Update on Crystal-Induced Arthritis and Hyperuricemia From ACR 2007

Robert Terkeltaub, MD Disclosures

Introduction

A remarkable set of convergent factors has contributed to the increased prevalence and clinical complexity of gout in the United States in the last 15 years. These factors include increased population longevity, broad use of thiazide and loop diuretics and of low-dose aspirin that promote hyperuricemia, surges in the prevalence of hypertension, metabolic syndrome, obesity, and end-stage renal disease. Concurrently, there has been a shortfall in new US Food and Drug Administration (FDA)-approved therapies for treatment of hyperuricemia and gouty inflammation.[1] The 2007 American College of Rheumatology (ACR) meeting featured a variety of studies pertinent to clinical practice via refinements in understanding of the impact of hyperuricemia and gout and evolving antihyperuricemic and anti-inflammatory treatments, which are discussed below.

Gout and Chronic Kidney Disease

Rocca Rey and colleagues[2] assessed the impact of kidney function on the prevalence and costs of gout in a privately insured population, analyzing the related healthcare costs of gout according to the degree of impaired kidney function in a large, national managed care database of more than 1,500,000 adults. Annual average all-cause and estimated gout-related medical costs were compared with the glomerular filtration rate (GFR). Gout prevalence increased with kidney impairment from 4.3/1000 with a GFR of ≥ 90 mL/minute to 55.7/1000 with stage V chronic kidney disease (CKD) defined by a GFR < p =" .16).">

Potential Direct Pathogenic Contribution of Hyperuricemia to Vascular Disease

A major issue for current medical practice rests in the potential vascular ramifications of asymptomatic hyperuricemia. Several studies have demonstrated serum urate to be a risk factor for cardiovascular disease or poor outcome in cardiovascular disease. Most of these studies have employed univariate analyses. Several of the past studies with multivariate analyses have seen serum urate dropout as an independent risk factor, but the potential roles of elevated urate as an independent risk factor, pathogenic mediator, and biomarker of vascular disease remain a somewhat contentious issue. In this regard, serum urate is one of the solutes sensed within the juxtaglomerular apparatus, and it has been suggested in animal studies that hyperuricemia can promote an increase of plasma renin and systemic depletion of nitric oxide.[3] However, effects of urate on the human vascular endothelium function are controversial. Oxidative stress can promote hypertension, and serum urate, generated by the oxidized form of xanthine oxidase in a reaction that produces superoxide anion, can be a biomarker for oxidative stress.[4] Of interest, xanthine oxidase inhibition may diminish hypertension[5] and reduce endothelial dysfunction primarily by antioxidant effects of the xanthine oxidation rather than by lowering serum urate.[6]

The meta-analysis by Kim and colleagues[7] investigated the link between hyperuricemia and coronary heart disease. They conducted a MEDLINE search from January 2003 to April 2007. Hyperuricemia was designated as > 7.0 mg/dL, and a total of 9 prospective cohorts were analyzed with a fixed-effects model. After adjustment for age and other risk factors, hyperuricemia was associated with higher coronary heart disease, with a ratio of 1.356 in women and 1.121 in men. The study authors concluded that elevation of serum urate was an independent risk factor for coronary artery disease in both sexes, but that the effect was greater in women.

Contrasting findings were obtained by Neogi and colleagues[8] in analyzing the multicenter National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study on risk factors for heart disease. Carotid atherosclerotic plaques and their thickness were assessed by ultrasound, and coronary artery calcification by helical computed tomography. Sex-specific logistic regression was done to account for familial clustering, and the study authors assessed the association of serum urate with the presence of carotid plaques or coronary artery calcification adjusting for age, body mass index, race, and hypertension and its treatments, as well as diabetes and its treatments, renal insufficiency, aspirin use, alcohol intake, smoking, and education. Several thousand subjects were analyzed in each study. The association of serum urate with carotid atherosclerotic plaque demonstrated a dose-response relationship for men (including men without hypertension) but not for women overall. Nevertheless, postmenopausal women had a similar association between serum urate and atherosclerotic plaque as men. Serum urate was not associated with coronary artery calcification in men or women, an issue not yet resolved in the current literature. The association of the serum urate with carotid atherosclerotic plaque in men and postmenopausal women but not with coronary artery calcification is informative, especially because the relationship was seen in the absence of hypertension. This study argues for potential mechanisms by which serum urate may be related to certain adverse cardiovascular outcomes, but importantly, serum urate is not related to all such events and not in all human subsets. This is consistent with the lack of an evidence-based approach to treating asymptomatic hyperuricemia for presumed cardiovascular risk-lowering benefit.

