- Email: [email protected]
Premature mortality and chronic alcoholism: Medical examiner cases, New Jersey
Sot. Sci. Med. Vol. 29, No. 6, pp. 729-732. 1989 Printed in Great Britain. All rights reserved
0277-9536189 Copyright
Q 1989 Maxwell
53.00 + 0.00
Pergamon Macmillan plc
PREMATURE MORTALITY AND CHRONIC ALCOHOLISM: MEDICAL EXAMINER CASES, NEW JERSEY PAUL W. HABERMAN’* and GEETHA NATARAJAN~ ‘New Jersey Department of Health, Division of Alcoholism, Trenton, New Jersey and *New Jersey Office of State Medical Examiner, Newark, New Jersey, U.S.A. Abstract--To obtain estimates of premature mortality from nonviolent causes associated with chronic alcoholism, median ages at death were calculated for 994 Essex County, New Jersey Medical Examiner cases aged 25 or older classified as (1) ‘nonabusers’ or (2) alcoholics whose underlying cause of death was natural disease and (3) those who died of chronic alcoholism. Alcohol-associated mortality accounts for considerable years of potential life lost. Differences in its life-shortening effects according to sex, race, and natural disease versus chronic alcoholism as underlying and/or contributing cause(s) of death are analyzed. Criteria for classifying cases as alcoholics were: (1) autopsy findings attributable to alcoholism; (2) case information that the decedent was a ‘known alcoholic’, or ever had any health or other related problems because of drinking; or, (3) alcohol-specific disease or condition attributable to alcoholism as underlying or contributing cause of death. Nonabusers were cases not classified as alcoholics and/or other drug abusers. Nonabusers had the oldest median age at death, followed by alcoholics who died of (1) natural disease and (2) chronic alcoholism. Female nonabusers were older than their male counterparts, whereas female alcoholics whose underlying cause of death was natural diseases or chronic alcoholism were younger than male alcoholics with the same underlying cause. Median age at death was considerably lower for blacks than whites in all subgroups, with consistently greater race than sex differences. Evidence is presented which supports the accelerated development of alcoholism symptoms and associated illness among women. Aspects of increased mortality risk among alcoholics with natural disease and/or chronic alcoholism as underlying or contributing cause(s) of death are discussed. Key words-chronic
alcoholism, premature mortality, medical examiner cases, natural disease
INTRODtJCl-ION An increased risk of mortality among alcoholics, shortening their life span substantially, has been consistently demonstrated in numerous studies [l, 21. The excess mortality among alcoholics has been reported in terms of observed versus expected deaths
and comparative survival rates for alcoholics and nonalcoholic controls. Alcohol-associated mortality accounts for considerable years of potential life lost (YPLL) per reported death, using national Multiple Cause of Mortality tapes [3]. In this study, estimates of premature mortality from nonviolent causes associated with chronic alcoholism are presented in terms of sex- and race-specific median ages at death and YPLL for medical examiner cases classified as alcoholics compared to those classified as nonabusers. METHODS The study sample consists of 994 Essex County, New Jersey Medical Examiner cases aged 25 years or older within a 4-year time frame, October 1981September 1985. All had chronic alcoholism or a natural disease as a nonviolent underlying cause of death and were classified as alcoholics or ‘nonabusers’. These decedents comprise 44% of all 2255 *Address correspondence to: P. W. Haberman, New Jersey Division of Alcoholism, CN362, Trenton, NJ 08625, U.S.A.
age-eligible Medical Examiner cases in Essex County during the study period. The primary reasons for medical examiner investigation of nonviolent deaths of adults in this study and among other such medical examiner cases are deaths which occurred either suddenly, unattended at home or within 24 hr of hospital admission. Cases were classified as alcoholics according to the following criteria: (1) autopsy findings of moderate or severe fatty liver, cirrhosis of the liver, pancreatitis not attributable to any other disease or condition, or any alcohol-specific condition, such as alcohol cardiomyopathy; (2) any indication in the Medical Examiner case record that the decedent was a ‘known alcoholic’ or ever had any health or other problem because of drinking; or (3) alcoholism indicated as the manner of unnatural death, or as the underlying or a contributing cause of death [4]. As partial validation of study criteria, follow-up microscopic findings in all of 24 systematically selected cases confirmed gross anatomical (autopsy) findings of alcoholic liver change, irrespective of manner of death [4]. Drinking-related problems incorporate most of Keller’s [S] often-cited definition of alcoholism: “a chronic disease manifested by repeated implicative drinking so as to cause injury to the drinker’s health or to his social or economic functioning.” Toxicology findings of alcohol, that is, a positive blood or brain alcohol level (BAL) was not a criterion for classification as an alcoholic. Decedents were classified as narcotics abusers or other drug abusers according to equivalent criteria. 729
PAULW. HABERMAN and GEETHA NATARAJAN
730
Table I. Characteristics of medical examiner cases classified as alcoholics or nonabusers according to chronic alcoholism or natural disease as underlying cause of death: Essex Countv. New Jerscv. October 1981-Seotember 1985 (%) Classification and underlying cause of death
All persons’ (N = 994)
Men (672)
Women (322)
Blackst (507)
Whitest (437)
64.0
59.8
72.7
54.6
75.1
21.3 14.7
26.2 14.0
Il.2 16.1
25.3 20.1
16.9 8.0
Nonabusers
Natural disease Alcoholics :
Natural disease Chronic alcoholism§
*Excludes decedents under age 25 or with unknown age. *Excludes Hispanics. :Excludes decedents also classified as narcotics abusers. §Includes decedents whose cause of death was chronic and acute alcoholism.
