Postservice Mortality of US Air Force Veterans Occupationally Exposed to Herbicides in Vietnam: 15-Year Follow-up

Transcription

1 American Journal of Epidemiology Copyright 1998 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 148, No. 8 Printed in U.S.A. Postservice Mortality of US Air Force Veterans Occupationally Exposed to Herbicides in Vietnam: 15-Year Follow-up Joel E. Michalek, 1 Norma S. Ketchum, 1 and Fatema Z. Akhtar The US Air Force continues to assess the mortality of veterans of Operation Ranch Hand, the unit responsible for aerially spraying herbicides in Vietnam. The authors of this study found that the cumulative all-cause mortality experience of these veterans was not different from that expected (standardized mortality ratio (SMR) = ). Overall, cause-specific mortality did not differ from that expected regarding deaths from accidents, cancer, or circulatory system diseases, but the authors found that there was an increased number of deaths due to digestive diseases (SMR = 1.7, 95% confidence interval (Cl) -.). When analyzing by military occupation, they found an increase in the number of deaths caused by circulatory system diseases (SMR =, 95% Cl -.) among enlisted ground personnel, the subgroup with the highest dioxin levels. Most of the increase in the number of deaths from digestive diseases was caused by chronic liver disease and cirrhosis, and more than half of the increase in the number of deaths from circulatory system diseases was a result of atherosclerotic heart disease. In the subgroup of Ranch Hand veterans who had survived more than 0 years since their military service in Southeast Asia, the authors found no significant increase in the risk of death due to cancer at all sites (SMR = ) and a nonsignificant increase in the number of deaths due to cancers of the bronchus and lung (SMR = 1.). Am J Epidemiol 1998;148: dioxins; herbicides; mortality Debate is ongoing concerning the susceptibility of humans to the toxic effects of,,7,8-tetrachlorodibenzop-dioxin (dioxin), the contaminant found in Agent Orange and other herbicides sprayed during the Vietnam War. Several studies have examined the postservice mortality experience of Vietnam veterans (1-7). A study of US Army Chemical Corps veterans () and a study of Australian Army veterans (4) reported an increased risk of death due to accidents and digestive diseases but no increase due to cancer; however, a study of women veterans () found an increased risk of deaths caused by specific cancers. The relation between dioxin exposure and mortality has been studied in industrial populations (8-1). A study of US chemical workers who were exposed to dioxin reported an increased cancer mortality and, in workers who survived more than 0 years since their exposure, an increased mortality due to soft-tissue Received for publication January 1, 1998, and accepted for publication March 5, Abbreviations: Cl, confidence interval; ICD-9, International Classification of Diseases, Ninth Revision; SMR, standardized mortality ratio. 1 Air Force Research Laboratory, Brooks Air Force Base, TX. Vista Technologies, Inc., San Antonio, TX. Reprint requests to Dr. Joel E. Michalek, AFRL/HEDB, 606 Doolittle Road, Building 807, Brooks Air Force Base, TX sarcoma and cancer of the respiratory system (9). However, a study of workers who were exposed as a result of a trichlorophenol process accident found no association between dioxin exposure and cancer mortality (11). A study of workers at a herbicide-producing plant in Hamburg, Germany, reported a dose-dependent relation between dioxin exposure and mortality caused by cancer and ischemic heart disease (1). The Air Force Health Study is a 0-year prospective study of the health (1, 14), mortality (15), and reproductive outcomes (16-18) of veterans of Operation Ranch Hand, the unit responsible for the aerial spraying of herbicides in Vietnam. The study began in 198 and will conclude in 00. This paper updates our first report (15) by summarizing current all-cause and cause-specific postservice mortality among veterans of Operation Ranch Hand. MATERIALS AND METHODS Population definition and the process by which mortality was determined are discussed elsewhere (15). Briefly, we contrast cumulative Ranch Hand (n = 1,61) mortality through December 1, 199 (verified as of December 1995) with that expected on the basis of the mortality of a comparison population of 19,080 US Air Force veterans who flew or serviced C

