Obstructive Sleep Apnea and the Risk of Gout

A Population-Based Case-Control Study

Caroline van Durme; Bart Spaetgens; Johanna Driessen; Johannes Nielen; Manuel Sastry; Annelies Boonen; Frank de Vries


Arthritis Res Ther. 2020;22(92) 

In This Article


Our study showed that the almost doubled risk of gout with OSA disappeared after adequate statistical adjustment for BMI, renal function, heart failure, and recent use of diuretics. Notwithstanding, subgroups of patients with OSA, more specifically women and those with a history of heart failure, who had recently used diuretics, who had an eGFR between 59 and 15 ml/min, or who had a BMI above 30 kg/m2, still had a 2- to 4-fold increased risk of gout.

The absence of an overall association between OSA and gout in our study contrasts with a 1.5-fold increased risk of gout with OSA that was reported by two previous studies[4,9] which had used the same data source, i.e., the CPRD GOLD database, or a data source (THIN) that partly overlaps with CPRD.[19] Table 3 shows that this difference can be largely explained by more comprehensive statistical adjustment for potential confounding in the current study, in particular for renal impairment, as measured by eGFR and READ codes, and heart failure. When renal function declines, less uric acid is excreted which leads to hyperuricemia and eventually gout[20] (Figure 2).

Figure 2.

Possible biological pathways linking OSA to gout. XOR, xanthine oxidoreductase; OSA, obstructive sleep apnea

Heart failure is probably under-recorded in large observational studies based on diagnostic codes.[21] Specific for the UK primary care databases, under-diagnosis might be related to manual coding of cardiology discharge letters by general practice staff. Therefore, we statistically adjusted our analyses for proxy indicators of heart failure such as recent use of diuretics. We also statistically adjusted our analysis for the use of other medications that are commonly used for heart failure, including beta-blockers, ACE-inhibitors, and ARBs.[22] Since (repeated) prescriptions, including outpatient prescriptions of cardiologists, are generally issued by GPs every 4 weeks, this proxy indicator is likely to be better captured and therefore more likely to further reduce the level of residual confounding. Notwithstanding, heart failure remained associated with gout, even after full adjustment. The explanation might be found in insufficient adjustment for the known increased xanthine oxidoreductase (XOR) activity in the myocardium of the failing heart which leads to an elevation of uric acid[23] (Figure 2). Filippatos et al. demonstrated in their study that hyperuricemia was associated with poor outcomes in patients with heart failure without CKD but not in those with CKD, confirming the hypothesis that hyperuricemia in patients with heart failure could not only be explained by reduced uric acid excretion because of a poor kidney function.[24] Otaki et al. also demonstrated an association between XOR activity and severity and clinical outcomes in patients with heart failure.[25] Further evidence on a possible role of an increased XOR activity in heart failure can be found in studies demonstrating a beneficial effect of adding allopurinol to the treatment of patients with heart failure.[26,27] This beneficial effect was not demonstrated with benzbromarone, which is a uricosuric drug and therefore decreases the uric acid concentration by increasing its excretion.[28] An alternative explanation for the independent contribution of heart failure to OSA can also be found in the influence of overnight rostral fluid shifts to the neck and lungs in patients with heart failure.[29]

Unexpectedly, we found that the risk of gout with OSA disappeared in men after adjustment for confounders, while in women, the risk remained elevated. It is widely accepted that females are at lower risk of gout, as a result of the uricosuric effect of estrogens in women before menopause.[30] However, even after menopause, the risk in females remains lower and we can therefore assume that other causal factors probably play a role. Sex differences have also been noted in the prevalence and severity of OSA, with women presenting with less severe and less prevalent disease. These differences are decreased after menopause.[31] Differences in fat distribution, upper airway anatomy (in particular the posterior tongue region), mechanisms affecting ventilatory stability, and sex hormones might explain the differences between men and women in OSA.[31] In this line, it is of note that a study by Wang et al. showed that fat accumulation around the head measured by dual-energy X-ray absorptiometry was positively correlated with uric acid levels in women but not in men.[32]

Obesity, which itself is associated with hypoxemia, is the main risk factor for the development of OSA. In more obese patients, OSA is aggravated with more severe oxygen desaturations and hypoxemia, which may explain why the risk of gout remains high in the highest BMI category[33,34] (Figure 2).

Our study had several limitations. First, there probably is underreporting of both OSA and gout, especially in the less severe cases.[17,35] Misclassification of both exposure (OSA) and our outcome of interest, i.e., gout, is probably random and may therefore lead to regression towards null.[36] It could have masked a true association between OSA and gout among men. Among women, it could have masked a higher true association. With respect to confounders, although our data regarding renal function were more accurate, renal function is not routinely measured in primary care. This could lead to residual confounding. Another limitation in our study, which is present in all epidemiological studies where researchers try to estimate the total causal effect of an exposure on an outcome of interest, is the problem of potential over-adjustment. Ideally, one should not control for factors which lie in the causal pathway between exposure and outcome, as it leads to a regression of the risk towards null.[37] In our case, renal function could also be influenced by OSA itself as nocturnal hypoxemia present in OSA could accelerate the decline in kidney function and therefore reduce uric acid excretion and induce or exacerbate hyperuricemia and eventually the risk of gout.[38] Renal function would then be in the causal pathway from OSA to gout. Another limitation concerns the limited number of patients present in some subgroups, especially women and the groups with the worst CKD (CKD 4 and 5). The conclusion drawn from those results should therefore be interpreted with caution.

Our study had several strengths. First, we were able to include a large number of patients with gout and controls. The findings of this study are therefore likely to be generalizable to patients with gout and OSA in the total UK population.[14] Second, the large amount of clinical information routinely and longitudinally collected in clinical practice allowed us to statistically adjust for many potential confounders such as age, sex, smoking status, alcohol use, kidney function, comorbidity, and use of medication.