Mortality of National Service Vietnam Veterans

(Executive Summary)

A report of the 1996 retrospective cohort study of Australian Vietnam Veterans
Issued by the Department of Veterans' Affairs, Canberra,  on 11 November 1997
Crane PJ, Barnard DL, Horsley KW, Adena MA.  ISBN 0 642 38363 X>

Table of Contents

Introduction
Background
Design of this study
Methods used in the analysis
Results
Table 1 - Relative Risk of Death
Summary of findings
Strengths and weaknesses of this study
Conclusions
Recommendations>

Introduction

This study is part of a retrospective examination of Australian veterans who served in the Vietnam War.  It is a mortality study of Australian national servicemen who were conscripted into the army during the Vietnam War.  Only deaths between 1 January 1982 and 31 December 1994 are analysed in this study because an earlier mortality study conducted as part of the Australian Veterans Health Studies (AVH Study) has already reported on deaths that occurred between the end of the men's national service and 1982.   In both the current study and the earlier AVH Study, the death rates and causes of death of national servicemen who served in Vietnam were compared with those of national servicemen who did not serve in Vietnam.

Background

From January, 1965 to December, 1972 men were randomly selected from among the Australian male population of 20 year olds to be conscripted for national service.   All were enlisted into the army.  Therefore, when conscripted these men were all of similar age and were in excellent health, with no chronic diseases or serious congenital anomalies.  Furthermore, the recruits did not include men with criminal records or overt personality problems.

For 12 months after enlisting, all national servicemen received their basic training followed by some specialist training.  Once training was complete the army decided which units or men were required in Vietnam. National servicemen were generally sent to Vietnam only once and for up to 12 months.  The men who were not selected to go to Vietnam continued to serve in the army usually in Australia.  Generally, all national servicemen were discharged two years after they enlisted.  Because, the national servicemen who served in the army in Australia were similar to those who served in Vietnam it is appropriate to compare these groups.  By studying a control group whose health status closely matched the veterans pre-Vietnam health status, it is more likely that any alteration in rates or causes of mortality that are detected may have resulted for army service in Vietnam and not be as a result of the 'healthy worker effect'.

The 'healthy worker effect' describes the effect of a selection bias which results in workers having lower death rates than the general population.  The selection bias arises because the general population includes persons who are not fit enough to work or to remain in work.  Death rates in employed populations are typically 70 to 90 per cent of those in the general population, but the effect usually diminishes with increasing age of the study cohort or time from entry into the cohort.

Compiling and publishing the Nominal Roll of Vietnam Veterans was the first part of this retrospective examination of Australian veterans who served in the Vietnam War.  The second was publishing, in 1997, the report entitled Mortality of Vietnam veterans: the veteran cohort study which describes a retrospective cohort study of Australian Vietnam veterans.  It reports the causes and rates of death among Australian male personnel who served in the army, navy, air force or Citizen Military Forces in Vietnam or Vietnamese waters during the Vietnam war.  The deaths included in the study occurred between 1962 and 1994 (excluding deaths in Vietnam) however, the study concentrated on the period 1980-94.  These deaths were compared with national statistics on the deaths of Australian males during this period.  The national servicemen who served in Vietnam were part of the veteran cohort study cohort.   However, the aims and methods of the study differed.

Design of this study>

The aim of this study was to compare the mortality rates of the national servicemen who served in Vietnam at any time during the war (veterans) with those of the national servicemen who did not serve in Vietnam (non-veterans).  The population examined consisted of former national servicemen who were alive on 1 January 1982.  These men were followed-up until the date of their death or 31 December 1994, whichever came first.

The data were collected by matching government and public records with the roll of national servicemen either to discover a date and cause of death, or to confirm that the person was still alive.  Since the data needed for the study were already on record the protocol did not involve contacting veterans or their families.  The records searched were the electoral roll, National Death Index, and the Medicare database.   In Australia, electoral enrolment is compulsory, as is registration of death.   Medicare benefits are not means tested and are available to all Australian citizens, so that there is considerable incentive for all people who are alive to be recorded on the Medicare database.

Methods used in the analysis

The analysis of the national service cohort used statistical methods which are standard for epidemiological studies of cohorts.  The death rates obtained from the data base searches were compared with expected death rates which were calculated from the Australian national male rate, taking account of age and calendar year of death.  For each cause of interest, the observed death rate was divided by the expected death rate to give a Standardised Mortality Ratio (SMR).  A ration greater than one would indicate a death rate that was higher than the adjusted Australian male average.