2 Mortality in Veterans Exposed to Herbicides 787 cargo aircraft in Southeast Asia during the same time that the Ranch Hand unit was active in Vietnam ( ). Comparison veterans were not involved in spraying herbicides in Vietnam but are demographically similar to Ranch Hand veterans. This report updates our previous analysis (15), for which the dates of military service in Southeast Asia were unavailable for 179 comparison subjects. Rather than exclude those veterans, we replaced the missing dates with randomly generated dates. By using this method, we were able to include all comparison veterans in our previous report (15). Since then, we retrieved the actual service dates for 148 of these 179 veterans from military personnel records but were unable to find dates for 1 veterans because the records were missing; therefore, we excluded those veterans from this report. Additionally, we excluded one comparison veteran who was erroneously included in our last report and added 11 comparison veterans whom we found and verified since our last report. Therefore, the total number of comparison subjects is now 19,080, 1 fewer than reported previously (15). All Ranch Hand and comparison veterans are male. The number of veterans at risk and the number of person-years are summarized in table 1 by military occupation (pilots and navigators, administrative officers, enlisted flight engineers, and enlisted ground personnel). All pilots and navigators were officers. We used military occupation as a surrogate to adjust for differences in socioeconomic characteristics and inferred dioxin exposure among Ranch Hand veterans. Most enlisted personnel were not college educated, and most officers were college graduates. Furthermore, dioxin assay results (1) suggest that among Ranch Hand veterans, enlisted personnel were more heavily exposed than officers and, that among enlisted veterans, ground personnel were more heavily exposed than flight engineers. We classified underlying causes of death in accordance with the rules and conventions outlined in the International Classification of Diseases, Ninth Revision (ICD-9) (19). Veterans who survived until December 1, 199, the cutoff date for these analyses, contributed the time, in years, between the date on which they entered follow-up (the date they began their service in Southeast Asia) and the cutoff date; those known to have died before the cutoff date contributed the time, in years, between the date on which they entered follow-up and their date of death. We computed the standardized mortality ratio, the ratio of the observed to the expected number of deaths (0). We adjusted this ratio by stratifying on age (in 5-year intervals), follow-up time (in 5-year intervals), and military occupation and by summing the observed and expected numbers of deaths across strata. The expected number of deaths within each stratum was the product of the number of Ranch Hand person-years and the comparison death rate. We used the Mid-P method derived from a Poisson model to compute confidence intervals for the standardized mortality ratio (1). Although the confidence interval formula was based on the assumption that the comparison death rates were constants, the comparison rates were subject to random variation; therefore, the widths of our confidence intervals were slightly understated. In some tables in this report, because of small numbers of veterans, we combined the data on pilots, navigators, and administrative officers into a single occupational category called "officers." We were not able to adjust for race because too few blacks (6.1 percent) were included. We tabulated the number of observed and expected cancer deaths by length of survival time since military service in Southeast Asia (<0 years, ^0 years) and by military occupation. The most recent spraying missions took place in Because we defined the end of the follow-up period as December 1, 199, all veterans who survived to the end of that period had survived >0 years since their first tour in Vietnam. All veterans were included in the <0-year analysis, whereas only those veterans who had survived 0 years since their first tour of duty in Southeast Asia were included in the s0-year analysis. The statistical power of this study was limited by the size of the Ranch Hand group. Based on two-sided testing, this study had a power of 99.8 percent for TABLE 1. Number of US Air Force veterans and person-years at risk, by military occupation: Air Force Health Study Military occupation Total Operation Ranch Hand veterans (no.) At risk ,61 Person-years 11, ,55 14,845 1,94 Comparison veterans (no.) At risk 5,4 84,89 10,75 19,080 Person-years 14,075 7,46 71,90 77, ,79

