Association of Hyperuricemia With Conventional Cardiovascular Risk Factors in Elderly Patients

Abstract

The relationship between serum uric acid (UA) and cardiovascular risk profile was investigated in 557 outpatients (415 women) aged 60 years and older. Patients were grouped according to a UA cutoff level of 5.5 mg/dL. Prevalence of obesity, hypertension, and impaired glucose metabolism was increased in women with higher UA, who had higher body mass index (37.7±6.9 vs 33.1±5.9 kg/m², P<.001), waist circumference, and serum glucose and triglyceride concentrations than women with lower UA levels. Conversely, men with higher UA levels showed lower high‐density lipoprotein cholesterol and higher left ventricular mass than men with lower UA levels. Estimated glomerular filtration rate was reduced in patients with high UA levels of both sexes (65±17 vs 72±16 mL/min/1.73 m², P<.001, for women; 70±16 vs 76±15 mL/min/1.73 m², P<.03, for men). Grouping patients by sex‐specific median UA concentrations produced similar results. These data indicate that, even in the elderly, UA clusters in a sex‐specific fashion with features of metabolic syndrome and signs of target organ damage.

Data about the relationship between serum uric acid (UA) concentration and cardiovascular risk are conflicting and not uniform between sexes; therefore, complex and uneven explanations about the role of UA in the pathophysiology of cardiovascular disease have been proposed.1, 2, 3, 4, 5, 6 Serum UA elevation in patients at high cardiovascular risk has been considered either as a marker of oxidative stress in conditions of local ischemia1 or as a consequence of insulin resistance, since insulin reduces UA excretion.7 According to another view, UA has a key role in inducing hypertension and insulin resistance itself, in particular during early life; thereafter, when these metabolic derangements are already established, the association between serum UA levels and cardiovascular risk profile could become no longer statistically detectable.4 Indeed, little information is available about the relationship between serum UA concentration and cardiovascular risk in the elderly.

The level at which serum UA should be considered harmful is also debated. Several observations have shown a relationship between serum UA and cardiovascular disease even at concentrations lower than the usual cutoff levels of 6 mg/dL for women and 7 mg/dL for men. Hence, it has been suggested that the above limits could be kept as far as the risk of gout and urolithiasis is concerned, but a concentration of 5.2 mg/dL to 5.5 mg/dL in both sexes may be the preferred cutoff level in the evaluation of cardiovascular risk.4 A recent meta‐analysis has shown that the occurrence of new‐onset diabetes mellitus or impaired fasting glucose is associated with UA levels according to a nonlinear relationship. The sharpest increase for men and women occurs at a concentration of about 5.5 mg/dL.8

In the present study, we evaluated the cardiovascular risk profile according to serum UA levels above or below 5.5 mg/dL in an outpatient population older than 60 years. The prevalence of cardiovascular risk factors according to the sex‐specific median serum UA level was also evaluated.

Patients and Methods

The population of the Improvement of Cardiovascular Risk Profile in Older Neapolitans (ICON) study9 was included in this analysis. All outpatients older than 60 years seen from September 1996 to October 2005 in the Internal Medicine Clinic of two facilities of the Italian National Health Service comprised the original study group. The clinics were located in two low‐income districts of the city of Naples: one in the historical center and the other in the suburban district of Scampia. The present analysis was extended to patients who had a visit up to December 31, 2012, with the same selection criteria. In both clinics, evaluations were conducted by the same physician (TM). On most occasions, patients were referred by their general practitioners for metabolic disorders, mainly obesity, hypertension, and dyslipidemia.

Methods for data collection have been previously described in detail.9 Briefly, a structured questionnaire about lifestyle was administered at the first visit. Six levels of smoking habit were considered: 1 (never smoker), 2 (former smoker), 3 (<5 cigarettes per day), 4 (5–10 cigarettes per day), 5 (11–20 cigarettes per day), and 6 (>20 cigarettes per day). Alcohol intake was categorized into 6 levels: never consumer, former consumer, <15 mL/d of alcohol, 15 mL/d to 40 mL/d, >40 mL/d, and addicted with physical dependence. A glass of wine, the alcoholic drink far more frequent in this population, was considered equivalent to 14 mL of alcohol (125 mL×11.5% alcohol). Coffee was consumed exclusively as espresso, mostly homemade: three levels of intake were considered, from 1 (no coffee at all or former consumer) to 3 (3 cups per day or more). As far as physical activity, patients were assigned to level 1 (sedentary, light housework), 2 (regular housework, slow/short walking), or 3 (hard manual job, brisk walking for more than 1 hour per day, regular gym attendance or sport practicing).

