Dr. Krishnan has received consulting fees, speaking fees, and/or honoraria from Savient Pharmaceuticals and UCB Pharma (less than $10,000 each) and from Takeda Pharmaceuticals (more than $10,000); he has served as a paid consultant to Cowan and Associates.
Despite the recent doubling of the incidence of gout among women and its substantial prevalence particularly in the aging female population, the risk factors for gout among women remain unknown. We undertook this study to evaluate purported risk factors for incident gout among women and to compare them with those among men.
Using prospective data from the Framingham Heart Study, we examined over a 52-year period (1950–2002) the relationship between purported risk factors and the incidence of gout in 2,476 women and 1,951 men.
We documented 304 incident cases of gout, 104 of them among women. The incidence rates of gout for women per 1,000 person-years according to serum uric acid levels of <5.0, 5.0–5.9, 6.0–6.9, 7.0–7.9, and ≥8.0 mg/dl were 0.8, 2.5, 4.2, 13.1, and 27.3, respectively (P for trend < 0.0001). The magnitude of this association was lower than that among men (P for interaction = 0.0002). Multivariate relative risks conferred by increasing age (per 5 years), obesity (body mass index ≥30 kg/m2), alcohol intake (≥7 ounces of pure alcohol/week), hypertension, and diuretic use were 1.24, 2.74, 3.10, 1.82, and 2.39, respectively (all P < 0.05), for women.
These prospective data with long-term followup provide evidence that higher levels of serum uric acid increase the risk of gout in a graded manner among women, but the rate of increase is lower than that among men. Increasing age, obesity, alcohol consumption, hypertension, and diuretic use were associated with the risk of incident gout among women.
Gout, a common and excruciatingly painful inflammatory arthritis (1), has historically been considered a male disease, and most gout research has focused on men (2–9). However, growing evidence suggests a substantial disease burden of gout among older women, whose representation in the general population has grown with increasing longevity. The prevalence of self-reported physician-diagnosed gout based on the Third National Health and Nutrition Examination Survey (NHANES-III) was 3.5% in women ages 60–69 years, 4.6% in women ages 70–79 years, and 5.6% in women ages ≥80 years (10). The incidence of gout has doubled among women over the past 20 years, according to the Rochester Epidemiology project study (11). Despite this substantial and increasing disease burden, the risk factors for gout among women remain unknown.
Given the significant role of estrogen in serum uric acid levels as well as the substantial sex difference in the incidence of gout and perhaps in uric acid metabolism (12), extrapolation of data on the risk factors for gout from men to women should be done with caution. The serum uric acid level, which is the most prominent predictor and precursor of gout, has been prospectively studied in relation to the risk of incident gout among men (2), but no corresponding data among women are available. The widely cited Normative Aging Study, exclusively based on men (2), showed annual incidence rates of gout per 1,000 person-years of 0.8, 0.9, 4.1, 8.4, and 49.0 for serum uric acid levels of <6.0, 6.0–6.9, 7.0–7.9, 8.0–8.9, and ≥9.0 mg/dl, respectively. It remains unknown whether these levels of serum uric acid pose the same levels of risk of gout among women.
Previous case series studies of gout in women (12–15), which have been the exclusive source for teaching and textbooks, have demonstrated higher mean uric acid levels among women with gout than among men with gout. For example, a previous study found that the mean serum uric acid levels were 9.3 mg/dl in female patients and 8.4 mg/dl in male patients (P < 0.03), and this difference was not explained by diuretic use, hypertension, or renal insufficiency (12). While this difference appears substantial, particularly given the lower baseline serum uric acid level among women than among men, no prospective data are available on this topic. To fill the gap in knowledge of the risk factors for gout in women, we investigated the associations between serum uric acid levels, other purported risk factors, and the risk of incident gout specifically among women in the Framingham Heart Study cohort and compared them with those among men.
PATIENTS AND METHODS
We conducted a cohort study using the prospectively collected data in the Framingham Heart Study original cohort, using the data set obtained from the National Heart, Lung, and Blood Institute (NHLBI). The Framingham Heart Study is an ongoing longitudinal study of 5,209 men and women from the town of Framingham, MA, ages 29–62 years at the time of recruitment in 1948. Subjects have been followed up biennially, and at each study examination data from a detailed medical history, physical examination, and laboratory tests were collected. Additional details of the Framingham Heart Study cohort are described elsewhere (16–18). We limited our analysis to 2,476 women and 1,951 men for whom we had complete followup data and who were free of gout at baseline. Ethics approval for this study was obtained from the University of British Columbia Behavioural Research Ethics Board.