What Is the Clinically Meaningful Extent of Effects of Low-Dose Aspirin on Serum Urate?

Low-dose aspirin use increases serum urate by inhibiting renal elimination of uric acid. In an analysis of the effects of low-dose aspirin use on serum urate levels and prevalence of hyperuricemia in the NHLBI Family Heart Study, Zhang and colleagues[9] showed that the relationship of aspirin use to both serum urate and hyperuricemia was similar in men and women. However, prospective studies need to be performed to further assess the impact of this association.

Increased Risk for Diabetes Mellitus in Men With Gout

Choi and colleagues[10] presented a study that investigated the impact of gout on the risk for type 2 diabetes. This is a topic of great interest because of the strong relationship between insulin resistance and increased urate reabsorption in the nephron, the clinical linkages between metabolic syndrome and gout, and the preliminary description of a linkage of fructose metabolism and insulin resistance with hyperuricemia itself.[11] In a separate study, Choi and colleagues[10] prospectively analyzed the relationship between a history of gout and incident type 2 diabetes in 11,138 male participants. They documented 1210 new cases of type 2 diabetes in this cohort. Men with gout had a 41% increased risk for incident type 2 diabetes compared with nondiabetic men. The association remained significant even after adjusting for serum urate levels, but increasing serum urate levels updated annually were associated with an increase in risk for type 2 diabetes. Similar associations were seen for baseline urate levels. This large prospective study suggests that men with gout have a substantially higher risk for type 2 diabetes, underlining the complexity of managing subjects with gout.

Uric Acid as a Biomarker of Pulmonary Artery Hypertension in Scleroderma

Gaffo and colleagues[12] looked at the possibility that brain natriuretic peptide (BNP) and serum urate are predictors (and biomarkers) of pulmonary artery hypertension in patients with systemic sclerosis. BNP is clearly a useful predictor in patients with ventricular failure, and serum urate levels have been reported to correlate with pulmonary artery pressures in patients with idiopathic arterial hypertension of the pulmonary artery tree.[13] Both BNP and uric acid were found to predict pulmonary hemodynamic properties in patients with scleroderma, although it is not yet clear whether or how serum urate can be employed in diagnosis or monitoring therapy and disease progression.

Clinical Trials of *Febuxostat for Managing Hyperuricemia in Gout Patients

Urate-lowering pharmacotherapy with febuxostat and allopurinol was further studied by Becker and colleagues[14] in 182 African-American subjects with gout. Patients received febuxostat up to 240 mg a day and allopurinol up to 300 mg a day. Response to febuxostat in the African-American subjects was comparable to the response seen in the overall phase 3 drug study population,[15] demonstrating a durable serum urate-lowering response in these patients. Data from the febuxostat comparative extension longer-term (EXCEL) study reinforced that more than half of the patients randomized to allopurinol treatment at a maximum dose of 300 mg daily failed to achieve or maintain target serum urate <>

Handheld Meter for Serum Urate Monitoring

Blood serum urate measurements were accomplished with a handheld cholesterol chemical urate meter and were comparable to the measurements obtained with standard laboratory testing.[18] The ease of use of this device offers the opportunity to improve the quality of care for patients with hyperuricemia and gout, particularly enabling the use of more intensive serum urate-lowering regimens, such as *PEGylated uricase or *febuxostat -- both of which are currently under clinical development.

