DISCUSSION Sudden Natural Deaths Robinson 11 Sudden deaths at work represented about 1 per cent of all deaths in Allegheny County. This represents deaths among the presumed healthy and fit working population who were not expecting to die. The highest sudden death rate vas observed for men employed in the services occupations, including school custodians, janitors, maintenance workers, security guards, and firemen. Other studies have reported increased cardiovascular mortality or risk factor levels, but not sudden death, for men employed in similar services occupations. These include Washington State school custodians (16), prison guards (17), and messengers in the British civil service (18). Our data suggest that sudden death may be a significant portion of the excess cardiovascular risk experienced by men in lower status occupations. Medical screening at entry into an occupation may only be partially effective in identifying potential heart attack victims. Epidemiologic research has shown that only 50 per cent of all sudden cardiac death victims have a previous clinical history of heart disease (20, 21). Based on coroner's data, only 21 per cent (6/29) of the men who died suddenly at work in services occupations had histories of heart disease or hypertension known to spouses or other survivors. Robinson 12 Service occupations, such as that of security guard, custodian or night watchman, may be taken as secondary jobs by persons who have left jobs as firemen, policemen, or other mentally or physically demanding or stressful occupations. Studies of the U.S. steel industry (19) suggested that health selection by management explained excess cardiovascular mortality observed for Janitors and plant protection workers. Some service occupations, such as security guards, may involve stressful, sedentary or solitary work. This may present special hazards for individuals who have left physically demanding or stressful jobs. A study of Finnish foundry workers (22) found that men who vent from heavy to light occupations had the highest risk of cardiovascular mortality. Occupational stress (23, 24) has been related previously to coronary heart disease incidence. Vigorous physical activity was reported in another study (25) in association with a two to threefold increased risk of sudden cardiac death. Solitary work may present hazards for individuals who may work alone or during evening hours. Sudden deaths at work, particularly fatal heart attacks, might be prevented by better management of stressful, sedentary or solitary work. For example, a portable, automatic signaling system could be devised for use by the solitary worker who may need rapid emergency first aid for a heart attack. An alert system would protect solitary workers against life-threatening situations. Surveillance of solitary workers may reduce the risk of sudden deaths at work. Three sudden deaths at work in our study involved firefighters. One was electrocuted when the boom of the cherry picker he was operating while fighting a fire hit a high voltage wire. The other two both died of heart Robinson 13 attacks in the fire station after fires. All three had trace amounts of carboxyhemoglobin found in their blood at the time of death. The role of exposure to combustion products and exertion in carboxyhemoglobin accumulation in firefighters has been documented in previous studies (26, 27). Cardiovascular and respiratory system diseases are recognized occupational diseases of firefighters (28). However, a mortality study of Boston firefighters concluded that their mortality rates were not strongly associated with firefighting as an occupation; although, the authors felt their survival experience may have been influenced by entry selection, ethnic derivation, or sociocultural attributes (29). Several occupational exposures have been shown to cause sudden death, such as nitroglycerin, fluorocarbons, carbon disulfide, and several other solvents listed in table 1. The extent of occupational exposure to these agents in our study population is unknown since no cardiotoxic agents were mentioned in coroner's records or on the death certificates as possible factors in the sudden cardiovascular work deaths. Future investigations of sudden deaths at work should seek information on any potential exposures to cardiotoxic agents used industrially. That the coroner's records identified more sudden deaths at work than death certificates did was observed in a previous study (14). The county coroner's records may represent an underutilized surveillance system for sudden deaths which occur at work. Sudden natural deaths at work can not usually be recognized or identified from death certificates. There is no public health office which receives reports of heart attacks which occur at Robinson 14 vork. The Occupational Safety and Health Administration routinely receives reports of fatal work injuries. It does not routinely receive reports of all heart attacks at work. Heart attacks at work, which occurred twice as often as fatal injuries at work in Allegheny County, are only investigated by the Occupational Safety and Health Administration if they occurred under highly unusual circumstances or were clearly work-related. Fatal Injuries Work accidents involving nontransport vehicles were the biggest single cause of fatalities at work. Table 4 shows 19 per cent (13/68) of work fatalities involved nontransport motor vehicles. Falls from open vehicles, and backing up and running over operators not in the vehicle were the most common circumstances of injury. Seat belt use was not mentioned in the records. In two of the injuries, broken backup lights were cited. Proper maintenance and operation of nontransport motor vehicles could reduce the high rate of fatality associated with their operation. Baker reported road transport vehicles to be the single largest cause of fatalities on the job in Maryland in 1978, with nontransport vehicles being the second highest cause of worker deaths (14). In our study, the construction industry involved the largest number of workers killed and had the highest age-adjusted fatality rate of all industries; however, it represented only 5 per cent of the employed population. This industry was ranked third nationally in 1984 for fatal injuries at work by the National Safety Council (2). The construction Robinson 15 industry vas also ranked third by a death certificate study of fatal injuries at work in Texas (30). Fifty-four per cent (7/13) of the fatalities involving multiple injuries occurred in the construction industry in our study (table 5). Prevention strategies aimed within the construction industry, may be particularly appropriate and effective. Better worker safety training, education and equipment may be needed to prevent potentially catastrophic incidents of multiple fatality and injury at vork. Table 5 shows at least 25 per cent (17/68) of fatalities involved the injury or death of fellow workers or innocent bystanders who were not included among the 68 fatal work injuries we studied. For example, multiple injuries occurred in our study when a co-worker tried to rescue the victim. In three of the 17 incidents, persons attempting to help burn victims were severely burned themselves. Two separate incidents were reported of rescue workers, including one policeman, who became unconscious and were asphyxiated after descending without a respirator into a pit or a minehole. Death rates for occupational fatalities are gathered and calculated in a variety of ways. Although increased or decreased rates have been reported for different geographic areas, the rates themselves may not be truly comparable. Their incomparability may obscure any real differences between the rates or make any apparent differences difficult to interpret. For example, our annual age-adjusted fatal injury rate in Allegheny County 1979-1982 was 5.5 per 100,000 actively employed white males. The National Safety Council estimated the age-adjusted U.S. death rate for fatal injuries at work to be 13 per 100,000 workers in 1980 (2). A different study of fatal occupational injuries |