John, S. M. B.H.B.; Koelmeyer, T. D. M.B., B.S., F.R.A.C.S., F.R.C.P.A.
There is a common conception that the very old die of old age. This is particularly prevalent among younger practitioners who lack specific geriatric training. It is possible that the tendency toward atypical and/or asymptomatic presentations in the elderly (1) is mistaken for no specific cause of death. This article reviews forensic autopsy reports of the very aged. With this information, the attitude that the aged die of old age can be evaluated.
Forensic autopsies are indicated in both natural and unnatural deaths: natural when a physician is uncertain of the cause of death because he or she either was not present at the death or was present and had no diagnosis. An unnatural death is directly attributable to accident, suicide, or homicide. In this unnatural category, death may be instantaneous or may occur after a time period with or without treatment.
When the cause of death is unknown, even in the oldest of the old, a forensic autopsy is required. Unfortunately, there is an unawareness among doctors of the exact role forensic autopsies hold in the system of coroners (2).
The use of a study population receiving forensic autopsies is biased toward cases in which there is an element of diagnostic uncertainty and hence a tendency toward the “clinical diagnosis” of old age. Given that the majority of forensic autopsies in this age group are performed after natural deaths, they provided the best source of information for this study.
The forensic autopsy reports of patients aged 91 years or older were reviewed. The autopsies took place at the Department of Forensic Pathology, Auckland School of Medicine, Auckland, New Zealand. This city of ∼ 1.3 million people is served by a single coroner and four forensic pathologists. The time period of the study was January 1, 1988 to December 11, 1998.
The cases were separated into natural and unnatural causes of death. This division is consistent with the law governing death certification in New Zealand. The data were analyzed and then tabulated. Table 1 illustrates the unnatural deaths. Table 2 categorizes both natural and unnatural deaths. A separation, although legally valid, negates some of the relevant clinical information this table provides. Table 3 categorizes the factors complicating the cause of death.
Three hundred nineteen subjects met the study criteria. The mean age was 92.5 years. There were 214 women (67%), whose mean age was 92.2, and 105 men (33%), whose mean age was 93.2.
Two hundred seventy-two (85%) deaths were natural; the remaining 47 (15%) were unnatural (Tables 1 and 2). Of the natural deaths, only 13 (5%) were “written off” as being attributable to old age or senile debility.
A total of 61 (19%) deaths were multifactorial or complicated (Tables 3 and 4). Complicating factors occurred in all 28 fracture cases, in 25% of gastrointestinal and infectious causes of death, and 7% of cardiovascular deaths. No central nervous system or neoplastic deaths were associated with complicating factors (Table 4).
Twenty-eight fractures were recorded as the underlying cause of death, third only to ischemic heart disease (IHD) and bronchopneumonia.
The selection bias in a series of forensic autopsies will affect absolute mortality data; however, these statistics were not the aim of the study. Rather, the emphasis was on deaths in which there was uncertainty or a tendency toward the diagnosis of old age. These are the very cases an autopsy series selects.
Paterson et al. (3) evaluated the diagnostic accuracy of specialist geriatricians and found that they made the correct diagnosis in 74% of geriatric cases. In an earlier study, nonspecialists achieved a correct diagnostic rate of 47%(4). Pulmonary embolism and bronchopneumonia are two common diagnoses missed most often (5,6), and cerebral vascular accident (CVA) the least (6). Gee reviewed 3,000 of his own autopsies on institutionalized geriatric patients; acute myocardial infarction (AMI) was clinically undiagnosed in 74% of cases, pulmonary thromboembolism in 74%, and ruptured abdominal aortic aneurysm in 59%(7). These are all common causes of death in the very elderly, but clinically they may be missed more often than not.
In this study, the most common causes of death were IHD, 74 (23%); bronchopneumonia, 37 (12%); fractures, 28 (9%); AMI, 25 (8%); CVA, 19 (6%); and ruptured aneurysm, 17 (5%). Cardiovascular causes accounted for 50% of deaths; infectious diseases 17%; central nervous system, gastrointestinal, and neoplastic causes all <7% each; and other causes, 18%. These findings vary slightly from those of Klima et al. (8), who found that in patients aged >80, 24% to 38% died of infection and 26% to 34% of cardiovascular disease. All the other categories of cause of death were comparable.
