Death Records Search

 

Of the 522 reports evaluated, we found that 78 were misclassified on the CMMF. Some misclassifications were coding errors; others were judged so by the evaluators on the basis of extensive review of the investigation reports. Misclassifications included homicides, suicides, poisonings by other gases (e.g., natural gas, gasoline), and burn victims, among others. Among the 444 authenticated cases of unintentional CO poisoning:, 174 (39%) deaths were caused by vehicular, fire-related, or unknown sources. The remaining 270 deaths were caused by residential and other nonvehicular sources. To determine death rates and relative risks, we compared numbers and percentages of CO-related deaths to population and demographic data from both the California Census and the Department of Finance estimates. Results Demographics. The demographic characteristics of the 270 cases of deaths caused by nonvehicular sources are presented in Table 1. The overall male to female ratio was 2.2; there were 187 (69%) male and 83 (31%) female cases. Consistent with previous findings, there were proportionately more deaths among African Americans than other ethnic groups (i.e., 15.9% of cases, compared with 7.4% of African Americans in the state population during the study period). The death rate among African Americans was more than 3 times higher than for Hispanics and Asians--who had the lowest death rates. Death rates were high among people aged 20-40 y and 60+ y, and the highest rate was found for the elderly who were between the ages of 70 y and 79 y: Table 1.--Demographic Information on Unintentional Nonvehicular Carbon Monoxide (CO) Decedents in California During 1979-1988   No. of CO deaths Percentage of Parameter (1979-1988) CO deaths Sex Male 187 69.3 Female 83 30.7 Ethnic group Caucasian 177 65.3 Hispanic 37 13.7 African American 43 15.9 Asian 11 4.1 Native American 3 1.1 Age (y)(*) < 10 22 8.1 10-19 20 7.4 20-29 66 24.4 30-39 50 18.5 40-49 21 7.8 50-59 27 10.0 60-69 34 12.6 70-79 23 8.5 80-89 7 2.6 Percentage of Death rate in Parameter state census given population Sex Male 49.60 1.62 x [10.sup.-5] Female 50.40 0.65 x [10.sup.-5] Ethnic group Caucasian 67.00 1.17 x [10.sup.-5] Hispanic 19.00 0.67 x [10.sup.-5] African American 7.50 2.30 x [10.sup.-5] Asian 4.20 0.65 x [10.sup.-5] Native American 0.84 1.78 x [10.sup.-5] Age (y)(*) < 10 15.58 0.59 x [10.sup.-5] 10-19 14.06 0.66 x [10.sup.-5] 20-29 17.97 1.34 x [10.sup.-5] 30-39 17.01 1.23 x [10.sup.-5] 40-49 11.03 0.80 x [10.sup.-5] 50-59 9.15 1.17 x [10.sup.-5] 60-69 8.05 1.76 x [10.sup.-5] 70-79 4.92 2.02 x [10.sup.-5] 80-89 2.22 1.24 x [10.sup.-5]

 