Sitting blood pressure (BP) was used for the present analysis. Measurements were performed with a mercury sphygmomanometer after at least 5 minutes of rest and repeated until stabilization. The average of the last 2 values was considered. Biochemical analyses were executed with standard techniques. All patients with a baseline serum UA measurement available were included in the analysis. Glomerular filtration rate was estimated according to the Chronic Kidney Disease Epidemiology Collaboration formula.10

Obesity was defined as body mass index (BMI) ≥30.0 kg/m². Arterial hypertension, impaired glucose metabolism (diabetes mellitus or impaired glucose tolerance or impaired fasting glucose) and dyslipidemia were diagnosed according to current guidelines. In particular, arterial hypertension was defined as BP ≥140/90 mm Hg or ongoing pharmacologic treatment; diabetes mellitus as fasting plasma glucose ≥126 mg/dL or plasma glucose ≥140 mg/dL 2 hours after a 75‐g oral glucose load or ongoing pharmacologic treatment; impaired glucose tolerance as plasma glucose ≥140 mg/dL and <200 mg/dL 2 hours after the glucose load; impaired fasting glucose as fasting plasma glucose ≥110 mg/dL and <126 mg/dL; and dyslipidemia as fasting total plasma cholesterol ≥200 mg/dL or low‐density lipoprotein cholesterol ≥130 mg/dL or total plasma triglyceride ≥150 mg/dL or ongoing pharmacological treatment.

In a subsample of 160 patients, left ventricular mass (LVM) was measured by echocardiography. Echocardiography was performed using phased‐array commercial echocardiographs with M‐mode, two‐dimensional, and Doppler capabilities. Left ventricular dimensions and septal and posterior left ventricular wall thickness were measured by the American Society of Echocardiography recommendations.11, 12 LVM was obtained by an anatomically validated formula and normalized for body surface area or body height in m^2.7.13

Two groups, according to baseline serum UA levels, were formed. Patients were assigned to group A if their UA was <5.5 mg/dL and to group B if their UA was equal to or above that level. Unpaired t test, one‐way analysis of variance, χ² test, and Wilcoxon rank sum test were performed when appropriate. Pearson correlation analysis was also performed. Two‐tailed significance was calculated for all analyses. SPSS version 14 (SPSS, IBM, Armonk, NY) was used for calculations. Data are expressed as mean±standard deviation.

Results

Five hundred fifty‐seven patients (142 men and 415 women, mean age 67±6 years) were included in the study. Group A (lower UA level) comprised 330 patients and group B (higher UA level) 227 patients. Serum UA concentration was 4.2±0.9 mg/dL and 6.6±0.9 mg/dL in the two groups, respectively. Women represented 84% of group A and 60% of group B (χ² significant at the 0.001 level). Because of this uneven sex distribution, we performed separate analyses for either sex.

Smoking and alcohol consumption were not common in our population. Few patients smoked more than 20 cigarettes per day and no alcoholics with physical dependence were found. Most patients were sedentary and only a few performed regular physical activity during leisure time or were involved in heavy work as part of their employment. In no case did between‐group differences attain statistical significance with the Wilcoxon rank sum test. Lifestyle habits in each group are summarized in Table 1.

Table 1.

