Cause-Specific Mortality in Patients with Gout in the United States Veteran’s Health Administration: A Matched Cohort Study

Abstract

Objective:

Compare all-cause and cause-specific mortality risk between gout and non-gout patients in the Veteran’s Health Administration (VHA).

Methods:

We performed a matched cohort study, identifying patients with gout in the VHA from 1/1999–9/2015 based on the presence of ≥2 ICD-9 codes for gout (274.X). Gout patients were matched up to 1:10 on birth year, sex, and year of VHA enrollment with non-gout patients and followed until death or end-of-study (12/2017). Cause of death was obtained from the National Death Index. Associations of gout with all-cause and cause-specific mortality were examined using multivariable Cox regression.

Results:

Gout (n=559,243) and matched non-gout controls (n=5,428,760) had a mean age of 67 years and were 99% male. There were 246,291 deaths over 4,250,371 patient-years in gout patients and 2,000,000 deaths over 40,441,353 patient-years of follow-up in controls. After matching, gout patients had an increased risk of death (HR 1.09; 95% CI 1.08–1.09) which was no longer present after adjusting for comorbidities (HR 0.98; 95% CI 0.97–0.98). The strongest association of gout with cause-specific mortality was observed with genitourinary conditions (HR 1.50; 95% CI 1.47–1.54). Gout patients were at lower risk of death related to neurologic (e.g., Alzheimer’s, Parkinson’s) (HR 0.63; 95% CI 0.62–0.65) and mental health (HR 0.66; 95% CI 0.65–0.68) conditions.

Conclusion:

A higher risk of death among gout patients in the VHA was related to comorbidity burden. While deaths attributable to neurologic and mental health conditions were less frequent among gout patients, genitourinary conditions were the most overrepresented causes of death.

Gout is the most common type of inflammatory arthritis and is characterized by hyperuricemia that triggers recurrent flares separated by asymptomatic inter-critical periods of variable duration. Almost 10 million individuals in the U.S. are afflicted by gout. The most recent estimates of the burden of gout in the U.S. general population come from the National Health and Nutrition Examination Survey, estimating an approximate prevalence of 4% based on self-reported diagnoses of gout [1]. In the U.S. Veteran’s Health Administration (VHA), an older, male predominant population, the prevalence of gout was previously found to approach 6% based on administrative diagnoses [2]. Slightly lower frequencies of gout have been reported from other parts of the world [3–6].

While an episode of acute inflammatory arthritis from gout prompts patients to seek attention for diagnosis and treatment, gout is recognized to contribute to several poor long-term outcomes. Patients who suffer from gout are more than 16-times as likely to simultaneously suffer from multiple other chronic health care conditions compared to individuals without gout [7]. These chronic health conditions, which constitute an excess comorbidity burden, include cardiovascular disease, chronic kidney disease, and metabolic diseases such as diabetes mellitus [7]. Accompanying a higher burden of comorbid conditions, gout has been associated with increased health care costs and mortality rates [8].

While it has been established that gout patients have increased rates of comorbidity and all-cause mortality, what has yet to be better understood is the cause-specific mortality risk in this patient population. Historically, cardiovascular disease (CVD) has been thought to be the primary driver of excess mortality in patients with gout [9, 10]. CVD is a common comorbid condition in gout, and mechanistically it has been hypothesized that increased serum urate concentrations lead to systemic inflammation and endothelial dysfunction that contributes to atherosclerosis [11]. While CVD represents a major cause of death in gout patients, few prior studies have rigorously examined gout-related risk for other specific causes of death. In one population-based cohort study from Sweden, investigators observed that gout patients were at a higher risk of death from several causes beyond CVD including genitourinary and digestive diseases, as well as infections [12]. Interestingly, the authors also found that patients with gout were at a significantly lower risk of dying from neurodegenerative diseases such as dementia. These latter findings were consistent with earlier studies showing that patients with high serum urate levels demonstrated higher cognitive function and were at a lower risk for the development of Parkinson’s and Alzheimer’s diseases [13, 14]. It remains unknown, however, whether these intriguing findings are generalizable to other populations of patients with gout and if they are reproducible with alternative study designs and data sources. Similarly, initial efforts to estimate cause-specific mortality in gout examined broad categories of death composed of many heterogeneous causes. For example, the risk of death among more specific causes of death such as those attributable to chronic kidney disease or acute renal failure (broadly defined as genitourinary causes) remains poorly defined. Determining the gout-specific risk of death from these causes is exceptionally important to inform future efforts aimed at closing the premature mortality gap that exists in gout [15].

