischaemia. such as constant post-
',
Br. J . Surg., Vol. 79, No. 12, December 1992
operative tachycardia induced by pain and surgical stress' h . 2 4 , disruption of fluid and electrolyte homoeostasisZ5,and arterial hypoXaemia1h.1 8 . 2 6 . 2 7 . Episodes of arterial hypoxaemia may be harmful for the heart'h.'s.26,since they are often accompanied by rises in pulse rate15.i6.2hand blood p r e ~ s u r e ~ ' . ' ~leading , to a marked increase in the rate-pressure product. Severe episodic changes in arterial oxygen saturation are especially seen in the sleep apnoea syndrome that may prove to be an important surgical risk f a c t ~ r ~ ~ The , ~ ~prevalence ,~'. of this syndrome in the surgical population is unknown, but may be as high as 10 per cent, especially in old age3'. Lying supine, as in the postoperative period because of pain, aggravates the ventilatory disturbance with pronounced apnoea and worsened episodic oxygen d e ~ a t u r a t i o n ~Snoring ~. is associated with considerable morbidity and r n ~ r t a l i t y ' ~ . ~and ~ -patients ~ ~ , with obstructive sleep apnoea are more likely to die during sleep". An undiagnosed sleep apnoea syndrome could thus have been a contributing pathogenetic factor in some of the present patients with sudden postoperative death. Patients with obvious or possible causes of death were excluded from analysis to allow an examination of deaths in the light of unopposed postoperative hypoxaemia and sleep-related cardiopulmonary problems. These were less likely to occur beyond the first week after surgery3' 40. The 36 patients so excluded in Figure 1 may have been affected by an unexpected cardiac event that could have been attributed to postoperative hypoxaemia and/or sleep-related cardiopulmonary problems, but because they had several competing severe organ dysfunctions it was impossible to determine the exact cause of death or time of occurrence of the initial event leading to complications and death. It may be argued that the events leading to unexpected death would have been discovered and successfully treated if the patients had been monitored more intensely, especially during the night. Routine nocturnal surveillance late after operation in the majority of surgical departments is by hourly checks by a nurse usually recording whether the patient is breathing or not. In the light of the present results and recent studies demonstrating severe nocturnal hypoxaemia and related cardiac eventsl"'6.2h , routine monitoring with pulse oximetry or continuous electrocardiography may be indicated for several days after surgery.
Acknowledgements This work was supported by the Danish Heart Foundation. the P. Carl Petersen Foundation, the Velux Foundation and the Foundation of 17 December 1981.
References I. 7 -.
3. 4.
5. 6.
7. 8.
Fielding LP, Phillips RKS, Hittinger R . Factors influencing mortality after curativc resection for large howcl cancer in elderlj patients. Ltitic~,r 1989: i : 595-7. Kind P. Outcome measurement using hospital activity data: deaths aftcr surgical procedures. BrJSurcg 1990;77: 1399-402. Mitler M M , Carskadon MA, Czeisler CA, Dement WC. Dinges DF. Graeber RC. Catastrophes. sleep and public policy: cmseiiws report. Slwp 1988; 11: 100-9. Willich S N . Levy D, Rocco MB, Tofler G H , Stone PH. Muller J E . Circadian variation in the incidence of sudden cardiac death in the Framingham heart study population. A i i i J Cttrcliol 1987; 60: 801-6. Muller JE, Ludmer PL. Willich SN t't ril. Circadian variation in the frequency of sudden cardiac death. Circ,ulrrtioti 1987: 75: 131 8. Muller JE, Stone PH. Turi ZG ( ' I t i / . Circadian variation in thc frequency of onset of acute myocardial infarction. N Eiiyl J M t ~ l 1985; 313: 1315 22. Thompson DR, Sutton TW, Jowett NI. Pohl JEF. Circadian variation in the frequency of onset of chest pain in acute myocardial infarction. Br H w r t J 1991: 65: 177 X. Totler GH, B r u i n s k i D, Schaefer A1 e / t i / . Concurrent morning increase in platelet ageregability and the rish of myocardial infarction and suddcn cardiac death. N En;]/ J M d 1987; 316: 1514 18.
1301
Circadian variation in postoperative death: J. Rosenberg et al.
9. 10.
11. 12. 13. 14. 15.
16. 17.
18. 19.
20.
21. 22.
23.
24.
