One of the interesting issues to consider in studying gender disparities in SLE is the nature of its initial presentation in male and female patients. The greater awareness of SLE as a potential diagnosis, particularly in fertile females, may lead to a greater delay in diagnosis in men with similar symptoms. Alternatively, if men displayed an atypical phenotype at presentation, a delay in diagnosis, and thus treatment, might result. The consequence would be a greater burden of inflammation and subsequent damage over time.
Demographically it has been reported that male SLE patients are older at disease onset. However, this is not strongly supported by the literature. In fact, there is little evidence to support a discrepancy in either disease onset (time of first SLE symptoms) or diagnosis (fulfilment of four ACR criteria) by gender. While there is considerable variability, males have been reported to experience first SLE-related symptoms at a mean age range of 26–38.4 years in comparison with females (26.3–31.9 years). The mean age range at diagnosis in males was 26–55 years, with females being diagnosed at a mean age of 27.9–42.6 years. Only two studies reported a significantly older age of onset in male subjects, one of which was a study of army veterans, which seems likely to be biased by its recruitment policy. There is some suggestion that the age of onset in males and females may differ in certain ethnicities. Specifically, two studies have reported an older onset of SLE in males compared with females in the Caucasian subpopulation of their cohorts. Interestingly, the overall trend was to a shorter delay in diagnosis of male patients, with a range of 6–46 months between symptom onset and diagnosis in males and 8–58 years in females.
To address the clinical phenotype at presentation, a number of groups have specifically assessed organ involvement at disease onset. Although significant heterogeneity exists, the most consistent findings are a lower incidence of musculoskeletal symptoms, RP, alopecia and photosensitivity in men at diagnosis, with the suggestion of more prevalent serositis and discoid lupus. Reduced arthralgia/arthritis has been reported in Chinese (P = 0.02), Greek (P = 0.004) and Caucasian (P < 0.03) populations. Serologically only anti-Ro antibodies have been shown to have a gender-specific association at diagnosis, being more frequently positive in females (P = 0.049). Only one study has found a greater incidence in nephropathy in men at disease onset, inconsistent with findings from an earlier Greek cohort. Stefanidou et al. also found higher frequencies of thromboses (P < 0.01) and gastrointestinal symptoms (P < 0.01) in men. Similarly, no difference in renal biopsy frequency at diagnosis or the resultant histology was shown in a Chinese SLE cohort. Interestingly, a study from Latin America suggested that males may display more constitutional symptoms at diagnosis; this, however, has not been reported in other geographic regions. Overall therefore, despite the suggestion that men less frequently display the typical mucocutaneous features of SLE at disease onset, this has not been reflected by a delay in diagnosis as outlined above.
The relative paucity of male patients with SLE has contributed to the difficulty in analysing the disease course in males. This problem has been circumvented by the use of case–control studies and a multicentre approach with more recent reports describing larger inception cohorts. Both ethnicity and age have a strong influence on SLE expression; in particular, patients of African American extraction frequently exhibit more significant renal disease, hypertension and discoid lupus and less photosensitivity. In comparison with Caucasians, Hispanics also have a greater incidence of arthritis. Differences are also noted within ethnic subpopulations. For example, Hispanics from Texas (of Mexican/Central American ancestry) more frequently suffer from serositis, renal involvement, psychosis and thrombocytopenia and less frequently from photosensitivity and malar or discoid rash than Puerto Rican Hispanics. While this may reflect the complex interplay of genetics and environment, it highlights the great difficulties in dissecting the role of gender in altered disease phenotypes from what are typically heterogeneous populations.
It seems prudent to distinguish studies from different geographical regions and ethnic populations to make cohort comparisons more meaningful. This approach may enable a better assessment of the role that gender plays in disease expression within particular population subgroups.
