Over 50% of patients with systemic lupus erythematosus (SLE) have clinically significant renal involvement during the course of their disease. Approximately 8•20% of these are found on renal biopsy to have membranous lupus nephropathy (MLN).1,2 Pure MLN has distinct histologic features compared to proliferative lupus nephritis.3 This difference is reflected in the clinical presentations of proliferative and membranous lupus nephritis, with the former being dominated by a nephritic process whereas MLN is characterised predominantly by nephrotic syndrome.1,4,5

Few studies have addressed the course and outcome of MLN directly,4,6•14 and their results are contrasting. Most of our knowledge is derived from extrapolations of data of idiopathic membranous nephritis and/or proliferative lupus nephritis. Research on MLN frequency, natural history, prognosis and treatment was further limited due to a lack of uniform definition over the past several decades, making it difficult to get consistent information from the published literature about MLN.1

An Italian study15 has reported the outcome in 103 patients with MLN, but a third of them had mixed forms. More recently, Mejía-Vilet et al.13 have reported the results of a cohort of 60 Hispanic and Mexican-mestizo patients with pure MLN suggesting an impact of ethnicity on the response to different immunosuppressants (azathioprine, cyclophosphamide or mycophenolate).

Large cohorts of patients with pure MLN are lacking, and to date the influence of ethnicity and social conditions on pure MLN outcome has rarely been studied.

The aims of the present study are to evaluate the clinical features, course, outcome and prognostic indicators in pure MLN and to explore the association of ethnicity, socio-economic level, country of residence and the type of health insurance with the course and prognosis of pure MLN.

In this study, we retrospectively analysed the baseline and long-term follow-up characteristics of a cohort of 150 patients with pure MLN. Unlike many other studies, we excluded patients with mixed forms of membranous nephritis. There are only a few studies concerning pure MLN, and the conclusions about the presentation and natural course are unclear. This is possibly a consequence of the difficulties to identify histologically pure forms of MLN according to the different classifications and also of the limitations for the follow-up given that an optimal treatment for this type of lupus nephritis has not been established yet. Making inferences about pure MLN from cohorts that include different subtypes of class V lupus nephritis can be deceptive.10,19,20 To the best of our knowledge, this is the largest cohort on pure membranous lupus nephritis to be reported in the literature. It is also the first to compare two large groups of patients with MLN from different sociocultural settings.

It is noteworthy the high percentage of patients with positive anti-dsDNA antibodies in our series (73.3% of patients). MLN may be present with many, few or no other clinical or serologic manifestation of SLE.21,22 Anti-dsDNA antibodies are highly specific of SLE and their high prevalence in our cohort supports these were confirmed SLE diagnoses. This high percentage of anti-dsDNA positivity is similar to other series.13

Patients
tm) characteristics at MLN presentation in our cohort are in agreement with those reported by other authors (Table 4). As in our cohort, MLN usually presents in women in their fourth decade of life and after a mean of four years after SLE diagnosis. Renal function is usually preserved at the beginning; impairment has been reported in 0•20% across different cohorts.6,8•11,19,20,23•28 Nephrotic proteinuria is found at diagnosis in a variable proportion of patients, and it usually evolves favourably with low percentages of death and/or ESRD at follow-up.13,22,24,27,29 The percentage of poor outcomes does not seem to vary with the length of the follow-up, since the renal survival is similar independently of the follow-up of the cohort. What seems to adversely influence the outcome is the presence and persistence of a nephrotic syndrome and having mixed proliferative lesions.4•6,8,9,30 In very old and short-term studies18,31 MLN was felt to exhibit an indolent and relatively benign course. However, subsequent long-term series have shown contrasting data about the prognosis of this disorder. Baldwin et al.32 found 71% of their 24 MLN patients to be persistently nephrotic; 25% progressed to renal failure over a period of 6 years. González-Dettoni et al.33 observed one the highest percentages of ESRD amongst published cohorts with 25% of their 16 patients after a 4.5 years follow-up. In contrast, ours and other cohorts13,18,19,21 have found a lower rate of progression to ESRD (Table 4). Moreover, a number of recent cohorts have even reported no patients progressing to ESRD.11,23,24,26,34 It is likely that different patient populations with varying epidemiological features such as ethnicity and socioeconomic level, different degrees of proteinuria and renal insufficiency, differing histologic features of MLN and diverse treatment protocols may account for these divergent results. Establishing these differences was precisely one of the major points of our comparison study between two international cohorts.