In 1996, 6,567 people aged >85 years died in New Zealand (9). Of these deaths, only 72 (1.1%) had forensic autopsies. There were a similar proportion of cardiovascular and respiratory deaths in the groups in which autopsy was and was not performed: 52% versus 48.6%, respectively, in the cardiovascular category, and 18% versus 16.7%, respectively in the respiratory deaths. There were more deaths of gastrointestinal causes in the autopsy group, 9.7% versus 3.5%, perhaps because of the often vague nature of abdominal conditions, necessitating autopsy. Deaths resulting from central nervous system conditions were also more common in the autopsy group, 5.6% versus 1.3%, which may have reflected the increased ratio of hemorrhagic to thrombotic strokes observed in this study. Both neoplastic and infectious causes of death were less prevalent in the autopsy group than in the nonautopsy group: 1.3% versus 12.2% and 2.8% versus 6.0%, respectively.
To make the diagnosis of IHD, the pathologists at this institute required either 80% single vessel stenosis or >60% multiple vessel stenosis with myocardial scarring. In the absence of these or any other significant findings, old age or senile debility was entered on the death certificate. Even given these criteria, a valid argument exists that these pathologic changes would have developed in the deceased long before death. So how can a pathologist state with any certainty that these preexisting changes were the cause of death at a particular time? Therefore, IHD may be seen by some as a “soft” autopsy diagnosis, somewhere to “dump” cases when no other cause of death can be found. We acknowledge this point, but we urge that it should not detract from the overwhelming finding of the study. The elderly die of disease not old age. In fact, the portion of persons aged >85 years dying of cardiovascular causes is similar in nonautopsy and autopsy groups (9).
A high proportion of deaths in this series (22%) were multifactorial. Mackenbach et al. (10) analyzed competing causes of death on 5,975 death certificates. Competing causes were most common in respiratory disease, relative risk (RR) of 1.42, followed by cardiovascular (RR 1.08), and least common in neoplastic deaths (RR 0.54). The relative scarcity of respiratory and neoplastic deaths, and the high number of complicated fracture cases, prevents a direct comparison with these findings. Between one sixth and one quarter of cardiovascular, infectious, and gastrointestinal causes of death were multifactorial. All 28 fracture cases were complicated. Eleven cases of pneumonia (8 bronchial and 3 lobar) and 10 of IHD accounted for 75% of the factors complicating fracture deaths.
Falls accounted for all 29 deaths associated with fractures. In this series, 65% of the accidental deaths resulted from falls. In the United States, falls are the leading cause of death resulting from accidents in those aged >79 years (11).
In only six cases was the cause of death cancer; in a further five it was a complicating factor. This low rate is consistent with other studies (12,13). A study of cancer in patients aged ≥85 demonstrated a decreased rate of cancer, cancer-related mortality, and metastatic disease with advancing age. The rate of diagnostic accuracy also fell with age (12). Although cancer in the very elderly decreases, it is underdiagnosed and its significance is therefore underestimated (12).
Of the 25 cases of AMI in this series, eight (32%) ruptured. This is a high rate, although not inconsistent with some other studies that also found high rates of rupture in the elderly (13–15).
The striking feature of the CVAs in this study is the abnormal ratio between hemorrhagic and thrombotic stroke: 90% were hemorrhagic. This reversal of normal (13) is most likely caused by selection bias. Patients with hemorrhagic strokes are more likely to die instantly, hence without a diagnosis, and therefore require an autopsy.
Three men committed suicide in this series. Suicide is most commonly associated with young men; however, its incidence is actually biphasic. The second rise occurs in geriatric men (13,16,17). Their rate exceeds that of young men and continues to further increase with age (13).
Only two deaths were caused by chronic obstructive airways disease (COAD) in this series; both were emphysema. There are two likely explanations for this minimal contribution from a common group of diseases. First, they are unlikely to be referred to the coroner given their chronic nature. Second, it is likely that predisposition to COAD will result in death before the age of 90.
Fifteen deaths in this series were due to ruptured abdominal aortic aneurysm (AAA). The definitive therapy is surgery. Elective repair of an AAA has a good prognosis, especially when compared with the dismal outlook of a ruptured aneurysm. A low prerupture diagnosis rate, however, denies most patients this chance. The survival rate after rupture was found to be 12.5% in one series of 112 patients (18).
