Chapter 6 - Results

6.1 Overview of Analysis
6.2 Deaths from diseases of a Priori interest
6.3 Mortality of Korean War veterans
6.3.1 Conditions with a statistically lower SMR
6.3.2 Conditions with a statistically elevated SMR
6.4 Mortality by Branch of Service
6.4.1 Deaths of Navy Korean War veterans
6.4.2 Deaths of Army Korean War veterans
6.4.3 Deaths of Air Force Korean War veterans
6.4.4 Summary
6.5 Neoplasms
6.5.1 Navy veterans
6.5.2 Army veterans
6.5.3 Air Force veterans
6.5.4 Smoking-related cancers
6.6 Effect of Service in Korea
6.6.1 Days served in Korea
6.6.2 Period of service in Korea
6.7 Summary

This chapter presents the results of the mortality analysis. The chapter discusses veteran mortality for broad disease groups of the International Classification of Diseases (ICD) chapters and specific diseases of interest within some of the groupings. It also reviews mortality by branch of Service. Mortality from cancers is discussed in detail with special reference to smoking-related cancers. Finally the chapter explores the mortality for those veterans who experienced different durations and periods of service in Korea.

6.1 Overview of analysis

The study has investigated the mortality of the Korean War veterans for the period 1950 through 2000, as per the protocol (Appendix A). Results are presented as standardised mortality ratios (SMRs). A SMR is the ratio of the observed number of deaths among the Korean War veterans to the expected number of deaths within the same aged Australian male population.

As detailed in section 5.6.1 of the preceding chapter, the population at risk includes 886 people whose vital status was 'unknown'. It was not possible to determine whether the unknown veterans were still alive and living in Australia. To reflect the uncertain status of the unknown veterans, results are presented using two scenarios:

  • Scenario 1 excludes all unknown veterans. This assumes the veterans lost to follow-up have the same rate of death as the other Korean War veterans. This is the usual scenario used in mortality studies and is the preferred option. If the death rate of those lost to follow-up is substantially different, then the SMR using this scenario may be an over or under-estimate of the true situation.
  • Scenario 2 includes the unknown veterans and assumes that all the veterans lost to follow-up are still alive and residing in Australia. The analysis using Scenario 2 will result in unlikely estimates of SMRs that are lower than the 'true' situation.

Calculation of the SMR by itself is insufficient for determining whether veterans experienced higher or lower rates of death than might be expected. A statistical test is required to test whether the actual number of deaths experienced by veterans was statistically different from those expected. The test involves calculating a 95% confidence interval around the SMR. This provides a range between upper and lower values within which the true SMR is likely to be. The range of values is likely to be narrower for those diseases with greater numbers of deaths, and conversely wider for deaths from conditions with fewer numbers.

Given that a SMR of 1.0 means that there is no difference in mortality between Korean War veterans and the Australian community, a confidence interval which does not include 1.0 indicates a significant difference. For example, a SMR of 1.21 with a confidence interval of 1.18 to 1.24 is significantly different because the interval does not include 1.0. If the confidence interval was 0.92 to 1.52, the difference would not be statistically significant because the confidence interval includes 1.0.

Back to Top

6.2 Deaths from diseases of a priori interest

In planning this study, several specific causes of death were considered to be of particular interest because they were suggested by the review of the current literature or of concern to veterans' organisations. These specific a priori causes of death are listed in Table 1-1 of this report. Minor changes have been made to the list of diseases to compensate for changes in coding practices and availability of data over the past 50 years. Specifically, chronic airways obstruction (ICD9 code 496 and ICD10 code J44) was added to the item 'chronic bronchitis and emphysema' and is presented as chronic obstructive pulmonary disease (COPD). Cirrhosis of the liver was reclassified as alcoholic liver disease for consistency with the ICD10 coding. Malignant neoplasms of the oropharynx was expanded to include head and neck cancers. Hepatocellular carcinoma data was first available from 1968 and this category was changed to liver and gallbladder cancer for which data existed from 1950.

Table 6-1 details the mortality results for the revised coding of a priori diseases of interest. Mortality from a majority of the diseases of a priori interest was elevated. This included mortality from all causes, chronic obstructive pulmonary disease, ischaemic heart disease, stroke, alcoholic liver disease and external causes. Among the neoplasms of interest, cancer of the oesophagus, gastrointestinal and colo-rectal cancers, head and neck, lung, genito-urinary and prostate cancers were statistically significantly increased. Mortality from two a priori diseases (tuberculosis and peptic ulcer disease) and four cancers (liver and gallbladder, mesothelioma, melanoma and leukaemia) did not differ from that of Australian males. None of the a priori diseases showed a statistically significant decrease in mortality.