New Developments in Understanding and Treating Gouty Inflammation

Basic research studies of clinical interest included a study that observed uric acid to directly promote T-cell proliferation and activation in an antigen-independent system,[19] and observations of heightened osteoclast development promoted by urate crystals that may be relevant to tophus-associated bone erosion.[20] In addition, novel inflammatory mediators for experimental gouty inflammation were described, specifically, the neutrophil chemotactic cytokines IP-10 and CXCL16; the proangiogenic vascular endothelial growth factor (VEGF),[21] and the histamine-generating histidine decarboxylase.[22] Nevertheless, interleukin (IL)-1beta activation and secretion driven by the NALP3 inflammasome[23] remains the most topical subject of clinical development for biologic therapy of crystal-induced inflammation. Hoffman and colleagues[24] employed novel models of NALP3 knockout and NALP3 leucine-rich repeat domain deletion to report that the leucine-rich repeat domain of the NALP3 inflammasome was required for experimental gouty inflammation.

A recent, open-label, uncontrolled pilot study of 10 patients suggested clinical benefit of *anakinra for gouty inflammation.[25] At ACR 2007, Torres and colleagues[26] presented their assessment of the capacity to reduce pain and inflammation in animal models of gout of mouse IL-1 Trap. Molecules of IL-1 Trap are constituted with a soluble recombinant receptor and accessory protein chimeric molecule linked to immunoglobulin (Ig)G Fc, and IL-1 Trap binds and neutralizes both IL-1 alpha and IL-1beta. This group first used IL-1 receptor 1 and control knockout mice. They injected monosodium urate (MSU) and calcium pyrophosphate dehydrate (CPPD) crystals into the peritoneum, as well as in the subcutaneous synovium-like air pouches and the ankle joint, demonstrating a major contribution of IL-1 signaling to inflammation in each model.[26] Administration of mouse IL-1 Trap suppressed neutrophil infiltration in both the MSU and CPPD crystal injection models, despite distinctions in these systems of crystal-induced inflammation. IL-1 Trap was just as effective in the intra-articular model as the highest tolerated dose of colchicine, and IL-1 Trap was effective in reducing pain and ankle swelling when it was administered 1 day after the intra-articular urate crystal injection. However, it was noted that a substantial amount of the inflammation in the intra-articular crystal injection model employed was IL-1-independent,[26] and this could well be the circumstance in many cases of human gouty arthritis.

Pilot Study of *Rilonacept (IL-1 Trap) in Gouty Arthritis

A pilot trial of rilonacept in 10 subjects with gout was reported.[27] This study employed a placebo-controlled, multicenter, nonrandomized, single-blind, monosequence crossover design. The study population comprised subjects with a diagnosis of chronic refractory gouty inflammation of more than 6 months' duration, with many of these subjects having frequent or prolonged arthritis flares. After a 2-week placebo run-in, a loading dose of two 160-mg subcutaneous injections of rilonacept or placebo in the first week was followed by a blinded switch in all 10 patients to active treatment of 5 weekly injections (1 injection per week) of 160 mg rilonacept, with final assessment at 14 weeks. An important aspect of the dosing in the trial design is that rilonacept is long-acting compared with anakinra. Rilonacept was generally well tolerated, and the blinded switch from placebo to rilonacept promoted significant decreases of clinical activity (pain visual analogue score response and subject and physician global scores) and highly sensitive C-reactive protein levels relative to placebo. This study reinforces the potential therapeutic value of IL-1 antagonism for gouty inflammation, although this study design, as in the prior study of anakinra,[25] did not rule out a component of spontaneous improvement in gouty inflammation. Larger, randomized, controlled, double-blind studies are clearly warranted for IL-1 antagonism in gouty arthritis.