Some significant differences were found between our two cohorts. The American cohort had higher poor prognostic indicators than the Spanish cohort as a higher percentage of patients had nephrotic syndrome, high blood pressure and impaired renal function at disease presentation. SLE duration at MLN diagnosis was longer in the American cohort. This needs to be cautiously interpreted as it may reflect a more difficult access to health care rather than a milder disease in this population. The two cohorts were also quite different regarding the ethnic groups distribution. The North-American group included mainly Hispanic-American patients and patients of African descent whereas the majority of the Spanish patients were Caucasian. Our multivariable analysis did not show prognostic differences among the different ethnicities, probably due to the small number of patients of some of the ethnic groups in the sample. Nevertheless, it is well established that there are considerable racial differences in the prevalence, severity, and prognosis of lupus nephritis.35,36 Patients of African descent are more severely afflicted with SLE and have worse outcomes when compared with Caucasian patients.36,37 The reported prevalence of SLE is seven to eight times higher in Afro-Caribbeans, and two to four times higher in Asians than in Caucasians, while biopsy-proven lupus nephritis is 25 times more common in both these groups compared with Caucasians.38,39 It is also known that the response to therapy varies among different ethnicities.13 A higher rate of response to therapy has been described in Asiatic cohorts.40 On the other hand, Hispanics may have a higher response to mycophenolate than to cyclophosphamide as induction therapy and a lower rate of relapses when they receive mycophenolate as a maintenance treatment as compared with azathioprine.41•43

In addition to ethnicity, low socioeconomic status has also been associated with progression of SLE, mainly through increased cardiovascular risk.44 If we consider the type of health insurance as a proxy of the socioeconomic status for the North-American cohort, it could explain why this cohort had a more severe disease at presentation and a poorer prognosis. Nevertheless, due to the characteristics of the health system in Spain, which has universal health coverage, the type of health insurance cannot be taken as indicative of socioeconomic status for this population. The entire Spanish cohort had public health insurance whereas only 2 patients from the North-American cohort had this type of health insurance. Furthermore, half the North-American cohort had no health insurance at all. This fact may have delayed the diagnosis of MLN in USA patients and also limited the therapeutic possibilities influencing the control of the disease. We found significant differences in the use of oral steroids and several immunosuppressants in favour of the Spanish cohort. Although there is no standard immunosuppressive regimen for MLN, it is known that the addition of immunosuppressive drugs in combination with steroids may improve renal survival.45 The wide variation of treatment combinations in our cohort did not allow us to reach any conclusion about the optimal regimen, however our data suggests that an aggressive therapeutic approach influences the course of the apparently benign MLN. Patients from the Spanish cohort, who more often received immunosuppressive therapy, had a better prognosis, reflected by a higher percentage of patients achieving a 24h-proteinuria under 0.5g at the end of follow-up and lower percentage of patients developing high blood pressure. In particular, there was a notable difference in the use of cyclophosphamide (28.4% of patients in the Spanish cohort vs 18.8% in the American cohort) and azathioprine (39.2% vs 14.6%) between the two cohorts. It is likely that doctors in Spain prescribed immunosuppressants more frequently trying to avoid poor outcomes given that these treatments were easily available. Another explanation can be that these drugs were given to patients who either had relapses after treatment with only corticosteroids or were steroid-dependent.

