The Relationship between Military Combat and Cardiovascular Risk: A Systematic Review and Meta-Analysis

Table 1

Description of individual studies and their outcomes and findings.

Author year	Population	Numbers and type of exposure	Study design	Age in years	Male, % in combat group	Follow up	Outcomes	Key finding/covariate adjustment
Combat + traumatic injury
Hrubec and Ryder 1980 [9]	US military WWII (1944–45) veterans	3890 proximal amputees	Retrospective cohort	>80% <30 years old at time of injury	100%	>30 years	All-cause and disease specific mortality	↑ adjusted all-cause (RR : 1.36 : 1.25–1.48) CVD (RR : 1.58 : 1.40–1.79) and CHD related death (RR : 1.56 : 1.36–1.79) among proximal amputees vs. injured. ↑ risk of all-cause (1.29 : 1.18–1.41), CVD (1.44 : 1.26–1.64) and CHD (1.45 : 1.24–1.68) death among proximal vs distal amputees and vs general population.
		2917 distal amputees
				3 groups age matched
		3890 injured		Ages at analysis not provided
		US population (age matched)
Labouret et al. 1987 [35]	French veterans	106 with combat related amputation (49 AKA)	Cross-sectional	Compared by age decades from 40–89 years	100%	>15 years	Systolic and diastolic blood pressure	Higher unadjusted prevalence of systolic (not diastolic) HTN in the amputees vs controls (56% vs. 29%; ) and significant for each age decade comparison.
	WWI (1914)	184 age matched controls without HTN
	WWII (1939)
	Other
Rose et al. 1987 [36]	US Vietnam War veterans	19 AKA	Cross-sectional	20–22 at injury and 35–36 years at analysis	100%	≥;15 years	Insulin response to glucose infusion	↑ unadjusted rate of HTN (10/19) in amputees vs controls (1/12; ); no difference lipid levels.
		12 age matched controls
Vollmar et al. 1989 [34]	German WWII (1939–1945) veterans	329 veterans with AKA	Cross-sectional	67.2 years AKA	100%	43.8 years from injury	Ultrasound diagnosis of infrarenal abdominal aortic aneurysms	↑ AKA in amputees vs controls (5.8% vs. 1.1%); no differences in risk of HTN, hyperlipidemia and DM (comparative data not reported)
		702 nonamputee veterans
				68.1 years controls with comparable burden of CVD risk factors
Yekutiel et al. 1989 [26]	Israeli War Veterans wars (1948–9, 1956, 1967, 1973)	53 traumatic lower limb amputees	Cross-sectional	57.2 years	100%	>20 years from injury	Hypertension, CHD and DM	↑ unadjusted prevalence of CHD in amputees vs controls (32.1% vs. 18.2%; ) and DM (22.6% vs. 9.4%); no difference in HTN (35.8% vs. 35.2%)
		159 age and sex-matched controls
Lorenz et al. 1994 [25]	German population conflicts not stated	226 veterans with traumatic lower limb amputations	Cross-sectional	Age not reported (short report)	Not reported	Unreported but >1 year	Ultrasound diagnosis of abdominal aortic aneurysms	No difference in prevalence of aortic aneurysms among amputees (4.4%) vs controls (4%). No difference in risk of hypertension, diabetes or hyperlipidemia.
		199 controls
Peles et al. 1995 [43]	Israel defence force veterans 1948–1974	52 Amputees	Cross-sectional	Amputees 52 years controls 53 years	100%	33 years after injury	Insulin resistance and autonomic function	Age adjusted ↑ in insulin levels among amputees vs controls; No unadjusted difference in glucose, lipids and blood pressure
		53 nonmilitary controls
Modan et al. 1998 [19]	Israeli army wounded 1948–1974	Cohort 1 201 veterans + traumatic lower limb amputation 1832 general US population	Retrospective cohort study	50% <40 years	100%	24-year	All-cause CVD and non CVD mortality	Two fold ↑ (amputees vs. controls) in unadjusted risk of all-cause (21.9% vs. 12.1% among older) and CVD-related death (8.9% vs. 3.8%,).
		Cohort 2 101 amputees 96 controls (matched by age and ethnicity)	Cross-sectional
							CV risk factors	Cohort 2 ↑ plasma insulin levels (2 hour post oral glucose load) in amputees; No differences in unadjusted CHD (19.8% vs. 16.7%), cerebrovascular disease (3.0% vs. 5.2%), obesity, DM, HTN (43.6% vs. 35.4%), hyperlipidemia (37.6% vs. 30.2%)
Shahriar et al. 2009 [37]	Iranian wars	327 bilateral lower limb amputees	Cross-sectional	42 years at analysis with age of 20.6 years at injury control group age not reported	100%	Mean 22.3	Obesity and CVD risk factors	↑ unadjusted risk of HTN (28.5% vs. 20.4%: ), total and LDL cholesterol () obesity (31.8% vs. 22.3%) and smoking (31.8% vs. 22.3%; ) versus control
		Iranian general population (demographics undefined) [5]
Kunnas et al. 2011 [24]	Finnish Military WWII veterans	102 injured combat veterans	Prospective cohort study	55 years	100%	28 years	CHD mortality	(↑ adjusted risk of CHD (HR 1.7 : 1.1–2.5; ) death among injured/wounded vs control. No difference in total cholesterol or DM.
		565 non injured veterans
Stewart et al. 2015 [27]	US Military Iraq and Afghanistan wars 2002–2011	3846 severe traumatic injuries	Retrospective cohort	25–29.2 years	≥98%	1.1–4.3 years	Armed Forces Medical Examiner System (AFMES) database of outcomes	Each 5-point ↑in the ISS linked to a 6%, 13% and 13% ↑ in the adjusted risk of HTN (OR 1.06; 1.02–1.09; ), CAD (1.13; 95% CI 1.03–1.25; ), DM (1.13; 1.04–1.23; ). ↑ Risk versus control population
		Millennium cohort [30, 41]
Ejtahed et al.2017 [46]	Iran veterans of Iran-Iraq War	235 veterans with bilateral traumatic lower limb amputations vs general population	Cross-sectional	31.5 years at injury and 52 years at follow up	100%	32.1 years form injury	Metabolic syndrome	2-fold ↑ in metabolic syndrome, including HTN, insulin levels, hyperlipidemia and obesity (amputees (62.1%) vs general Iranian population (27.5% )
				Age for comparator not reported
Uninjured combat
Bullman et al. 1990 [20]	US Vietnam War veterans	6668 high-combat veterans deaths	Retrospective cohort	Similar ages in both groups	100%	Median follow up >5 years	ICD8 8 codes	↓ in proportionate CVD mortality vs control group (mortality ratio 0.93 : 0.88–0.98).
		27917 low combat veteran deaths
O’Toole et al. 1996 [40]	Australian Vietnam War veterans	641 army veterans (10.8% injured) vs age-sex matched population expected	Cross-sectional	29.5 years at military discharge	100%	>15 years	Self-reported physical health status	↑ adjusted risk of HTN (RR 2.17 : 1.71–2.62), DM (2.71 : 1.32–4.09) and lipids (2.73 : 1.94–3.52); CVD (RR 1.98 : 0.52–2.33) not significant. No relationship between increasing combat burden to any CVD outcomes or risk factors.
MacFarlane et al. 2000 [21]	UK Military veterans of Gulf War I (1990–91)	53416 war veterans	Retrospective cohort	71.5% <30 years at study enrolment	97.7%	8 years	Multiple	No significant difference in all-cause (MRR 1.05 : 0.91–1.21) and CVD mortality (0.74 : 0.49–1.12) among deployed vs nondeployed veterans mortality.
		53450 nondeployed military
Eisen et al. 2005 [42]	US Military Gulf War (1991)	1061deployed war veterans	Cross-sectional	30.9 years deployed	78% in both groups	10 years	Physical health and QOL	No significant difference in adjusted risk of DM (1.52 : 0.81–2.85) or hypertension (0.90 : 0.60–1.33).
		1128 nondeployed		32.6 years non deployed
Granado et al. 2009 [41]	US Military (2001–2003) (25% Iraq and Afghanistan)	4385 combat	Prospective cohort	Not reported	74.8–86%	2.7 years	SF-36 questionnaire arterial hypertension	↑ adjusted incidence of HTN among multiple combat veterans vs. nondeployers (OR 1.33 : 1.07–1.65:).
		4444 deployed noncombat		But grouped by birth decades
		27232 nondeployed
Kang et al. 2009 [28]	US Gulf War (1991) veterans	6111 war veterans	Cross-sectional analysis of prospective cohort	31.5 years for war veterans	79.9% active 78.2% control	14 years	Health questionnaires	↑ adjusted self-reported prevalence of HTN (RR 1.11 : 1.04–1.19), stroke (RR 1.32 : 1.14–1.52), CHD (RR 1.22 : 1.08–1.39) and obesity. No significant difference in DM (RR 1.11 : 0.99–1.25).
		3859 veterans not deployed to Persian Gulf
				33.6 years for control (in 1991)
Johnson et al. 2010 [33]	US Veterans World War II 40.6% (1941–1945), Korean War 34.6% (1950–1953) Vietnam Conflict 16.8% (1961–1975)	1178 combat (13.1% veterans) 2127 noncombat (deployed) veterans	Prospective cohort	19–20 years at enrolment	100%	36 years after military entry	Carotid intima-media thickness (CIMT) and carotid plaque	↑ age-adjusted CIMT in combat veterans (Risk difference 12.79 µm : 0.72–24.86) noncombat veterans. No significant difference in carotid plaque noted.
				57.3 years combat veterans
		2,042 nonmilitary
				51.8 years non veterans
				54.1 years non-combat veterans
Johnson et al. 2010 [44]	US Veterans World War II 40.6% (1941–1945), Korean War 34.6% (1950–1953) Vietnam Conflict 16.8% (1961–1975)	1178 combat veterans (13.1% injured)	Prospective cohort	19–20 years at enrolment	100%	36 years after military entry	Myocardial infarction unstable angina or CHD-related death	No significant differences in adjusted CHD between combat (13.2%) and noncombat veterans (11.3%), and nonveterans (11.6%); similar ischaemic stroke risk (7.76% vs. 5.22% vs. 6.43%). ↑ prevalence of DM combat vs noncombat but no significant difference in HTN, lipid profiles.
				57.3 combat veterans
		2127 noncombat (deployed) veterans
				51.8 non veterans
		2,042 nonmilitary		54.1 non-combat veterans
Crum-Cianflone et al. 2014 [30]	US Military Iraq and Afghanistan wars 2001–2009	12280 deployed combat	Prospective cohort	34.4 years at baseline and mean age at CHD diagnosis 43.1 years (comparative ages not reported)	84.4%	5.6 years	Coronary heart disease	Combatants ↑ adjusted (age, sex, race) risk of CHD (OR 1.63 : 1.11–2.40) vs deployed noncombat servicemen but ↓ unadjusted risk of DM and hypertension.
		10602 deployed noncombat
		37143 nondeployed military
Schlenger et al. 2015 [22]	US Vietnam War veterans	1632 theatre veterans	Retrospective cohort	41.5 years theatre veterans	>95%	>10 years	ICD codes for causes of death	No significant difference in all cause (16.79% vs. 16.61%), CVD (5.23% vs. 3.81%) or CHD-related (3.02% vs. 2.33%) deaths.
		716 Era (noncombat) veteran controls		40.9 years control
Barth et al. 2016 [23]	UK Gulf War (1991)	621901 Gulf War veterans 746247 noncombat veterans	Retrospective Cohort	28 years – war veterans	93% active	13.6 years	All cause and disease specific mortality (ICD-9)	No difference in adjusted CVD mortality among Gulf War vs noncombat veterans (0.99 : 0.093–1.05) but ↓ all-cause mortality (RR 0.97 : 0.95%–0.99%). ↓ risk of all cause (RR 0.49 : 0.48–0.50) and CVD (RR 0.43 : 0.42–0.45) related mortality in Gulf War veterans vs. US population.
				30 years – noncombat veterans	86.7% control
		US general population			Significant
Sheffler et al. 2016 [32]	US Vietnam War veterans 1959–1973	107 combat veterans	Cross-sectional	45.4 years – combat	100%	10 years	Multiple health outcomes	↓ adjusted (OR 0.25 : 0.09–0.63; ) rate of diabetes among noncombat servicemen. No difference in unadjusted CHD, hypertension, heart attacks or stroke.
		620 noncombat controls		46.0 years – noncombat
Thomas et al. 2017 [31]	US Military veterans Vietnam war (43.6%)	564 combat veterans (29.2% injured)	Cross-sectional	59.0 years – combat	87.6–93%	>20 years	Validated health questionnaires	↑ adjusted risk of stroke (OR 1.38 : 1.03–3.33); no difference in adjusted risk of heart attacks, high cholesterol HTN and other heart disease.
				61.3 years – noncombat
		916 noncombat veterans
Hinojosa 2018 [29]	US Military Iraq and Afghanistan Wars 2012–2015	14932 combat veterans	Cross-sectional	56.1 years – veterans 48.8 years – control	66.3% in military group vs 42% in nonmilitary controls	>1 year	CVD outcomes	↑ adjusted prevalence of HTN in veterans (OR 1.49 : 1.23–1.81), CHD (OR 1.55 : 1.0–2.40), and heart attacks (2.26 : 1.41–3.62); ↑ rates of stroke among male only veterans (OR 3.32 : 2.03–5.47).
		135135 civilians

CHD, coronary heart disease; DM, diabetes mellitus; CVD, cardiovascular disease; HTN, hypertension, Results presented in brackets as odds ratio, relative risk and 95% confidence intervals unless stated; CHD, coronary heart disease; DM, diabetes mellitus; CVD cardiovascular disease; HTN, hypertension; AKA, above knee amputation; ISS, injury severity Score. Results presented in brackets as odds ratio (OR), relative risk (RR), mortality rate ratio (MRR), hazard ratio (HR) and 95% confidence intervals unless stated; refers to studies where proportion with traumatic injury <50%. ^∗Detailed demographics for this population either not fully defined or disclosed.