Assessment of exposure variables.
We evaluated serum uric acid levels and purported risk factors for gout that were obtained in the Framingham Heart Study, including age, education, body mass index (BMI), alcohol consumption, hypertension, medication use (diuretics, hormone replacement therapy), clinical laboratory measures (blood glucose level, blood cholesterol level), and menopause status. These factors are closely consistent with those included in the widely cited Normative Aging Study of gout among men (2). Serum creatinine levels were first measured at the 14th examination in the Framingham Heart Study and thus were not included in our primary analysis.
Serum uric acid levels were measured in the Framingham Heart Study for the first 4 examinations and the 13th examination with an autoanalyzer using a phosphotungstic acid reagent (19). Age was ascertained by self-report at the baseline Framingham Heart Study examination, and then subsequent age was calculated from the time between examinations and the baseline age. Information about education was obtained at the baseline examination. Height and weight were measured at each examination, and BMI was calculated as the weight in kilograms divided by the square of the height in meters. The information on medication use and alcohol consumption was ascertained by self-report (20). Systolic and diastolic blood pressures were measured twice in the left arm, with the patient in a sitting position, using a mercury-column sphygmomanometer positioned near eye level. The average of the 2 readings was used for each blood pressure variable. A person was classified as hypertensive if he or she had a systolic blood pressure ≥140 mm Hg, a diastolic blood pressure ≥90 mm Hg, or was receiving antihypertensive drugs (21). Random blood glucose levels were determined using Nelson's method (22). Cholesterol levels were determined according to the Abel-Kendall method (23). Menopause was defined as the absence of menstruation for ≥1 year (24).
Ascertainment of incident cases of gout.
Our end point was the first recorded event of clinically diagnosed gout. At the Framingham Heart Study examination, a clinical diagnostic impression of gouty arthritis was noted if the subject had experienced acute joint pain, accompanied by swelling and heat, lasting from a few days to 2 weeks, and followed by complete remission of symptoms (25). Evidence of gout was further confirmed when an attack of arthritis promptly responded to therapeutic doses of colchicine or other antigout medications. Since study participants were rarely seen during an acute attack of arthritis, aspiration of joint fluids for examination of uric acid crystals, considered a diagnostic gold standard, was not performed (25). Subjects with a diagnosis of gout at the baseline examination were defined as prevalent cases; subjects who were free of gout at baseline and received a diagnosis of gout at any subsequent Framingham Heart Study examination were defined as incident gout cases.
We calculated the incidence rates of gout according to serum uric acid level categories in the Framingham Heart Study. We computed person-time of followup from our study baseline (1950) to the clinical diagnosis of gout at a subsequent examination, day of the last attended examination, or end of the study period (26th biennial examination, conducted in 2000–2002), whichever came first. Individuals who died or in whom gout was diagnosed at any Framingham Heart Study examination were excluded from subsequent followup.
We used time-dependent Cox proportional hazards modeling to estimate the relative risk (RR) of incident gout associated with the most recent prior serum uric acid levels during the cohort followup. Our primary RRs for the association with serum uric acid levels were unadjusted, because serum uric acid is the direct causal intermediate for other examined risk factors to contribute to the risk of gout (e.g., alcohol causes hyperuricemia, which in turn leads to an increased risk of gout). We categorized serum uric acid levels into the categories of <5.0, 5.0–5.9, 6.0–6.9, 7.0–7.9, and ≥8.0 mg/dl to appropriately reflect their range among women. We used ≥8.0 mg/dl of uric acid as the top category, since only 7 women in this cohort had uric acid levels ≥9.0 mg/dl at least once over the entire followup period. We assessed the trend, using the median values of uric acid level in each category, to minimize the influence of outliers.