Toxicology findings of narcotics or other moodaltering drugs, except if medically prescribed, were substituted in the criteria for autopsy findings attributable to alcoholism. All decedents not classified as alcoholics and/or other drug abusers were classified as nonabusers. The following cases were excluded from this study: homicide, suicide or accident victims and those classified as narcotics or other drug abusers. Also, alcoholics whose underlying or contributing causes of death indicated acute alcholism only were excluded because the analytical focus is on the life-shortening effects of chronic alcoholism. The sample was limited to decedents aged 25 years or older because autopsy findings attributable to chronic alcoholism are very infrequent before this age. The role of alcohol in violent deaths and those involving narcotics abuse among these Essex County Medical Examiner cases have been previously presented [4,6, 71. Among all 2255 age-eligible cases during the 4-year time frame, 743 (36%) were classified as alcoholics, the same proportion as in the study sample (Table 1). For routine use at the Centers for Disease Control [S], YPLL is calculated as death occurring before a predetermined end point, set at 65 years of age. Deaths of persons aged 65 or older are not counted using this procedure, although the remaining life expectancy for a person aged 65 in 1983 was 16.7 years [9]. Using an alternative procedure, YPLL is calculated as death occurring before remaining life expectancy at that age from an abridged annual National Center for Health Statistics life table. YPLL rates also differ considerably according to sex and race so that sexand race-specific YPLL are useful for many health applications. The median is preferable to the arithmetic mean to determine midpoints of age at death
for these Medical Examiner cases because of the disproportionately large number of nonabusers aged 75 or older at death, 133 of 636 (21%). The manner of death of all study cases is natural; for analysis purposes, diseases not associated with alcohol abuse or alcoholism are considered as natural underlying or contributing causes of death and are differentiated from chronic alcoholism or alcoholassociated conditions. All ICD-9-CM coding of the cause(s) of death in New Jersey is done by the Bureau of Vital Statistics, Department of Health after the death certificates have been forwarded by the certifying physicians. Chronic alcoholic liver disease and cirrhosis, alcohol dependence syndrome, and alcohol abuse (ICD-g-CM codes 571.0-3, 303 and 305.0, respectively) in that order, accounted for almost all alcohol-specific causes of death [lo]. More than fourfifths (82%) of 105 cases with autopsy findings of acute or chronic pancreatitits (ICD-9-CM codes 577.0 or 577.1) also had autopsy findings of chronic alcoholic liver disease, irrespective of the manner of death. RESULTS
Of the 994 Essex County, New Jersey Medical Examiner cases in Table 1, 67% were men; 51%, non-Hispanic blacks and 44%, non-Hispanic whites. Forty-five percent of the blacks and two-fifths of the men were classified as alcoholics compared to about one-quarter of the whites and women. Almost three-fifths of the female alcoholics compared to only about one-third of their male counterparts had chronic alcoholism listed as the underlying cause of death. As shown in Table 2, nonabusers whose underlying cause of death was natural disease had the oldest median age at death, followed in order by alcoholics
Table 2. Median age of medical exarmner cases classified as alcoholics or nonabusers according to chronic alcoholism or natural disease as underlying cause of death (COD): Essex County, New Jersey, October 198l-September 1985 Classification: Underlying COD: Characteristics’ All persons Men Women Blacks (non-Hispanics) Whites (non-Hispamcs) *Excludes decedents tExcludes decedents IMedian test, 2 df, @Includes decedents
Nonabuserst Natural disease
Alcohoiicst Chronic alcoholism5
Natural disease
N
Age
N
Age
YPLL
N
Age
YPLL
x2:
636 402 234 277 328
59.8 58.3 62.8 54.6 64.2
212 176 36 128 74
53.4 53.7 51.9 50.7 57.2
6.4 4.6 10.9 3.9 7.0
146 94 52 102 35
44.9 45.8 43.6 43.4 54.9
14.9 12.5 19.2 I I.2 9.3
63.8 30.0 37.9 31.7 20.3
under age 25 or with unknown age. Black and white races exclude Hispanics. classified as narcotics abusers. P < 0.001. whose cause of death was acute and chrome alcoholism.