Septic Arthritis Associated With Underlying Crystalline Arthritis

Ahuja and colleagues[28] described the association of concomitant septic arthritis with crystal arthropathy in a large case series. A total of 440 patients with septic arthritis were identified; in 40 of these patients crystals were also seen on synovial fluid analysis. Mean age of the patients was approximately 70 years, and they were 60% male and 40% female, with urate crystals present in 55%, CPPD crystals in 42.5%, and 1 patient (2.5%) having both MSU and CPPD crystals. The knee joint was most frequently affected, and Staphylococcus aureus was the most common microorganism, which was identified in 50% (with 40% of this group having methicillin-resistant S aureus [MRSA]). It is likely that crystalline arthritis, such as asymptomatic CPPD deposition in the medial meniscal fibrocartilage, was an incidental finding in some of these patients. However, this study underlines the fact that a variety of abnormalities in joints, not simply rheumatoid arthritis, can predispose to infectious arthritis and, further, that the finding of crystals is not uncommon in infected joint fluid due to "enzymatic strip-mining" triggered by the infection.

The Use of Ultrasound to Screen for Chondrocalcinosis

Thiele and Schlesinger[29] explored the use of ultrasound to detect CPPD crystals in hyaline articular cartilage. Ultrasound examined for hyperechoic (bright) bands embedded in the center of the anechoic (dark)-appearing hyaline cartilage that paralleled the hyperechoic contour of subchondral bone. Fibrocartilage of the wrist also was scanned, with assessment for hyperechoic, oval-shaped, or rounded particles with irregular borders within hyperechoic triangular fibrocartilage. Conventional radiography detected chondrocalcinosis in 7 out of the 8 knees and 2 out of 2 elbows and wrists. MRI described chondrocalcinosis in 2 of 2 wrists but not in either the elbows and knees. The study authors observed that conventional radiography and ultrasound are superior to MRI for detecting chondrocalcinosis in hyaline cartilage but equally sensitive in the detection of calcifications in fibrocartilage of the wrist. Ultrasound appeared more sensitive than conventional radiography in detecting chondrocalcinosis of the knees. MRI was not useful in detecting chondrocalcinosis in either the elbows or the knees, in agreement with past findings. Because ultrasound is relatively easy to use in the office, is safe, and not particularly costly, it could be useful as a screening tool for assessing CPPD crystal deposition disease.

Conclusions

Data presented at the 2007 ACR meeting argued for the use of serum urate as a biomarker for some conditions, such as adverse cardiovascular outcomes and diabetes, and only in selected patient groups because the overriding data supporting this argument remain contradictory. Such observations indicate a lack of evidence to treat asymptomatic hyperuricemia for presumed cardiovascular risk-lowering benefit. Small animal experimental results and pilot human clinical trial data presented reinforced the potential therapeutic value of IL-1 antagonism for gouty inflammation. However, rilonacept and other IL-1 inhibitors will require larger, randomized, double-blind, controlled studies for gouty arthritis, in large part because of the natural tendency of gouty inflammation to spontaneously improve after onset. The association of underlying crystal-associated arthritis with septic arthritis, and the potential value of ultrasound as a diagnostic tool for chondrocalcinosis, were presented.

*Febuxostat, PEGylated uricase, anakinra, and rilonacept have not been US Food and Drug Administration (FDA)-approved for gout.

This activity is supported by an independent educational grant from Bristol-Myers Squibb, Forest, and Genentech.

         

Gracias Dr. José Manuel Ferrer Guerra!

Low HDL Cholesterol, Even When LDL Levels Are Low, Is Cardiovascular Dynamite, New TNT Analysis Shows

September 28, 2007 — Among patients treated with statins, including those who achieved very low levels of cholesterol with high-dose statin therapy, high-density lipoprotein (HDL)-cholesterol levels are still predictive of major cardiovascular (CV) events [1]. These are the new findings from the Treating to New Targets (TNT) study, a trial previously covered by heartwire.

"In this post hoc analysis from the TNT trial, HDL cholesterol was a significant predictor of major cardiovascular events across the entire study cohort, even after all other baseline risk factors, including baseline low-density lipoprotein (LDL)-cholesterol level, had been taken into account," writes Dr Philip Barter (Heart Research Institute, Sydney, Australia) on behalf of the TNT investigators in the September 27, 2007 issue of the New England Journal of Medicine.