Asia. Previous studies have reported that Asian patients have higher rates of renal involvement, more active renal disease and higher rates of nephritis-associated autoantibodies in comparison with predominant white populations. To characterize the gender-specific symptomatology in Asian patients, Mok et al. performed a case–control study of 51 male and 201 female patients followed in rheumatology or nephrology outpatients. After a median disease duration of 103.6 months in males (101.6 months in females), no significant differences in major organ involvement were identifiable, despite a trend towards less RP, alopecia and arthritis in men. Importantly, there was no evidence of increased damage using the SLICC/ACR index and treatment modalities were not influenced by gender. Interestingly, male subjects appeared to have a lower number of disease flares per patient-year (P = 0.04). Serologically only anti-Ro antibody status differed by gender, present in 62% of females and 47% of men (P = 0.05). Some similarities are found between these results and an earlier case–control study by Koh et al. In this group, the incidence of renal disease and serositis was similar between Oriental male and female patients with SLE, but fewer males were seropositive for anti-Ro antibodies (P < 0.001). This group demonstrated an increased frequency of arthritis and leucopenia in females (P < 0.04). A more recent study of a Chinese cohort failed to detect a significant difference in the clinical phenotype of male patients with SLE, although a trend towards higher rates of renal involvement was reported and SLEDAI scores were significantly high in the male subgroup.
Thai males have also been shown to have less arthralgia than females in addition to less RP and alopecia and more prevalent thrombocytopenia (P < 0.05 for all). While the incidence of overt nephropathy did not differ, there were significantly more male subjects with increased creatinine levels over an average follow-up of 2 years (P = 0.006). A smaller study aimed primarily at dissecting the impact of gender on cutaneous manifestations of SLE found no difference between male and female patients in non-cutaneous manifestations by retrospective chart review.
Europe. Many large European studies have addressed the modifying effect of gender on disease expression. Even within the same population, conflicting results are reported. Both Stefanidou et al. and Voulgari et al. examined the clinical phenotype of SLE in Greek patients. The latter found that men had significantly more serositis (P < 0.01), less photosensitivity (P < 0.05), oral ulcers (P < 0.01), RP (P < 0.05), thrombocytopenia (P < 0.05) or increased ESR (P < 0.01) in comparison with women. There was no overall increase in renal involvement in men. In contrast, Stefanidou et al. found that male SLE patients had higher rates of nephropathy (P = 0.009), myositis (P = 0.023) and tendonitis (P = 0.007) at follow-up, with no discernible difference in the prevalence of cutaneous or mucosal features as reported by Voulgari et al.. These differences may be somewhat biased by the recruitment process, with Stefanidou et al. basing their study on patients attending rheumatology or nephrology clinics, and by both disease duration and study type. It is interesting, however, that despite using a patient group specifically recruited from nephrology outpatients, which might be expected to result in an overall higher prevalence of nephropathy, both Voulgari et al. (24.6% males) and Stefanidou et al. (27.1% males) report rates of renal involvement much lower than many reported studies. In terms of overall disease activity, assessed by the ECLAM index, and damage, assessed by the SLICC/ACR index, Voulgari et al. showed no significant gender difference.
Focusing on Western European populations, four additional studies have been published addressing gender and SLE. Lopez et al. examined the demographic and immunological features of a homogeneous cohort of SLE patients. Females were more commonly anti-Ro antibody positive (P = 0.047) and younger (P < 0.01) at diagnosis. In patients of similar ethnicity, Font et al. failed to show a significant difference in onset age or serology, but demonstrated a lower incidence of arthritis and malar rash at follow-up (P < 0.01), with increased frequencies of discoid and subacute cutaneous lupus (P < 0.03). No difference in the incidence of nephropathy was detected. In contrast, a report from a Danish cohort described an increased incidence of nephropathy, including end-stage renal failure. The prevalence of serositis was also higher and photosensitivity less. Again this study used a recruitment approach that included nephrology outpatients, which must be borne in mind when interpreting these results.
UK populations, which are more ethnically diverse, are represented in a study by Renau et al. using a retrospective approach. In contrast to the above, a trend towards an increase in renal failure and death was noted in females with SLE, with a significant increase in oral ulceration (P < 0.05) and IgM aCLs in women (P < 0.05).
