Comorbidity in gout at the time of first diagnosis: sex differences that may have implications for dosing of urate lowering therapy

Gout is the most common form of inflammatory arthritis and is caused by deposition of monosodium urate (MSU) crystals in peripheral joints and surrounding tissues as a result of chronic elevation of serum urate (SU) levels. The incidence and prevalence of gout are increasing in many parts of the world and the prevalence is > 1% in North America and Europe [1].

Gout has been linked with an increased occurrence of several comorbidities, which all may influence the overall effect on quality of life and increased mortality [2–4]. Some of these comorbidities, such as use of diuretics [5, 6], renal disease [7–14], obesity [15, 16], transplantation [17–20], psoriasis [21, 22] and alcohol overconsumption [23, 24], have also been suggested to increase SU levels and hence be part of the causal pathway for developing hyperuricemia and clinical gout. Epidemiological studies examining the relationship between gout and risk for cardiovascular diseases (CVD) overall and in men and women separately report conflicting findings, with a significant association reported by some studies [3, 4, 25, 26], but not others [27–29]. There is only limited evidence for positive associations with cerebrovascular disease [30–33] or peripheral vascular disease (PVD) [34, 35].

Current clinical guidelines and management pathways endorsed by the European League against Rheumatism [36], American College of Rheumatology [37] and British Society for Rheumatology [38] agree on the importance of assessing comorbidities before deciding to initiate treatment of gout with urate lowering therapy (ULT). In addition, there are also results supporting the need for higher doses of ULT if patients have comorbid conditions, such as renal disease, use of diuretics or obesity [39]. This effect of decreased eGFR and diuretic use is probably explained by their ability to increase SU levels and not by any effect on the dose–response relationship between given allopurinol dose and change in urate levels [40].

It is well known that several comorbidities occur more commonly in patients with gout compared to the general population [20, 41–43], although there are relatively few population representative studies from Europe [43]. Furthermore, the timing of comorbidity occurrence relative to gout diagnosis has not always been addressed. In particular, there are very few studies examining the prevalence of comorbidities in patients with gout by sex [44–47] and none of them have presented results at the time of gout diagnosis. If such differences between sexes were found, they could indicate that pathological pathways differ in importance in men and women. In addition, such differences could be of importance to take into consideration when initiating and dosing ULT [39].

Using healthcare register data in the Western Swedish Health Care Region (VEGA) to identify cases and population registers to identify matched controls, and linking these groups to several national registers, this study aimed to: examine the occurrence of several comorbidities at the time of first diagnosis of gout compared with matched controls, overall and by sex; and examine the crude and age-standardized prevalence of these comorbidities in men and women with gout at first diagnosis.

In this large, population-based study of patients with gout, recruited from both primary and specialized health care, we found that both comorbidities suggested to increase SU level and other comorbidities were more common in gout cases vs population-based controls, both overall and for men and women separately. All comorbidities, except for alcoholism, were or tended to be more frequent in women than in men at the time of first gout diagnosis. When standardizing for age, these sex differences were attenuated, and diagnoses of CVD were actually more prevalent in men. Since gout may be diagnosed several years after the first attack [50], both hyperuricemia and gout may have been present for a considerable time before and a causal relation to CVD or other comorbidities is not possible to deduce from the present study. Regarding occurrence of comorbidities suggested to increase SU level, there were only minor sex differences after age standardization, which suggests large similarities in the pathophysiologic pathways leading to gout in men and women.

The prevalence of some comorbidities in our study, such as hypertension, CHF, diabetes, PVD and renal disease, was higher than that in the study by Kuo et al. [43], despite the fact that both studies are register based, include a large group of patients and report occurrence of comorbidities at the time of first gout diagnosis. The prevalence of psoriasis, however, was almost the same in both studies. Possible explanations, apart from the different geographical setting, could be that the mean age of gout patients in our study was 5 years higher and that our study had a slightly higher proportion of women (32.6% vs 27.5%), who were found to have higher occurrence of associated comorbidities than men.