Discussing the rationale of the study, the authors note that while low levels of HDL cholesterol are known to be a strong predictor of increased CV risk, it is unknown whether low HDL cholesterol remained a significant risk factor in subjects treated to very low LDL-cholesterol levels. Some have argued, Barter and colleagues point out, that if LDL levels were reduced sufficiently, "the level of HDL cholesterol might be irrelevant."

TNT: Part of the lower-is-better posse

The TNT trial was one of several studies that made the "lower-is-better" case for lowering LDL cholesterol beyond the then-current recommended guidelines. TNT, first presented at American College of Cardiology (ACC) 2005 Scientific Sessions and published in the New England Journal of Medicine, showed that lowering LDL-cholesterol levels in stable coronary heart disease (CHD) patients substantially below the current targets resulted in better clinical outcomes [2].

Intensive lipid lowering with atorvastatin (Lipitor, Pfizer) 80 mg daily provided greater protection from major CV events compared with low-dose atorvastatin in stable CHD patients. High-dose atorvastatin reduced the primary composite end point of death from CHD, nonfatal myocardial infarction (MI), resuscitation after cardiac arrest, and fatal or nonfatal stroke 22% compared with patients treated with atorvastatin 10 mg.

In this newest post hoc analysis, the investigators assessed the predictive value of HDL cholesterol in 9770 patients, with the primary outcome measure being the time to a first major CV event. All subjects were stratified into quintiles based on their HDL-cholesterol levels determined at month 3 of the double-blind treatment phase.

In the determination of the five-year risk of major CV events across the different quintiles, univariate analysis showed the event rate to be reduced by 40% in the highest quintile when compared with subjects with the lowest HDL-cholesterol levels. After adjustment for covariates, HDL cholesterol remained a significant predictor of CV events; there was a 25% reduction in risk when those in the highest quintile were compared with those in the lowest. When the analysis was stratified according to LDL-cholesterol levels in patients receiving statins, the relationship between HDL-cholesterol levels and major CV events was of borderline significance (p=0.05).

Multivariate analysis of the relationship between HDL-cholesterol levels at month 3 and risk of major cardiovascular events

Comparison with HDL quintiles Q5 (HDL cholesterol ≥ 55 mg/dL) Q4 (HDL cholesterol 47 to <>

Hazard ratio (95% CI) vs Q1 (HDL <>

The TNT investigators also report that when they examined subjects who achieved very low levels of cholesterol, less than 70 mg/dL, those with HDL-cholesterol levels 55 mg/dL or greater were still at significantly less risk for a major CV event than those in the lowest quintile.

Commenting on the results of the study for heartwire, Dr Monty Krieger (Massachusetts Institute of Technology, Cambridge, MA) said that this is yet another study showing that HDL cholesterol is protective, this time at various LDL levels, and will reinforce the concept that doing something to treat low levels of HDL cholesterol may be beneficial, regardless of the LDL-cholesterol numbers.

Pfizer sponsored the TNT study. Dr. Barter reports receiving consulting fees from Pfizer and AstraZeneca; lecture fees from Pfizer, AstraZeneca, and Merck; and grant support from Pfizer.

Sources

Barter P, Gotto AM, LaRosa JC, et al. HDL cholesterol, very low levels of LDL cholesterol, and cardiovascular events. N Engl J Med. 2007;357:1301-1310.

LaRosa JC, Grundy SM, Waters DD, et al. Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N Engl J Med. 2005;352:1425-1435.

The complete contents of Heartwire, a professional news service of WebMD, can be found at www.theheart.org, a Web site for cardiovascular healthcare professionals.

Clinical Context

Population studies have consistently shown that HDL cholesterol levels are a strong and independent predictor of cardiovascular disease (CVD), and in several studies, each increase of 1 mg/dL of HDL cholesterol level was associated with a 2% to 3% reduction in future risk for CHD, whereas each decrease of 40 mg/dL in LDL cholesterol level corresponded to a 24% reduction in CV events. A low baseline level of HDL cholesterol is predictive of CV events even during treatment with statins, but whether this predictive value applies to all levels of LDL cholesterol is uncertain.

