Uric Acid and CKD Progression Matures with Lessons for CKD Risk Factor Discovery

• chronic kidney disease
• uric acid
• allopurinol
• febuxostat
• epidemiology and outcomes
• risk factors

Higher circulating levels of uric acid are associated with new and progressive CKD (1–3). As a modifiable metabolite, uric acid is a potential target for interventions to slow CKD. Medications, diet, and other lifestyle approaches are regularly used in practice to lower uric acid in patients with gout and could be easily translated to CKD care. However, the benefit of treating hyperuricemia to slow CKD has been debated. The most recent Kidney Disease Improving Global Outcomes Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease concluded there was “insufficient evidence to support or refute the use of agents to lower serum uric acid concentrations” as a strategy to slow CKD (www.kdigo.org). Since then, the field has called for more rigorous trial evidence to determine if uric acid should be a target of therapy (1). Recently, the New England Journal of Medicine published two randomized controlled trials focused on the lowering of uric acid with allopurinol to slow CKD, helping to resolve this uncertainty (4,5). This perspective will briefly review evidence linking uric acid and CKD progression and consider broader lessons for the field.

Uric acid is produced from purine nucleotide metabolism. Humans lack the major catabolic enzyme that degrades uric acid, thus, circulating levels are governed by rates of production and excretion. Uric acid is produced at greater rates in patients with high dietary purine, alcohol, and fructose intake. Excretion depends on kidney function, resulting in strong correlation with falling GFR (2). Insulin resistance and metabolic syndrome, diuretics, and volume contraction each augment kidney reabsorption and raise levels. In sum, uric acid levels are affected by diverse factors including diet, insulin resistance/metabolic syndrome, kidney function, volume status, medications, and genetic variation in kidney transporters, each of which can confound observational studies (1).

Basic studies have defined potential pathways linking uric acid with CKD. Mechanisms include endothelial dysfunction, activation of the renin-angiotensin-aldosterone system, and oxidative stress (1). However, several trials of intermediate outcomes cast doubt on the relevance of this physiology in patients. In one trial (n=149), halving of serum uric acid with allopurinol or probenecid did not improve endothelial function or reduce renin-angiotensin-aldosterone system activation by 8 weeks (6,7). Further, associations between uric acid and CKD progression in cohorts with CKD weaken substantially when adjusting for GFR (2). In cohorts without CKD at baseline, associations with new-onset CKD often persisted modestly despite GFR adjustment (3).

Many observational studies in CKD epidemiology evaluate multivariable adjusted associations between biomarkers and outcomes and have not used more rigorous causal inference designs that could better select candidates for clinical trials, such as Mendelian randomization. Mendelian randomization studies can reduce bias in epidemiologic studies by functioning as “instruments.” In these studies, investigators isolate the variation in a trait that is induced purely by genetics. When evaluating unrelated individuals from the same ancestral background, genes sort independently and are not related to environmental confounders like diet or body composition. Several large Mendelian randomization studies have been conducted evaluating uric acid and CKD. For instance, in one study, using data from >400,000 participants and 26 single nucleotide variants influencing uric acid levels, there was no causal association between uric acid and presence of CKD or GFR (8). This was despite replicating known causal associations with gout.

The most definitive results guiding practice are obtained from clinical trials. In 2010, Goicoechea et al. (9) studied the role of allopurinol in preventing CKD, cardiovascular and mortality events, and slowing progression over a median of 2 years in 113 participants. They showed a reduction in event rates, but the extremely small sample size and low event numbers required confirmation in larger, multisite trials. More recently, the New England Journal of Medicine published the results of two larger, randomized controlled clinical trials evaluating the effect of uric acid reduction using allopurinol on the progression of CKD. In the Preventing Early Renal Loss in Diabetes (PERL) Trial (n=530), allopurinol was tested in patients with type 1 diabetes and relatively early, but high-risk, kidney disease (mean GFR of approximately 70 ml/min per 1.73 m²) (4). The Controlled Trial of Slowing of Kidney Disease Progression from the Inhibition of Xanthine Oxidase (CKD-FIX) Trial (n=363) included participants with diabetic and nondiabetic CKD but lower baseline GFR (mean of approximately 30 ml/min per 1.73 m²) (5). Both trials were conducted across multiple sites and enrolled high-risk populations with either albuminuria or evidence of rapid decline in kidney function (>3 ml/min per 1.73 m² per year decline before enrollment). In both trials, declining GFR was meaningful during the trial (approximately −2.5 and −3.3 ml/min per 1.73 m² per year, respectively). This is important because it suggests the patients studied were those in whom slowing CKD progression pharmacologically was strongly warranted. Both trials also demonstrated compelling evidence that the drug affected the target therapeutic pathway, with both trials achieving strong and sustained reduction in uric acid in the treatment arm compared with control (2–3 mg/dl). Participants in CKD-FIX had higher uric acid on average at baseline (mean, 8.2 mg/dl) compared with PERL (mean, 6.1 mg/dl). Despite studying the right populations with a drug that meaningfully affected the target, both trials showed convincingly negative results, with virtually no difference in the loss of GFR over follow-up in allopurinol versus control in either trial. Although both trials were still relatively small, the estimated difference in GFR across groups was so small that trials two to three times larger would be highly unlikely to overturn the results.

