A recent study suggested that hyperuricemia itself predicts the development of several cardiac risk factors, including hypertension and hyperlipidemia (see doi: 10.1161/HYPERTENSIONAHA.116.08998.)
Details:
— 5899 Japanese subjects were enrolled who at baseline did not have overweight/obesity (BMI>25), hypertension (>140/90 after resting quietly for 5 min), diabetes (meds or A1c>6.4), and dyslipidemia (LDL >140, HDL <40, and/or TG >150), as well as any history of gout or hyperuricemia on medications, or chronic kidney disease with the eGFR <60. [ie, a pretty normal cohort medically]
— 282 men and 133 women had hyperuricemia defined as serum uric acid (SUA) > 7 mg/dL in men or >6 mg/dL in women
— Mean age 47, 1864 men, there was small but statistically significant differences between those with hyperuricemia versus normal SUA at baseline (e.g. in men, BMI 22.4 vs 21.8, blood pressure 116/73 vs 114/72, eGFR 87 vs 82, albumin 4.5 vs 4.4, with similarly small differences in women) though there was a more convincing difference in drinking habits at 72% vs 62%).
— SUA on average was 7.65 in hyperuricemic men vs 5.59 in those with SUA <7; 6.44 in hyperuricemic women vs 4.20 in those with SUA < 6
— patients had an initial exam in 2004, and a follow-up exam in 2009
Results:
— hyperuricemia was associated with an increased cumulative incidence of (all OR’s expressed as per SUA increase of 1 mg/dL):
— hypertension, 14.9% vs 6.1% (p<0.001), odds ratio (OR) of 1.5 [ie, the OR is much higher in those with much higher SUA levels]
— dyslipidemia, 23.1% vs 15.5% (p<0.001), OR of 1.3
— chronic kidney disease, 19.0% vs 10.7% (p<0.001), OR of 1.3
— overweight/obesity, 8.9% vs 3.0% (p<0.001), OR of 1.5
— diabetes, 1.7% vs 0.9% (p<0.001), OR 1.5
–in the above, there were some differences between men and women: there was no increase in diabetes in men (though in women was 2.3% vs 0.5%, with p=0.011); and in women no increase in dyslipidemia (though in men 5.2% vs 19.2%, p=0.020) or overweight/obesity (trend in women with p=0.08, but in men was 11.0% vs 4.9%, p<0.001)
— multivariate analysis:
–model 1 (controlling for age, sex, and smoking/drinking), all looking at OR and p value per SUA increase of 1gm/dL:
–hypertension increased with OR of 1.4, p<0.001
–dyslipidemia increased with OR of 1.3, p<0.001
–CKD increased with OR of 1.3, p<0.001
–model 2 (controlling also for eGFR), all looking at OR and p value per SUA increase of 1gm/dL:
–hypertension increased with OR of 1.5, p<0.001
–diabetes increased with OR of 1.5, p=0.004
–dyslipidemia increased with OR of 1.3, p<0.001
–CKD increased with OR of 0.9, p=0.006 [ie, lower odds ratio]
–overweight/obesity increased with OR of 1.4, p<0.001
–model 3 (also controlling for BMI), all looking at OR and p value per SUA increase of 1gm/dL:
–hypertension increased with OR of 1.4, p<0.001
–diabetes increased with OR of 1.4, p=0.01
–dyslipidemia increased with OR of 1.2, p<0.001
–CKD increased with OR of 0.9, p=0.004
–overweight/obesity increased with OR of 1.1, p nonsignificant
Commentary:
— this study is particularly interesting because it isolates any of the cardiometabolic issues from hyperuricemia, which have historically been conflated with them, by choosing people who had hyperuricemia initially but none of these issues at baseline and following them 5 years later. It has been quite unclear what the directionality is (or if it exists): does hyperuricemia in fact lead to these medical problems, or do these issues lead to hyperuricemia through its effects on insulin resistance (leading to the various components of the metabolic syndrome), and renal vasoconstriction and reduced GFR (leading to decreased renal uric acid excretion). Or are both hyperuricemia and cardiometabolic risk factors both related to a third entity, perhaps insulin resistance. Animal studies have supported the role of uric acid as causal in these conditions. This study would have been stronger if they had intermediate exams, not just at the beginning and 5 years later, but still would not answer the issue of causality definitively.
— one big issue with multivariate adjustment (as a general issue), which really comes to the fore in this study, is that it really depends on the variables being independent. For example, controlling for eGFR even within the <60 range and finding no significant relationship between SUA and developing CKD does not necessarily mean that there really is no relationship. Perhaps a mild reduction of eGFR within the normal range is leading to reduced excretion of SUA (and higher blood levels) from this mild decrease in GFR itself (so controlling for eGFR may create the perhaps erroneous impression that SUA is unrelated). Or, controlling for BMI within the normal range may similarly not show that SUA elevation is not predictive of subsequent overweight/obesity, since mild increases of BMI may be associated with some increased insulin resistance leading to higher BMI in the future. So, I would not attach much significance to models 2 or 3 above.
so, this study complements and adds to the previously noted association between hyperuricemia and cardiovascular disease (see blogs noted below). However, one can still not definitively show causation, because the hyperuricemia could be an innocent bystander associated with the real cause.
— But, based on these studies, it seems reasonable to me to check SUA levels for 2 reasons (and I have been doing so pretty regularly):
— there are important lifestyle risk factors for increased SUA levels, especially fructose intake. I have been successful in a few cases of working with patients to decrease their consumption of sodas and other products with high fructose corn syrup, often finding pretty dramatic decreases in SUA levels (eg from the 8.5 range to the 7 range). And decreasing alcohol consumption
— In addition, I also think it would be reasonable to be even more aggressive in primary prevention of cardiac disease in those patients who have high SUA levels, both in terms of discussing the importance of healthy lifestyles and also having a lower threshold for starting meds. And the meds chosen might be different: eg, using losartan (but not other ARBs) or amlodipine/nifedipine for hypertension, since these lower SUA levels: see below.
—a Danish study suggested that treating hyperuricemia in those with allopurinol led to fewer cardiovascular events
— another study provides an interesting evolutionary perspective on hyperuricemia, as well as a Taiwanese study finding a dramatic decrease in cardiovascular events by treating hyperuricemia
–here is my very brief blog on antihypertensives and uric acid from 2012, predating the bmj website:
— large study of general practices in UK, looking at 25K pts with gout. I had seen some older studies finding that losartan (but NOT other arb’s or ace-I’s) lower uric acid levels. In this large UK database, they found a 19% dec risk of clinical gout with losartan (compared to other hypertensive pts), 13% with ccbs (21% dec with amlodipine, 13% with nifed, and 14% with dilt), with inc gout risk with diuretics, b-blockers, ace-I’s, other arb’s besides losartan. They note in their discussion some studies (which I looked at and are pretty small…) find that ccbs (esp nifed and amlod) and losartan are uricosuric and decrease serum uric acid levels. See doi:10.1136/bmj.d8190 ). there have been some recent reports that high fructose corn syrup is perhaps the largest (or close to it) dietary component which increases uric acid.