(*) One case with unstated age was excluded from the table. Seasonal variation and longitudinal trend. Seasonal variations were clearly observed in this study, as well as in earlier investigations.[1-6,10] Three-quarters (75.2%) of the deaths occurred during a 5-mo period: November to March. The highest number of deaths occurred in December (n = 65), and the lowest number of deaths occurred in August (n = 4) (Fig. 1). A sharp increase in number of deaths per month was very apparent during the months that approached winter. The greatest number of deaths occurred in December, which is not necessarily the coldest month of the year, but most likely the month when most combustion appliances are initially fully used during the heating season. During the study years, the annual number of unintentional deaths caused by CO poisoning from nonvehicular sources varied from 16 to 39, whereas the incidence rates ranged from 0.59 to 1.5 deaths/100,000 [multiplied by] y. Although a decreasing trend was observable over the 10-y period, the trend did not reach statistical significance, perhaps the result of the small sample size. [Figure 1 ILLUSTRATION OMITTED] Association with types of residences. As is shown in Table 2, the majority of nonvehicular CO-related deaths occurred in residential settings: 89 occurred in single-family homes, and 72 took place in multifamily dwellings. Given the similarity in building structure, we included the 12 deaths that occurred in hotels or motels in the multiunit category, Twenty victims were found in mobile homes/trailers, and 62 were identified in temporary shelters (e.g., cabins, shacks, converted garages, tents, campers, recreational vehicles, vans, houseboats). A total of 25 deaths occurred in nonresidential locations (e.g., bakeries, clubhouses, root cellars, mines). Locations of demise were unavailable for 2 deaths. Table 2.--Number and Percentage of Unintentional Nonvehicular Carbon Monoxide (CO) Deaths in California During 1979-1988, by Residence Type, and Percentages and Relative Risks of Various Residence Types   No. of Percentage CO deaths of CO Location (1979-1988) deaths Single-family homes 89 33.0 Multi-unit dwellings 72 26.7 Mobile/trailer homes 20 7.4 Temporary shelters (e.g., cabins, tents) 62 23.0 No. used as permanent residence 19 7.0 Other than living quarters (e.g., mines) 25 9.3 Unknown 2 0.7 Total 270 100.0 Percentage of households Relative Location in census(*) risk Single-family homes 69.9 1 Multi-unit dwellings 26.4 2.1 Mobile/trailer homes 3.3 4.7 Temporary shelters (e.g., cabins, tents) -- -- No. used as permanent residence 0.5 29.8 Other than living quarters (e.g., mines) NA NA Unknown NA NA Total 100.0([dagger])

 

Note: NA = not available. (*) For the census data, we took means for the 1980 and 1990 census to give approximate percentages for 1985. ([dagger]) as a consequence of rounding, percentages of all categories do not total up to 100%. Although the largest number of deaths (89) occurred in single-family homes, people who lived in single-family houses actually were at the lowest risk for unintentional CO-related deaths. On the basis of an extrapolation for the 1985 census, approximately 70% of the California population lived in single-family houses--in which only 33% of the unintentional CO deaths occurred. In contrast, 26% of Californians lived in multiunit dwellings, where approximately 27% of the deaths occurred. Approximately 3.3% of the California population lived in mobile homes and/or trailers, where 7.4% of the deaths took place. If we take the population who lived in single-family houses as the reference group, the relative risks for living in multiunit dwellings and mobile homes/trailers are approximately 2 and 5, respectively. The most striking observation was that approximately one-fourth (23%) of unintentional CO deaths occurred in temporary shelters. The common types of temporary shelters were campers, recreational vehicles, vans, cabins, shacks, tents, and houseboats--all of which were less likely to be served by standard natural gas or electricity than in single-family or multiunit dwellings. Among the 62 deaths in temporary shelters that could be classified, 24 victims stayed temporarily, whereas 19 stayed in the temporary shelters permanently. If we assume that the 19 victims known to have died while using temporary shelters as permanent residences belonged to the 0.5% of the California population that used temporary shelters as permanent residences, the relative risk is 30. If we base our analysis on the 43 African Americans in population, 14 (32.6%) deaths occurred in single-homes, 2 (4.7%) in mobile homes, 20 (46.5%) in multiunit dwellings, and 7 (16.3%) in temporary shelters and other place. Among the total study group of 270 victims, 89 (33%) deaths occurred in single-family homes, 72 (26.7)% in multiunit dwellings, 20 (7.4%) in mobile homes/trailers, and 89 (33%) in temporary shelters and other places. If we contrast African Americans with the total population, the same proportion of deaths occurred in single-family homes--our reference group. There were relatively more deaths of African Americans in multiunit dwellings, but fewer deaths in temporary shelters, than in the total study group. Association with combustion appliances and combustion fuels. Two hundred and sixteen deaths (in 157 episodes) were caused by conventional heating and cooking appliances. Charcoal grills and hibachis caused 58 deaths in 46 episodes. One death in 1 episode in which a wood-burning grill (not a charcoal grill) was used was reported in previous studies,[9,12] but it was not included in this analysis. In at least 22 of the grill/hibachi deaths, individuals reported that the appliance was used for heating, instead of cooking. As shown in Table 3, the largest number of deaths (n = 80) and episodes (n = 52) occurred in residences with wall or floor heaters. Portable room heaters, stoves, and fireplaces were associated with 47 deaths, among which 15 resulted from the misuse of stoves for heating. Central forced-air furnace and room heaters with flues--the most common gas heating appliances--claim a 58% market share in California. Only 6.9% (15 deaths) of the total deaths were caused by the malfunctioning or misuse of the aforementioned common heating appliances.