To make the comparison between the national service Vietnam veterans and non-veterans, the SMR for veterans was divided by the SMR for the non-veterans.  This produced a Relative Rate (RR) of death for each cause.  A ration greater than one would be indicative of a higher rate of death for veterans than for non-veterans.  The advantage of using ratios to compare the mortality rates is that any biases due to some veteran deaths not being ascertained and included in the study would probably be cancelled out by the likelihood of a similar bias in the ascertainment of non-veteran deaths.

Some care needs to be taken when interpreting the RR and SMR figures, as neither is a direct reflection of increases in rates of disease and mortality when compared with the Australian population.  The 95% confidence interval (95%CI) is also an important statistic: in cases where the confidence interval includes 1, the death rates in the two groups are not statistically different.

Results

At the start of this study, 1 January 1982, there were 43 595 national servicemen in the cohort, 18 949 veterans and 24 646 non-veterans.  To determine their vital status on 31 December 1994, the end of the study, their records were matched against the National Death Index, the electoral roll and the Medicare data bases.  The final vital status results for the period 1982-94 were as follows:

• National service veterans
  
  • 17 844 were determined to be alive, representing 94.2% of the group;
    
  • 371 (2.0%) were determined to be dead; and
    
  • 725 (3.8%) remained 'lost to follow up'.
    
• National service non-veterans
  
  • 23 269 were determined to be alive, representing 94.4% of the group;
    
  • 416 (1.7%) were determined to be dead; and
    
  • 957 (3.9%) remained 'lost to follow up'.

This study of the Vietnam veteran national servicemen found an elevated overall mortality of borderline statistical significance.  Compared with the non-veteran group, they had an overall relative risk of mortality of 1.15.  This elevation of overall mortality reflected elevations in cancers and digestive diseases.

Mortality from diseases of the digestive system was twice that among the cohort which did not serve in Vietnam; the relative risk was 2.1.  Cirrhosis of the liver alone accounted for this elevation because it was the only elevated cause of death in this group.

Mortality for all cancers showed an elevation of borderline significance, with a relative risk of 1.28.  This elevation primarily resulted from an elevation in lung cancer mortality (which was of a priori interest to the study) and to a much lesser degree, unexpected elevation in pancreatic and brain cancer mortality.

Lung cancer mortality showed a relative risk of 2.2.  This was based on 27 deaths from lung cancer among the Vietnam veterans, and 16 lung cancer deaths in the comparison cohort.

Suicide, which has been of particular interest in Vietnam veterans, was not significantly elevated, with a relative risk of 1.13.

For the period 1982-94, there was little evidence of any difference in mortality between the various corps groupings (infantry; engineers; armour and artillery; corps with a minor presence in the field; and non-field corps) for the two groups of national servicemen.  This was in contradiction to the AVH Study which found that for engineers mortality was statistically significantly higher among veterans than non-veterans.

Summary of findings

The death rate from all causes for Vietnam veteran national servicemen relative to other national servicemen who did not go to Vietnam was estimated to be 1.15 (95% CI) 1.00 to 1.33) for 1982-94.  This estimated RR takes account of the known variation in death rates by age group and calendar year.  This RR provides some slight evidence of excess mortality among national service veterans compared with non-veterans, the VCS showed that the overall level of mortality for veterans was at most the same as and was probably lower than would have been expected compared with the Australian male population.

This study found that national service veterans had a death rate from lung cancer that was more than twice that of the non-veterans.  This elevation may be related to some influence or exposure that occurred during service in Vietnam.  Deaths from soft tissue and other sarcoma's, non-Hodgkin's lymphoma, and Hodgkin's disease cancers which are consistently associated with exposure to herbicides or dioxin, were too few to justify any conclusion as to whether the risks of death were elevated among the veterans compared with the non-veterans.

National service veterans had a death rate from cirrhosis of the liver that was nearly three times that of non-veterans.  Although the causes of cirrhosis are well established - mainly excessive alcohol consumption or Hepatitis B infection - this study did not have any data on alcohol consumption or Hepatitis B infection in either the deceased or the living national servicemen.  However, 20 of the 21 cirrhosis deaths among veterans were recorded in a manner which indicated that excess alcohol consumption was the cause of the cirrhosis.

Although deaths from external causes accounted for one third of the deaths in the cohort, the death rate among national service veterans was not statistically significantly different from that for the non-veterans.  Furthermore, death rates for veterans from suicide, motor vehicle accidents or other external causes were not statistically significantly elevated when compared with those of non-veterans.