Premature mortality and chronic alcoholism who died of natural disease and those who died of chronic alcoholism. This order of median age at death for nonabusers and alcoholics was consistent among men, women, blacks and whites. These differences in median age at death are highly significant overall and according to sex and race, using the median test [l 11. The longest-lived Medical Examiner subgroup, nonabusers who died of natural causes, have a much shorter life expectancy than do all persons aged 25 or older in the United States. The median age at death for nonabusers in the study was 59.8 years, compared to a national life expectancy for all persons aged 25 in 1983 of 76.3 years [9]. YPLL for these Medical Examiner cases arguably should use as end points sex- and race-specific ages at death of the nonabusers, which range from 54.6 years for blacks to 64.2 years for whites. Factors contributing to the short life expectancy of these Medical Examiner cases and other comparable cases are: (1) very low standards of living, (2) inadequate health care, and (3) overrepresentation of blacks among cases investigated by medical examiner offices. With age at death of nonabusers as end points, the median YPLL for alcoholics with natural disease as the underlying cause was 6.4 and for those with chronic alcoholism as the underlying cause was 14.9. By sex and race, the median YPLL of alcoholics ranged from 3.9 for blacks, whose underlying cause of death was natural disease, to 19.2 for women whose underlying cause was chronic alcoholism (Table 2). Female nonabusers were older than their male counterparts, whereas female alcoholics, whose underlying cause of death was natural disease or chronic alcoholism, were younger than male alcoholics with the same underlying cause. These sex differences in median ages at death. borne out by the xz values in Table 2, and perhaps the relatively large percent of female alcoholics who died of chronic alcoholism seemingly support the ‘telescoping of drinking effects’ or accelerated development of alcoholism symptoms and physical illness among women [12, 131. In all subgroups. the median age at death was considerably lower among blacks than among whites. Likewise, the race differences in median age at death were consistently greater than were the sex differences. There were too few Hispanics in the sample for separate analysis according to underlying cause of death. As shown in Table 3, alcoholics with natural disease as the underlying cause of death were younger
731
if chronic alcoholism was a contributing cause. Similarly, alcoholics whose underlying cause was chronic alcoholism died at a slightly younger age if natural disease was a contributing cause. The median YPLL ranged from 5.2 for alcoholics with natural disease as the underlying cause without chronic alcoholism as a contributing cause to 15.5 for those with chronic alcoholism as the underlying cause and natural disease as a contributing cause. The number of alcoholics with chronic alcoholism or natural disease listed as a contributing cause was too small to present meaningful findings by sex and race. Only one-third of 146 with chronic alcoholism and one-fifth of 212 alcoholics with natural disease as the underlying causes had the other condition listed as a contributing cause. DISCUSSION
Adult deaths due to natural causes investigated by Medical Examiner Offices most often occur suddenly, unattended at home or within 24 hr of hospital admission. Cardiovascular disease is the underlying cause of most such deaths, including about seventenths for nonabusers and alcoholics with natural disease as underlying cause in the study sample. Alcoholics have been found to have excess mortality, not only from fatty liver or cirrhosis of the liver, but also from cardiovascular diseases [2, 14, 151. However, specific evidence was necessary to classify as alcoholics those Medical Examiner cases who died of cardiovascular disease other than alcoholic cardiomyopathy (ICD-9-CM code 425.5). The reported methodology to estimate premature mortality from alcoholism can be used to obtain comparable data in other medical examiner jurisdictions. Its limitations for the general population are that deaths not investigated by medical examiners or coroners are much less likely to have autopsies performed and alcohol-related conditions recorded. Of the 994 study cases, 84% were autopsied, including 77% of 636 nonabusers and 95% of 358 alcoholics, compared to only 11% of all decedents aged 25 or older in the United States in 1983 [16]. Of these alcoholics, 78% had positive autopsy findings related to chronic alcoholism, including 92% of 188 with alcoholism indicated in the manner or cause of death. Just 18% were classified as alcoholics only on the basis of reported alcohol-related problems in the Medical Examiner case record. However, any cases who died of natural disease without autopsy findings attributable to alcoholism, but with unreported
Table 3. Median age of medical examiner cases classified as alcoholics or nonabusers according to chronic alcoholism and/or natural disease as underlying and contributing cause of death: Essex Countv. New Jersev. October 1981~Scotember 1985 Causes of death Underlying Nonabusers: Alcoholics:
Natural Natural Natural Chronic Chronic
disease disease disease alcoholism@ alcoholism6
Contributing
Chronic alcoholism# Natural disease
All persons* N
Aget
YPLL
636 170 42 97 49
59.8 54.6 50.4 45.4 44.3
5.2 9.4 14.4 15.5
*Excludes decedents under age 25 or with unknown age. +Medtan test. x’ = 68.6, 4 & P c 0.001. ZExcludes decedents classified as narcotics abusers. Vncludes decedents whose cause of death was acute and chronic alcoholism.