3 788 Michalek et al. detecting a relative risk of for all-cause mortality. Regarding deaths caused by cancer, the power to detect a relative risk of and.0 was 77.4 and 99.6 percent, respectively. This study had no power to detect small or moderate increases in Ranch Hand death rates specific to rare cancers, such as soft-tissue sarcoma and lymphoma. RESULTS Demographic characteristics of all veterans are presented in table. Birth year and race were similar for Ranch Hand and comparison veterans. On a percentage basis, more Ranch Hand veterans (5.0 percent) than comparison veterans (7.5 percent) were pilots or navigators, and more comparison veterans (56. percent) than Ranch Hand veterans (46.6 percent) were enlisted ground personnel. We were unable to adjust for smoking, a risk factor for cardiovascular disease, or for alcohol consumption, a risk factor for liver disease, because risk factor information was available only for the subgroup of veterans who participated in the medical follow-up arm of the study (1). Ranch Hand mortality data are summarized in table. Overall, the observed and expected numbers of Ranch Hand deaths did not differ significantly (standardized mortality ratio (SMR) =, 95 percent confidence interval (CI) ). Because of small numbers, we could not analyze the deaths caused by suicide, homicide, infectious or parasitic diseases, endocrine diseases, or respiratory diseases or those resulting from ill-defined or unknown causes. There were no significant excesses of deaths caused by cancer (SMR = ) or circulatory diseases (SMR = ). The number of deaths caused by digestive diseases (observed number = 9) was greater than the expected number (5.1) (SMR = 1.7, 95 percent CI -.). Data on all-cause mortality and mortality due to cancer and circulatory diseases are summarized in TABLE. Demographic characteristics of US Air Force veterans exposed to herbicides in Vietnam: Air Force Health Study Birth year Median Range Characteristic Black race (%) Military occupation (%) Operation Ranch Hand veterans Comparison veterans table 4 by military occupation. Within-stratum relative risks of death from any cause ranged from to, and there were no significant stratum-specific differences between the observed and the expected numbers of deaths. The risk of death from cancer was not increased among pilots and navigators (SMR = ), enlisted flight engineers (SMR = ), or enlisted ground personnel (SMR = 0.8). The risk of death caused by diseases of the circulatory system was increased among Ranch Hand enlisted ground personnel (SMR =, 95 percent CI -.). Data on cancer mortality by survival time since service in Southeast Asia (<0 years, ^0 years) and by military occupation are summarized in table 5. Eleven of the 0 Ranch Hand cancer deaths (7 percent) occurred within the first 0 years after the start of service, whereas the remaining 19 deaths (6 percent) occurred at least 0 years afterward. The overall risk of death from cancer was not significantly increased among Ranch Hand veterans who survived at least 0 years after service in Southeast Asia (SMR =, 95 percent CI -1.6), and the risks among officers (SMR = ), enlisted flight engineers (SMR = 1.), and enlisted ground personnel (SMR = ) were not significant. The risk of death from cancer among Ranch Hand veterans within 0 years of their military service in Southeast Asia was decreased in all three occupational strata, although none of the deficits was significant. For each military occupation, we enumerated cancer deaths by the primary anatomic site and by the length of survival time (data not shown). Among Ranch Hand veterans who survived at least 0 years since their service in Southeast Asia, nine deaths were caused by cancers of the bronchus and lung (expected number = 7.) (SMR - 1., 95 percent CI 0.6-.). No other increases were apparent, although the numbers were too small to analyze. After serving in Southeast Asia, one Ranch Hand enlisted ground veteran died of malignant lymphoma.9 years later (expected number = 0.6), one Ranch Hand enlisted flight engineer died of multiple myeloma 4. years later (expected number = 0.8), and one Ranch Hand officer died of soft-tissue sarcoma 19.8 years later (expected number = 0.). No Ranch Hand deaths were caused by non-hodgkin's lymphoma, Hodgkin's disease, myeloid leukemia, or lymphoreticulosarcoma. We tabulated the numbers of observed and expected deaths, by underlying cause, among enlisted ground personnel who died of circulatory diseases (table 6). More than half of the excess noted in table 4 was due to atherosclerotic heart disease (SMR = 1.4, 95 percent CI ). We also enumerated, by underlying