Table 6-1: Standardised Mortality Rates for a priori causes of death for all Korean War veterans
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All causes
7514
1.21 1.18-1.24 1.11 1.09-1.14
COPD*
362
1.49 1.33-1.64 1.34 1.21-1.48
Ischaemic heart disease
1951
1.10 1.05-1.15 1.02 0.97-1.06
Stroke
451
1.17 1.06-1.28 1.06 0.97-1.16
Tuberculosis
12
0.95 0.40-1.49 0.88 0.38-1.39
Peptic ulcer disease
43
1.42 0.99-1.84 1.30 0.91-1.69
Alcoholic liver disease
109
1.36 1.11-1.62 1.28 1.04-1.52
External causes chapter
814
1.37 1.27-1.46 1.29 1.20-1.38
Neoplasms
       
Oesophagus
93
1.59 1.27-1.91 1.46 1.17-1.76
Gastrointestinal
390
1.18 1.07-1.30 1.09 0.98-1.20
Colo-rectal
295
1.18 1.05-1.32 1.09 0.97-1.21
Liver and gallbladder
63
1.30 0.98-1.62 1.20 0.90-1.50
Head and neck
114
1.96 1.60-2.32 1.82 1.48-2.15
Lung
802
1.47 1.37-1.58 1.36 1.27-1.45
Mesothelioma
4
0.51 0.01-1.00 0.46 0.01-0.91
Genito-urinary
286
1.24 1.10-1.39 1.13 0.99-1.26
Prostate
181
1.29 1.10-1.48 1.16 0.99-1.33
Melanoma
78
1.28 0.99-1.56 1.18 0.92-1.45
Leukaemia
60
0.99 0.74-1.24 0.91 0.68-1.14
*COPD = Chronic Obstructive Pulmonary Disease

Back to Top

6.3 Mortality of Korean War Veterans

There was a total of 7,514 deaths among all Korean War veterans. The most frequent causes of death were diseases of the circulatory system (2,894), cancer (1,895), external causes such as suicide and motor vehicle accidents (814), and diseases of the respiratory system (573).

Overall mortality for military Korean War veterans was elevated. The SMR for all causes of death using Scenario 1 was 1.21 (95% CI 1.18-1.24), that is, the overall death rate for male Korean War veterans was 21% higher than the Australian male population. Table D1 (Appendix D) shows the SMRs and their confidence intervals for all causes and for specific causes for all Korean War veterans using both Scenario 1 and Scenario 2.

The SMR and 95% CI for all causes of death, and 27 specific causes of death, are shown in Figure 6-1.

6.3.1 Conditions with a statistically lower SMR

When analysed by cause of death, two groups of diseases, skin disease and congenital malformations, showed SMRs that were statistically significantly less than one. However, there were only one and three deaths respectively from each of these causes.

6.3.2 Conditions with a statistically elevated SMR

Those disease groups or chapters in which there were statistically significantly more deaths of Korean War veterans than were expected given death rates in the Australian male population include circulatory diseases, neoplasms, respiratory diseases, digestive diseases and external causes. The results for the chapters as well as specific diseases of interest are shown in Table 6-2.

Specific cancers are detailed in Section 6.5, Neoplasms.

Table 6-2: Statistically significantly elevated causes of death for all Korean War veterans
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All causes
7514
1.21 1.18-1.24 1.11 1.09-1.14
Circulatory disease chapter
2894
1.13 1.09-1.17 1.04 1.00-1.07
Ischaemic heart disease
1951
1.10 1.05-1.15 1.02 0.97-1.06
Stroke
451
1.17 1.06-1.28 1.06 0.97-1.16
Neoplasms chapter
2476
1.31 1.26-1.36 1.20 1.16-1.25
Respiratory disease chapter
573
1.32 1.21-1.42 1.20 1.10-1.30
COPD*
362
1.49 1.33-1.64 1.34 1.21-1.48
Respiratory excluding COPD*
164
1.45 1.23-1.67 1.31 1.11-1.51
Digestive disease chapter
306
1.35 1.20-1.50 1.25 1.11-1.39
Liver, gallbladder, bile ducts
186
1.33 1.14-1.52 1.24 1.06-1.42
Alcoholic liver disease
109
1.36 1.11-1.62 1.28 1.04-1.52
External causes chapter
814
1.37 1.28-1.47 1.29 1.20-1.38
Suicide
211
1.31 1.14-1.49 1.23 1.07-1.40
*COPD = Chronic Obstructive Pulmonary Disease

Back to Top

Figure 6-1: Standardised Mortality Ratios (SMR) for all Korean War veterans

Figure 6-1: Standardised Mortality Ratios (SMR) for all Korean War veterans

Back to Top

6.4 Mortality by branch of Service

Mortality was also investigated by Service branch. Table 6-3 summarises the number of deaths and person years which contributed to the analysis by Service branch.