732
PAULW. HABERMAN and GEETHANATARAJAK
drinking problems, would have been ‘false negatives’ misclassified as nonabusers. Although death certificates are the primary source of mortality statistics, the accuracy of certificates without autopsy verification has been questioned, especially for alcoholism [17, 181. Autopsies are usually necessary to confirm the presence of alcoholassociated diseases or conditions and to classify decedents as alcoholics. Moreover, death certificates seriously understate alcohol-related morbidity and mortality because of denial or lack of knowledge by informants, generally relatives or friends of the decedent, and reluctance by the certifying physician to report a stigmatizing condition [19]. In conclusion, different aspects of the life-shortening effects of alcohol-associated illness have been examined in this medical examiner study by using autopsy findings, contributing as well as underlying causes of death, and informant reports of alcoholrelated problems. Much of these medical examiner data are unavailable or often inaccessible for general mortality so that the full extent of alcohol-related illness among decedents in the general population is very difficult to ascertain. Nonetheless, the sex and ethnic differences from alcohol-related disease in this study have considerable generalizability. Acknowledgement-This research was supported by a consultant agreement to the first author from the New Jersey Department of Health, Division of Alcoholism.
Taylor J. R. and Helzer J. E. The natural history of alcoholism. In The Biology of Alcoholism. Vol. 6. The of
Alcoholism:
Psychological
P. W. and Natarajan G. Trends in alcoholism and narcotics abuse fro& medical examiner data.
J. Stud. Alcohol 47. 316-321. 1986. 5. Keller M. Definition of alcoholism. Q. J. Stud. Alcohol
21, 125-134, 1960. 6. Haberman P. W. Alcohol and alcoholism in traffic and other accidental deaths. Am. J. Drug Alcohol Abuse 13,
475-484, 1987. P. W. Alcohol use and alcoholism among motor vehicle driver fatalities. Inr. J. Addict. 22,
7. Haberman
1119-l 128, 1987. 8. Centers for Disease Control. Premature mortality in the
United States: public health issues in the use of years of potential life lost. Morbid. Mortal. Wkiy Rep. Suppl. No. 2S 35, IS-I IS, 1986. 9. National Center for Health Statistics. Viral Sratisrics of the United States, 1983. Vol. II, Sec. 6, Life Tables, DHSS Publ. No. (PHS) 86- 1104, Government Printing Office, Washington, DC., 1986. 10. National Center for Health Statistics. The International Classification of Diseases, 9th Reuision, Clinical Mod@cation, Vol. 1. Commission on Professional and Hospi-
tal Activities, Ann Arbor, Mich., 1978. 11. Winkler R. L. and Hays W. L. Statistics: Probabiliry, Inference, and Decision. Holt, Rinehart & Winston, New York, 1975. 12. Ashley M. J., Olin J. S., Le-Riche W. H., Kormaczewski A., Schmidt W. and Rankin J. G. Morbidity in alcoholics. Archs intern. Med. 137, 883-887, 1977. 13. Bissell L. and Haberman P. W. Alcoholism in the Professions. Oxford University Press, New York, 1984. 14. Pell S. and D’Alonzo C. A five-year mortality study of alcoholics. J. occup. Med. 15, 120-125, 1973. 15. Taylor J. R., Combs-Ox-me T. and Taylor D. A. Alcohol and mortality: diagnostic considerations. J. Stud. Alcohol 44, 17-25,
REFERENCES
Pathogenesis
4. Haberman
Factors
(Edited by Kissen B. and Begleiter H.), pp. 17-65. Plenum Press, New York, 1983. Vaillant G. E. The Natural History of Alcoholism. Harvard University Press, Cambridge, Mass., 1983. Bertolucci D., Lowman C. and Dufour M. Surveillance and assessment of alcohol-related mortality-United States, 1980. Morbid. Mortal. Wkly Rep. 34, 161-163, 1985.
1983.
16. National Center for Health Statistics. Viral Statistics of the United Stales, 1983. Vol. II, Mortality, Part k, DHHS Publ. No. IPHS) 87-1101. Government Printine Office, Washington, D.C., 1987.’ 17. Roberts W. C. The autopsy: its decline and a suggestion for its revival. New Engl. J. Med. 299, 332-338, 1978. 18. Taylor J. R., Holmes S. J., Comb-Orme T. and Scott E. B. Alcohol and death certificates. J. Am. med. Ass. 248, 3096, 1982.