 

Table 3.--Number, Percentage, and Relative Risk of Unintentional Nonvehicular Carbon Monoxide (CO) Deaths in California During 1979-1988, by Heating Appliance   No of CO deaths Percentage Heating appliance (1979-1988) of CO deaths Floor, wall, or room heaters (unknown portion vented) 80 37.0 Portable room heaters and stoves used for heating and fireplace(*)([dagger]) 47 29.2 (mostly unvented) Portable room heaters 33 Stoves used for heating 15 Fireplaces 1 Central forced air or room heaters with flue (vented) 15 6.9 Hibachis/charcoal grills 58 26.9 Steam, hot water, or electric 0 0 Total 216 100.0 Percentage of households Heating appliance in 1980 census Relative risk Floor, wall, or room heaters (unknown portion vented) 24 13 Portable room heaters and stoves used for heating and fireplace(*)([dagger]) 4(*) 45 (mostly unvented) Portable room heaters Stoves used for heating Fireplaces Central forced air or room heaters with flue (vented) 58 1 Hibachis/charcoal grills --([dagger]) --([dagger]) Steam, hot water, or electric 14 0 Total 100 --   (*) In the 1980 census questionnaire, portable room heaters, stoves, and fireplaces are grouped as one choice for the kind of heat used most in the respondents' living quarters. To calculate the relative risk, we grouped heating sources to be consistent with this census category. ([dagger]) No census data available. The percentages of deaths caused by various heating appliances are compared in Table 3 with the percentages of different heating appliances used in California households (according to the 1980 census). If central furnace/room heaters with flues are taken as the reference group, the relative risk for wall/floor heaters, both unvented and vented, is 13. We used a conservative method that excluded undetermined causes to calculate the relative risk for portable heaters and stoves (49). These are typically unvented by design or use. The total number of deaths from charcoal grills/hibachis was 58. Given that charcoal grills/hibachis were not listed in the census questionnaire as heating appliances, we did not calculate a relative risk for this appliance. The risk of deaths was extremely large, given the likelihood that only a small percentage of the population used charcoal grills/hibachis for heating. The numbers and percentages of CO deaths, by heating fuel, are presented in Table 4. The 1980 census revealed that the majority of California homes used natural gas as their main heating fuel (79.4%). A total of 126 deaths occurred in homes in which natural gas from utilities was used. We used natural gas from utilities as a reference group, and we compared relative risks of heating fuels. Bottled/tank/liquefied petroleum (LP) gas caused 30 CO-related deaths. To compare the risk of using charcoal as the main heating fuel, we confirmed that only 22 cases (included in Table 4) used charcoal for heating. Among the other 27 deaths from charcoal grills/hibachis, the use of the appliance could not be determined. We confirmed that a total of 22 deaths were caused by charcoal (as heating fuel), thus constituting 12.1% of the nonvehicular deaths caused by heating fuels. Given that the percentage of households that used charcoal as the main heating fuel was unknown, a relative risk could not be calculated, inasmuch as it has only a small market share of 2.7% in California, the relative risk of bottled/tank/LP gas was approximately 7, compared with natural gas from utilities.

 