Other studies that have examined sex differences in occurrence of comorbidities by sex differ to some extent from ours. One previous study on established gout by Harrold et al. from 2017 [44] showed, in line with the present study, that women had higher prevalence of obesity, hypertension and diabetes mellitus and were more likely to take diuretics compared to men, although this study was based on a smaller sample of patients identified from an American network of rheumatologists. In another previous American study [45] based on 6133 patients diagnosed with gout as part of their health insurance plans, all evaluated comorbidities (obesity was not included) were more frequent in women with gout compared to men. On the other hand, these studies showed that women had higher prevalence of kidney disease compared to men, which is not confirmed in our study. One possible explanation could be that our study is conducted at the time of first diagnosis of gout, whereas the previous studies examined prevalent gout, and that the mean age of our male patients was 5 and 7 years higher respectively than in the two previous studies. There is also one small clinical study by De Souza et al. [47] based on only 58 patients that showed no differences in the occurrence of comorbidities between women and men, which could be explained by the small sample size of the study. We have also shown that thromboembolism and COPD were more frequent in women with gout compared to men. The association between these comorbidities and gout has not been studied before. The higher prevalence in women might suggest some differences in pathophysiologic pathways or risk profile between men and women, but it is not possible to study the underlying mechanisms in the present study.

Regarding CVD, we found a higher occurrence in men compared to women, when adjusting for age. Current literature on risk for cardiovascular events in men and women with gout reports conflicting results. In line with the present study that showed higher prevalence of CHD and CHF in men, when we standardized for age, Krishnan et al. [26] showed that men with hyperuricemia had an increased risk for myocardial infarction in a controlled trial examining the efficacy of risk reduction in men at high risk for vascular events. Results from the Framingham study [25] showed an increased risk for CHD and angina in men with gout, but not in women, independently of other cardiovascular risk factors. Choi and Curhan [4] showed in a large prospective study that men with gout had higher overall mortality compared to men without gout. In contrast, De Vera et al. [51] in a population-based cohort showed higher risk for myocardial infarction in women compared to men, and Clarson et al. [34] in a population representative cohort in the UK showed that female patients with gout were at greater risk for any vascular event compared to men with gout. These discrepancies could be explained by differences between studies in study design, sample size and selection of the study population. Regarding potential explanatory mechanisms for higher frequency of CVD in men, we showed that other conventional risk factors for CVD, such as obesity, hypertension or diabetes, were not higher in men, although smoking and dyslipidemia could not be taken into account due to lack of data.

A previous study by Wright et al. [39] showed that obesity, use of diuretics and, to a lesser extent, renal disease had a significant influence on plasma urate concentrations and allopurinol daily maintenance dose. According to this study, dose requirements to achieve treatment goals for SU were found to increase by 2-fold to over 3-fold with increasing total body weight and were 1.25–2 times higher in those taking diuretics, whereas renal function had a relatively modest impact on the required allopurinol dose. Another study [40] suggests that the effect of decreased eGFR and diuretic use is explained by their ability to increase SU levels and not by any effect on the dose–response relationship between the given allopurinol dose and the change in urate levels. Considering that women in our study and previous studies [44] have a trend toward a higher prevalence of obesity and use of diuretics, factors that increase urate levels in serum, these comorbidities should be considered when initiating and adjusting ULT in women.

Some possible limitations should be acknowledged. The identification of gout patients was based on ICD codes, which may have led to some misclassification bias. However, according to a previous validation study [49], a more strict definition of gout requiring one or more visits with an ICD-coded gout diagnosis by a rheumatologist or two or more visits in primary care was found to have a high positive predictive value for fulfilling the various classification criteria for gout. The comorbidity pattern of gout cases in the present study was very similar to the comorbidity pattern of those fulfilling the stricter case definition (Additional file 2: Table S2), suggesting similar validity for both case definitions. Regarding definition of comorbidities, the validity of ICD codes in the Swedish national patient registry is generally very high [48], although not all diagnoses used in this study have been validated. There is, thus, a possible risk for underestimation of some comorbidities, in particular obesity and alcoholism, when ICD codes are used for definitions. However, there is no reason to believe that such underestimation would be of different magnitude in cases compared to controls and in women compared to men. Smoking, which is a major risk factor for CVD, could not be taken into account because of a lack of data. Finally, the study was performed in the western part of Sweden, which may hamper generalizability. However, according to previous reports, the sociodemographic distribution and healthcare seeking in this region are very similar to Sweden as a whole [52, 53].

This study has also several strengths. First, it is a population-based study, including all gout cases in the region, which reduces the risk of selection bias. Second, patients were identified from both primary and specialized health care, which covers all of the different phenotypes of gout, from mild to severe disease. Third, several unrelated independent data sources were used. For instance, case identification and information regarding medication or education level were retrieved from completely independent sources. Fourth, the study includes a large number of subjects with gout and controls, which generates statistically robust estimates for the occurrence of the various comorbidities.

This large, population-based study shows that all patients with gout have higher occurrence of many comorbid conditions at the time of first diagnosis, compared to matched controls from the general population. The majority of these comorbidities are more common in women than in men with gout. In particular, the higher occurrence of diuretic use and obesity in women may need to be taken into account when initiating ULT.