P. W. and Baden M. M. Alcohol, Other Drugs and Violent Death. Oxford University Press, New
19. Haberman
York, 1978.
0277-9536189 Copyright
Q 1989 Maxwell
53.00 + 0.00
Pergamon Macmillan plc
PREMATURE MORTALITY AND CHRONIC ALCOHOLISM: MEDICAL EXAMINER CASES, NEW JERSEY PAUL W. HABERMAN’* and GEETHA NATARAJAN~ ‘New Jersey Department of Health, Division of Alcoholism, Trenton, New Jersey and *New Jersey Office of State Medical Examiner, Newark, New Jersey, U.S.A. Abstract--To obtain estimates of premature mortality from nonviolent causes associated with chronic alcoholism, median ages at death were calculated for 994 Essex County, New Jersey Medical Examiner cases aged 25 or older classified as (1) ‘nonabusers’ or (2) alcoholics whose underlying cause of death was natural disease and (3) those who died of chronic alcoholism. Alcohol-associated mortality accounts for considerable years of potential life lost. Differences in its life-shortening effects according to sex, race, and natural disease versus chronic alcoholism as underlying and/or contributing cause(s) of death are analyzed. Criteria for classifying cases as alcoholics were: (1) autopsy findings attributable to alcoholism; (2) case information that the decedent was a ‘known alcoholic’, or ever had any health or other related problems because of drinking; or, (3) alcohol-specific disease or condition attributable to alcoholism as underlying or contributing cause of death. Nonabusers were cases not classified as alcoholics and/or other drug abusers. Nonabusers had the oldest median age at death, followed by alcoholics who died of (1) natural disease and (2) chronic alcoholism. Female nonabusers were older than their male counterparts, whereas female alcoholics whose underlying cause of death was natural diseases or chronic alcoholism were younger than male alcoholics with the same underlying cause. Median age at death was considerably lower for blacks than whites in all subgroups, with consistently greater race than sex differences. Evidence is presented which supports the accelerated development of alcoholism symptoms and associated illness among women. Aspects of increased mortality risk among alcoholics with natural disease and/or chronic alcoholism as underlying or contributing cause(s) of death are discussed. Key words-chronic
alcoholism, premature mortality, medical examiner cases, natural disease
INTRODtJCl-ION An increased risk of mortality among alcoholics, shortening their life span substantially, has been consistently demonstrated in numerous studies [l, 21. The excess mortality among alcoholics has been reported in terms of observed versus expected deaths
and comparative survival rates for alcoholics and nonalcoholic controls. Alcohol-associated mortality accounts for considerable years of potential life lost (YPLL) per reported death, using national Multiple Cause of Mortality tapes [3]. In this study, estimates of premature mortality from nonviolent causes associated with chronic alcoholism are presented in terms of sex- and race-specific median ages at death and YPLL for medical examiner cases classified as alcoholics compared to those classified as nonabusers. METHODS The study sample consists of 994 Essex County, New Jersey Medical Examiner cases aged 25 years or older within a 4-year time frame, October 1981September 1985. All had chronic alcoholism or a natural disease as a nonviolent underlying cause of death and were classified as alcoholics or ‘nonabusers’. These decedents comprise 44% of all 2255 *Address correspondence to: P. W. Haberman, New Jersey Division of Alcoholism, CN362, Trenton, NJ 08625, U.S.A.
age-eligible Medical Examiner cases in Essex County during the study period. The primary reasons for medical examiner investigation of nonviolent deaths of adults in this study and among other such medical examiner cases are deaths which occurred either suddenly, unattended at home or within 24 hr of hospital admission. Cases were classified as alcoholics according to the following criteria: (1) autopsy findings of moderate or severe fatty liver, cirrhosis of the liver, pancreatitis not attributable to any other disease or condition, or any alcohol-specific condition, such as alcohol cardiomyopathy; (2) any indication in the Medical Examiner case record that the decedent was a ‘known alcoholic’ or ever had any health or other problem because of drinking; or (3) alcoholism indicated as the manner of unnatural death, or as the underlying or a contributing cause of death [4]. As partial validation of study criteria, follow-up microscopic findings in all of 24 systematically selected cases confirmed gross anatomical (autopsy) findings of alcoholic liver change, irrespective of manner of death [4]. Drinking-related problems incorporate most of Keller’s [S] often-cited definition of alcoholism: “a chronic disease manifested by repeated implicative drinking so as to cause injury to the drinker’s health or to his social or economic functioning.” Toxicology findings of alcohol, that is, a positive blood or brain alcohol level (BAL) was not a criterion for classification as an alcoholic. Decedents were classified as narcotics abusers or other drug abusers according to equivalent criteria. 729
PAULW. HABERMAN and GEETHA NATARAJAN
730
Table I. Characteristics of medical examiner cases classified as alcoholics or nonabusers according to chronic alcoholism or natural disease as underlying cause of death: Essex Countv. New Jerscv. October 1981-Seotember 1985 (%) Classification and underlying cause of death
All persons’ (N = 994)
Men (672)
Women (322)
Blackst (507)
Whitest (437)
64.0
59.8
72.7
54.6
75.1
21.3 14.7
26.2 14.0
Il.2 16.1
25.3 20.1
16.9 8.0
Nonabusers
Natural disease Alcoholics :
Natural disease Chronic alcoholism§
*Excludes decedents under age 25 or with unknown age. *Excludes Hispanics. :Excludes decedents also classified as narcotics abusers. §Includes decedents whose cause of death was chronic and acute alcoholism.
