From MEDLINE, we retrieved 178 original articles. The two observers (KD, TT) selected 69 articles on title and abstract (kappa 0.57). From EMBASE and the Cochrane Library, we retrieved another 20 articles. In the Web of Science, we searched for articles quoted by the authors of our top ten articles. No new articles were found. Out of the 89 full articles, we selected 14 original studies with specific information on sex differences of gout and a clear description about how they made the diagnosis of gout. These articles were scored according to the quality criteria mentioned in Table 1. The maximum score was 5. The minimum score for inclusion was set on 2.5. Nine articles were included, and five were excluded (Table 2) [11, 15].
Table 2 Selected studies with a score of less than 2.5 points
Two of the nine articles were conducted in a general population and seven in a hospital. The age of the male patients ranged from 53 to 70 years and of the female patients from 49 to 61 years. The ethnicity of the study population differed in the studies. In two studies [16, 17], the diagnosis of gout was made by synovial tap, in one by using both synovial tap and the ACR criteria [18], in four by the ACR criteria [19–22], and in two by ACR criteria or clinical improvement on treatment with colchicines [23]. The best way to classify a patient as having acute gout is to demonstrate characteristic sodium urate monohydrate crystals in the joint fluid. If the diagnosis is based on colchicine of ACR criteria, other causes of arthritis were also included in the studies. Because of this, we reviewed the studies with crystal-proven gout separately from the studies using the ACR criteria and the colchicine definition.
The diagnosis of gout based on synovial fluid analysis
Two studies are based on a crystal-proven gout [16, 17]. In the study of Puig et al., in 89% of the study population, the gout was crystal-proven (n = 40) [18]. Because of the high percentage, we include this study also in the group “crystal-proven gout” (Table 3).
Table 3 Sex differences according to the three selected articles with crystal-proven gout
Patient characteristics
The age of onset of gout as described in the three studies was on average 8.1 years later in women, and the duration of the gout at study entry was significantly higher in men. Gallerani et al. only studied the influence of the season on gout attacks and showed a significant peak of gout attacks in April in male patients, 36% of all the attacks [17]. Most of the women were postmenopausal at the onset of gout (86% and 92%).
Considering the use of diuretics, female gout patients received diuretics significantly more often than male patients, 57% of the women vs 14% of the men [18] and 83% of the women vs 47% of the men [16]. In the same studies, men with gout were more likely to drink alcohol than women, 14% of the women vs 55% of the men [18] and 10% of the women vs 45% of the men [16]. Puig et al. also analyzed sex differences in obesity in gout patients and found no differences between men and women.
Joint location
Puig et al. and Lally et al also described the location of the arthritis [16, 18]. Men seemed to have a higher prevalence of podagra at their first attack, 52% of the women vs 57% of the men [16]. Puig et al. found significantly more tophi in female gout patients compared to male, 27% in women vs 10% in men [18]. Lally et al. found a higher prevalence of polyarticular gout in male patients, 56% of the women vs 80% of the men [16]. They found no sex differences in tophaceous gout, but in their study the upper limb seemed to be more involved in male patients, 44% in women vs 47% in men [16].
Comorbidities
Finally, Puig et al. and Lally et al. reported the incidence of “gout-associated comorbidities” as diabetes, hypertension, and renal insufficiency [16, 18]. In both studies, an association was found between renal insufficiency and postmenopausal women, 54% of the women vs 11% of the men [18] and 30% of the women vs 12% in men [16]. Puig et al. considered that hypertension was more common in female than in male gout patients, 78% of the women vs 14% of the men [18]. The sex distribution on diabetes and gout was equal [18].
Gout according to the ACR criteria or clinical improvement on colchicine
Six studies did not base the diagnosis on a crystal-proven gout (Table 4).
Table 4 Sex differences according to the six selected articles with diagnose gout made by ACR criteria
Patient characteristics
The age of onset of gout was on average 9.2 years later in women (mean age 66 years) compared to men (mean age 54 years), and the duration of the gout at study entry was significantly higher in men. Most of the women were postmenopausal at the onset of the gout (66–95%). Three studies described the use of diuretics [19–21]. In two studies, female gout patients received diuretics significantly more often than male patients, 77% in women vs 40% in men [19] and 72% in women vs 48% in men [20]. Three studies reported the use of alcohol in gout patients [20–22]. They found that men with gout were more likely to drink alcohol than women, 2% of the women vs 20% of the men [20], 7% of the women vs 16% of the men [21], and 55% of the women vs 82% of the men. Tikley et al. found sex differences in the relation between obesity and gout. In women, they did not found any relation between the body mass index (BMI) and gout, but they found a higher prevalence of gout in men with a BMI > 25 (odds ratio 7.8) [22]. Harrold et al. studied the recurrence of gout attacks and reported a significantly higher number in men [19]. Gout in the family history was equally distributed between the sexes [21].
Joint location
The studies of De Souza et al. and Deesomchok et al. described the location of the arthritis [21, 23]. Men had a higher prevalence of podagra at their first attack, 23% of the women vs 45% of the men [23]. During the gout disease period, the sex differences in the location of the recurrent attack widened, with an increasingly higher prevalence of podagra in men [21]. Female patient more often had other joints involved, such as ankle, fingers, and upper limb [23]. Polyarticular gout seemed to be more related to female gout patients, 63% of the women vs 39% of the men [20] and 41% of the women vs 24% of the men [23]. The presence of tophi differed in the various studies, 39% of the women vs 26% of the men [20], 30% of the women vs 48% of the men [21], 18% of the women vs 31% of the men [24], and 34% of the women vs 18% of the men [23].
Comorbidities
Five studies reported the incidence of several “gout-associated comorbidities” [19–23]. In four of them, hypertension was more common in female than in male gout patients, 81% of the women vs 57% of the men [19], 81% of the women vs 77% of the men [21], 45% of the women vs 39% of the men [23], and 65% of the women vs 59% of the men [22]. In two studies, this difference was significant [19, 22]. Five studies considered diabetes [19–23]. In one study, diabetes was significantly more prevalent among female than among male gout patients, 30% of the women vs 17% of the men [19]. The other four were contradicting.
Four studies analyzed the relationship between gout and cardiovascular heart disease. These results were also inconsistent, 25% of the women vs 19% of the men [19], 26% of the women vs 57% of the men [20], 25% of the women vs 16% of the men [21], and 0% of the women vs 11% of the men [23]. Dyslipemia was more common in women (42% vs 38%) [18–21] and Deesomochok et al. did not find a sex difference in cerebral vascular accident in gout patients [23]. In this study, a significantly higher prevalence of hematologic malignancies was found in female gout patients, 22% of the women compared to 3% of the men [23].
Renal insufficiency was studied in five studies, and in all an association was found with gouty arthritis in postmenopausal women (in three of them, the association was significant), especially in those with preexisting joint disease, 18% of the women vs 10% of the men [19], 25% of the women vs 15% of the men [20], and 22% of the women vs 12% of the men [24].
We studied the conclusions of the five excluded studies and compared these with the nine included papers. No significant difference in outcome variables was found between the two groups.