The objective of the current study was to determine cause-specific mortality in a population of well-characterized patients with gout in the VHA relative to matched patients without gout. We hypothesized that in addition to CVD, gout would be associated with an increased risk of death related to genitourinary conditions (particularly chronic kidney disease [CKD]) and simultaneously associated with a lower risk of death from neurodegenerative disease.

Materials and Methods

Study Design and Participants

We performed a matched cohort study within the VHA. We utilized national Veterans Affairs (VA) administrative and electronic health record data from January 1999 through September 2015, which was accessed through the VA Corporate Data Warehouse (CDW) in the VA Informatics and Computing Infrastructure [16]. The study was approved by the VA Nebraska-Western Iowa Health Care System Institutional Review Board.

We identified prevalent and incident gout patients using an administrative algorithm that required patients to have an International Classification of Diseases, 9^th revision (ICD-9) code of 274.xx from at least two separate encounters at least 30 days apart [2]. We limited the time period for cohort creation to September 2015 to avoid misclassification related to ICD-10 implementation that started in October of 2015. Each gout patient was matched with up to 10 non-gout patients by birth year, sex, and year of VHA enrollment. Non-gout patients were defined as having no prior ICD-9 diagnostic codes of gout and no prior pharmacy dispensing of urate lowering therapy (allopurinol, febuxostat, probenecid, or pegloticase). The index date for gout patients was defined as the date corresponding to the 2^nd gout diagnostic code while non-gout patients were assigned an index date that corresponded to the same calendar date as their matched gout counterpart. Patients were followed from the index date until death or end of study follow-up (December 2017), whichever occurred first. Non-gout patients were also censored at the time they fulfilled the gout algorithm, after which they crossed over and contributed to follow-time and events to the gout cohort.

Vital Status and Cause of Death

Vital status and cause of death were obtained through linkage with the National Death Index (NDI). Only underlying cause of death was used in the analysis. ICD-10 codes for cause of death from the NDI were categorized according to the Centers for Disease Control and Prevention cause of death chapters. ICD-10 codes have been used for cause of death in the NDI since 1999. Analyses focused on 13 chapters for which the number of deaths allowed for meaningful analyses: genitourinary, skin, digestive, metabolic, blood, cardiovascular, musculoskeletal, respiratory, nervous system, mental health disease, malignancy, infection, and external causes. To date there have been no validation studies of the NDI for specific causes of mortality, however the NDI has been validated for its accuracy in the identification of cancer-related deaths [17].

Patient Demographics and Comorbidities

Patient demographics (age, sex, race) as well as diagnostic codes were obtained from VA administrative data. BMI and smoking status were obtained from electronic health record (EHR) data. Body mass index (BMI, kg/m²) was calculated from vital signs data using the most recent weight and the modal height [18]. Based on previous work, the determination of smoking status using VA EHR data has excellent accuracy [19]. The presence of comorbid conditions was assessed based on having at least two diagnostic codes from inpatient or outpatient encounters at any time before/on the index date. The Rheumatic Disease Comorbidity Index (RDCI) was calculated as an overall measure of comorbidity burden and has previously been shown to generate values that are strongly predictive of overall mortality and functional status in patients with arthritis and rheumatic diseases [20, 21]. Conditions included in the RDCI include lung disease, myocardial infarction, other cardiovascular disease, stroke, hypertension, depression, diabetes mellitus, ulcer or stomach problem, and cancer.

Statistical Analyses

After matching to account for age, sex, and year of VHA enrollment to account for secular trends, crude associations of gout with all-cause and cause-specific mortality were calculated as simple frequencies and proportions of gout and non-gout patients who died from each cause. Associations were further examined using Cox proportional hazards regression models, first without covariates to better understand the mortality burden in patients with gout. To understand whether mortality in gout was independent of other health conditions and/or patient factors, we subsequently examined multivariable Cox proportional hazards models, adjusting for race, BMI categories, smoking status, and the individual comorbidities that comprise the RDCI [20]. After assessing major categories, we further examined the associations of gout with specific causes of death within categories that conferred the highest risk of death in the gout population and had a sufficient number of deaths for sub-analyses (Supplemental Table 1). We also further examined the associations of gout with specific causes of death within the two categories that conferred the lowest risk of death including mental health and neurologic disease categories. In these sub-analyses, causes of death within the neurologic and mental health chapters were grouped based on the Healthcare Cost and Utilization Project Clinical Classification Software (HCUP-CCS) [22]. We chose to use stratified Cox proportional hazards models for primary analysis as the purpose of this study was to address the etiology of death in gout patients, rather than estimating incidence or predicting prognosis in the presence of competing risk. With this objective, the Cox proportional hazards model is recommended over the use of a Fine-Gray competing risks approach [23]. However, as the Fine-Gray adjustment has been undertaken in similar prior studies, a secondary analysis using a Fine-Gray model was undertaken on a 1% random sample of the study population. A 1% random sample was chosen initially for analysis given the large size of the data set and computational demands posed. Given similarities in estimates generated from this random sample (differences in HR of ≤0.06), Fine-Gray models were not fit using the full dataset. All analyses were completed using Stata v15 (StataCorp, College Station, TX) within the VA Informatics and Computing Infrastructure (VINCI).