Multivariate models were used to evaluate the following purported risk factors in a time-dependent manner: age (continuous, per 5 years), education level (below grade 12, grade 12 completed, above grade 12), BMI (<25 kg/m2, 25–29.9 kg/m2, and ≥30 kg/m2), alcohol consumption (abstinent/light [0–1 ounce, 29.57 ml, per week], moderate [2–6 ounces per week], and heavy [≥7 ounces per week] ), hypertension (yes or no), use of diuretics (yes or no), blood glucose level (continuous, per 10 mg/dl), blood cholesterol level (continuous, per 10 mg/dl), menopausal status (yes or no), and hormone replacement therapy (yes or no). The last 2 categories of BMI employed correspond to overweight and obesity according to the World Health Organization definition (26). The alcohol intake in ounces per week was obtained by multiplying the average amounts of alcohol in beer, wine, and spirits (mixed drinks) times the amount consumed. Total alcohol intake was categorized as abstinent/light, moderate, and heavy, as explained above (20). The same categories were used to classify intakes of beer and spirits, but the top 2 categories were collapsed for wine intake, because no woman with heavy wine intake developed gout in this cohort. We explored potential interaction by sex by testing the significance of interaction terms added to our final multivariate models. For all RRs, we calculated 95% confidence intervals (95% CIs). All P values are 2-sided. All statistical analyses were conducted using SAS software, version 9.1 (SAS Institute, Cary, NC).
The baseline characteristics of the Framingham Heart Study cohort in this study, according to sex, are shown in Table 1. There were 2,476 women, comprising 56% of the cohort. The mean age at baseline was 47 years for women and 46 years for men. The mean baseline serum uric acid level was 4.0 mg/dl for women and 5.1 mg/dl for men.
Table 1. Baseline characteristics of the subjects according to sex*
Women (n = 2,476)
Men (n = 1,951)
Except where indicated otherwise, values are the percentage of subjects. BMI = body mass index.
First assessed at Framingham Heart Study examination 7 (1962).
Over a 28-year median followup period, we identified 104 incident cases of gout in women and 200 in men. The incidence of gout was 1.4 per 1,000 person-years in women and 4.0 per 1,000 person-years in men. The incidence of gout increased with increasing levels of serum uric acid among women, but the rate of increase was lower than that among men (P for interaction = 0.0002) (Figure 1 and Table 2). Specifically, the incidence rates of gout for women per 1,000 person-years according to serum uric acid levels of <5.0, 5.0–5.9, 6.0–6.9, 7.0–7.9, and ≥8.0 mg/dl were 0.8, 2.5, 4.2, 13.1, and 27.3, respectively (P for trend < 0.0001) (Table 2). The RRs for gout among women according to the corresponding serum uric acid categories were 1.00, 3.29, 5.49, 22.09, and 45.75, respectively (Table 2). When we moved our cohort followup baseline to the 13th examination (n = 1,465 women, ages ≥53 years, and 99.6% postmenopausal), the corresponding incidence rates were 1.3, 2.6, 3.7, 13.5, and 28.5, respectively (P for trend < 0.0001), and the corresponding RRs for gout were 1.00, 2.38, 3.36, 17.75, and 34.98, respectively.
Table 2. Incidence rate and multivariate RRs for development of gout according to updated uric acid level (categorical and continuous for women and men)*
Uric acid level, categorical, mg/dl
Uric acid level, continuous, mg/dl
RR = relative risk; 95% CI = 95% confidence interval.
Overall incidence rate.
Adjusted for age (continuous, per 5 years), education level (below grade 12, grade 12 completed, above grade 12), body mass index (<25 kg/m2, 25–29.9 kg/m2, ≥30 kg/m2), alcohol consumption (abstinent/light [0–1 ounce, 29.57 ml, per week], moderate [2–6 ounces per week], heavy [≥7 ounces per week]), hypertension (yes or no), use of diuretics (yes or no), blood glucose level (continuous, per 10 mg/dl), blood cholesterol level (continuous, per 10 mg/dl), and menopausal status (yes or no).
Among the purported risk factors for gout, increasing age, obesity, alcohol consumption, diuretic use, and hypertension (all P < 0.05) were independently associated with the risk of incident gout in women (Table 3). The magnitudes of associations with these factors did not differ significantly from those among men, except for a stronger age effect among women than among men (P for interaction = 0.02) (Table 3). The age-adjusted RR of gout conferred by menopause was 4.18 (95% CI 0.55–31.86), and the multivariate RR was 3.54 (95% CI 0.46–27.15). When we limited our analysis to a subgroup of women who were postmenopausal at baseline (n = 1,249), the multivariate RR conferred by the use of hormone replacement therapy was 0.24 (95% CI 0.03–1.79). Furthermore, when we evaluated the potential effect of adding creatinine in our model using the creatinine values obtained at the 14th examination and later, the multivariate RRs associated with the identified factors and their significance levels did not change materially.