Toxicology findings of narcotics or other moodaltering drugs, except if medically prescribed, were substituted in the criteria for autopsy findings attributable to alcoholism. All decedents not classified as alcoholics and/or other drug abusers were classified as nonabusers. The following cases were excluded from this study: homicide, suicide or accident victims and those classified as narcotics or other drug abusers. Also, alcoholics whose underlying or contributing causes of death indicated acute alcholism only were excluded because the analytical focus is on the life-shortening effects of chronic alcoholism. The sample was limited to decedents aged 25 years or older because autopsy findings attributable to chronic alcoholism are very infrequent before this age. The role of alcohol in violent deaths and those involving narcotics abuse among these Essex County Medical Examiner cases have been previously presented [4,6, 71. Among all 2255 age-eligible cases during the 4-year time frame, 743 (36%) were classified as alcoholics, the same proportion as in the study sample (Table 1). For routine use at the Centers for Disease Control [S], YPLL is calculated as death occurring before a predetermined end point, set at 65 years of age. Deaths of persons aged 65 or older are not counted using this procedure, although the remaining life expectancy for a person aged 65 in 1983 was 16.7 years [9]. Using an alternative procedure, YPLL is calculated as death occurring before remaining life expectancy at that age from an abridged annual National Center for Health Statistics life table. YPLL rates also differ considerably according to sex and race so that sexand race-specific YPLL are useful for many health applications. The median is preferable to the arithmetic mean to determine midpoints of age at death
for these Medical Examiner cases because of the disproportionately large number of nonabusers aged 75 or older at death, 133 of 636 (21%). The manner of death of all study cases is natural; for analysis purposes, diseases not associated with alcohol abuse or alcoholism are considered as natural underlying or contributing causes of death and are differentiated from chronic alcoholism or alcoholassociated conditions. All ICD-9-CM coding of the cause(s) of death in New Jersey is done by the Bureau of Vital Statistics, Department of Health after the death certificates have been forwarded by the certifying physicians. Chronic alcoholic liver disease and cirrhosis, alcohol dependence syndrome, and alcohol abuse (ICD-g-CM codes 571.0-3, 303 and 305.0, respectively) in that order, accounted for almost all alcohol-specific causes of death [lo]. More than fourfifths (82%) of 105 cases with autopsy findings of acute or chronic pancreatitits (ICD-9-CM codes 577.0 or 577.1) also had autopsy findings of chronic alcoholic liver disease, irrespective of the manner of death. RESULTS
Of the 994 Essex County, New Jersey Medical Examiner cases in Table 1, 67% were men; 51%, non-Hispanic blacks and 44%, non-Hispanic whites. Forty-five percent of the blacks and two-fifths of the men were classified as alcoholics compared to about one-quarter of the whites and women. Almost three-fifths of the female alcoholics compared to only about one-third of their male counterparts had chronic alcoholism listed as the underlying cause of death. As shown in Table 2, nonabusers whose underlying cause of death was natural disease had the oldest median age at death, followed in order by alcoholics
Table 2. Median age of medical exarmner cases classified as alcoholics or nonabusers according to chronic alcoholism or natural disease as underlying cause of death (COD): Essex County, New Jersey, October 198l-September 1985 Classification: Underlying COD: Characteristics’ All persons Men Women Blacks (non-Hispanics) Whites (non-Hispamcs) *Excludes decedents tExcludes decedents IMedian test, 2 df, @Includes decedents
Nonabuserst Natural disease
Alcohoiicst Chronic alcoholism5
Natural disease
N
Age
N
Age
YPLL
N
Age
YPLL
x2:
636 402 234 277 328
59.8 58.3 62.8 54.6 64.2
212 176 36 128 74
53.4 53.7 51.9 50.7 57.2
6.4 4.6 10.9 3.9 7.0
146 94 52 102 35
44.9 45.8 43.6 43.4 54.9
14.9 12.5 19.2 I I.2 9.3
63.8 30.0 37.9 31.7 20.3
under age 25 or with unknown age. Black and white races exclude Hispanics. classified as narcotics abusers. P < 0.001. whose cause of death was acute and chrome alcoholism.