Both PERL and CKD-FIX studied allopurinol. A recent clinical trial of 443 patients with CKD stage G3, the Febuxostat versus Placebo Randomized Controlled Trial Regarding Reduced Renal Function in Patients with Hyperuricemia Complicated by CKD Stage 3 (FEATHER), used febuxostat to lower uric acid compared with placebo. FEATHER also showed no improvement in GFR. FEATHER was limited because the investigators did not select for high progression risk and, therefore, the decline in kidney function during the study was very modest in both groups. Thus it is more challenging to rule out an effect of febuxostat in a higher-risk population on the basis of this study alone (10). Interpreted along with PERL and CKD-FIX, evidence is accumulating that pharmacologic management of uric acid levels is not indicated for the purpose of slowing CKD.

How should the kidney community respond to these results? New trials on lowering uric acid provide a clear answer on these therapies that will affect practice recommendations. Not only did studies fail to demonstrate benefit, but they also had adequate power and enrolled high-risk populations that could exclude important effects and largely settle the question of the role of these therapies in secondary prevention in CKD. However, these results also pose a troubling question. CKD epidemiology is full of examples of biochemical measurements, like uric acid, that are both affected by CKD and hypothesized causes of CKD progression. Because many of these factors persist to some degree after adjustment for GFR, they are often implicated in a self-perpetuating cycle of kidney function decline. Purported associations, like this one with uric acid, should make us worry that many of these identified factors could be only by-products, and not causes, of CKD. Exclusively relying on trials to settle these questions is expensive, burdensome, and slow. Analytic innovation in our approach to risk-factor discovery in CKD, including better “control” for confounding by kidney function, is urgently needed. Earlier incorporation of causal inference designs and proof-of-concept trials in the discovery pipeline will also improve our vetting of putative targets.

Another consideration is that uric acid may be a marker, or correlate, of an important clinical cause of CKD progression. It is often easier to draw conclusions about easy-to-measure biomarkers instead of more latent clinical phenotypes that are increasingly recognized in medicine. In this example, metabolic syndrome and insulin resistance, diet, and body composition are all strong correlates of uric acid that may be markers of an adverse metabolic milieu that is both more important and more elusive than uric acid alone. In many specialty areas within epidemiology, these correlation patterns are explicitly interpreted. For instance, if a single nucleotide variant is discovered in genetic epidemiology, a region of genetic correlation, or “linkage disequilibrium,” may be investigated. If a metabolite is found in metabolomics, its pathways are considered. If a food or nutrient is identified in nutritional epidemiology, dietary patterns are often studied or advised. Making advances in CKD epidemiology will require us to use a less literal and more expansive view of what our biochemical clues may mean about the underlying phenomenon. So, although reduction of uric acid with targeted pharmacologic therapy may not reduce CKD progression, would lifestyle approaches focused on the metabolic syndrome be more effective?

In the end, the practical lesson from studies of uric acid and CKD progression is somewhat simple. Pharmacologic treatment to lower uric acid levels is not likely to slow CKD. The trial results are clear. But we can also learn about our need to deeply vet confounding by kidney function and to consider more elusive phenotypes as we move forward in the quest for secondary prevention in CKD. Uric acid may not be the answer, but new and unexpected breakthroughs (11) show that the future is bright.

• Copyright © 2021 by the American Society of Nephrology