Significant differences were also found regarding the use of antihypertensives, diuretics, statins and antiaggregants between both groups since they were more commonly used to treat patients in the Spanish cohort. Patients with MLN are at increased risk not only for ESRD but also cardiovascular complications45,46 and adjuvant therapy has an impact on the MLN prognosis.47 In fact, in our study, not only ESRD but also mortality was predicted mainly by cardiovascular comorbidities such as cardiac insufficiency, ischaemic cardiopathy and peripheral arteriopathy. Over the last few decades, with the decrease of early mortality due to uncontrolled disease, cardiovascular complications have emerged as important causes for late mortality.48 Since vascular complications are also accelerated in patients with renal failure, the risk of cardiovascular complications is compounded in patients with ESRD and SLE.49,50 Therefore, adjuvant therapy can have a positive influence on prognosis in patients with lupus nephritis.

We found significant differences in the time of follow-up. Spanish patients were followed three times longer than the USA cohort. Despite this notable difference, overall, the outcome was better for the Spanish cohort. It is possible that the USA cohort would have had even a worse prognosis if having been followed up for almost 10 years as the Spanish one. Nevertheless, there is no evidence that the length of follow-up conditions a poorer prognosis as presenting with nephrotic syndrome or having mixed lesions do.6,8,9

High basal proteinuria, as well as a high serum creatinine and a low creatinine clearance were all predictors of ESRD. This means that patients with a more severe basal nephritis had more chances of losing their renal function at follow-up. In the multivariable analysis, we only identified a low baseline serum creatinine and having health insurance as predictors of good prognosis in terms of achieving a 24h-proteinuria under 0.5g at the end of follow-up, but it is possible that in a larger sample factors such as ethnicity, cardiovascular conditions or nephrotic syndrome are identified as independent predictors. Nevertheless, no major prognostic indicators of MLN renal survival emerges from the literature. Mercadal et al.29 reported that a profound initial hypoalbuminaemia was a risk factor for ESRD and a sustained heavy proteinuria was a predictor of doubling of the serum creatinine. However other authors18,51 have found no correlation between the degree of proteinuria at the time of the renal biopsy and subsequent renal function deterioration among patients with MLN. Austin et al.52 identified proteinuria higher than 5g/d and treatment with prednisone alone as associated with a decreased probability of remission. Substantial data has established the prognostic importance of proteinuria in various glomerular diseases including idiopathic membranous nephropathy22,53 but it seems that in membranous lupus nephritis data are still contrasting.

Our study has several limitations. Although the Spanish MLN patients are well represented in our sample since they were followed up at 24 different participating centres, all the patients from the USA attended a single centre in New York. It is probable that this hospital does not represent the average standard of care in the USA. Thus, although our results lack of generalizability, the comparison between two international cohorts represents a strength as it brings up some determinants of pure MLN outcome not visible in more homogeneous cohorts. Since our aim was to make a description of MLN patients we did not record the date of events like death or ESRD and subsequently could not perform a survival analysis. The wide variety of therapies received by each patient prevented us from evaluating the effectiveness of each single immunosuppressant for pure MLN since the final outcome of the disease could not be associated with a specific drug. According to what has been previously reported6,8,9 we do not think that the difference in the mean follow-up between the two cohorts influenced the results; in any case, a longer follow-up for the USA cohort would have made the differences in prognosis between cohorts more obvious so it seems unlikely our findings can be explained by the different length of follow-up. A retrospective design is not the ideal methodology for prognostic studies. Ideally, a prospective study would identify true prognostic factors for MLN. Despite this, the large number of patients included in this cohort would enable probable prognostic factors to be studied in greater detail in future prospective cohorts. Furthermore, these preliminary results and our review of the literature draw attention to the influence of different types of health insurance and ethnicity in the prognosis of membranous lupus nephritis.

MLN usually begins with nephrotic syndrome, high proteinuria and normal serum creatinine. Prognosis is favourable in terms of the maintenance of renal function, although proteinuria usually persists over time. Patient and renal survival are high in patients with pure types of MLN but severity at presentation predicts a poor long-term outcome in terms of ESRD. Other factors such as baseline cardiovascular disease and not having a health insurance are also related with poor prognosis.

This study had no funding source or sponsor.

Ethics approval: Ethical approval for this study was obtained in February 2011 from the Hospital Universitario Puerta de Hierro Majadahonda of Madrid, Spain.

The authors declare no conflicts of interest.