4 Mortality in Veterans Exposed to Herbicides 789 TABLE. Cause-specific and all-cause mortality of US Air Force veterans exposed to herbicides in Vietnam: Air Force Health Study Cause of death Infectious or parasitic diseases Cancer Endocrine diseases Circulatory diseases Respiratory diseases Digestive diseases Ill-defined or unknown Accident Suicide Homicide All causes ICD-9* codes , Observed Expected U % Cl*,t * ICD-9, International Classification of Diseases, Ninth Revision; SMR, standardized mortality ratio; Cl, confidence interval. i No Operation Ranch Hand deaths were caused by neoplasms of an uncertain or unspecified nature (expected number = 0.), diseases of the blood and blood-forming organs (expected number = 0.1), mental disorders (expected number = ), diseases of the nervous system and sensory organs (expected number = 1.6), diseases of the genitourinary system (expected number = 0.5), or congenital anomalies (expected number = 0.). TABLE 4. All-cause, cancer, and circulatory disease mortality, by military occupation, among US Air Force veterans exposed to herbicides in Vietnam: Air Force Health Study Cause of death All causes Cancer Circulatory diseases Military occupation Observed Expected * SMR, standardized mortality ratio; Cl, confidence interval % Cl*,t cause, the nine Ranch Hand deaths caused by digestive diseases (data not shown). Most of the increase noted in table was caused by chronic liver disease and cirrhosis (SMR =.1, 95 percent Cl -4.1) (data not shown). DISCUSSION In our evaluation of all-cause mortality, we found no differences between the observed and the expected numbers of Ranch Hand deaths. This overall result agrees with our previous analysis (15) and with the findings from other studies of veterans (1,, 4, 5). Without regard to survival time, there were fewer Ranch Hand deaths due to cancer than expected. This finding of an overall cancer decrement is inconsistent with the findings from some studies but is consistent with others. A study of US Army and Marine Vietnamera veterans found an increased number of deaths due to laryngeal and lung cancer (7), and a study of veterans from Michigan reported an increased number of deaths due to non-hodgkin's lymphoma (6). A study of women Vietnam veterans found an excess risk of pancreatic cancer and cancer of the uterine corpus (). In addition, studies of dioxin-exposed industrial cohorts have reported an increased cancer mortality. A US study of workers at 1 chemical plants that made

5 790 Michalek et al. TABLE 5. Cancer mortality, by number of years since service in Southeast Asia and military occupation, among US Air Force veterans exposed to herbicides in Vietnam: Air Force Health Study Military occupation Officers All personnel Observed Expected <0 years >0 years %CI*,t Observed Expected * SMR, standardized mortality ratio; Cl, confidence interval SMR 1. 95% Cl TABLE 6. Number of circulatory disease deaths, by underlying cause, among US Air Force enlisted ground personnel exposed to herbicides in Vietnam: Air Force Health Study Underlying cause Atherosclerotic heart disease ICD-9* codes 410,411,41,4140, 4148, 4149, 484,4409, 444 Observed 16 Expected OGO/ r-i* f Cardiomyopathy 454, 455, 480, Cerebrovascular disease 40, 41, 49, 440, 449, 46, Hypertensive disease 4019, 409, 47, 4411, Other circulatory diseases 989, 4151,4160, 4169, 40, 475, 479, 49, 499, 449,446, 4511, Total All codes shown above * ICD-9, International Classification of Diseases, Ninth Revision; SMR, standardized mortality ratio; Cl, confidence interval. dioxin-contaminated products (9) and a study of exposed German workers (10) found an excess mortality from cancer 0 or more years after the first exposure. A separate study of exposed workers in Hamburg, Germany, reported a dose-dependent relation between cancer mortality and exposure to dioxin (1). However, a study of occupational exposure to dioxin caused by a trichlorophenol process accident found no increased cancer rates in workers exposed only to dioxin (11). To examine mortality in a way that provides adequate time for cancer to develop, researchers generally identify subjects who survived at least 0 years since their exposure (9, 10). Our analysis found a slight but nonsignificant increase in the risk of death from cancer among Ranch Hand veterans who survived at least 0 years since their service in Southeast Asia and a nonsignificant deficit of deaths from cancer within 0 years of service. Within-stratum results among veterans who survived at least 0 years did not suggest a dioxin effect, because there was no increase in risk among Ranch Hand enlisted ground personnel (SMR = ), the subgroup with the highest dioxin levels. We studied cancer deaths according to the primary anatomic site and found a small increase in the number of deaths caused by cancer of the bronchus and lung among Ranch Hand veterans who survived at least 0 years since service in Southeast Asia, which was not significant but was consistent with an increase in respiratory cancer mortality in an industrial cohort more than 0 years after exposure (9). Additional follow-up is needed to determine whether these slight increases persist. There were too few deaths caused by any of seven selected cancers (malignant lymphoma, multiple myeloma, soft-tissue sarcoma, non- Hodgkin's lymphoma, Hodgkin's disease, myeloid leukemia, or lymphoreticulosarcoma) to analyze or interpret.