Fat emboli would have been found in the majority of fracture cases because this is common when death occurs shortly after a major fracture (19). For fat emboli to be entered as a cause of death, they were of such an extent that they could not be ignored. Findings indicating the fat embolism syndrome aid in making this diagnosis. This was the case in 3 of the 29 fracture cases. The normal incidence of this syndrome is 0.5% to 5%(11). Our finding of 10% is high, but the numbers in this study are too small to call this significant.
The geriatric population accounts for 25% of house fire fatalities (20). Faulty or misused electrical appliances, especially electric blankets, are a common cause (20). One death resulted from house fire in this study.
This article demonstrates natural death to be caused by specific diseases in the very elderly. Given the tendency toward atypical and or asymptomatic presentations in this age group (1), diagnoses based on symptoms alone may be more difficult (3). If no diagnosis is apparent before death, the patient must receive a forensic autopsy. In this study, the cause of death was found in 95% of these patients. The high postmortem diagnostic rate in this clinically difficult age group validates its legal requirement.
1. France MJ, Vuletic JC, Koelmeyer TD. Does advancing age modify the presentation of disease? Am J Forensic Med Pathol 1992; 13:120–3.
2. Start RD, Delargy-Aziz Y, Dorries CP, et al. Clinicians and the coronial system: ability of clinicians to recognise reportable deaths. BMJ 1993; 306:1038–41.
3. Paterson DA, Dorovitch MI, Farquhar DL, et al. Prospective study of necropsy audit of geriatric inpatient deaths. J Clin Pathol 1992; 45:575–8.
4. Hofman WI. The pathologist and the geriatric autopsy. J Am Geriatr Soc 1975; 23:11–3.
5. Rossman I, Rodstein M, Bomstein A. Undiagnosed diseases in an aging population: pulmonary embolism and bronchopneumonia. Arch Intern Med 1974; 133:366–9.
6. Gross JS, Neufeld RR, Libow LS, et al. Autopsy study of elderly institutionalized patients: review of 234 autopsies. Arch Intern Med 1988; 148:173–6.
7. Gee WM. Causes of death in a hospitalized geriatric population: an autopsy study of 3000 patients. Virchows Arch A Pathol Anat 1993; 423:343–9.
8. Klima MP, Povysil C, Teasdale TA. Causes of death in geriatric patients: a cross-cultural study. J Gerontol Med Sci 1997; 52A:M247–53.
9. Ministry of Health (NZ). Mortality and demographic data 1996. Wellington: Ministry of Health, 1999.
10. Mackenbach JP, Kunst AE, Lautenbach H, et al. Competing causes of death: a death certificate study. J Clin Epidemiol 1997; 50:1069–77.
11. Hoskin AF. Fatal falls: trends and characteristics [abstract]. Stat Bull Metrop Insur Com 1998; 79:10.
12. Stanta G, Campagner L, Cavallieri F, Giarelli L. Cancer in the oldest old: what we have learned from autopsy studies. Clin Geriatr Med 1997; 13:55–68.
13. Evans JG, Williams, eds. Oxford textbook of geriatric medicine. Oxford: Oxford University Press, 1992, p. 277.
14. Reeder GS. Identification and treatment of complications of myocardial infarction. Lancet 1995; 70:880.
15. Baker GE, Koelmeyer TD. Deaths due to unrecognised myocardial infarction causing left ventricular rupture: can we improve the diagnostic rate? N Z Med J 1999; 112: 429–30.
16. Snowdon J. Suicide rates and methods in different age groups: Australian data and perceptions. Int J Geriatr Psych 1997; 12:253–8.
17. Galanos AN, Gardner Jr, WA Riddick L. Forensic autopsy in the elderly. South Med J 1989; 82:462–6.
18. Panneton JM, Lassonade J, Laurendeau F. Ruptured abdominal aortic aneurysm: impact of comorbidity and postoperative complications on outcome. Ann Vasc Surg 1995; 9:535–41.
19. Cotran RS, Robbins SL, Kumar V. Robbins pathological basis of disease, 6th ed. Philadelphia: W.B. Saunders, 1992, p. 130.
20. Elder AT, Squires T, Busuttil A. Fire fatalities in elderly people. Age Aging 1996; 25:214–6.
© 2001 Lippincott Williams & Wilkins, Inc.