Premature mortality and chronic alcoholism who died of natural disease and those who died of chronic alcoholism. This order of median age at death for nonabusers and alcoholics was consistent among men, women, blacks and whites. These differences in median age at death are highly significant overall and according to sex and race, using the median test [l 11. The longest-lived Medical Examiner subgroup, nonabusers who died of natural causes, have a much shorter life expectancy than do all persons aged 25 or older in the United States. The median age at death for nonabusers in the study was 59.8 years, compared to a national life expectancy for all persons aged 25 in 1983 of 76.3 years [9]. YPLL for these Medical Examiner cases arguably should use as end points sex- and race-specific ages at death of the nonabusers, which range from 54.6 years for blacks to 64.2 years for whites. Factors contributing to the short life expectancy of these Medical Examiner cases and other comparable cases are: (1) very low standards of living, (2) inadequate health care, and (3) overrepresentation of blacks among cases investigated by medical examiner offices. With age at death of nonabusers as end points, the median YPLL for alcoholics with natural disease as the underlying cause was 6.4 and for those with chronic alcoholism as the underlying cause was 14.9. By sex and race, the median YPLL of alcoholics ranged from 3.9 for blacks, whose underlying cause of death was natural disease, to 19.2 for women whose underlying cause was chronic alcoholism (Table 2). Female nonabusers were older than their male counterparts, whereas female alcoholics, whose underlying cause of death was natural disease or chronic alcoholism, were younger than male alcoholics with the same underlying cause. These sex differences in median ages at death. borne out by the xz values in Table 2, and perhaps the relatively large percent of female alcoholics who died of chronic alcoholism seemingly support the ‘telescoping of drinking effects’ or accelerated development of alcoholism symptoms and physical illness among women [12, 131. In all subgroups. the median age at death was considerably lower among blacks than among whites. Likewise, the race differences in median age at death were consistently greater than were the sex differences. There were too few Hispanics in the sample for separate analysis according to underlying cause of death. As shown in Table 3, alcoholics with natural disease as the underlying cause of death were younger
731
if chronic alcoholism was a contributing cause. Similarly, alcoholics whose underlying cause was chronic alcoholism died at a slightly younger age if natural disease was a contributing cause. The median YPLL ranged from 5.2 for alcoholics with natural disease as the underlying cause without chronic alcoholism as a contributing cause to 15.5 for those with chronic alcoholism as the underlying cause and natural disease as a contributing cause. The number of alcoholics with chronic alcoholism or natural disease listed as a contributing cause was too small to present meaningful findings by sex and race. Only one-third of 146 with chronic alcoholism and one-fifth of 212 alcoholics with natural disease as the underlying causes had the other condition listed as a contributing cause. DISCUSSION
Adult deaths due to natural causes investigated by Medical Examiner Offices most often occur suddenly, unattended at home or within 24 hr of hospital admission. Cardiovascular disease is the underlying cause of most such deaths, including about seventenths for nonabusers and alcoholics with natural disease as underlying cause in the study sample. Alcoholics have been found to have excess mortality, not only from fatty liver or cirrhosis of the liver, but also from cardiovascular diseases [2, 14, 151. However, specific evidence was necessary to classify as alcoholics those Medical Examiner cases who died of cardiovascular disease other than alcoholic cardiomyopathy (ICD-9-CM code 425.5). The reported methodology to estimate premature mortality from alcoholism can be used to obtain comparable data in other medical examiner jurisdictions. Its limitations for the general population are that deaths not investigated by medical examiners or coroners are much less likely to have autopsies performed and alcohol-related conditions recorded. Of the 994 study cases, 84% were autopsied, including 77% of 636 nonabusers and 95% of 358 alcoholics, compared to only 11% of all decedents aged 25 or older in the United States in 1983 [16]. Of these alcoholics, 78% had positive autopsy findings related to chronic alcoholism, including 92% of 188 with alcoholism indicated in the manner or cause of death. Just 18% were classified as alcoholics only on the basis of reported alcohol-related problems in the Medical Examiner case record. However, any cases who died of natural disease without autopsy findings attributable to alcoholism, but with unreported
Table 3. Median age of medical examiner cases classified as alcoholics or nonabusers according to chronic alcoholism and/or natural disease as underlying and contributing cause of death: Essex Countv. New Jersev. October 1981~Scotember 1985 Causes of death Underlying Nonabusers: Alcoholics:
Natural Natural Natural Chronic Chronic
disease disease disease alcoholism@ alcoholism6
Contributing
Chronic alcoholism# Natural disease
All persons* N
Aget
YPLL
636 170 42 97 49
59.8 54.6 50.4 45.4 44.3
5.2 9.4 14.4 15.5
*Excludes decedents under age 25 or with unknown age. +Medtan test. x’ = 68.6, 4 & P c 0.001. ZExcludes decedents classified as narcotics abusers. Vncludes decedents whose cause of death was acute and chronic alcoholism.