6 Mortality in Veterans Exposed to Herbicides 791 Among all Ranch Hand veterans, the risk of death from diseases of the circulatory system was not increased (SMR = ); however, among enlisted ground personnel, we found a nonsignificant excess (SMR = ), and nearly half of the increase was due to atherosclerotic heart disease (SMR = 1.4). As a group, these personnel have the highest current dioxin levels, and those levels correlate with skin contact with herbicides in Vietnam (). The increased risk of death caused by diseases of the circulatory system is consistent with the findings from three studies and inconsistent with two others. A study of Australian Army veterans (4) found increased circulatory system disease mortality in Vietnam veterans. A study of the Seveso, Italy, population after a chemical plant accident () reported an increased number of deaths from diseases of the circulatory system in the most contaminated area, and a study of exposed workers at a Hamburg, Germany, plant found a dose-related increased risk of death caused by ischemic heart disease (1). Two studies of US Army veterans found fewer than the expected number of deaths caused by circulatory system diseases (1, ). The suggested increased risk of death from circulatory system diseases among Ranch Hand enlisted ground personnel should be interpreted cautiously, because we were unable to adjust for known risk factors, such as smoking and family history. The number of Ranch Hand deaths caused by digestive diseases was small (observed number = 9) but borderline significantly increased (SMR = 1.7). Seven of these deaths were caused by chronic liver disease and cirrhosis. As with the increased risk of death due to circulatory diseases, this finding should be interpreted with caution, because we were unable to adjust for known risk factors, such as alcohol consumption. This increase in the number of deaths caused by digestive diseases is consistent with the results of the Australian Army study (4), which also found elevated digestive system disease mortality in Vietnam veterans, mainly related to alcohol consumption. However, a study of US Army veterans (1) found no significant difference between Vietnam veterans and non-vietnam veterans regarding digestive disease mortality. More recently, a study of a Taiwan population poisoned by cooking oil that was contaminated with heat-degraded polychlorinated biphenyls (PCBs) reported an increased mortality from chronic liver disease and cirrhosis (4). A study of pentachlorophenol manufacturing workers found an increased mortality from cirrhosis of the liver among the most exposed subgroup (5). The strengths of our study include a large comparison population that was demographically similar to the Ranch Hand group and a complete determination of the mortality status of all subjects. As a group, the Ranch Hand veterans were probably among the most herbicide-exposed cohorts of Vietnam veterans. Thus, our study offers the best available opportunity to address a hypothetical relation between dioxin exposure and mortality in Vietnam veterans. Our analysis of deaths caused by diseases of the circulatory system was limited by our inability to adjust for smoking and for family history of circulatory system diseases, and our analysis of digestive disease mortality was limited by our inability to adjust for alcohol consumption. Available information for the subgroup of Ranch Hand and comparison veterans who participated in the physical examinations that were given as part of this study (1, 14) suggests that Ranch Hand veterans do not differ substantially from comparison subjects in terms of smoking or alcohol consumption. Our analysis of deaths caused by cancer before 0 years or after 0 years since military service in Southeast Asia was limited because we could not examine risks that might occur after a longer delay. Exposure misclassification was also a potential weakness. Serum dioxin measurements in the subgroup of Ranch Hand veterans who have been examined revealed that 40 percent of them have background levels similar to those of comparison veterans (). We are uncertain about the exposure status of Ranch Hand veterans with low dioxin levels; some could have had elevated dioxin levels in Vietnam but their body burden decayed to background levels in the intervening time period, and some may not have had elevated levels during their service in the Ranch Hand unit. A study of skin exposure to herbicides in Ranch Hand enlisted veterans () showed that veterans assigned to administrative duties and those reporting no skin exposure had the lowest dioxin levels (approximately 60 percent had background levels) and that more than 75 percent of those reporting high levels of skin exposure had above-background levels of dioxin. Thus, it seems that some Ranch Hand veterans were probably unexposed or received minimal exposure to herbicides in Vietnam. More specific statements about Ranch Hand exposure are not possible based on available data. Our last mortality report (15), which assessed cumulative Ranch Hand mortality as of December 1, 1987, also found no difference between observed and expected mortality from all causes combined (SMR =, 95 percent CI ). In this report, we did not analyze deaths among enlisted ground personnel. By using the cutoff date from that report (December 1, 1987) with the current data, we found a nonsignificant increased risk of death caused by diseases of the