Lifestyle According to Serum UA Concentration

	Women		Men
	Group A (Low Serum UA)	Group B (High Serum UA)	Group A (Low Serum UA)	Group B (High Serum UA)
Smoking habit, %
Never/former smokers	85.3	92.0	78.8	83.1
1–10 cigarettes per d	5.8	3.0	11.6	5.6
11–20 cigarettes per d	7.5	3.5	3.8	7.9
>20 cigarettes per d	1.4	1.5	5.8	3.4
Alcohol intake, %
Never/former drinkers	60.4	62.0	36.5	25.6
<14 mL alcohol per d	31.6	29.3	23.1	33.3
15–40 mL alcohol per d	8.0	8.0	36.5	32.2
>40 mL alcohol per d	0.0	0.7	3.8	8.9
Alcoholics	0.0	0.0	0.0	0.0
Coffee intake, %
Never/former consumers	12.3	8.8	7.7	8.9
<3 cups per d	54.3	50.0	61.5	64.4
≥3 cups per d	33.4	41.2	30.8	26.7
Physical activity, %
Sedentary	52.2	55.7	51.0	56.8
Light activity	45.9	42.8	45.1	42.0
Heavy activity	1.9	1.5	3.9	1.2

Abbreviation: UA, uric acid. Each lifestyle habit has been categorized in levels, as explained in Patients and Methods. In this Table, some levels have been joined together. Data are expressed as percentage of patients in each subgroup. All analyses were not significant at the Wilcoxon rank sum test.

A high prevalence of overweight and obesity were found in the study population, as shown in the Figure. Most patients were hypertensive and/or dyslipidemic and more than a fourth had impaired glucose metabolism.

Figure 1.

Prevalence of metabolic abnormalities according to serum uric acid concentration. Group A: low serum uric acid. Group B: high serum uric acid. Impaired glucose metabolism indicates diabetes mellitus or impaired glucose tolerance or impaired fasting glucose.

Obesity, impaired glucose metabolism, and arterial hypertension were more frequent in group B among women; at variance, prevalence of dyslipidemia in women was not different in the two groups. No between‐group difference in the distribution of these metabolic abnormalities was found in men (Figure).

BMI, waist circumference, and serum glucose and triglyceride levels were higher in women in group B, as compared with their group A counterparts (Table 2). Again, male participants did not show any between‐group difference in these variables; however, men in group B had lower high‐density lipoprotein (HDL) cholesterol levels than men in group A (Table 2). BP was similar in the two groups, both in women (group A, 153/84±24/12 mm Hg; group B, 153/86±22/12 mm Hg) and in men (150/84±23/11 mm Hg and 149/84±25/12 mm Hg, respectively). Estimated glomerular filtration rate (eGFR) was significantly reduced in group B patients of both sexes, as compared with group A patients (Table 2). Chronic kidney disease (CKD) (National Kidney Foundation stage 3 or higher) was present in 20% of women in group A and in 39% of women in group B (χ² significance <.001); the corresponding figure for men was 12% and 32%, respectively (χ² significance <.02). When present, CKD was mild in most cases. Very few patients had stage 4 (women, 3 cases in group A and 1 in group B; men, 1 and 2 cases, respectively) and none had stage 5 CKD.

Table 2.

Clinical and Metabolic Parameters According to Serum UA Concentration

	Women			Men
	Group A (Low Serum UA)	Group B (High Serum UA)	P Value	Group A (Low Serum UA)	Group B (High Serum UA)	P Value
Body mass index, kg/m2	33.1±5.9	37.7±6.9	<.001	29.9±5.7	30.8±4.1	ns
Waist, cm	106±11	112±13	<.001	114±10	109±10	ns
Glucose, mg/dL	104±30	112±31	<.01	113±40	107±18	ns
Triglycerides, mg/dL	133±65	150±71	<.02	126±61	141±65	ns
HDL cholesterol, mg/dL	55±13	55±13	ns	53±18	48±12	<.04
eGFR, mL/min/1.73 m2	72±16	65±17	<.001	76±15	70±16	<.03

Abbreviations: eGFR, estimated glomerular filtration rate; HDL, high‐density lipoprotein; UA, uric acid.

LVM adjusted by height was significantly greater in men in group B as compared with men in group A (62.5±18 g/m^2.7 vs 54.6±11 g/m^2.7, P<.04; N=61). A similar difference was found when LVM was adjusted by body surface. At variance, no between‐group difference in LVM was seen in women.

Serum UA was positively correlated with BMI and with serum triglyceride levels and was inversely correlated with eGFR in both sexes. In women, UA was also positively correlated with waist circumference, diastolic BP, and blood glucose, while in men a negative correlation with HDL cholesterol was found (Table 3). A direct correlation was present in both sexes between serum UA and triglyceride/HDL cholesterol ratio, an index of insulin resistance.14, 15 Moreover, serum UA was correlated with serum alanine aminotransferase activity in women (r=0.107, P<.05) but not in men.