Rocco MB, Barry J, Campbell S et al. Circadian variation of transient myocardial ischemia in patients with coronary artery disease. Circulation 1987;75: 395-400. Mulcahy D, Keegan J, Cunningham D et al. Circadian variation of total ischaemia burden and its alteration with anti-anginal agents. Lancet 1988;ii: 755-8. Marler JR, Price TR, Clark G L et a[. Morning increase in onset of ischemic stroke. Stroke 1989;20: 473-6. Ballard RD, Saathoff MC, Pate1 DK, Kelly PL, Martin RJ. Effect of sleep on nocturnal bronchoconstriction and ventilatory patterns in asthmatics. J Appl Physiol 1989;67: 243-9. Hetzel MR, Clark TJH, Branthwaite MA. Asthma: analysis of sudden deaths and ventilatory arrests in hospital. BMJ 1977;i : 808- 11. McNicholas WT, Fitzgerald MX. Nocturnal deaths among patients with chronic bronchitis and emphysema. BMJ 1984; 289: 878. Rosenberg J, Dirkes WE, Kehlet H. Episodic arterial oxygen desaturation and heart rate variations followingmajor abdominal surgery. Br J Anaesth 1989;63: 651-4. Rosenberg J, Rasmussen V, von Jessen F, Ullstad T, Kehlet H. Late postoperative episodic and constant hypoxaemia and associated ECG abnormalities. Br J Anaesth 1990;65: 684-91. Catley DM, Thorton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced episodic oxygen desaturation in the postoperative period: its association with ventilatory pattern and analgesic regimen. Anesthesiology 1985;63: 20-8. Pateman JA, Hanning CD. Postoperative myocardial infarction and episodic hypoxaemia. Br J Anaesth 1989;63: 648-50. Seigel S, Castellan NJ. Nonparametric Statistics for the Behavioral Sciences. 2nd ed. New York: McGraw-Hill, 1988. Mangano DT, Hollenberg M, Fegert G et al. Perioperative myocardial ischemia in patients undergoing noncardiac surgery. I: Incidence and severity during the 4 day perioperative period. J Am Coll Cardiol 1991; 17: 843-50. Geft IL, Fishbein MC, Ninomiya K et al. Intermittent brief periods of ischemia have a cumulative effect and may cause myocardial necrosis. Circulation 1982;66: 1150-3. Mangano DT, Wong MG, London MJ, Tubau JF, Rapp JA. Perioperative myocardial ischemia in patients undergoing noncardiac surgery. 11: Incidence and severity during the first week after surgery. J Am Coll Cardiol 1991; 17: 851-7. Mangano DT, Browner WS, Hollenberg M, London MJ, Tubau JF, Tateo IM. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery. N Engl J Med 1990;323: 1781-8. Kehlet H. Modification of responses to surgery by neural blockade: clinical implications. In: Cousins MJ, Bridenbaugh
1302
PO, eds. Neural Blockade in Clinical Anesthesia and Management of Pain. 2nd ed. Philadelphia: JB Lippincott, 1988: 145-88.
25. Shires GT. Important factors in the management of homeostasis in the surgical patient. Acta Chir Scand Suppl1988;550: 29-35. 26. Reeder MK, Muir AD, Foex P, Goldman MD, Loh L, Smart D. Postoperative myocardial ischaemia: temporal association with nocturnal hypoxaemia. Br J Anaesth 1991;67: 626-31. 27. Tirlapur VG, Mir MA. Nocturnal hypoxemia and associated electrocardiographic changes in patients with chronic obstructive airways disease. N Engl J Med 1982;306: 125-30. 28. Reeder MK, Goldman MD, Loh L, Muir AD, Casey KR, Gitlin DA. Postoperative obstructive sleep apnoea: haemodynamic effects of treatment with nasal CPAP. Anaesthesia 1991; 46: 849-53. 29. Reeder MK, Goldman MD, Loh L, Muir AD, Casey KR. Haemodynamic effects of periodic ventilation: abolition with supplementary oxygen. Br J Anaesth 1991;67: 326-8. 30. Rosenberg J, Kehlet H. Postoperative episodic oxygen desaturation in the sleep apnoea syndrome. Acta Anaesthesiol Scand 1991;35: 368-9. 31. Hanning CD. Obstructive sleep apnoea. Br J Anaesth 1989;63: 477-88. 32. Guilleminault C. Obstructive sleep apnea syndrome: a review. Psychiatr Clin North Am 1987;10: 607-21. 33. Miki H, Hida W, Kikuchi Y, Takishima T. Effect of sleep position on obstructive sleep apnea. TohokuJExp Med 1988;156: 143-9. 34. Rees J. Snoring. BMJ 1991;302: 860-1. 35. Guilleminault C, Connolly SJ, Winkle RA. Cardiac arrhythmia and conduction disturbances during sleep in 400 patients with sleep apnea syndrome. Am J Cardiol 1983;52: 490-4. 36. D’Alessandro R, Magelli C, Gamberini G et al. Snoring every night as a risk factor for myocardial infarction: a case-control study. BMJ 1990;300: 1557-8. 37. Ancoli-Israel S, Klauber MR, Kripke DF, Parker L, Cobarrubias M. Sleep apnea in female patients in a nursing home: increased risk of mortality. Chest 1989;96: 1054-8. 38. Knudsen J. Duration of hypoxaemia after uncomplicated upper abdominal and thoraco-abdominal operations. Anaesthesia 1970; 25: 372-7. 39. Jones JG, Sapsford DJ, Wheatley RG. Postoperative hypoxaemia: mechanisms and time course. Anaesthesia 1990;45: 566-73. 40. Knill RL, Moote CA, Skinner MI, Rose EA. Anesthesia with abdominal surgery leads to intense REM sleep during the first postoperative week. Anesthesiology 1990;73: 52-61. Paper accepted 24 May 1992
Br. J. Surg., Vol. 79, No. 12, December 1992