732
PAULW. HABERMAN and GEETHANATARAJAK
drinking problems, would have been ‘false negatives’ misclassified as nonabusers. Although death certificates are the primary source of mortality statistics, the accuracy of certificates without autopsy verification has been questioned, especially for alcoholism [17, 181. Autopsies are usually necessary to confirm the presence of alcoholassociated diseases or conditions and to classify decedents as alcoholics. Moreover, death certificates seriously understate alcohol-related morbidity and mortality because of denial or lack of knowledge by informants, generally relatives or friends of the decedent, and reluctance by the certifying physician to report a stigmatizing condition [19]. In conclusion, different aspects of the life-shortening effects of alcohol-associated illness have been examined in this medical examiner study by using autopsy findings, contributing as well as underlying causes of death, and informant reports of alcoholrelated problems. Much of these medical examiner data are unavailable or often inaccessible for general mortality so that the full extent of alcohol-related illness among decedents in the general population is very difficult to ascertain. Nonetheless, the sex and ethnic differences from alcohol-related disease in this study have considerable generalizability. Acknowledgement-This research was supported by a consultant agreement to the first author from the New Jersey Department of Health, Division of Alcoholism.
Taylor J. R. and Helzer J. E. The natural history of alcoholism. In The Biology of Alcoholism. Vol. 6. The of
Alcoholism:
Psychological
P. W. and Natarajan G. Trends in alcoholism and narcotics abuse fro& medical examiner data.
J. Stud. Alcohol 47. 316-321. 1986. 5. Keller M. Definition of alcoholism. Q. J. Stud. Alcohol
21, 125-134, 1960. 6. Haberman P. W. Alcohol and alcoholism in traffic and other accidental deaths. Am. J. Drug Alcohol Abuse 13,
475-484, 1987. P. W. Alcohol use and alcoholism among motor vehicle driver fatalities. Inr. J. Addict. 22,
7. Haberman
1119-l 128, 1987. 8. Centers for Disease Control. Premature mortality in the
United States: public health issues in the use of years of potential life lost. Morbid. Mortal. Wkiy Rep. Suppl. No. 2S 35, IS-I IS, 1986. 9. National Center for Health Statistics. Viral Sratisrics of the United States, 1983. Vol. II, Sec. 6, Life Tables, DHSS Publ. No. (PHS) 86- 1104, Government Printing Office, Washington, DC., 1986. 10. National Center for Health Statistics. The International Classification of Diseases, 9th Reuision, Clinical Mod@cation, Vol. 1. Commission on Professional and Hospi-
tal Activities, Ann Arbor, Mich., 1978. 11. Winkler R. L. and Hays W. L. Statistics: Probabiliry, Inference, and Decision. Holt, Rinehart & Winston, New York, 1975. 12. Ashley M. J., Olin J. S., Le-Riche W. H., Kormaczewski A., Schmidt W. and Rankin J. G. Morbidity in alcoholics. Archs intern. Med. 137, 883-887, 1977. 13. Bissell L. and Haberman P. W. Alcoholism in the Professions. Oxford University Press, New York, 1984. 14. Pell S. and D’Alonzo C. A five-year mortality study of alcoholics. J. occup. Med. 15, 120-125, 1973. 15. Taylor J. R., Combs-Ox-me T. and Taylor D. A. Alcohol and mortality: diagnostic considerations. J. Stud. Alcohol 44, 17-25,
REFERENCES
Pathogenesis
4. Haberman
Factors
(Edited by Kissen B. and Begleiter H.), pp. 17-65. Plenum Press, New York, 1983. Vaillant G. E. The Natural History of Alcoholism. Harvard University Press, Cambridge, Mass., 1983. Bertolucci D., Lowman C. and Dufour M. Surveillance and assessment of alcohol-related mortality-United States, 1980. Morbid. Mortal. Wkly Rep. 34, 161-163, 1985.
1983.
16. National Center for Health Statistics. Viral Statistics of the United Stales, 1983. Vol. II, Mortality, Part k, DHHS Publ. No. IPHS) 87-1101. Government Printine Office, Washington, D.C., 1987.’ 17. Roberts W. C. The autopsy: its decline and a suggestion for its revival. New Engl. J. Med. 299, 332-338, 1978. 18. Taylor J. R., Holmes S. J., Comb-Orme T. and Scott E. B. Alcohol and death certificates. J. Am. med. Ass. 248, 3096, 1982.
P. W. and Baden M. M. Alcohol, Other Drugs and Violent Death. Oxford University Press, New
19. Haberman
York, 1978.