While these studies vary considerably by design and duration of follow-up and are open to many biases, including selection bias, no clear patterns of gender-specific disease expression have emerged in European populations.
USA. The significant ethnic diversity of the US population makes discerning the impact of gender, independent of ethnicity, more difficult. This problem is further complicated by differences in age of onset between ethnic populations, given the known modifying effect this has on disease expression. However, several large studies have been reported that have attempted to circumvent these issues and correct for the effect of ethnicity on disease expression.
Two recent cohorts have suggested that male patients have an increase in renal disease independent of ethnicity and age. Both used systematic follow-up and included a large number of patients (618 and 1979 patients). Andrade et al. correlated male gender with reduced musculoskeletal symptoms and higher frequencies of LA antibodies. Tan et al. showed that men had significantly increased disability, lymphopenia, thrombocytopenia and a spectrum of manifestations of renal disease (nephritic syndrome, proteinuria, haematuria, renal insufficiency and failure) (P < 0.05). In addition to corroborating the lower risk of musculoskeletal involvement, in the latter study men were protected from malar rash, photosensitivity, oral ulcers, alopecia and RP (P < 0.05). Serologically, in the Tan et al. study, men had a greater prevalence of Coombs positivity, lupus anticoagulant, low C3 anti-Sm, anti-dsDNA, anti-RNP and anti-La antibodies than females (P < 0.05). These results are notably out of line with any other study of male lupus. Both Tan et al. and Andrade et al. also highlighted a potential association between gender and damage. Andrade et al. reported a higher mean SDI in men at baseline and showed a positive correlation between male sex and further damage accrual. Tan's group showed evidence of increased neuropsychiatric, renal and cardiovascular damage in African American and white males in comparison with females. Interestingly, African American men were more likely than Caucasians to have discoid lupus, alopecia, renal involvement, anti-Sm antibodies and greater damage in renal and cardiovascular domains. Looking specifically at recently diagnosed disease (median duration from diagnosis to enrolment 13 months), the Carolina Lupus study showed an increase in proteinuria and haematological dyscrasias in men.
These data contrast notably with earlier studies that failed to identify an association between gender and renal disease. One of these studies included 2188 patients identified through the US Veterans Administration database. As would be expected, this cohort was significantly older than those in other studies. While no difference in renal disease was observed, male patients had higher rates of myocardial infarction and neoplasm. Ward et al. found that only an increased risk of seizures could be attributed to male gender, while Miller et al. associated serositis with male SLE and females had a greater incidence of neurological involvement, thrombocytopenia and alopecia.
Clearly these studies have reported significantly differing results. It is of interest that it is the more recently performed studies that have suggested that renal disease is more prevalent in men, even after adjusting for ethnicity. Whether this truly reflects better detection or whether it is the result of selection, referral bias or differences in other demographic details is unclear.
Latin America. There have been two large studies examining the question of disparate clinical phenotypes in men with lupus in Latin America. The first, a multicentre study (GLADEL cohort), found that men were significantly more likely to experience constitutional symptoms, e.g. fever and weight loss, in addition to having more cardiovascular disease (hypertension), proteinuria, cellular casts and haemolytic anaemia (P < 0.05). There was a nonsignificant trend towards a higher frequency of glomerulonephritis and no difference in the incidence of nephrotic syndrome or serositis. Women, on the other hand, suffered from more alopecia, photosensitivity, arthralgia and RP. Serologically only IgG aCLs and low C3 demonstrated any gender predilection (more common in men). Despite 68% of male patients having aCLs, there was no significant increase in vascular thromboses. Importantly, no significant difference in the objective assessments of disease activity (SLEDAI index), damage (SLICC index) or mortality was detected following adjustment for confounders.
Interestingly, an earlier cross-sectional study also found an elevated incidence of renal disease in male SLE patients (58% vs 44%) and suggested an increase in vascular thrombosis and protection from RP (P < 0.03). While no difference in the prevalence of IgG aCLs was detectable in this cohort, they found that anti-dsDNA antibodies appeared to be more prevalent in the male subgroup (P = 0.002).