Results

Baseline Characteristics

We identified 559,243 gout patients, which were age, sex, and enrollment year matched to 5,428,760 non-gout controls. Baseline patient characteristics are shown in Table 1. Both gout and non-gout patients had a mean age of 67 years and were predominantly (99%) male. White, non-Hispanic race was most frequent in gout (62%) and non-gout (59%) patients. Gout patients were more likely to be overweight or obese (93% vs. 85%), defined by a BMI greater than 25 kg/m². Overall, gout patients had a greater burden of comorbidity based on the RDCI score and individual comorbidities.

Table 1.

Baseline characteristics of gout and matched non-gout patients^*

Characteristic	Gout (n=559,243)	Non-gout (n=5,428,760)
Demographics
Age, years	67 (12)	67 (12)
Male sex, %	99	99
Race/ethnicity, %
White non-Hispanic	62	59
Black non-Hispanic	16	10
Hispanic	3	4
Other	4	3
Missing	15	23
Health factors
Body mass index, %
<20 kg/m2	1	2
20 to <25 kg/m2	6	13
25 to <30 kg/m2	28	37
≥30 kg/m2	65	48
Smoking Status, %
Never	19	17
Former	40	32
Current	36	35
Missing	6	16
RDCI score (range 0–9)	2.0 (1.5)	1.3 (1.5)
MI or atherosclerosis, %	4	2
Other CVD, %	36	22
Stroke, %	3	2
Hypertension, %	73	43
Lung disease, %	12	9
Depression, %	15	12
Diabetes mellitus, %	30	19
Ulcer or stomach problem, %	5	3
Cancer, %	12	9

Values: mean (standard deviation) or %

^*Gout and non-gout cases matched on year of birth, sex, and VHA enrollment year. All other patient characteristics shown differed significantly (p < 0.05) between groups

Abbreviations: CVD, cardiovascular disease; MI, myocardial infarction; RDCI, rheumatic disease comorbidity index

All-Cause Mortality

There were 246,291 deaths over 4,250,371 patient-years in patients with gout and 2,000,000 deaths over 40,441,353 patient-years of follow-up in controls. After matching, gout was associated with a 9% increase in risk of death from all causes (Table 2). Following adjustment for BMI, race, smoking status, and comorbidities, gout patients no longer had an increased risk of all-cause mortality compared to non-gout patients (HR 0.98; 95% CI 0.97–0.98) (Figure 1).

Table 2.

All-cause and cause specific mortality in gout vs. non-gout (unadjusted)

	No. of deaths
Cause of Death	Gout (n=559,243)	Non-gout (n=5,428,760)	Unadjusted HR (95% CI)
All-Cause	246,291	1,995,866	1.09 (1.08, 1.09)
Specific Causes of Death
Genitourinary	11,930	54,876	1.88 (1.84, 1.92)
Skin	476	2,884	1.43 (1.29, 1.57)
Metabolic	13,214	86,629	1.34 (1.32, 1.37)
Digestive System	9,665	65,546	1.34 (1.31, 1.37)
Blood	967	6,345	1.32 (1.23, 1.42)
Musculoskeletal	1,076	7,320	1.29 (1.21, 1.38)
Infection	6,760	48,771	1.24 (1.21, 1.27)
Cardiovascular	98,735	676,850	1.29 (1.28, 1.29)
External Causes	8,872	81,119	0.98 (0.96, 1.00)
Malignancy	51,258	492,660	0.96 (0.95, 0.97)
Respiratory	24,859	249,195	0.87 (0.86, 0.88)
Mental health	7,615	86,577	0.69 (0.68, 0.71)
Nervous System	8,461	115,982	0.60 (0.59, 0.61)

Hazard ratios (HRs) and 95% confidence intervals (CIs) generated using conditional Cox regression model with matching on year of birth, sex, and VHA enrollment year; total follow-up 4,250,371 patient-years in gout and 40,441,353 patient-years in non-gout

Figure 1:

Associations of Gout with All-Cause and Cause-Specific Mortality in the Veteran’s Health Administration; forest plot showing hazard ratios (HRs) and 95% confidence intervals (CIs) for all cause and cause specific mortality in patients with gout relative to matched non-gout controls. Calculated using multivariable cox proportional hazards regression, adjusting for body mass index, race, smoking status, and comorbidities. Patients were matched on year of birth, sex, and year of VHA enrollment.