Table 6-3: Number of deaths and person years contributed by branch of Service
Service Branch Number of veterans contributing to the analysis Number of unknowns Number of deaths   Person Years contributed
Unknowns Excluded Unknowns Included
Navy
5,742
237
2226
231,799
243,199
Army
10,482
628
4795
393,087
422,480
Air Force
1,157
21
493
47,150
48,173
Total
17,381
886
7514
672,036
713,852

6.4.1 Deaths of Navy Korean War veterans

Table D2 (Appendix D) shows SMRs for Navy Korean War veterans, by cause of death. There were 2,226 deaths observed in this group of 5,742 men. The most common causes of death were the same as for all Korean War veterans, that is, diseases of the circulatory system (875), neoplasms (767), external causes (211) and then respiratory diseases (156).

Navy veterans had an 11% increase in overall mortality with the SMR for all causes of death being 1.11 (95% CI 1.06-1.16). Figure 6-2 shows the SMRs for Navy veterans for the analysed causes of death.

Conditions with a statistically lower SMR

There were two disease groups for which Navy veterans had a statistically significant reduced risk of mortality; infectious disease and genito-urinary disease. Two specific diseases, tuberculosis and motor vehicle accidents, also showed a reduced risk of mortality when compared to the age-matched male Australian population. Table 6-4 shows the SMRs and 95% confidence intervals for these diseases using both Scenarios.

Table 6-4: Causes of death for Navy Korean War veterans which were statistically significantly lower than the male Australian population
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
Infectious disease chapter 8 0.56 0.18-0.94 0.52 0.17-0.87
Tuberculosis 1 0.27 0.00-0.80 0.26 0.00-0.75
Genito-urinary disease chapter 10 0.61 0.24-0.99 0.56 0.21-0.90
Motor vehicle accidents 53 0.73 0.54-0.93 0.70 0.51-0.89

Back to Top

Figure 6-2: Standardised Mortality Ratios for Navy Korean War veterans

Figure 6-2: Standardised Mortality Ratios for Navy Korean War veterans

Conditions with a statistically elevated SMR

Navy veterans had a statistically significantly elevated standardised mortality risk for circulatory diseases including stroke, neoplasms and chronic obstructive pulmonary disease. Table 6-5 shows the SMRs and 95% confidence intervals for these diseases.

Table 6-5: Causes of death for Navy Korean War veterans which were statistically significantly higher than the male Australian population
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
Circulatory diseases chapter 875 1.09 1.02-1.16 1.02 0.95-1.08
Stroke 141 1.21 1.01-1.41 1.11 0.93-1.30
Neoplasms 767 1.22 1.14-1.31 1.14 1.06-1.23
COPD * 101 1.30 1.05-1.56 1.20 0.97-1.43
*COPD = Chronic Obstructive Pulmonary Disease

Back to Top

6.4.2 Deaths of Army Korean War veterans

Of the three Service branches, Army veterans had the highest mortality. Table D3 (Appendix D) shows the SMRs for Army Korean War veterans, by cause of death. There were 4,795 deaths observed in this group of 10,482 men.

As for all Korean War veterans, the most frequent causes of death for Army veterans were diseases of the circulatory system (1,819), neoplasms (1,555), external causes (552) such as suicide and motor vehicle accidents, and diseases of the respiratory system (380).

The SMR for all causes of death for Army veterans was 1.31 (95% CI 1.27 - 1.35). The SMR and 95% CI for all causes of death, for disease groups and some specific diseases of interest among Army Korean War veterans are shown in Figure 6-3.

Conditions with a statistically lower SMR

The only condition for which Army veterans had a statistically significant reduced risk of mortality was congenital malformations, with only one death being recorded.

Conditions with a statistically elevated SMR

Those diseases for which there were statistically significant more deaths of Army Korean War veterans than were expected given death rates in the Australian male population in both Scenarios are shown in Table 6-6.

Table 6-6: Statistically significant elevated causes of death for Army Korean War veterans
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All causes
4797
1.31 1.27-1.35 1.19 1.16-1.22
Neoplasms chapter
1555
1.41 1.34-1.48 1.28 1.21-1.34
Circulatory disease chapter
1819
1.20 1.14-1.25 1.08 1.04-1.13
Ischaemic heart disease
1230
1.18 1.11-1.24 1.07 1.01-1.13
Stroke
280
1.22 1.08-1.36 1.10 0.97-1.22
Respiratory disease chapter
380
1.48 1.33-1.63 1.33 1.19-1.46
COPD*
241
1.69 1.48-1.79 1.51 1.32-1.70
Respiratory less COPD
104
1.57 1.27-1.87 1.40 1.13-1.66
Digestive disease chapter
209
1.57 1.36-1.79 1.44 1.24-1.63
Liver, gallbladder & bile ducts
123
1.51 1.24-1.77 1.38 1.14-1.63
Peptic ulcer disease
34
1.90 1.26-2.53 1.72 1.14-2.30
Alcoholic liver disease
69
1.47 1.12-1.82 1.36 1.04-1.68
External causes chapter
552
1.58 1.45-1.71 1.47 1.35-1.59
Suicide
145
1.53 1.28-1.78 1.43 1.19-1.66
*COPD = Chronic Obstructive Pulmonary Disease

Back to Top

Figure 6-3: Standardised Mortality Ratios for Army Korean War veterans

Figure 6-3: Standardised Mortality Ratios for Army Korean War veterans

Back to Top

6.4.3 Deaths of Air Force Korean War veterans

Table D4 (Appendix D) shows SMRs for Air Force Korean War veterans, by cause of death. There were 493 deaths observed in this group of 1,157 men. The most common causes of death were from diseases of the circulatory system (199), neoplasms (153), external causes such as suicide and motor vehicle accidents (53), and diseases of the respiratory system (38).

Compared with the Australian male population, Air Force Korean War veterans had a statistically significant 11% lower death rate for all causes of death. The SMR for all causes of death was 0.89 (95% CI 0.81- 0.96). The SMRs and 95% CIs for the causes of death analysed for Air Force veterans are shown in Figure 6-4.

Conditions with a statistically lower SMR

Table 6-7 shows those causes of death that were statistically significantly lower when compared to the male Australian population.

Table 6-7: Causes of death for Air Force Korean War veterans which were statistically significantly lower than the male Australian population
Cause of death Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All causes
493
0.89 0.81-0.96 0.85 0.77-0.92
Circulatory disease chapter
199
0.82 0.71-0.93 0.78 0.67-0.89
Ischaemic heart disease
139
0.84 0.70-0.98 0.80 0.67-0.94
Alcoholic liver disease
2
0.35 0.00-0.83 0.34 0.00-0.82

Conditions with a statistically elevated SMR

There were no causes of death which were elevated amongst Air Force veterans.

Back to Top

6.4.4 Summary

In summary, when comparing the mortality of Korean War veterans with the Australian male population, Army Korean War veterans had the highest mortality rate of the three Service branches. Mortality from all causes was 31% elevated and 13 causes of death were also increased. Army veterans showed a reduced mortality for only one cause of the 27 causes of death, congenital malformations. Navy veterans experienced an overall increased mortality rate of 11% with an increased mortality rate from circulatory system diseases including stroke, neoplasms and chronic obstructive pulmonary disease. Navy veterans had a reduced mortality for four of the 27 causes of death. Air Force veterans showed a 11% lower overall mortality rate, with a lower mortality rate in four of the causes of death analysed. However, because of the small number of Air Force veterans, it is difficult to reach definitive conclusions about the other diseases in this group.

Figure 6-4: Standardised Mortality Ratios for Air Force Korean War veterans

Figure 6-4: Standardised Mortality Ratios for Air Force Korean War veterans

Back to Top

6.5 Neoplasms

Mortality due to all neoplasms was investigated by individual primary site. Table D5 (Appendix D) shows observed and expected deaths from 27 different cancers for all Korean War veterans for Scenario 1 and Scenario 2.

The most frequently occurring cancer deaths are from lung (802), gastrointestinal (390) (mainly colo-rectal), and genito-urinary cancers (286). The SMRs of selected cancers for all Korean War veterans and by branch of Service are shown in Figure 6-5.

Several cancers showed a significantly increased mortality compared with a similarly aged Australian male population and are detailed in Table 6-8.

Table 6-8: Elevated causes of death from cancer for all Korean War veterans
Causes of cancer deaths Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All neoplasms
2476
1.31 1.26-1.36 1.20 1.16-1.25
Gastrointestinal
390
1.18 1.07-1.30 1.09 0.98-1.20
Colo-rectal
295
1.18 1.05-1.32 1.09 0.97-1.21
Genito-urinary
286
1.24 1.10-1.39 1.13 0.99-1.26
Prostate
181
1.29 1.10-1.48 1.16 0.99-1.33
Head and Neck
114
1.96 1.60-2.32 1.82 1.48-2.15
Larynx
55
1.95 1.43-2.46 1.80 1.33-2.28
Lung
802
1.47 1.37-1.58 1.36 1.27-1.45
Oesophagus
93
1.59 1.27-1.91 1.46 1.17-1.76
Unknown
150
1.51 1.27-1.75 1.39 1.17-1.61

6.5.1 Navy veterans

Table D6 (Appendix D) shows the 767 deaths from cancer for Navy Korean War veterans. The most frequently occurring cancer deaths are from lung (228), gastrointestinal (130) and genito-urinary cancers (98).

One cancer, lymphoid leukaemia, demonstrated a statistically significantly lower mortality when compared with the male Australian population, with an SMR of 0.40 (95% CI 0.00-0.094). However, this SMR has been derived from only two cases.

Three cancers were statistically significantly elevated among Navy veterans. These are detailed in Table 6-9.

Table 6-9: Elevated causes of deaths from cancer for Navy Korean War veterans
Cause of cancer deaths Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All neoplasms 767 1.22 1.14-1.31 1.14 1.06-1.23
Genito-urinary 98 1.34 1.07-1.60 1.23 0.99-1.47
Head and Neck 41 2.09 1.45-2.72 1.97 1.37-2.56
Lung 228 1.27 1.10-1.43 1.19 1.03-1.34

Back to Top

6.5.2 Army veterans

Overall, Army veterans had a higher mortality from cancers than veterans from other Service branches. Table D7 (Appendix D) shows the mortality ratios for the 1,555 deaths using both scenarios for the 27 cancers of interest compared to the Australian population. As for all veterans, the most frequently occurring cancer deaths are from lung (536), gastrointestinal (238) and genito-urinary cancers (154).

Only one death from mesothelioma was observed among the Army veterans where five were expected. This was statistically significantly less than expected.

Eight cancers, listed in Table 6-10, are statistically significantly elevated among Army veterans compared with a similarly aged male Australian population.

Table 6-10: Elevated causes of deaths from cancer for Army Korean War veterans
Cause of cancer deaths Number of deaths Scenario 1
Unknowns excluded
Scenario 2
Unknowns included
SMR 95% CI SMR 95% CI
All neoplasms
1555
1.41 1.34-1.48 1.14 1.06-1.23
Gastrointestinal
238
1.24 1.08-1.40 1.11 0.92-1.30
Colo-rectal
177
1.22 1.04-1.40 1.08 0.87-1.30
Head and Neck
64
1.90 1.43-2.37 1.97 1.37-2.56
Larynx
40
2.44 1.68-3.19 1.29 0.59-1.99
Liver and gallbladder
46
1.67 1.19-2.14 0.88 0.44-1.33
Lung
536
1.69 1.55-1.83 1.19 1.03-1.34
Oesophagus
62
1.82 1.36-2.27 1.21 0.74-1.68
Unknown
102
1.77 1.42-2.11 1.20 0.84-1.55

Back to Top

6.5.3 Air Force veterans

Table D8 (Appendix D) shows the 153 deaths from cancer for Air Force Korean War veterans. The most frequently occurring cancer deaths for Air Force veterans were lung (38), genito-urinary (33), mainly prostate, and gastrointestinal cancers (22).

Two cancers showed a significant reduction in mortality among Air Force veterans using both Scenarios: liver and gallbladder cancer and stomach cancer. However, the mortality ratio was based on one and three deaths respectively.

There was only one statistically significantly elevated cause of cancer death among Air Force veterans. This was prostate cancer with the SMR being 1.71 (95% CI 1.03-2.28). Death from prostate cancer was not statistically significantly elevated in any other branch of the Services.

Figure 6-5: Standardised Mortality Ratios for selected cancers by branch of Service

Figure 6-5: Standardised Mortality Ratios for selected cancers by branch of Service

Back to Top

6.5.4 Smoking-related cancers

Many of the cancers for which Korean War veterans show an increased mortality are smoking-related cancers. An analysis was conducted to examine if smoking could account for all of the elevation in the smoking-related cancer deaths in Korean War veterans.

The level of smoking during the war and afterwards for Korean War veterans is not known. There was no systematic measurement of smoking rates during the war nor has there been an assessment of smoking rates of Korean War veterans in the years following the war. Anecdotal evidence suggests a high level of smoking during the war and cigarettes were freely available in large quantities to all ADF personnel during the Korean conflict. Furthermore, Army personnel in particular were living in confined and poorly ventilated conditions in the latter part of the war.

Given that no estimates of smoking prevalence rates exist, an analysis investigating a range of smoking prevalence rates from 30% to 100% was undertaken. This analysis generates a hypothetical number of expected cancer deaths based on these rates. The previous chapter, section 5.8.1, provides an explanation of the method used to derive the estimated cancer mortality. Table 6-11 provides the results of the analysis.

An analysis of six types of cancer revealed an elevation in mortality from two cancers, head and neck cancer and cancer of the larynx. These results cannot be explained by smoking alone as in both scenarios, even if 100% of the veterans smoked, the observed number of cancer deaths exceeds that expected.

In the case of a further two smoking-related cancers, cancer of the pancreas and cancer of the stomach, the cancers themselves did not show an elevated rate of mortality compared with the Australian community.

For the other two smoking-related cancers, cancer of the lung and cancer of the oesophagus, the elevated mortality among veterans is consistent with high levels of smoking. Levels of smoking prevalence that could explain the elevated mortality rates among veterans for each smoking-related cancer are as follows:

  • Lung. For the expected number of cancer deaths to equal the observed number of deaths, smoking rates of veterans would have been about 69%.
  • Oesophagus. For the expected number of cancer deaths to equal the observed number of deaths, smoking rates of veterans would have been about 94%.

These levels of smoking prevalence are higher than those of the Australian community.

Table 6-11: Expected deaths from cancer among Korean War veterans assuming various levels of smoking prevalence
Type of cancer Observed Expected SMR 95% Confidence Interval Period examined Smoking prevalence (%)
30 40 50 60 70 80 90 100
  Scenario 1* Expected Deaths
Head and neck
114
58
1.96
1.60-2.32
1968-2000
47 53 60 67 74 80 87 94
Larynx
55
28
1.95
1.43-2.46
1950-2000
22 26 30 34 38 42 47 51
Lung
802
544
1.47
1.37-1.58
1950-2000
386 491 597 702 807 913 1018 1124
Oesophagus
93
59
1.59
1.27-1.91
1950-2000
49 56 63 69 76 83 90 96
Pancreas
94
84
1.13
0.90-1.35
1950-2000
77 82 86 90 94 99 103 107
Stomach
98
90
1.08
0.87-1.30
1950-2000
87 89 91 94 96 99 101 103
  Scenario 2**                
Head and neck
114
63
1.82
1.48-2.15
1968-2000
50 58 65 72 79 87 94 101
Larynx
55
31
1.80
1.33-2.28
1950-2000
23 28 32 37 41 46 50 55
Lung
802
590
1.36
1.27-1.45
1950-2000
419 534 648 763 878 992 1107 1222
Oesophagus
93
64
1.46
1.17-1.76
1950-2000
53 61 68 75 83 90 97 105
Pancreas
94
91
1.04
0.83-1.25
1950-2000
84 89 93 98 103 107 112 116
Stomach
98
98
1.00
0.80-1.20
1950-2000
94 97 99 102 104 107 109 112
* Scenario 1: Excludes veterans of unknown status.
** Scenario 2: Assumes veterans of unknown status are alive in Australia

Back to Top

6.6 Effect of service in Korea

Exposure assessment has been a weak aspect of most cohort studies of veterans due to lack of information on a range of environmental and individual risk factors. This study is no exception. There was little data available to validate anecdotal accounts of exposure to occupational and environmental hazards during veterans' periods of service in Korea. Three surrogate measures were used in this study as an indication of the intensity of exposure to the hazards of the Korean War experience and to assess whether any dose-response relationship existed. They were:

  • branch of Service;
  • duration of service in Korea; and
  • period served in Korea.

However, none of these surrogate measures can be directly linked to a particular disease.

The rationale for using branch of Service to classify exposure was that the activities and location of service of the Navy, Army and Air Force varied and therefore the potential for exposure may also have varied. Chapter 4 highlights some of these differences with respect to the nature of service in Korea. However, men performing a similar job in each of the Services may all have been similarly exposed to a hazard and not differently as this argument would suggest.

The preceding section has shown that there are differences in mortality between the Service branches. Army had a higher mortality than the other Service branches, followed by Navy. RAAF did not have significantly elevated mortality.

To explore differences between Services further, the study also assessed mortality by both duration and period of service in Korea for Navy and Army. The Air Force cohort was too small for this analysis.

Nevertheless, each of these measures has limited ability to characterise the degree and duration of exposure to specific aspects of Korean service that may have affected subsequent death rates. The measures are unlikely to discriminate between veterans who were exposed to hazardous agents or environments and those who were not.

6.6.1 Days served in Korea

As detailed in Chapter 4, the distribution of the days of service in Korea varied between the Service branches. The mean number of days of service for Navy was 249 days with a range of seven to 880 days, whereas Army personnel served an average of duration of service of 311 days with a range of one to 1,467 days. Air Force personnel served an average of 145 days with a range of one to 1,161 days.

Days served in Korea were grouped into natural divisions within each Service branch as described in Chapter 5, section 5.6.5. The numbers of RAAF veterans were insufficient to allow for a meaningful division into service duration for mortality analysis.

Navy veterans

Tables D9 to D11 (Appendix D) show the observed and expected number of deaths and the SMRs of Navy veterans who served for short, medium or long duration categories for selected causes of death.

Figure 6-6 illustrates the relationship between the SMRs for the duration categories among Navy personnel for all deaths and four main disease groups. Three categories demonstrated a significantly different mortality when compared with the Australian community. Navy Korean War veterans who served for a short duration (1-174 days) had a statistically lower mortality for digestive diseases. Navy Korean War veterans who served between 175 and 294 days had statistically elevated mortality for all causes of death and for all cancers in both Scenarios. However, there was no trend evident between any of the duration of service categories.

Back to Top

Figure 6-6: Standardised Mortality Ratios for major disease groups among Navy veterans by duration of service in Korea

Figure 6-6: Standardised Mortality Ratios for major disease groups among Navy veterans by duration of service in Korea

Back to Top

Figure 6-7 shows the relationship between duration of Navy service during the Korean conflict and mortality for specific diseases of interest. Those Navy personnel who served in the medium duration category had a significantly increased mortality rate from head and neck cancer (110% increase) and lung cancer (32% increase) compared with the Australian community mortality rates. No other diseases tested for Navy veterans demonstrated a statistically significant difference from the Australian community among the duration categories. Again, there was no trend evident between the duration of service categories.

Figure 6-7: Standardised Mortality Ratios for specific diseases among Navy veterans by duration of service in Korea

Figure 6-7: Standardised Mortality Ratios for specific diseases among Navy veterans by duration of service in Korea

Back to Top

Army veterans

Army veterans demonstrated statistically significantly increased mortality for all causes of death in all three duration categories, peaking for the medium duration of 346 to 389 days. Tables D12 to D14 (Appendix D) show the observed and expected number of deaths and the SMRs for the three duration periods.

Figure 6-8 illustrates the SMRs for all deaths and four main disease groups by duration category for the Army veterans. For all deaths, all neoplasms and respiratory system diseases, all duration categories demonstrated elevated mortality compared to the Australian community. Mortality in the medium and long duration periods is not significantly higher than that for the short duration period. An increased mortality is noted for circulatory and digestive system diseases in the short and medium duration categories. However for all diseases analysed, the 95% confidence intervals for the three duration categories overlap, indicating a lack of significant difference between the discrete categories.

Figure 6-8: Standardised Mortality Ratios for major disease groups among Army veterans by duration of service in Korea

Figure 6-8: Standardised Mortality Ratios for major disease groups among Army veterans by duration of service in Korea

Back to Top

Figure 6-9 shows the standardised mortality ratios for specific diseases among Army veterans by duration of service category. Only lung cancer was statistically significantly elevated for the three duration categories with an elevation of 87% in the long duration category. Head and neck cancer was elevated 135% for the medium duration period. Ischaemic heart disease was elevated by 20% for those who served for a short duration. Finally suicide was elevated for the short and medium duration categories by 44% and 85%, respectively. However, as with the other analyses, the 95% CI for the three duration categories overlap, indicating a lack of significant difference between the discrete categories.

Figure 6-9: Standardised Mortality Ratios for specific diseases among Army veterans by duration of service in Korea

Figure 6-9: Standardised Mortality Ratios for specific diseases among Army veterans by duration of service in Korea

Back to Top

Survival time post Korea

The study further investigated the effect that duration of service had on the survival time of the servicemen after they left Korea. The Cox Proportional Hazards Model, in which duration of service was analysed as a continuum (months of service) rather than as discrete categories, showed that longer service was associated with a small but statistically significant shortening of survival when the servicemen were analysed as one group but the results of the analysis on each separate Service branch were not significant (Appendix D, Table D18).

The analysis took into account that the age of the servicemen in the war was an important factor in survival time after the war, but did not cater for other important influences such as combat experiences, war wounds or other health risk factor behaviours. War wounds could, for instance, directly shorten duration of service in Korea and have a detrimental effect on survival time.

6.6.2 Period of service in Korea

As detailed in Chapter 2, the Korean conflict had two distinct phases. The beginning of the war from 1950 to the end of 1951 was an offensive and counter-offensive phase where troops were on the move. By 1952 the war entered a defensive phase where the front remained relatively stable. Each of these phases exposed troops to different hazards which may have influenced their health. Mortality was investigated among Army veterans for the three period categories:

  • Period 1: Service in Korea was completed on or before 31 December 1951 (Early)
  • Period 2: Service in Korea commenced prior to 31 December 1951 and finished after 31 December 1951 (Both)
  • Period 3: Service in Korea commenced after 31 December 1951 (Late)

For those who served in Period 1, 780 deaths were observed among the 1,395 Army veterans; for those who served in Period 2, 572 deaths were observed among 1,148 veterans; and for those who served in Period 3, there were 3,436 deaths among 7,939 veterans. Tables D15 to D17 (Appendix D) give the details of mortality for the three periods of service. Figure 6-10 shows the SMRs and 95% confidence intervals for the major disease groups by period of service and Figure 6-11 shows the SMRs and 95% confidence intervals for specific diseases by period of service.

For all periods of service, Korean War Army veterans had increased mortality from all causes of death, all neoplasms, lung cancer, circulatory diseases, digestive diseases and external causes. For those who commenced Korean service after 31 December 1951 (Period 3), mortality was also elevated from head and neck cancer, respiratory disease, ischaemic heart disease and stroke. With the exception of suicide, the 95% confidence intervals of the mortality results between the different periods overlap indicating the lack of statistically significant difference in mortality among the periods of service.

There was a significant difference in mortality from suicide between the periods of service. Those who completed service in Korea prior to 1952, (Period 1), experienced a 42% lower mortality due to suicide compared to the male Australian population. However those who served in Korea after 31 December 1951, (Period 3), experienced a 65% increase in mortality from suicide.

Analysis of the suicide deaths since the Korean War indicates that the increased mortality was consistently elevated throughout the post-war period with no particular years showing a dominant mortality period.

Back to Top

Figure 6-10: Standardised Mortality Ratios for major ICD disease chapters among Army veterans by period of service in Korea

Figure 6-10: Standardised Mortality Ratios for major ICD disease chapters among Army veterans by period of service in Korea

Back to Top

Figure 6-11: Standardised Mortality Ratios for specific diseases among Army veterans by period of service in Korea

Figure 6-11: Standardised Mortality Ratios for specific diseases among Army veterans by period of service in Korea

Back to Top

6.7 Summary

Overall, Korean War veterans experienced a 21% higher mortality rate than an equivalent Australian male population. Of the three Services, Army veterans experienced the highest level of mortality followed by Navy, then Air Force. Air Force veterans showed a lower mortality rate than the Australian male population.

Mortality from a majority of the diseases of a priori interest were elevated. They included mortality from all causes, chronic obstructive pulmonary disease, ischaemic heart disease, stroke, alcoholic liver disease and external causes. Among the neoplasms of interest, cancer of the oesophagus, gastrointestinal and colo-rectal cancers, head and neck, lung, genito-urinary and prostate cancers were statistically significantly increased. Mortality from two a priori diseases (tuberculosis and peptic ulcer disease) and four cancers (liver and gallbladder, mesothelioma, melanoma and leukaemia) did not differ from that of Australian males. None of the a priori diseases showed a statistically significant decrease in mortality.

The most common causes of death overall and in each Service were from diseases of the circulatory system, neoplasms, external causes such as suicide and motor vehicle accidents, and diseases of the respiratory system. Compared to the Australian male population, Army Korean War veterans had a 31% increased mortality rate for all cause mortality and an additional 13 of the 27 causes of death which were analysed. Army veterans showed a reduced mortality for only one cause of the 27 causes of death, i.e. congenital malformations. Navy veterans experienced an 11% overall increased mortality rate and an increased mortality for neoplasms, circulatory diseases, stroke and chronic obstructive pulmonary disease. Navy veterans had a reduced mortality for four of the 27 causes of death. Air Force veterans had an 11% decreased mortality overall and demonstrated a lower mortality rate for three of the causes of death which were analysed.

Mortality from specific cancers was investigated where it was found that there was a 31% elevated mortality from cancer for all Korean War veterans. The most frequently occurring cancer deaths in descending frequency are from lung, gastrointestinal, genito-urinary, head and neck, oesophagus, and larynx, and their mortality rates were all significantly elevated compared to a similarly aged Australian male population.

The most common deaths from cancer in the Navy and Army are similar to those of all Korean War veterans. However, in the Air Force, genito-urinary cancers were more common than gastrointestinal cancers.

In the Navy, significantly elevated mortality rates were found for lung, genito-urinary, and head and neck cancers, while death from lymphoid leukaemia was lower than the Australian male population. Elevated mortality rates were found in the Army for deaths from gastrointestinal, colo-rectal, head and neck, larynx, liver and gallbladder, lung and oesophagus cancers, and cancers of unknown primary site. Only mesothelioma had a reduced mortality rate. Among Air Force veterans, only prostate cancer was elevated. Two cancers categories, liver and gallbladder cancer, and stomach cancer, showed a lower mortality rate among Air Force veterans.

Many cancers for which there was elevated mortality among the Korean War veterans are cancers associated with smoking. Modelling suggested that even a 100% smoking rate would not explain all the excess cancers. However, in the absence of information about the percentage of veterans who smoked and the numbers of cigarettes smoked per day, it is not possible to determine precisely now much of the excess could be attributed to smoking.

Given the absence of quantitative data on exposure to occupational and environmental hazards, the effect of duration of service and the period of service in Korea were investigated for Navy and Army veterans. The number of Air Force veterans were too small for this analysis. Mortality by duration of service in Korea for Navy and Army veterans did not show any clear pattern between the different duration categories or within each Service branch. However investigating mortality by period of service among Army veterans reveals that those who completed their Korean service prior to 1952 (marking the end of the offensive / counter-offensive phase of the conflict), had a significantly lower mortality rate from suicide compared to the Australian male community. Those who served in Korea after 1952 had a higher mortality from suicide compared to the community norm. Apart from this unexpected and possibly chance post hoc finding, there was little evidence of any other trend from the effect of Korean service.

Chapter 7 discusses the results from the mortality study.