Table 4.--Number and Percentage of Unintentional Nonvehicular Carbon Monoxide (CO) Deaths in California During 1979-1988, by Main Heating Fuel   No. of CO deaths Percentage Heating fuel (1979-1988) of CO deaths Natural gas from utilities (i.e., utility gas) 126 69.2 Bottled, tank, and liquefied petroleum gas (butane and propane) 30 16.5 Electricity -- Fuel oil (kerosene) 1 0.6 Charcoal 22 12.1 Coal -- -- Wood 1 0.6 Other -- -- Unknown 2 1.1 None -- -- Percentage of households Relative Heating fuel in 1980 census risk Natural gas from utilities (i.e., utility gas) 79.44 1 Bottled, tank, and liquefied petroleum gas (butane and propane) 2.74 6.9 Electricity 15.49 0 Fuel oil (kerosene) 0.34 1.9 Charcoal Unknown Unknown Coal 0.01 -- Wood 1.96 0.32 Other 0.11 -- Unknown -- -- None 0.83 --   (*) For the census data, we obtained means for the 1980 and 1990 census to give approximate percentages for 1985. Discussion Comparison with other studies of unintentional CO deaths. Prior studies on CO poisonings and deaths have been based on death certificates,[1-6] on data from medical examiners,[3-6] and on reports from hyperbaric chambers and laboratories.[6] The range of ICD-9 E-codes that are used for describing the causes of death varied from study to study. Cook, Simon, and Hoffman[6] and Baron, Backer, and Sopher[3,4] considered deaths mostly from lCD-9 E codes 867-868.9, which cover accidental poisonings by gas or CO. There were, however, too few recorded deaths available for elaborate statistical analyses. Likewise, the national study by Cobb and Etzel[2] included more than 11,000 deaths listed under many E-codes--including motor vehicle, water, and air transport accidents--with CO poisoning as an underlying cause of demise. In the Gas Research Institute's topical report, investigators tried to duplicate the findings from the Cobb study, but their numbers were consistently lower by 10%.[1] The report's authors attributed the difference to coding variations. The likelihood of misclassification is further supported by the recent study by Moolenaar, Etzel, and Parrish,[5] who looked at cases identified from death certificates that were confirmed by the data of medical investigations, and they found that 12% of the cases were misclassified. This result is similar to what we found in our study: 15% of the cases identified from death certificates were misclassified.

 

Although both the present and earlier studies included different sets of E-codes, common trends were found in the demographic data and seasonal variations, in all the studies, there were proportionally more males, African-Americans, young adults (in their late teens to early 20's), and elderly ([is greater than or equal to] 60 y of age) individuals among the decedents. Also, the greatest number of CO-related deaths occurred during the winter months. However, the CO sources identified varied, reflecting the range of E-codes included in different studies. The death rate of unintentional CO poisoning in California is among the lowest in the United States. Nevertheless, the total number of 444 cases collected during a 10-y period allows a close investigation of factors associated with these deaths. The national data showed that the majority of CO-related deaths (i.e., 57%) were caused by vehicular sources, whereas in California there were proportionally fewer deaths from vehicular sources (41%). The large number of deaths in California caused by the misuse of charcoal grills/hibachis for heating was reported earlier.[12] Contrast with vehicular sources. The results of the present study showed that there were proportionally fewer males among the CO-related nonvehicular deaths than among the CO-related vehicular deaths. The male/female ratio was 2.3 for nonvehicular sources and 3.7 for vehicular sources. There were also far more multiple victims involved in the episodes caused by nonvehicular sources than in those caused by vehicular sources. Six episodes involving 4 or 5 victims or survivors each were all related to nonvehicular sources. In contrast, none of the 129 episodes caused by vehicular sources involved more than 2 victims, and 94% involved only 1 victim.[9] Factors associated with elevated CO concentrations in California Residential Study. In their study of indoor air quality of California residences, Colome et al.[13] identified risk factors for high CO exposure that were similar to those identified for CO deaths in this study. In a statewide survey of various indoor air pollutants, CO was monitored in 277 randomly selected homes. Elevated CO concentrations occurred most often in (a) smaller multiunit dwellings, (b) households in which gas ranges were used for cooking, and (c) dwellings heated by gas wall furnaces. Homes with low CO concentrations were more likely representative of single-family structures and dwellings with forced-air furnaces and electric cooking appliances. Long-term trends of gas heating appliances. During recent years, deaths nationally decreased steadily for both vehicular sources and nonvehicular sources. A decreasing trend was observed in California between 1979 and 1988, but it was not significant statistically. Investigators have attributed the decrease of CO-related deaths from vehicular sources to the improved exhaust systems of motor vehicles.(2) The drop of CO-related deaths from nonvehicular sources may be explained by improvements in residential heating systems and by homes built during the 1940s, more homes built since the 1970s have been equipped with central forced-air furnaces or room heaters with flues for heating. According to the census of 1980, approximately 40% of the California dwellings built between 1940 and 1949 used central forced-air furnaces or room heaters with flues, whereas these heating appliances were found in 75% of the dwellings built between 1975 and 1980. For the same periods, the percentage of wall, floor furnaces, and room heaters without flues has decreased from 43% to 5%.