Table 3. Multivariate RRs of risk factors for incident gout according to sex*
Adjusted for age (continuous); the relative risk (RR) for age is the univariate RR only.
Adjusted for age (continuous, per 5 years), education level (below grade 12, grade 12 completed, above grade 12), body mass index (BMI) (<25 kg/m2, 25–29.9 kg/m2, ≥30 kg/m2), alcohol consumption (abstinent/light [0–1 ounce, 29.57 ml, per week], moderate [2–6 ounces per week], heavy [≥7 ounces per week]), hypertension (yes or no), use of diuretics (yes or no), blood glucose level (continuous, per 10 mg/dl), blood cholesterol level (continuous, per 10 mg/dl), and menopausal status (yes or no).
We also evaluated the association between individual types of alcoholic beverages and the risk of incident gout. In women, the multivariate RRs of heavy consumption were 7.10 (95% CI 1.70–29.62) for beer and 2.66 (95% CI 1.24–5.72) for spirits, whereas the corresponding RRs among men were 2.00 (95% CI 1.26–3.19) and 1.82 (95% CI 1.26–2.63). The multivariate RR of moderate or heavy wine intake was 1.46 (95% CI 0.80–2.65) among women and 1.24 (95% CI 0.84–1.84) among men.
In this prospective cohort study, based on 52 years of followup, we found that higher levels of serum uric acid increase the risk of gout among women in a graded manner. The risk of gout among women with a serum uric acid level ≥8 mg/dl was 46 times higher than that among women with a serum uric acid level <5 mg/dl. However, the magnitude of this association among women was significantly lower than that among men, indicating that at a given level of serum uric acid >5 mg/dl, women have a significantly lower risk of developing gout than men. Furthermore, we found that increasing age, obesity, intake of beer and spirits, diuretic use, and hypertension were associated with the risk of incident gout in women. The magnitude of associations with these factors among women did not differ from that among men, except for a stronger age effect among women than among men (P for interaction = 0.02). These associations were mutually independent of these and other potential risk factors for incident gout.
To our knowledge, our findings provide the first prospective, quantitative data on the relationship between serum uric acid levels and the risk of incident gout among women. The overall smaller magnitude of association, compared with that for men, was consistent with the aforementioned study that found a significantly higher serum uric acid level among female gout patients than among male gout patients (12). The possibility of differential ascertainment of gout between sexes is unlikely, since the sex ratio of the incidence of gout in this cohort (2.8 for men:women) was similar to that in previous studies (2.7 for the UK General Practice Research Database study  and 3.3 for the Rochester Epidemiology Project ). Similarly, the consistency between the results from the entire cohort followup of all women and those from the analysis limited to the 13th examination or later (99.6% of women were postmenopausal) does not support the presence of systematic differential misclassification of serum uric acid levels associated with aging or the menopausal transition among women. If confirmed by future studies, these findings would suggest additional reasons for the lower background incidence of gout among women that go beyond their lower baseline serum uric acid levels.
We found that obesity was associated with the risk of incident gout among women, as was the case among men in this cohort and in previous studies (28–32). Hyperuricemia has been associated with obesity via both increased production and decreased renal excretion of uric acid (33, 34). Increased insulin levels associated with obesity likely play a role in the link between obesity and the increased risk of gout, since higher insulin levels are known to reduce renal excretion of urate (35–38). For example, exogenous insulin can reduce the renal excretion of urate in both healthy and hypertensive subjects (30, 37, 38). Specifically among women, the Coronary Artery Risk Development in Young Adults study showed that white women with a BMI >23.5 kg/m2 have a 5.7-fold increased likelihood of hyperuricemia compared with white women with a BMI <20.8 kg/m2 (32). In the NHANES-III, a marked increase in the prevalence of the metabolic syndrome and obesity was observed at higher serum uric acid levels (39). Other factors not related to uric acid, including chronic joint trauma due to excess body weight, have also been speculated upon as potential explanations for the link between obesity and gout (34, 40).