This is a post hoc analysis of data from the TNT trial in which patients with previous CVD were randomized to 2 doses of statin (10 vs 80 mg of atorvastatin daily) to examine the impact of quintile of HDL cholesterol level on CV morbidity and mortality.

Study Highlights

Included were men and women aged 35 to 75 years with clinical evidence of coronary disease, such as previous MI, previous or current angina, or revascularization.

Excluded were those with liver disease, nephrosis, pregnancy, congestive heart failure, abnormal creatine kinase, or revascularization within 1 month.

A washout period of 1 to 8 weeks was followed by lipid measurement, and after a run-in with 10 mg atorvastatin daily for 4 weeks, those with mean LDL cholesterol levels of less than 130 mg/dL were randomized to either 10 or 80 mg atorvastatin daily.

Assignment was double-blinded.

Primary endpoint was time to first occurrence of a major CV event defined as death from CHD, nonfatal MI, resuscitation after cardiac arrest, or fatal or nonfatal stroke.

The expected 5-year risk for CVD was calculated for each quintile of HDL cholesterol.

HDL levels were less than 38 mg/dL for quintile 1, 38 to less than 43 mg/dL for quintile 2, 43 to less than 48 mg/dL for quintile 3, 48 to less than 55 mg/dL for quintile 4, and 55 mg/dL or more for quintile 5.

9770 subjects had available HDL cholesterol levels determined at month 3 of the double-blind phase.

Subjects were categorized by quintiles of HDL cholesterol.

94% were white, 65% to 91% were men, mean age was 60 years, mean body mass index was 28 kg/m², 60% were former smokers, and 23% were never smokers.

60% had a history of MI, 50% bypass grafting, 53% angioplasty, 81% angina, 55% hypertension, and 15% diabetes.

Those with higher HDL cholesterol levels were older, more likely to be women, less likely to be current smokers, and were thinner.

Prevalence of diabetes was double in the lowest vs the highest quintile (21.2% vs 10.6%).

The 5-year event rate was reduced by 40% in the highest vs the lowest quintile of HDL cholesterol.

After adjustment for covariates, quintile of HDL cholesterol remained a significant predictor with a reduction in CV events from 95% in the lowest to 7.1% in the highest quintile, a 25% risk reduction (hazard ratio [HR], 0.75).

Risk for CVD differed significantly across HDL cholesterol quintiles (P = .04).

Risk for CVD was significantly lower in those assigned 80 vs 10 mg atorvastatin daily in all quintiles.

In each atorvastatin group, the frequency of CV events increased with decreasing levels of HDL cholesterol.

Among those assigned 10 mg of atorvastatin daily, those in the highest quintile were significantly less likely to have a CV event (HR, 0.71).

Among those assigned 80 mg of atorvastatin daily, the difference in CV events between the highest and lowest quintiles did not reach significance (HR, 0.81).

When stratified by LDL cholesterol level, quintile of HDL cholesterol was of borderline significance for predicting CVD (P = .05).

For quintiles 2 to 5, the HRs for CV events were 1.00, 0.80, 0.92, and 0.75, respectively.

Even in the lowest LDL cholesterol stratum, the quintile of HDL cholesterol remained a predictor of CV outcome with an HR of 0.61 (P = .03).

In analysis by continuous variable, each 1-mg/dL increment in HDL cholesterol level was associated with a 1.1% risk reduction in a CV event (P = .003).

The protective effect of HDL cholesterol was independent of sex, age, smoking status, body mass index, and other covariates.

The authors concluded that HDL and LDL cholesterol levels were independently predictive of CV events and that even at very low LDL levels, HDL level remained predictive of CV events.

Pearls for Practice

A higher HDL cholesterol level is independently associated with a lower risk for CV events in patients taking a statin.

In patients taking a statin, the protective effect of higher HDL cholesterol applies to all LDL cholesterol levels, even at levels of less than 70 mg/dL.

        

Gracias Dr. José Manuel Ferrer Guerra!!