 

Carbon monoxide-related morbidity and mortality. In a study on unintentional CO poisonings in Colorado, Cook et al.[6] determined a ratio of nonfatal to fatal poisonings of 4.6 (807/174). The authors indicated that nonfatal CO poisonings were likely underreported, because not all clinical laboratories were contacted. Their conclusion implied that: the ratio of nonfatal/fatal CO poisonings was even higher than their estimate of 4.6. Ratios varied greatly among different CO sources. For vehicular sources the ratio was 3, whereas the ratio was 19 for nonvehicular sources (e.g., furnaces). The authors also indicated that there was no reporting bias to affect the distribution of CO poisonings with respect to different sources. This conclusion suggests that in Colorado, for every 1 death of CO poisoning from a furnace, there could be more than 20 nonfatal CO poisonings. In the U.S. Consumer Product Safety Commission's 1997 annual report it was noted that between 5,700 and 10,000 people were seen in emergency rooms for suspected CO poisoning between 1992 and 1994, whereas there were only about 200 CO-related fatalities per year during the same period.[11] More studies are needed to clarify the relationship between morbidity of CO poisoning and different CO sources. Clarification is important not only for estimating morbidity of CO poisoning, but also in setting priorities for prevention. Carbon monoxide alarms. Required installation of CO alarms in residences for the prevention of CO poisoning is a controversial topic. Clearly, reliable properly installed devices can provide warning of dangerous levels of CO. Furthermore, alarms, which are now widely available, are reasonably adequate for use by the general population, even though they may be subject to interference from high levels of certain gases (e.g., alcohol, ammonia). Previously, problems with CO detectors giving "false alarms" were largely a consequence of devices that sounded alarms at CO levels that were not acutely hazardous. For example, during the fall and early winter of 1994, fire officials in Chicago, Illinois, responded to more than 8,500 CO alarms.[14] Most problems stemmed from improper placement of detectors, which were too close to gas appliances, and perhaps from weather inversions, coupled with, CO alarms that were too sensitive. Following the detector problem in Chicago, Underwriter's Laboratories standards that apply to CO alarms were revised and new standards were adopted. The devices available today are no longer triggered by CO levels in the ambient range or by short-term peaks, but, instead, are triggered only by sustained higher levels of acute concern. In the present study, 161 (60%) of the 270 deaths occurred in single-family homes, and multiunit dwellings that could possibly have been protected by CO alarms. Despite progress with CO alarms and the desirability of their use in residences, it must be emphasized that the devices offer only limited secondary protection against CO poisoning. The best protection is through control of the primary CO sources.

 