Table 3.

Correlation of Serum Uric Acid Concentration With Some Clinical and Metabolic Parameters

	Women		Men
	r	P Value	r	P Value
Body mass index	0.352	<.001	0.181	.031
Waist	0.255	<.001	–	ns
Diastolic BP	0.113	.021	–	ns
Glucose	0.138	.005	–	ns
HDL cholesterol	–	ns	−0.242	.006
Triglyceride	0.142	.004	0.280	.001
TG/HDL‐CHOL	0.154	.003	0.264	.003
eGFR	−0.279	<.001	−0.274	.002

Abbreviations: BP, blood pressure; eGFR, estimated glomerular filtration rate; HDL, high‐density lipoprotein; ns, not significant; TG/HDL‐CHOL, triglyceride/HDL cholesterol ratio.

We also performed all the above‐mentioned statistical analyses by splitting the population according to sex‐specific median serum UA level, which was 5.75 mg/dL in men and 4.80 mg/dL in women. Even with this approach, cardiovascular risk profile was differently associated with serum UA in the two sexes. Women whose serum UA level was above the median had a higher chance of being obese and having an impaired glucose metabolism as compared with women with low UA level, while this difference was not found in men. Moreover, BMI and waist circumference were increased in women with serum UA above the median, but not in the corresponding male subgroup. At variance, men with serum UA levels above the median exhibited higher triglyceride and lower HDL cholesterol concentrations compared with their under‐the‐median counterparts. eGFR was reduced in patients of both sexes when UA was above the median. LVM adjusted by body surface was higher in men with serum UA above the median as compared with the other men. All these differences were statistically significant.

Discussion

Our main finding is the association between high serum UA levels and several cardiovascular risk factors in elderly patients, with a different pattern between sexes. In our view, the uneven between‐group distribution of men and women required a separate analysis of the data for the two sexes. In fact, in a pooled analysis, any difference between high‐ and low‐UA groups could have been easily imputed to the sex imbalance, given the obviously different cardiovascular risk profile and lifestyle habits in either sex. Conversely, between‐group differences could have been missed when their trend in men and women was not concordant.

The cutoff value we chose to split UA levels into “high” and “low” is in agreement with epidemiological and clinical observations showing an increase in cardiovascular event rate over 5.2 mg/dL to 5.5 mg/dL in both sexes.4, 8 Moreover, using the sex‐specific serum UA median concentration as a cutoff value, the results remain substantially unchanged.

Conflicting epidemiological evidence about the occurrence of an independent association between serum UA levels and cardiovascular risk16, 17, 18 (see literature revision in reference 1) has cast doubts about the real pathophysiological role of UA. It has been proposed that UA plays a role in priming many metabolic disorders, which could be sustained by other mechanisms later in life, when their link with UA becomes less apparent.4 Indeed, it has been shown that hyperuricemia can predict the development of hyperinsulinemia,19 type 2 diabetes,20 hypertension,21, 22 and even obesity,22 although this phenomenon has not been confirmed in all populations.23 On the other hand, the association of high UA levels with metabolic derangements and markers of atherosclerosis has been poorly explored in elderly patients. In fact, most studies on this issue have been performed in young or adult populations.17, 23, 24, 25, 26, 27, 28 Here, we show that even in the elderly a clear relationship is still ongoing between high serum UA and several cardiovascular risk factors and disease markers.

In particular, we observed the clustering of high serum UA levels with the features of metabolic syndrome in women, while in men high UA was only associated with low HDL cholesterol. Triglyceride/HDL cholesterol ratio, an index of insulin resistance, was correlated with UA levels in both women and men.