The association between hypertension and gout among women in this study is consistent with previously reported associations between hypertension and incident gout in men (2, 4, 28). Previous retrospective case series of gout in women showed that a higher proportion of women than men with gout had hypertension and were treated with diuretics, suggesting that these risk factors may be more important among women (12, 13, 41). Our prospective data also showed that these factors are more common among women with gout than among men with gout (diuretic use, 47% versus 29%; hypertension, 82% versus 69%), suggesting that these risk factors may lead to a larger population-attributable risk of the development of gout among women. Renal urate excretion was found to be inappropriately low relative to glomerular filtration rates in patients with essential hypertension (42, 43). Reduced renal blood flow with increased renal and systemic vascular resistance may also contribute to increasing serum uric acid levels (44). Finally, diuretics also play an important role in the pathogenesis of gout through the combined effect of increased tubular reabsorption of uric acid and volume depletion (29, 45, 46). The results from our study support a significant role of diuretics in the risk of gout among women, independent of the presence of hypertension.
We found that overall alcohol intake was associated with the risk of gout in women. Among individual types of alcoholic beverages, beer and spirits were significantly associated with the risk among women. These findings were similar among men in this cohort as well as in the Health Professionals Follow-up study (5). Furthermore, our results were consistent with the associations with serum uric acid levels among US women in the NHANES-III (47). Interestingly, the same NHANES study also showed that the increase in serum uric acid levels associated with beer or spirit intake tended to be greater among women than among men (47), which was the case in the current study with incident gout as the outcome. Taken together, these data suggest potential effect modification by sex—that is, women may be more susceptible to these alcoholic beverages in terms of risk of hyperuricemia or gout—and call for confirmation by future studies.
A number of mechanisms have been implicated in the pathogenesis of alcohol-induced hyperuricemia, including both decreased urate excretion (48–51) and increased production (52, 53), although the magnitude of association appeared larger among women (47). Ethanol administration has been shown to increase uric acid production by net ATP degradation to AMP (52, 53). Additionally, decreased urinary excretion associated with dehydration and metabolic acidosis (34, 54) may contribute to hyperuricemia associated with ethanol ingestion. Our data provide support for a substantial role of alcohol intake in the risk of incident gout among women, independent of the presence of other risk factors.
The strengths and limitations of our study deserve comment. This study was performed using community-based prospective data from the Framingham Heart Study, and thus findings are likely to be generalizable to women and men in the US. Potential recall bias was avoided in this study, because the exposure assessment of purported risk factors was done before the diagnosis of gout. The baseline mean age of 47 years for women and the infrequent occurrence of gout in the premenopausal period limited our investigation of the potential effect of menopause on the risk of gout. Nevertheless, our point estimates, although nonsignificant, appeared to be consistent with the notion of an increased risk of gout in postmenopausal women and a decreased risk associated with hormone replacement therapy (55, 56).
The definition of gout was based on a clinical diagnosis of gout by examining physicians and did not require observation of urate crystals in joint fluid, as was the case in other large epidemiologic studies of gout (2–9, 25, 57, 58), including the widely cited Normative Aging Study of men (2). However, it is unlikely that potential misclassification of the diagnosis would explain the strong associations observed in this population study. It remains conceivable that the results may be even more striking with more specific case definitions of gout, as was the case in our prospective epidemiologic study of gout for suspected associated factors (3, 5). Nonetheless, confirming these results using such specific case definitions of gout would be valuable. The unavailability of creatinine data between the study baseline and the 13th examination precluded adjustment for reduced renal function in our primary analysis. However, our analysis using the creatinine values obtained at the 14th examination and later suggests that this factor did not materially confound the observed associations. Finally, our study was observational; thus, we cannot rule out the possibility that unmeasured factors might contribute to the observed associations.
In this prospective study over 50 years of followup, higher levels of serum uric acid increased the risk of gout in a graded manner among women, but the rate of increase was lower than that among men. Increasing age, obesity, hypertension, alcohol consumption (beer and spirits), and diuretic use are also associated with the risk of incident gout among women.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Choi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Bhole, de Vera, Choi.
Acquisition of data. De Vera, Krishnan, Choi.
Analysis and interpretation of data. Bhole, de Vera, Rahman, Krishnan, Choi.
The authors thank the Framingham Heart Study coordinators for access to the data set. The Framingham Heart Study is conducted and supported by the NHLBI in collaboration with the Framingham Heart Study investigators. The manuscript was reviewed by the NHLBI for scientific content and consistency of data interpretation with previous Framingham Heart Study publications, and significant comments were incorporated prior to submission for publication. The NHLBI had no role in the design, conduct, analyses, and reporting of the study or in the decision to submit the manuscript for publication. The authors would also like to acknowledge Mr. James Rankin for critical review of the manuscript.