Conclusion Major risk factors for unintentional CO deaths in California from residential and other nonvehicular sources were identified in this analysis of a 10-y record of coroners' examination reports. As was the case in other studies, death, rates were highest in winter. Risk was highest among males, African Americans, and the elderly; risk communication efforts should target these groups. Furthermore, the analysis indicates that individuals who live in mobile home/trailers, multiunit residences, and temporary shelters (e.g., tents, campers) are at substantially greater risk of CO poisoning than residents in single-family homes. These higher-risk individuals should also be targeted for educational outreach, and the general public should receive effective warnings about the hazards of unvented or poorly vented combustion heating appliances, charcoal grills/hibachis used in confined areas, and gas stoves used for heating. Use of CO alarms offers some secondary protection in permanent residences, but the best protection is control of CO sources. We acknowledge the support of Ed Becker, David Behrens, and Irwin, Billick, as well as the contributions of Steven Hayward (deceased), Yu-Lin Chang, Christina Fu, Fan-Yen Huang, and Elizabeth Huang. We especially thank the California County Medical Examiners/Coroners and their staff for their cooperation. This work was supported under the auspices of the Public Health Foundation of Los Angeles County, Inc., by funds provided by Southern California Gas Company and Pacific Gas & Electric Company. This study was not supported by the U.S. Environmental Protection Agency and does not necessarily represent the views of the same. Submitted for publication February 15, 1999; revised; accepted for publication October 6, 1999. Requests for reprints should be sent to Kai-Shen Liu, Ph.D., M.P.H., Environmental Health Laboratory Branch, California Department of Health Services, 2151 Berkeley Way, Berkeley, CA 94704. References [1.] Koontz MD, Niang LL. Unintentional Carbon Monoxide-Related Deaths Between 1979 and 1993. Germantown, MD: Gas Research Institute, June 1997. [2.] Cobb N, Etzel RA. Unintentional carbon monoxide-related deaths in the United States, 1979 through 1988. JAMA 1991; 266:659-63. [3.] Baron RC, Backer RC, Sopher IM. Unintentional deaths from carbon monoxide in motor vehicle exhaust: West Virginia. Am J Public Health 1989; 79:328-31. [4.] Baron RC, Backer RC, Sopher IM. Fatal unintentional carbon monoxide poisoning in West Virginia from nonvehicular sources. Am J Public Health 1989; 79:1656-58. [5.] Moolenaar RL, Etzel RA, Parrish RG. Unintentional deaths from carbon monoxide poisoning in New Mexico, 1980 to 1988--a comparison of medical examiner and national mortality data. West J Med 1995; 163:431-34. [6.] Cook M, Simon PA, Hoffman RE. Unintentional carbon monoxide poisoning in Colorado, 1986 through 1991. Am J Public Health 1995; 85:988-90.

 

[7.] Caplan YH, Thompson BC, Levine B, et al. Accidental poisonings involving carbon monoxide, heating systems, and confined spaces. J Forensic Sci 1986; 31:117-21. [8.] Centers for Disease Control. Unintentional carbon monoxide poisonings in residential settings--Connecticut, November 1993-March 1994. JAMA 1995; 274:1579-81. [9.] Girman JR, Chang Y-L, Hayward SB, et al. Causes of unintentional carbon, monoxide deaths in California. West J Med 1988; 168:158-65. [10.] Centers for Disease Control and Prevention. Deaths from motor-vehicle-related unintentional carbon monoxide poisoning--Colorado, 1996, New Mexico, 1980-1995, and United States. Morbidity and Mortality Weekly Report 1979-1992, 1997; 45(47): 1029-32. [11.] Ault K. Estimates of Non-Fire Carbon Monoxide Poisoning Deaths and Injuries. Washington, DC: United States Consumer Product Safety, December 1997. [12.] Liu K-S, Girman JR, Hayward SB, et al. Unintentional carbon monoxide deaths in California from charcoal grills and hibachis. J Expos Anal: Environ Epidemiol 1993; 3(Suppl 1):143-51. [13.] Colome S, Wilson AL, Tian Y. California residential indoor air quality. II. Carbon Monoxide and Air Exchange Rate: A Univariate and Multivariate Analysis. Irvine, CA: Gas Research Institute, Pacific Gas and Electric Company, San Diego Gas and Electric Company amd Southern California Gas Company; July 1994. [14.] Raskin F, Veasey D (Eds). Carbon monoxide detectors' Alarm Vex Chicago Official. Indoor Pollut News 1995; 8:3. KAI-SHEN LIU MARIA KATRINA PAZ(*) PETER FLESSEL JED WALDMAN JOHN GIRMAN([dagger]) Environmental Health Laboratory Branch California Department of Health Services Berkeley, California (*) Maria Katrina Paz is currently affiliated with the Johns Hopkins University School of Public Health, Department of Mental Hygiene, Baltimore, MD. ([dagger]) John Girman is currently affiliated with the Indoor Air Division, U.S. Environmental Protection Agency, Washington, DC.

-------------------------------------------------------------------------------

COPYRIGHT 2000 Heldref Publications in association with The Gale Group and LookSmart. COPYRIGHT 2001 Gale Group

-------------------------------------------------------------------------------