A stronger association of UA levels with cardiometabolic comorbidities and cardiovascular outcomes has often been shown in women as compared with men.1, 5, 26, 29, 30 A clear interpretation of this finding is still lacking. In our study population, women are more severely obese than men (see Table 2). This can help to explain our findings. In fact, during a fructose‐rich diet, women exhibit a larger decrease in insulin sensitivity as compared with men.31 Fructose represents 50% of the sucrose molecule and is mainly assumed with sugar‐containing beverages and candies. It can induce oxidative stress and insulin resistance, paving the way for hypertension, renal damage, fat accumulation, and liver steatosis. All these effects might be mediated by UA, which is synthetized in large amounts as a by‐product of fructose phosphorylation.3, 6 Endothelial function is also impaired in obese people; therefore, they could be more exposed to the effects of a sucrose‐rich, urate‐generating diet.6, 32 Older women, who have lost the protective, urate‐decreasing effect of estrogens,33 are likely even more vulnerable. Their propensity to develop liver steatosis is indicated by the correlation between serum UA and alanine aminotransferase.

On the other hand, the association of high UA levels with other metabolic abnormalities could be masked by the higher prevalence of other cardiovascular risk factors in men. Our finding of lower HDL cholesterol levels in men with high UA adds a new element to the previous observation of decreased HDL cholesterol in young men who consume sucrose‐rich diets.6

Reciprocal influences of UA and renal function34 may be involved in the reduction of eGFR observed in men and women with high levels of UA. It is well‐known that serum UA is increased in patients with CKD, particularly in stages 4 and 5, which indeed are rare in our population. Conversely, UA can induce renal damage both in the experimental model and in clinical settings.3, 35

Finally, our finding of increased LVM in men with high UA is, to our knowledge, new in the literature. As far as target organ damage is concerned, an increase in internal carotid artery resistivity index has been described in hyperuricemic women,27 while a higher incidence of atrial fibrillation is associated with elevated UA levels in diabetic patients36 and in the general population.37 At variance, hyperuricemia was not associated with target organ damage in a population with recently diagnosed hypertension.38 Although our own finding has to be considered with caution because of our small number of cases (61), it is consistent with a recent observation performed in a prospective cohort study involving about 5700 people aged 25 years or older. In that population, a 24% increase in incident ischemic stroke and a 13% increase in all‐cause mortality were observed per each standard deviation rise in serum UA levels.39

Study Limitation

An important limitation of this study was the data source. Patients were referred by their general practitioners for a help in diagnosis or treatment, mainly in the field of obesity and metabolic diseases. Therefore, they were not the best representation of the general population of the same age. However, the two clinics where the study was carried out are the only public internal medicine clinics in the districts of the city where they are located, with the exception of hospitals. No hospital is present in the suburban district of Scampia. Since our patients were of low socioeconomic status9 and therefore not likely to attend private practices, it is likely that our study sample is a reasonably good representation of elderly, low‐income Neapolitan people with metabolic problems.

The large predominance of women in the sample represents another limitation of our study. One reason could be the very high prevalence of obesity among women in the Neapolitan area, with more than 40%, compared with about 30% in the other areas of Italy. In addition, male obesity in the Neapolitan territory does not exceed 30%, similar to that in other Italian geographical areas.40 Moreover, it has been shown that men are less inclined than women to seek medical advice for dietary treatment of obesity41 and it is possible that this behavioral habit is enhanced by the low cultural level of our population.9 The separate analysis we performed for each sex can help to avoid the risk of results affected by this sex imbalance.

Conclusions

Available data about the link between serum UA and cardiovascular risk profile are conflicting and have been mainly obtained in young or middle‐aged populations. Current evidence indicates that serum UA is often increased in patients with metabolic syndrome and that the association between UA levels and cardiovascular outcomes is more consistent in women as compared with men. In the present study, we show that the association of high serum UA with several cardiovascular risk factors and markers of target organ damage is not limited to younger patients, but involves elderly people. In our setting, hyperuricemia clusters with obesity and other elements of metabolic syndrome in women, but this association is much weaker and more partial in men. Regarding target organ damage, renal impairment is present in hyperuricemic patients of both sexes, while LVM is increased in hyperuricemic men only. The sex‐related differences we found in this study may reflect different sensitivities to the metabolic derangements imputable to UA and/or different lifestyles in earlier years, which have led to uneven levels of obesity. Longitudinal studies on the effectiveness of therapeutic interventions aimed at lowering UA levels to reduce cardiovascular morbidity and mortality in either sex are warranted.

J Clin Hypertens (Greenwich). 2015;17:27–32. DOI: 10.1111/jch.12434. © 2014 Wiley Periodicals, Inc.