From the end of service until 1994 the death rates of national service veterans were generally higher than those of non-veterans, although the differences were progressively becoming smaller.  Also, over this period the death rates for both veterans and non-veterans were increasing relative to the Australian male population.

Unlike the AVH Study, this study found that corps groupings did not have a confounding effect on death rates.  While none of the results from the national servicemen comparison were affected by this finding, the results of the AVH Study may need to be reviewed, considering this new evidence.

Strengths and weaknesses of this study>

The study had various strengths and some unavoidable weaknesses that affect the interpretation of the results.  Its strengths include:>

• the mortality of national service veterans was compared with a group of non-veterans who differed from the veterans in as few respects as possible, other than Vietnam service;
• follow-up or tracing was for 13 years from 1 January 1982 to 31 December 1994 - the total length of follow-up for Australian national servicemen is now approximately 22 to 29 years;
• vital status was established for 96.1 per cent of the national servicemen in the study;
• there was no apparent bias in the determination of vital status; and
• the national servicemen studied were very similar - all were male and of similar age and virtually all were Caucasian.

The weaknesses of this study include:

• insufficient statistical power to detect modest differences in mortality rates, particularly for the rarer causes of death in spite of including all eligible national servicemen;
• the limited numbers of deaths expected and observed because the national servicemen were relatively young, on average 47 years old at the end of follow-up;
• probable under-ascertainment of deaths when using linkage to the various databases;
• generally national servicemen served for one year in Vietnam so that it is not possible to compare mortality outcome with length of service; and
• lack of measurement of exposure to risk factors that might confound the observed associations, such as cigarette smoking, alcohol intake, Hepatitis B, and herbicide and dioxin exposures in Vietnam.

Conclusions

1. This study found some evidence that national service veterans had an elevated mortality from all causes of death combined when compared with non-veterans.
2. This study found that national service Vietnam veterans died from lung cancer at a rate that was twice that of national servicemen who did not service in Vietnam.
3. The rates of death for most other malignancies were unremarkable.  In particular, those cancers most closely linked to phenoxy herbicides and its contaminate, 2,3,7,8-tetrachlorodibenzo-p-dioxin, did not show elevated mortality.  The numbers of observed and expected deaths for most of these cancers were small.
4. For brain cancer and pancreatic cancer, which were not of a priori interest, unexpected elevations in mortality were found when national service veterans were compared with non-veterans.
5. This study found that national service veterans had a significantly elevated mortality from cirrhosis of the liver when compared with non-veterans.
6. The rate of death from suicide was not elevated for national service veterans compared with non-veterans.
7. Unlike the AVH Study, this study found no evidence that corps groups had an effect on mortality rates, either within or between the nation service veterans and non-veterans.

Recommendations

1. This study should be repeated shortly after the year 2000, to assess the national servicemen's mortality experience from 1995 to 2000.  In the meantime, the mortality of form national servicemen should be periodically monitored to detect any emerging trends.  In the year 2000 study, an effort should be made to reduce the number of national servicemen lost to follow-up.
2. This study had to reconstruct the cohort of national servicemen from several sources because the computer tape that contained the AVH Study data was damaged.  Therefore, the information from the current study, including the date when each national servicemen was last known to be alive or dead, should be stored safely.  This would facilitate the conduct of future studies.  Future studies should include an assessment of the effect of updated follow-up on the conclusions of this study.
3. Date of birth data were missing from the National Death Index for most deaths that occurred before 1990.  This was the principal reason why less than the true number of deaths were ascertained and analysed by this study.  To improve the utility of the National Death Index those responsible for supplying and updating the data it contains should ensure that the date of birth is recorded for all deaths.
4. As with the Veteran Cohort Study, this study suggests that a series of smaller studies of Vietnam veterans should be considered.  These studies could further investigate mortality from causes of death such as; lung cancer, brain cancer and pancreatic cancer.
5. This study produced evidence that veterans are at risk of increased mortality due to lung cancer.  The veterans experienced a variety of exposures that could cause lung cancer, but this study was unable to determine which of these caused this elevation.   Whatever the reason was, participation by veterans in smoking cessation programs is likely to help moderate this risk
6. This study has produced evidence that veterans are at risk from increased mortality due to cirrhosis of the liver.   Excessive alcohol consumption may be a contributing factor in these elevated rates but this study did not measure exposures and no specific conclusions can be drawn on causality.  Additional programs aimed at alcohol moderation in veterans should be considered by the Department of Veterans' Affairs to help reduce the risk.