Cause-Specific Mortality

The frequencies of different causes of death among gout and non-gout patients, along with hazard ratios (HR) and 95% confidence intervals after matching are shown in Table 2. Cardiovascular disease, cancer, and respiratory disease deaths were the most common causes of death in our cohort. The top ten causes of death in each major disease category for the total study population are outlined in Supplementary Table 2. After matching by age, sex, and year of VA enrollment and adjusting for race, BMI, smoking status, and comorbidities, genitourinary disease was the most over-represented cause of death (HR 1.50, 95% CI 1.47–1.54), followed by digestive disease (HR 1.26, 95% CI 1.23–1.29), blood disorders (HR 1.20, 95% CI 1.12–1.29), musculoskeletal disease (HR 1.19, 95% CI 1.11–1.27), skin disease (HR 1.14, 95% CI 1.03–1.26), infection (HR 1.10, 95% CI 1.07–1.13), and cardiovascular disease (HR 1.07, 95% CI 1.06–1.08). After multivariable adjustment, a lower risk of death among patients with gout was observed for nervous system disease (including Alzheimer’s and Parkinson’s Disease) (HR 0.63, 95% CI 0.62–0.65), mental health disorders (including unspecified dementia and vascular dementia) (HR 0.66, 95% CI 0.65–0.68), respiratory disease (HR 0.81, 95% CI 0.80–0.83), malignancy (HR 0.93, 95% CI 0.92–0.94), external causes (HR 0.97, 95% CI 0.95–1.00) and metabolic disease (HR 0.96, 95% CI 0.94–0.98) (Figure 1). The results of the secondary analysis using a Fine-Gray model on a random 1% sample of the study population showed no substantive difference in hazard ratios from the primary analysis (data not shown).

We then assessed associations of gout with individual causes of death among the most overrepresented categories, genitourinary and digestive diseases. These were also the categories comprised of the most heterogeneous set of individual causes of death. Results of matched only and fully adjusted analyses showed a similar trend. Unadjusted analysis results are available for review in Supplementary Table 3. Following the same multivariable adjustments used in the previous model, gout was associated with an increased risk of nephritis (HR 1.91; 95% CI 1.69–2.15), chronic kidney disease (HR 1.71; 95% CI 1.65–1.76), acute renal failure (HR 1.54; 95% CI 1.49–1.60), and urinary calculus (HR 1.10; 95% CI 0.79–1.53) (Figure 2), although this latter estimate did not reach statistical significance. Within the broad category of digestive diseases, gout was associated with a higher risk of hepatic disease (HR 1.43; 95% CI 1.30–1.58), gastritis (HR 1.51; 95% CI 0.99–2.30), GI hemorrhage (HR 1.13; 95% CI 1.07–1.20), and GI perforation (HR 1.10; 95% CI 0.99–1.22) but not with GI ulcer (HR 0.96; 95% CI 0.79–1.17) (Figure 3).

Figure 2:

Associations of Gout with Genitourinary Disease Mortality in the Veteran’s Health Administration; forest plot showing hazard ratios (HRs) and 95% confidence intervals (CIs) for specific genitourinary causes of mortality in patients with gout. Calculated using multivariable cox proportional hazards regression, adjusting for body mass index, race, smoking status, and comorbidities. Patients were matched on year of birth, sex, and year of VHA enrollment.

Figure 3:

Associations of Gout with Digestive System Disease Mortality in the Veteran’s Health Administration; forest plot showing hazard ratios (HRs) and 95% confidence intervals (CIs) for specific digestive system/gastrointestinal causes of mortality in patients with gout. Calculated using multivariable cox proportional hazards regression, adjusting for body mass index, race, smoking status, and comorbidities. Patients were matched on year of birth, sex, and year of VHA enrollment.

We also assessed associations of gout with individual causes of death among the categories which conferred the lowest risk of death, including mental health and neurologic disease categories, to better assess whether this relationship was attributable to dementia (including Alzheimer’s, vascular dementia, and unspecified dementia) or Parkinson’s disease. Results of matched only and fully adjusted analyses were similar to the estimates obtained for overall neurologic and mental health related survival (see Supplementary Table 3.). Gout was associated with a lower risk from dementia (including Alzheimer’s, vascular dementia, and unspecified dementia) (HR 0.60; 95% CI 0.59–0.61) and Parkinson’s disease (HR 0.57; 95% CI 0.54–0.60).

Discussion

To date, this is among the first of studies to examine cause-specific mortality risk in gout focused on a large, US based patient population. While previous studies in other populations have focused largely on all-cause mortality or mortality related to cardiovascular disease (CVD) in patients with gout, our study demonstrates a significant increase in mortality risk from causes other than CVD. Specifically, we observed significant reductions in survival related to genitourinary conditions, namely acute or chronic renal failure as well as nephritis, followed by other causes including digestive disease, blood disorders, musculoskeletal conditions, skin disease, infection, and cardiovascular disease. Our study also found that gout patients were less likely to die from neurologic diseases, particularly neurodegenerative conditions, such as dementia, vascular dementia, Parkinson’s, or Alzheimer’s disease.

Although an association between gout and renal disease has been well established, few studies have examined mortality risk in gout patients attributable to this specific cause of death. Questions remain, however, concerning the direction of this relationship. Does excess uric acid and/or inflammation related to recurrent gout flares lead to damage of the genitourinary system or is it simply that patients with damaged genitourinary systems (i.e., reduced renal function) are more likely to accumulate uric acid in the context of reduced urate clearance and then proceed to develop gout? While uric acid is predominantly excreted through the urine and a decrease in renal elimination is associated with hyperuricemia, earlier studies have suggested mechanisms by which uric acid directly induces oxidative stress, endothelial dysfunction, and vascular smooth muscle cell proliferation [24, 25]. These changes then may lead to increased glomerular pressure as well as remodeling to the renal vasculature. Obermayr and colleagues further supported these findings by showing that gout patients with the same eGFR were more likely to have progression of their kidney disease in a dose-dependent fashion based on serum urate concentration. In this study, the authors were also able to show that the decrease in renal function was preceded by elevations in serum urate [26]. Taken together, results from these prior studies suggest a relationship between gout and renal disease that may be bidirectional, with the perpetuation of a “vicious cycle” that likely contributes to the reduced survival that we and others have shown to be characteristic of gout.

Findings of increased mortality related to genitourinary disease are consistent with results of a recent Swedish population-based study [12]. In our study of the U.S. VHA, we expanded on these observations by examining associations of gout with specific causes of death comprising the genitourinary category. Our findings demonstrating a more than 2.3-fold increased risk of death from nephritis were unexpected and may provide indirect support that gouty nephropathy represents a direct disease manifestation, a concept in gout that remains controversial [27]. It is important to note that this increased risk of death from nephritis could be due to misclassification or because death from nephritis is overall uncommon in the general population. There have been a number of reports, which have implicated hyperuricemia as an independent risk factor for the development of CKD [28, 29]. Despite these intriguing epidemiologic observations, other studies have shown that the treatment of asymptomatic hyperuricemia with xanthine oxidase inhibition does not delay the progression of CKD (27, 28). An important caveat to these studies, however, is that patients enrolled into these trials were already experiencing progressive renal disease (Stage 3 CKD or greater) and therefore the initiation of urate-lowering therapy may have been too late in the disease process to prevent progression [30].

Our study also found that gout patients were more likely to die from causes related to digestive or gastrointestinal diseases. Hepatic disease and gastrointestinal hemorrhage were the most overrepresented causes of death within this broad category. Although we were unable further delineate between different causes of liver disease, we anticipate that a proportion of these were likely to be related to excess use of alcohol, which is overrepresented in the context of gout [31]. Likewise, both hyperuricemia and gout have been linked to the risk of non-alcoholic fatty liver disease (NAFLD), raising speculation that uric acid itself drives an inflammatory process within the gastrointestinal system (30). Using data from the National Health and Nutrition Examination Survey (NHANES), Afzali and colleagues found that patients with higher serum urate concentrations had a higher risk of cirrhosis related hospitalization or death and a greater probability of having elevation of serum transaminase concentrations [32]. Another important factor involved in the development of hepatic disease, specifically NAFLD, is the presence of obesity, particularly since multiple studies have demonstrated positive correlations between serum urate and BMI [33–36]. We did, however, adjust our models for BMI and we continued to show an increased risk of death from hepatic disease in gout patients, suggesting that the increased risk of death is not likely from obesity alone. This notion is supported by work done by Yang and colleagues, which evaluated 800 NAFLD patients and showed that non-obese patients with NAFLD were less likely to go into NAFLD remission if they were hyperuricemic, whereas the remission rates were similar for obese patients with NAFLD regardless of their serum urate levels [37]. Although mechanisms underpinning the association of mortality from gastrointestinal perforation or hemorrhage in gout are unknown, it is possible that these associations could potentially be related adverse effects from medications commonly used to treat gout flares and chronic joint pain, including non-steroidal anti-inflammatory drugs (NSAIDs) and/or glucocorticoids.

Similar to findings from the Swedish population-based study, U.S. veteran patients with gout were also less likely to die from neurologic diseases including dementia, vascular dementia, Parkinson’s, and Alzheimer’s disease compared to individuals without gout. Despite the close similarities in findings between these two studies, there are conflicting reports suggesting that hyperuricemia or gout may predispose to the development of at least some of these neurodegenerative diseases. For example, a prospective cohort study by Serdarevic and colleagues showed that serum urate concentrations were higher in vascular dementia patients compared to controls [38]. Singh and Cleveland found in that gout was independently associated with a 15% higher risk of incident dementia in the elderly using US Medicare claims data [39]. A prospective cohort study of a population of black and white urban-dwelling adults showed that higher serum urate concentrations at baseline were associated with faster cognitive decline over time [40]. Importantly, these studies did not examine different types of dementia separately in their analysis [39, 40]. Mechanistically, it has been suggested that uric acid may act as an important antioxidant in neuronal tissues, while uric acid elsewhere in the body might paradoxically promote oxidative stress. This theory suggesting distinct compartmentalized roles for uric acid has been partially tested in at least one study demonstrating that the administration of urate post-stroke led to improved clinical outcomes at 90 days in women, although the same finding was not observed in men [41]. More recently, others have shown that the administration of inosine, a purine nucleoside, results in increased urate production in both serum and the central nervous system and may lead to slower clinical decline in Parkinson’s disease [42]. Overall, our findings coupled with those of other investigators suggest that further studies of this relationship are warranted as this may have important implications in gout management, particularly for target uric acid levels for patients on urate lowering therapy.

There are limitations to our study, which include the possibility of misclassification, both in terms of gout diagnosis as well as different causes of death. To minimize gout misclassification, we required at least two separate diagnostic codes separated in time, and it is expected this would bias our results towards the null. Misclassification of specific causes of death may be related to ICD-10 coding of causes of death. In select circumstances, it is possible this misclassification of cause of death may be different between gout and non-gout patients. For example, conditions such as cellulitis and osteomyelitis can mimic gout [43, 44]. With its focus on the U.S. VHA, our findings may not be generalizable to non-Veteran populations. However, given the high frequency of gout risk factors such as obesity, diabetes, hypertension, diuretic use and chronic kidney disease, this population is particularly impacted by gout [2].

While hyperuricemia with or without the associated diagnosis of gout has been shown to be associated with increased mortality in several studies, we chose to study patients with a diagnosis of gout [45–48]. Given its retrospective cohort design and the fact that serum urate concentrations are not routinely assessed in real-world care (even among those with gout) or are disproportionately measured in those with more severe gout, we chose not to examine associations between this laboratory measure and mortality given concerns over study bias [49, 50].

In conclusion, we found that patients with gout receiving care in the VHA are more likely than their non-gout counterparts to suffer from mortality related to a number of chronic health conditions beyond CVD, including those related to genitourinary and digestive diseases. We also found that gout patients were at a decreased risk of death from neurodegenerative diseases such Parkinson’s and Alzheimer’s dementia. Together, these findings emphasize the importance of better understanding the interplay between gout and related comorbidity as means of providing a more holistic approach to gout management and improving long term outcomes in this at-risk patient population.

Significance and Innovation.

• This is among the first studies to comprehensively evaluate cause-specific mortality in a U.S. gout population.

• In the Veteran’s Health Administration (VHA) from 1999–2015 excess mortality was present among gout patients and fully explained by their heightened comorbidity burden.

• Relative to matched controls, genitourinary causes including chronic kidney disease and acute renal failure were the most overrepresented cause of death in gout patients.

• Both nervous system and mental health causes of death were less frequent among gout patients than matched controls in the VHA.