7 79 Michalek et al. circulatory system among enlisted ground personnel (SMR = 1.6, 95 percent CI -.6) based on 15 deaths and 9.4 expected. Currently, the risk of circulatory disease deaths among enlisted ground personnel is still increased, although not significantly (SMR =, 95 percent CI -.), based on 4 deaths and 16.1 expected. Since our earlier report, the number of Ranch Hand deaths caused by cancer remains fewer than expected, although the standardized mortality ratio has increased from 0.8 to. Our earlier analysis also found an increased risk of death due to digestive diseases (SMR =.5, 95 percent CI -5.4), which was based on 6 deaths and.4 expected. Currently, the risk of death from diseases of the digestive system is still increased (SMR = 1.7, 95 percent CI -.), although not significantly, based on 9 deaths and 5.1 expected. Thus, the patterns we now see are similar to those we observed in REFERENCES 1. Centers for Disease Control. Postservice mortality among Vietnam veterans. JAMA 1987;57: Thomas TL, Kang HK. Mortality and morbidity among Army Chemical Corps Vietnam veterans: a preliminary report. Am J Ind Med 1990;18: Dalager NA, Kang HK, Thomas TL. Cancer mortality patterns among women who served in the military: the Vietnam experience. J Occup Environ Med 1995;7: Fett MJ, Nairn JR, Cobbin DM, et al. Mortality among Australian conscripts of the Vietnam conflict era. II. Causes of death. Am J Epidemiol 1987;15: Bullman TA, Kang HK, Watanabe KK. Proportionate mortality among US Army Vietnam veterans who served in Military Region I. Am J Epidemiol 1990;1: Visintainer PF, Barone M, McGee H, et al. Proportionate mortality study of Vietnam-era veterans of Michigan. J Occup Environ Med 1995;7: Watanabe KK, Kang HK, Thomas TL. Mortality among Vietnam veterans: with methodological considerations. J Occup Med 1991;: Bond GG, McLaren EA, Lipps TE, et al. Update of mortality among chemical workers with potential exposure to the higher chlorinated dioxins. J Occup Med 1989;1: Fingerhut MA, Halperin WE, Marlow DA, et al. Cancer mortality in workers exposed to,,7,8-tetrachlorodibenzo-pdioxin. N Engl J Med 1991;4: Zober A, Messerer P, Huber P. Thirty-four-year mortality follow-up of BASF employees exposed to,,7,8-tcdd after the 195 accident. Int Arch Occup Environ Health 1990;6: Collins JJ, Strauss ME, Levinskas GJ, et al. The mortality experience of workers exposed to,,7,8-tetrachlorodibenzop-dioxin in a trichlorophenol process accident. Epidemiology 199;4: Flesch-Janys D, Berger J, Gurn P, et al. Exposure to polychlorinated dioxins and furans (PCDD/F) and mortality in a cohort of workers from a herbicide-producing plant in Hamburg, Federal Republic of Germany. Am J Epidemiol 1995;14: Wolfe WH, Michalek JE, Miner JC, et al. Health status of Air Force veterans occupationally exposed to herbicides in Vietnam. I. Physical health. JAMA 1990;64: Henriksen GL, Ketchum NS, Michalek JE, et al. Serum dioxin and diabetes mellitus in veterans of Operation Ranch Hand. Epidemiology 1997;8: Michalek JE, Wolfe WH, Miner JC. Health status of Air Force veterans occupationally exposed to herbicides in Vietnam. II. Mortality. JAMA 1990;64: Wolfe WH, Michalek JE, Miner JC, et al. Paternal serum dioxin and reproductive outcomes among veterans of Operation Ranch Hand. Epidemiology 1995 ;6: Henriksen GL, Michalek JE, Swaby JA, et al. Serum dioxin, testosterone and gonadotropins in veterans of Operation Ranch Hand. Epidemiology 1996;7: Henriksen GL, Michalek JE. Serum dioxin, testosterone, and gonadotropins in veterans of Operation Ranch Hand. (Letter; comment). Epidemiology 1996;7: World Health Organization. International classification of diseases. Manual of the international statistical classification of diseases, injuries, and causes of death. Ninth Revision. Geneva: World Health Organization, Breslow NE, Lubin JH, Marek P, et al. Multiplicative models and cohort analysis. J Am Stat Assoc 198;78: Kulkarni PM, Tripathi RC, Michalek JE. Maximum (Max) and Mid-P confidence intervals and p values for the standardized mortality and incidence ratios. Am J Epidemiol 1997; 147:8-6.. Michalek JE, Wolfe WH, Miner JC, et al. Indices of TCDD exposure and TCDD body burden in veterans of Operation Ranch Hand. J Exp Anal Environ Epidemiol 1995;5:09-.. Bertazzi PA, Zocchetti C, Pesatori AC, et al. Ten-year mortality study of the population involved in the Seveso incident in Am J Epidemiol 1989;19: Yu M, Guo Y, Hsu C, et al. Increased mortality from chronic liver disease and cirrhosis 1 years after the Taiwan "Yucheng" ("Oil Disease") Incident. Am J Ind Med 1997;1: Ramlow JM, Spadacene NW, Hoag SR, et al. Mortality in a cohort of pentachlorophenol manufacturing workers, Am J Ind Med 1996;0: