10. Laboratory investigations
Participants who undertook medical assessments provided several blood samples for haematological, biochemistry and serological analysis, and this chapter reports the results of these laboratory investigations.
One laboratory was used in this study to minimise sources of variation, and to standardise the testing methodology and reference intervals for blood specimens that were collected from participants across Australia. The reason for this is that equipment, testing methodologies and reference intervals can vary between laboratories and in different states and territories. The laboratory used was the Institute of Medical and Veterinary Science (IMVS) in Adelaide, South Australia. The laboratory investigations, the reference intervals and the testing methodologies used by IMVS are detailed in appendix 6. Early in the study, the laboratory investigations for 18 study participants were performed by Dorevitch Pathology (also a NATA and RCPA Registered Laboratory) in Melbourne, and the test results have been included in the analysis, where appropriate, utilising that laboratory’s own reference intervals.
10.1 Aim
The aim of this analysis is to investigate whether male Australian Defence Force personnel who served in the Gulf War have a higher prevalence of laboratory test results that are indicative of adverse health effects than the comparison group; and, if so, whether these effects are associated with exposures and experiences that occurred in the Gulf War?
10.2 Research questions
1. Do Australian Gulf War veterans have significantly more haematological test results that are indicative of anaemia or inflammation than the comparison group?
2. Do Australian Gulf War veterans have significantly more biochemical test results that are indicative of renal impairment or liver disease or of raised blood glucose than the comparison group?
3. Do Australian Gulf War veterans have serological test results that are indicative of significantly more exposure to viral infections than the comparison group?
4. Do Australian Gulf War veterans have more abnormalities on urinalysis testing than the comparison group?
10.3 Definitions
10.3.1.1 Overview of parameters for analysis
The parameters used to compare the study groups in the analysis were decided prior to the data analysis commencing, and included those parameters considered to have either a scientific rationale through biological plausibility or where results of previous clinical or epidemiological studies indicated their importance. The numbers and proportions of results that were within the reference intervals are routinely reported in the tables for each test. For parameters, such as haemoglobin, where values either higher or lower than the reference intervals may be clinically important, both relationships to the reference interval are reported. Where the clinical importance of a test result related primarily to an elevated value, eg bilirubin, only that relationship to the reference interval is reported. Erythrocyte Sedimentation Rate (ESR) and C-reactive protein (CRP) have a detection limit, and only results that were within or higher than the reference interval could be reported.
10.3.1.2 Laboratory reference intervals
The primary analysis compared the results of the Gulf War veteran group with those of the comparison group. As a framework for this comparison, the reference intervals of IMVS were used to categorise the test results. In their determination of reference intervals, the distributions of parameters were not assumed to be Normal; and IMVS reported 95th percentile reference intervals with a 2.5% cut-off at both ends; ie 2.5% of the population would be expected to have a value higher, and 2.5% lower, than this interval. These reference intervals were based on the testing of 230 “laboratory normal” staff.
Reference intervals have some limitations for interpreting the clinical significance or meaning of results. If a test result was outside the reference interval it did not necessarily mean that it was abnormal. For example, haemoglobin of 130 g/L was lower than the reference interval of 135-175 g/L for an adult male. This would not usually be considered “abnormal” or of clinical importance, as the person would not be expected to have symptoms of anaemia, such as tiredness, until the haemoglobin was less than 100 g/L. However, test results just above or below a reference value may suggest pre-clinical conditions. Although reference intervals have limitations, they are an established way of presenting and comparing test results between groups. In this study, this limitation in reference intervals is overcome by having a comparison group which is also compared against the same set of reference values.
10.3.1.3 Definitions of laboratory investigation outcomes
Laboratory investigation outcome definitions were developed for the purposes of this study to assist in comparing the health outcomes of the Gulf War veterans and comparison group. The following definitions were based on the IMVS laboratory reference intervals:
- haematology test results indicative of anaemia were defined as Hb <135 g/L.
- haematology test results indicative of inflammation were defined as an Erythrocyte Sedimentation Rate (ESR) >10mm (or >15mm for participants age >50 years) or a C-reactive Protein (CRP) >10 mg/L or a White Cell Count (WCC) >11.0 x 109/L.
- biochemical tests results indicative of renal impairment were defined as a creatinine >0.12 mmol/L.
- biochemical tests results indicative of an elevated random plasma glucose were defined as a random plasma glucose >11.0 mmol/L.
- liver function test results indicative of liver disease were defined
as:
- ALP >110 U/L and GGT >60 U/L, which was indicative of obstructive (cholestatic) liver disease
- ALT >55 U/L and AST >45 U/L, which was indicative of inflammatory (hepatitic) liver disease
- serological test results indicative of prior exposure to viral infections
were defined as a:
- positive EBV antibody test result which was indicative of prior exposure to EBV
- positive CMV antibody test result which was indicative of prior exposure to CMV
- positive hepatitis C core antibody test result which was indicative of prior exposure to hepatitis C.
10.4 Results
Of the 2825 participants who completed the medical assessment, 2810 had blood taken. Of these, 1409 (50.1%) were Gulf War veterans and 1401 (49.9%) were comparison group subjects, and 2748 (97.8%) were males and 62 (2.2%) were females. Fifteen (0.5%) did not have blood taken, either because they did not consent or for logistical reasons. This total number of 2810 males includes the 18 participants for whom the Dorevitch Pathology reference intervals have been used. The Dorevitch data was excluded from the calculation of summary statistics such as means (or medians) and standard deviations for individual parameters. The different testing methodologies used by IMVS and Dorevitch Pathology meant that data for these statistics were not directly comparable.
Therefore, in this analysis, the comparisons for males in relation to reference intervals are based on 1379 male Gulf War veterans and 1369 male comparison group participants, and the comparisons on mean (or median) and standard deviation are based on 1365 male Gulf War veterans and 1365 male comparison group participants.
10.4.1 Quality control
The proportion of laboratory investigations for which there was one or more missing value was one of the quality control measures used to monitor progress in the study (Table 10.1).
| >= 1 missing value (N=2810) | ||
|---|---|---|
| n | (%) | |
|
Haematological tests
|
||
| Complete Blood Examination (CBE)* |
45
|
(1.6)
|
| Erythrocyte Sedimentation Rate (ESR) |
67
|
(2.4)
|
|
Biochemical Analyses
|
||
| Urea and Electrolytes (U&Es) † |
6
|
(0.2)
|
| Random plasma glucose |
16
|
(0.6)
|
| Liver Function Tests (LFTs) ‡ |
2
|
(0.1)
|
| C-reactive protein (CRP) |
1
|
(0.0)
|
|
Serology tests
|
||
| Epstein-Barr virus antibody test (EBV IgG) |
4
|
(0.1)
|
| Cytomegalovirus antibody test (CMV IgG) |
4
|
(0.1)
|
| Hepatitis C antibody test (Hep C Ab) |
4
|
(0.1)
|
* CBE includes haemoglobin, red cell count, packed cell
volume, mean corpuscular volume, mean corpuscular haemoglobin, mean corpuscular
haemoglobin concentration, white cell count, differential white cell count
(neutrophils, lymphocytes, monocytes, eosinophils, basophils) and platelets.
† U&Es include sodium, potassium, chloride, bicarbonate, anion gap,
urea, creatinine, ionised calcium and phosphate.
‡ LFTs include total bilirubin, total protein, albumin, globulin, gamma-glutamyl-transferase
(GGT), alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate
aminotransferase (AST) and lactate dehydrogenase.
These quality control results were very good, with very low proportions of missing values for all tests, especially when it is considered that this was a national study with the bloods transported across Australia to the one laboratory. The proportions of missing values for some tests such as CBE or ESR were slightly higher than others, and this was attributed to the inherent sensitivity of these tests to handling procedures, and the time between blood collection and analysis. There was no difference in the rates of missing values for the Gulf War group and comparison group (data not tabulated).
10.4.2 Gulf War vs non-Gulf War comparisons
We reviewed the distribution curves for all parameters. The distribution curves were within expectations, with an essentially normal (bell-shaped) distribution for the central portion of the curve, with a right hand skew at the higher end of values. Distributions for those parameters such as ESR and CRP, which have a detection limit, were censored at the lower end. There were no peaks outside the reference intervals at either the higher or lower ends of values. There were no differences in the shapes of the distribution curves for Gulf War veterans and the comparison group. The distributions of two parameters, haemoglobin and ESR, are presented in Figure 10.1 to illustrate these general features.
Figure 10.1 Haemoglobin and ESR distribution curves for Gulf War veterans and comparison group
|
|
 |
 |
10.4.2.1 Haematological test results
The haematology results in Table 10.2 show that the mean (and median ESR, eosinophils and basophils) values for the individual parameters for the two groups were very similar.
|
Parameter
|
GWV (N=1355) mean (SD) |
Comp grp (N=1361) mean (SD) |
P value* |
|---|---|---|---|
| Haemoglobin (Hb), g/L | 153.4 (9.5) | 153.1 (9.1) | 0.181 |
| Mean corpuscular volume (MCV), fl | 91.6 (4.7) | 91.5 (4.5) | 0.474 |
| Mean corpuscular haemoglobin (MCH), pg | 30.4 (1.4) | 30.5 (1.3) | 0.024 |
| White cell count (WCC), x 109/L | 6.3 (1.7) | 6.2 (1.7) | 0.568 |
| Neutrophil count, x 109/L | 3.8 (1.3) | 3.7 (1.4) | 0.339 |
| Lymphocyte count, x 109/L | 1.9 (0.5) | 1.9 (0.6) | 0.970 |
| Monocytes, x 109/L | 0.4 (0.1) | 0.4 (0.1) | 0.095 |
| Platelets, x 109/L | 227.8 (44.4) | 231.3 (48.5) | 0.203 |
| Median (range) | median (range) | ||
| Eosinophils, x 109/L | 0.11 (0-1.47) | 0.13 (0-1.7) | <0.001 |
| Basophils, x 109/L | 0.03 (0-0.15) | 0.02 (0-0.20) | 0.003 |
| Erythrocyte sedimentation rate (ESR), mm/hour | 3 (1-106) | 3 (1-52) | 0.252 |
* P values obtained using Mann-Whitney / Wilcoxon rank sum tests
There was a higher proportion of Gulf War veterans than comparison group subjects with at least one CBE parameter outside the reference interval (Table 10.3). There was less than one percent of participants in each group with more than 4 parameters outside the reference interval. The mean total number (SD) of haematological test results outside the reference interval was similar in both study groups {Gulf War veteran 0.65 (1.05) vs comparison group 0.58 (1.02); crude ratio of means 1.10, adjusted ratio of means 1.10 95% CI 0.97-1.25, p value 0.130}.
The two groups fell within, or higher or lower than, the reference intervals in similar proportions for the individual haematological parameters (Table 10.3). More than 5% of the ESR results of both groups were greater than the reference interval, but the elevations were marginal in both groups with 98.3% of Gulf War veterans and 97.0% of comparison subjects falling within 5mm/hr of the reference interval. Such marginal elevations are not thought to be of clinical importance. A small proportion of MCV results were higher than the reference intervals and this was equally likely for the two study groups.
Table 10.3 Prevalences and odds ratios (OR) for haematological test results in relation to reference intervals
Go to table 10.310.4.2.2 Biochemical test results
Table 10.4 shows that the mean (and median CRP and random blood glucose) values for the individual parameters were very similar for the two groups.
|
Analyte
|
GWV N=1365 mean (SD) |
Comp grp N=1365 mean (SD) |
P value* |
|---|---|---|---|
| Sodium, mmol/L |
141.8 (2.0)
|
141.4 (2.1)
|
<0.001
|
| Potassium, mmol/L |
4.2 (0.3)
|
4.2 (0.4)
|
0.081
|
| Urea, mmol/L |
5.4 (1.2)
|
5.5 (1.1)
|
0.050
|
| Creatinine, mmol/L |
0.1 (0.0)
|
0.1 (0.0)
|
0.337
|
| Ionised calcium, mmol/L |
1.1 (0.0)
|
1.1 (0.1)
|
0.446
|
| Phosphate, mmol/L |
1.1 (0.2)
|
1.1 (0.2)
|
0.187
|
|
median (range)
|
median (range)
|
||
| Random plasma glucose, mmol/L |
4.7 (1.9-22.7)
|
4.7 (2.4-23.7)
|
0.953
|
| C-reactive protein, mg/L |
2.0 (0.5 – 63)
|
2.0 (0.5 -92)
|
0.164
|
* P values obtained using Mann-Whitney / Wilcoxon rank sum tests
The proportion of participants who had all parameters within the reference interval, or an increasing number of parameters outside the reference intervals, was very similar between the two groups (Table 10.5). The two study groups were equally likely to fall in to the category ‘1-2 outside the reference interval’, for biochemical test results. The mean (SD) of test results outside the reference interval was similar in both groups {Gulf War 1.13 (0.94) vs comparison group 1.15 (0.98), crude difference between the means 0.99, adjusted difference between the means 0.99; 95% CI 0.93-1.06, p=0.833}.
Similar proportions of the two groups fell within, and higher or lower than, the reference intervals for most of the individual biochemical test parameters. Gulf War veterans were more likely than comparison group subjects to have results outside and above the reference interval for sodium and creatinine.
A random blood glucose was used to compare study groups for an elevated blood glucose, because routinely obtaining a fasting plasma glucose was not possible for logistical reasons. Although IMVS routinely reports a fasting blood glucose interval, IMVS does not report a reference interval for a random blood glucose because it is very difficult to define what is “normal”. Therefore, to compare study groups, the criteria used and the interpretation of these are those recently recommended by the NHMRC (http://www.health.gov.au/nhmrc/ advice/pdf/type2.pdf). The two study groups demonstrated similar patterns of random blood glucose. More than 80% of participants in both study groups had a random blood glucose <5.5 mmol/L, which suggests that diabetes mellitus is unlikely in these participants. Approximately 15% of participants in both study groups had a random blood glucose in the range of 5.5-11.0 mmol/L which suggests these participants may be at risk of diabetes mellitus if they have other risk factors, but further testing would be needed to confirm this. Less than 1% of participants in both study groups had a random blood glucose >11.0 mmol/L (an elevated random blood glucose) which suggests that Type 2 diabetes mellitus is likely in these participants, but further testing would be needed to confirm this.
Table 10.5 Prevalences and ORs of biochemical test result analysed in relation to reference intervals
Go to table 10.510.4.2.3 Liver function test results
Table 10.6 shows that the mean (or median) values for the liver function test results were very similar for the two study groups.
|
Parameter
|
GWV N=1365 mean (SD) |
Comp grp N=1365 mean (SD) |
P value* |
|---|---|---|---|
| Albumin, g/L | 42.2 (2.7) | 42.0 (2.7) | 0.039 |
| Globulin, g/L | 29.5 (3.8) | 29.2 (3.7) | 0.111 |
| Total protein, g/L | 71.7 (4.1) | 71.3 (4.0) | 0.007 |
| median (range) | median (range) | ||
| Total bilirubin, µmol/L | 10 (1-60) | 10 (1-73) | 0.001 |
| GGT, U/L | 26 (8-773) | 26 (7-990) | 0.542 |
| ALP, U/L | 74 (23-181) | 73 (28-187) | 0.099 |
| ALT, U/L | 28 (6-354) | 29 (4-238) | 0.858 |
| AST, U/L | 28 (7-155) | 28 (10-450) | 0.272 |
* P values obtained using Mann-Whitney / Wilcoxon rank sum tests
Furthermore, the proportion of participants from each group who had all parameters within, higher or lower than the reference intervals for individual parameters, or an increasing number of parameters outside the reference intervals, was very similar (Table 10.7). Gulf War veterans were not at increased risk of having more results outside the reference interval. The proportion with parameters outside the reference intervals fell smoothly above the ‘1-2 outside the reference interval’ category. The mean (SD) of liver function test results outside the reference interval was similar in both groups {Gulf War veterans 0.64 (0.91) vs comparison group 0.63 (0.91), crude difference between means 1.02, adjusted difference between means 1.0; 95% CI 0.90-1.12, P=0.995}.
| GWV | Comp grp | |||||||
|---|---|---|---|---|---|---|---|---|
|
Parameter
|
n | (%) | n | (%) | Crude OR | Adj OR | 95% CI | P value |
|
Liver function tests (LFTs)
|
||||||||
| All within ref intervals | 790 | (57.4) | 800 | (58.4) | - | - | - | - |
| 1-2 outside ref interval | 519 | (37.7) | 504 | (36.8) | - | - | - | - |
| >= 3 outside ref interval | 68 | (4.9) | 65 | (4.7) | - | - | - | - |
|
Albumin
|
||||||||
| within ref interval (34-48 g/L) | 1364 | (99.0) | 1352 | (98.8) | - | - | - | - |
| <34 g/L | 0 | (0) | 3 | (0.2) | 0.0 | 0.0 | - | - |
| >48 g/L | 14 | (1.0) | 14 | (1.0) | 1.0 | 0.9 | 0.4-2.0 | 0.852 |
|
Globulin
|
||||||||
| within ref interval (26-41 g/L) | 1185 | (86.0) | 1163 | (85.0) | - | - | - | - |
| <26 g/L | 184 | (13.4) | 199 | (14.5) | 0.9 | 0.9 | 0.7-1.2 | 0.531 |
| >41 g/L | 9 | (0.7) | 7 | (0.5) | 1.3 | 1.4 | 0.5-3.7 | 0.543 |
|
Total protein
|
||||||||
| within ref interval (65-85 g/L) | 1333 | (96.7) | 1310 | (95.7) | - | - | - | - |
| <65 g/L | 44 | (3.2) | 55 | (4.0) | 0.8 | 0.8 | 0.5-1.2 | 0.301 |
| >85 g/L | 1 | (0.1) | 4 | (0.3) | 0.3 | 0.2 | 0.02-2.0 | 0.178 |
|
Total bilirubin
|
||||||||
| within ref interval (6-24 µmol/L) | 1249 | (90.6) | 1257 | (91.8) | - | - | - | - |
| >24 µmol/L | 53 | (3.8) | 57 | (4.2) | 0.9 | 1.0 | 0.7-1.5 | 0.970 |
|
GGT
|
||||||||
| within ref interval (0-60 U/L) | 1248 | (90.6) | 1235 | (90.2) | - | - | - | - |
| >60 U/L | 130 | (9.4) | 134 | (9.8) | 1.0 | 0.9 | 0.7-1.2 | 0.584 |
|
ALP
|
||||||||
| within ref interval (30-110 U/L) | 1295 | (94.0) | 1300 | (95.0) | - | - | - | - |
| >110 U/L | 80 | (5.8) | 68 | (5.0) | 1.2 | 1.1 | 0.8-1.6 | 0.487 |
|
ALT
|
||||||||
| within reference interval (0-55 U/L) | 1206 | (87.5) | 1227 | (89.6) | - | - | - | - |
| >55 U/L | 172 | (12.5) | 142 | (10.4) | 1.2 | 1.2 | 0.9-1.5 | 0.181 |
|
AST
|
||||||||
| within ref interval (0-45 U/L) | 1286 | (94.3) | 1292 | (94.7) | - | - | - | - |
| >45 U/L | 78 | (5.7) | 73 | (5.3) | 1.1 | 1.0 | 0.7-1.4 | 0.946 |
Of particular note were the liver enzyme test results. A higher than expected, but similar, proportion of GGT and ALT results between the two study groups, and to a lesser extent, proportion of ALP and AST results, were higher than the reference intervals. To investigate these results further, we examined combinations of elevated parameters that may indicate underlying liver disease (Table 10.8).
A combination of raised AST and ALT can be indicative of hepatitis (inflammation of the liver) regardless of the aetiology. A raised GGT in combination with a raised ALP can indicate an obstructive pattern of liver disease. A raised GGT in combination with a raised ALT can be indicative of viral hepatitis. GGT was raised in combination with ALT in 4.9% of Gulf War veterans and 4.3% of the comparison group. The proportion of Gulf War veterans with each of these combinations was slightly greater than that of the comparison group, however the odds ratios were not significantly increased (Table 10.8).
An abnormal bilirubin can be indicative of any type of liver disease, and an elevated bilirubin in combination with an elevation of one other liver enzyme is suggestive of underlying liver disease. Very small numbers in both study groups had a raised bilirubin in combination with other liver enzymes, and no Gulf War veterans had an elevated bilirubin in combination with ALP. None of the odds ratios were significantly increased.
| GWV | Comp grp | |||||||
|---|---|---|---|---|---|---|---|---|
|
Combination of parameters
|
n | (%) | n | (%) | Crude OR | Adj OR | 95% CI | P value |
| AST and ALT >ref interval |
58
|
(4.3)
|
54
|
(4.0)
|
1.1
|
1.0
|
0.7-1.5
|
0.984
|
| GGT and ALP >ref interval |
22
|
(1.6)
|
14
|
(1.0)
|
1.6
|
1.6
|
0.8-3.1
|
0.208
|
| GGT and AST >ref interval |
36
|
(2.6)
|
28
|
(2.1)
|
1.3
|
1.2
|
0.7-1.9
|
0.563
|
| GGT and ALT >ref interval |
67
|
(4.9)
|
59
|
(4.3)
|
1.1
|
1.1
|
0.8-1.6
|
0.635
|
| Bilirubin and GGT >ref interval |
1
|
(0.1)
|
4
|
(0.3)
|
0.3
|
nc*
|
nc
|
nc
|
| Bilirubin and ALP >ref interval |
0
|
(0.0)
|
4
|
(0.3)
|
0.0
|
nc
|
nc
|
nc
|
| Bilirubin and AST >ref interval |
4
|
(0.3)
|
4
|
(0.3)
|
1.0
|
nc
|
nc
|
nc
|
| Bilirubin and ALT >ref interval |
4
|
(0.3)
|
6
|
(0.4)
|
0.7
|
0.7
|
0.2-2.5
|
0.564
|
* nc = not calculable
Very small numbers of participants had a positive hepatitis C core Ab status in conjunction with elevated levels of ALT {5 (0.1%) Gulf War veterans and 1 (0.1%) of the comparison group}, AST {3 (0.2%) Gulf War veterans and none of the comparison group} or bilirubin {none of the Gulf War veterans or comparison group}. This suggests that hepatitis C status was not an explanation for these raised liver function test parameters.
back to top10.4.2.4 Serological test results
Table 10.9 shows the prevalences and ORs for the serological test results, which are similar in the Gulf War veteran and comparison group.
| GWV | Comp grp | |||||||
|---|---|---|---|---|---|---|---|---|
|
Parameter
|
n | (%) | n | (%) | Crude OR | Adj OR | 95% CI | P value |
| Epstein-Barr virus antibody test, IgG Ab | ||||||||
| Not detected, Δsample < 0.100 |
56
|
(4.1)
|
65
|
(4.8)
|
-
|
-
|
-
|
-
|
| Equivocal, 0.100 - 0.200 |
48
|
(3.5)
|
31
|
(2.3)
|
1.8
|
1.6
|
0.9-2.8
|
0.132
|
| Detected, Δsample >0.200 |
1274
|
(92.5)
|
1270
|
(93.0)
|
1.2
|
1.1
|
0.7-1.5
|
0.797
|
| Cytomegalovirus antibody test, IgG Ab | ||||||||
| Not detected, <10 AU/ml |
671
|
(48.7)
|
666
|
(48.8)
|
-
|
-
|
-
|
-
|
| Equivocal, 10-15 AU/ml |
3
|
(0.2)
|
8
|
(0.6)
|
0.4
|
0.3
|
0.1-1.2
|
0.089
|
| Detected, >15 antibody units/ml (AU/ml) |
704
|
(51.1)
|
692
|
(50.7)
|
1.0
|
1.1
|
0.9-1.2
|
0.487
|
| Hepatitis C serology, hepatitis C core Ab | ||||||||
| Negative, S/CO <1.00 |
1367
|
(99.2)
|
1352
|
(98.9)
|
-
|
-
|
-
|
-
|
| Indeterminate, S/CO 0.80-0.99 |
0
|
(0)
|
6
|
(0.4)
|
0.0
|
0.0
|
-
|
-
|
| Positive, S/CO >=1.00 |
11
|
(0.8)
|
9
|
(0.7)
|
1.2
|
1.2
|
0.5-3.0
|
0.683
|
Note: The word 'sample' in the table above should be preceded by the Greek letter Delta. This may not display correctly in some browsers.
10.4.2.5 Urinalysis results
Urinalysis was performed by the HSA nurse during the medical assessment. Table 10.10 shows the urinalysis results. Urinalysis testing was performed by colorimetric assessment, and therefore there may have been some interobserver variability. ‘Traces’ are not considered ‘normal’ while ‘1-4+, positive or small-large’ are considered definitely abnormal. The Gulf War and comparison groups were not significantly different when compared for haematuria (blood in the urine), glycosuria (glucose in the urine), proteinuria (protein in the urine) or nitrites (which can indicate urinary infection). The proportions of both the Gulf War veteran and comparison group who had 1-4+ glycosuria was very similar, and similar to the proportions (0.8% vs 0.7%) that had an elevated random blood glucose. Although the numbers and proportion of participants who had positive results for blood or protein in their urine were small, these tests are not diagnostic and would generally require further investigation in clinical practice.
| GWV | Comp grp | |||||
|---|---|---|---|---|---|---|
|
Parameter
|
n | (%) | n | (%) | P value | |
|
Protein
|
||||||
| Negative |
1298
|
(94.0)
|
1303
|
(94.6)
|
 |
p=0.435
|
| Trace |
70
|
(5.1)
|
58
|
(4.2)
|
||
| 1-4+ |
13
|
(0.9)
|
17
|
(1.2)
|
||
|
Blood
|
||||||
| Negative |
1311
|
(94.9)
|
1306
|
(94.8)
|
|
p=0.878
|
| Trace |
44
|
(3.2)
|
48
|
(3.5)
|
||
| Small-large |
26
|
(1.9)
|
24
|
(1.7)
|
||
|
Glucose
|
||||||
| Negative |
1365
|
(98.8)
|
1363
|
(98.9)
|
|
p=1.000*
|
| Trace |
2
|
(0.1)
|
1
|
(0.1)
|
||
| 1-4+ |
14
|
(1.0)
|
14
|
(1.0)
|
||
|
Nitrites
|
||||||
| Negative |
1377
|
(99.8)
|
1369
|
(99.9)
|
 |
p=0.625*
|
| Positive |
3
|
(0.2)
|
1
|
(0.1)
|
||
* P value from Fisher’s Exact test.
10.4.3 Gulf War veteran group subanalysis
Overall, the patterns of results for Gulf War veterans and the comparison group were very similar, and further subanalysis in relation to laboratory investigations and exposures in the Gulf War group was not undertaken.
10.5 Discussion
The laboratory investigations component of the study involved the collection of blood from 2810 consenting participants at 10 HSA offices and two additional medical assessment sites, and its national transportation to a single laboratory, IMVS, for analysis. A number of haematological, biochemical and serological tests were used to compare the Gulf War and comparison groups, and definitions for adverse health outcomes were developed in order to address the research questions of the study.
The data collection for this aspect of the study required considerable and sustained effort and attention to detail by those involved in order to achieve such a high level of data completeness and quality. The standardisation of results achieved by using one national laboratory in this study was of prime importance. In this study, bloods were transported nationally, albeit with a protocol for the initial processing and transportation of specimens in order to minimise any adverse effects on the specimens, and IMVS aimed to have all bloods tested within 24 hours. However, some parameters are more sensitive than others with respect to the time between blood collection and analysis, and artefact may be an explanation for the elevation of some parameters, such as ESR and MCV in both groups.
The main finding is that the patterns of results for the Gulf War veteran and comparison groups are similar for most parameters. There are minor differences between the study groups on two individual parameters only, with a greater proportion of Gulf War veterans having a sodium or creatinine above the reference interval. There is also a suggestion that liver enzymes were more likely to be elevated in Gulf War veterans, but not significantly. In all of these results, the prevalences were quite small and they need to be interpreted with caution. Also, the clinical significance for these parameters falling above the reference range is uncertain, but may indicate an early pre-clinical phase.
Some test results such as the MCV were elevated in both groups. The MCV tends to increase at the rate of approximately 6% in the first twenty-four hours (Personal communication with IMVS haematologist, August 2000), and artefactual elevation due to the time between blood collection and analysis may be an explanation for the elevated level of MCV in both study groups. This is likely to have affected both groups in a similar manner. High levels of alcohol consumption may also be an explanation. There are other causes of macrocytosis such as nutritional folate and vitamin B12 deficiency, which are less likely in these groups.
An elevated ESR has been reported elsewhere to be present in between 4 and 8 per cent of patients.[356] The ESR tends to decrease with time after collection. For example, an ESR performed at 24 hours may be up to 20% lower than the same test performed on a "fresh" specimen (Personal communication with IMVS haematologist, August 2000), and thus a higher proportion of both study groups may have had an ESR above the reference interval, but again this is likely to have affected both groups in a similar manner. In this regard, the high prevalence of ESR above the reference interval is difficult to explain, unless the strictly defined reference limits used were too simplistic a way of assessing this parameter.
An estimation of blood glucose in the groups was limited to a random blood glucose because of the logistical difficulties of collecting a fasting specimen in this study. Type 2 diabetes mellitus affects approximately 6% of the Australian population aged 25 or older, and 50% of these individuals are undiagnosed and largely asymptomatic. Although a similar and small proportion of the study groups have an elevated random blood glucose that suggests diabetes mellitus is likely, 16.4 % of Gulf War veterans and 15.5% of the comparison group have a random blood glucose which suggests they may be at risk of diabetes mellitus if they have other risk factors. Further testing would be needed to confirm this.
According to the definitions developed for this study, the study groups have a similar prevalence of a combination of parameters that may indicate inflammation. However, a limit in the further interpretation of this is the non-specific nature of the parameters themselves in relation to causes of inflammation such as infection, vasculitis, malignancy and others.
The seroprevalence of CMV, EBV and hepatitis C found in both study groups are comparable to those reported in the scientific and medical literature. CMV infection is common in all human populations, reaching 60-70% in urban US cities.[357] Antibodies to EBV have been found in all population groups studied. By adulthood, 90-95% of most populations have demonstrable EBV antibodies.[358] In developed nations, the hepatitis C prevalence is typically 1-2% of the general population,[359] ie similar to the rates found in this study. Risk factors for hepatitis C include tattoos, blood transfusions and intravenous drug use, but we do not have data on these latter two risk factors from this study.
Although the US and UK clinical evaluation programs undertook further investigations such as laboratory tests in order to establish diagnoses for the self-referred veterans, only one cross-sectional epidemiological study has reported a medical assessment as part of the study. Ishoy et al[162] found minor differences between study groups. Of the parameters included in this study, Danish Gulf War veterans had a slightly elevated mean number of platelets {205 vs 211 x 109/l, p<.05}, a slightly lower eosinophil count (0.18 vs 0.20 x 109/l, P<0.01), and a very slightly elevated plasma creatinine (0.092 vs 0.091 mmol/l, P<.01). Of these previously reported findings, the only similar finding in our study was the significantly greater proportion of Gulf War veterans with an elevated creatinine.
We were unable to locate any published national Australian survey data for laboratory investigations (personal communication with Specialist Pathologist and Head, Diagnostic Services Laboratory, IMVS), and the generalisability of results from other studies may be limited because of the different testing methodologies used. Therefore, comparison of our results with those of the general population are limited to that achieved through the comparison of study groups using the IMVS reference intervals. Laboratory investigations are usually used to assess individuals in a clinical setting where a doctor wants to confirm, or exclude, possible diagnoses, but the use of reference intervals in this study provided a framework for comparing the study groups. The laboratory haematology, biochemistry and liver function test reference intervals were established by IMVS through the testing of 230 “laboratory normal” healthy individuals and comparing the results with those reported in the international scientific literature. Reference intervals have some limitations. The sample used by IMVS to develop their reference intervals was not representative of the general population. The high proportion of males in our study groups may mean that more than 2.5% of the groups would be expected to be outside the reference intervals. Results outside the reference intervals are not necessarily abnormal, and the comparisons made with the comparison group are very important in this respect.
Although the parameters used to compare the study groups in the analysis were decided a priori, a large number of analyses were undertaken to compare the two study groups on the laboratory test results, and this increases the risk of chance findings.
As part of the feedback of results to participants, a medical report was prepared by a HSA doctor who completed their medical assessment. This medical report was provided to each participant for follow up as required, and a copy was provided to their medical practitioner if they chose to nominate one.
10.5.1 Summary of findings
In summary, and in answer to the research questions, the laboratory investigation results of Gulf War veterans were very similar, with only minor differences between the study groups. The proportion of Gulf War veterans that are outside the reference intervals for individual parameters tends to be higher than the comparison group. Some of these differences may be due to artefact or due to technical difficulties.
Gulf War veterans do not have significantly more haematological test results that are indicative of anaemia or inflammation than the comparison group. A relatively high, but similar, proportion of Gulf War veterans and comparison group have haematological test results that are indicative of inflammation, according to the study definitions. The elevation of indicators of inflammation are based on non-specific parameters in both groups, and this finding may warrant further investigation in relation to possible exposure subgroups.
Gulf War veterans have a greater proportion of results that may indicate renal impairment, but the numbers are small and need to be interpreted with caution. Only one Gulf War veteran and no comparison group subjects had both urea and creatinine elevated (which is a more specific indicator of renal impairment). Gulf War veterans do not have more biochemical tests results that are indicative of raised blood glucose than the comparison group. Gulf War veterans have slightly greater proportions of combinations of liver enzymes that may be indicative of liver disease than the comparison group, but the differences are not statistically significant.
Some of the differences found, such as the slightly greater proportions of combinations of liver enzymes indicating liver disease and a raised creatinine indicating renal disease may need to be investigated further on an individual basis. Alcohol consumption may be a cause of the elevation of some haematological or liver function test parameters in both groups, although other explanations are possible. The high proportion of both study groups that have a random blood glucose within a range that suggests they may be at risk of diabetes warrants follow-up and if they have other risk factors, may need to be investigated further on an individual basis.
Gulf War veterans do not have significantly more serological test results that are indicative of prior exposure to viral infections with Epstein-Barr virus (EBV), Cytomegalovirus (CMV) or hepatitis C than the comparison group.
Australian Gulf War veterans do not have more have more abnormalities on urinalysis testing when compared for haematuria (blood in the urine), glycosuria (glucose in the urine), proteinuria (protein in the urine) or nitrites (which can indicate urinary infection) than the comparison group.
back to top