LUPUS MEMBRANOUS GLOMERULONEPHRITIS: DIAGNOSTIC PECULIARITIES AND LONG-TERM OUTCOME

Beatriz Domínguez-Gil, MD. Eduardo Hernández, MD. Agustín Carreño, MD. Esther González, MD. Milagros Ortiz, MD. Mª del Prado Sierra, MD. Jose Luis Rodicio, MD. Manuel Praga, MD.

Department of Nephrology.Hospital 12 de Octubre. Madrid. Spain.

CORRESPONDENCE:
Beatriz Domínguez-Gil, MD
Paseo de la Reina Cristina, nº 16-4º
28014 Madrid. Spain.
e-mail: jgranelln@seorl.org

KEY WORDS: Lupus nephritis; membranous glomerulonephritis;systemic lupus erythematosus; angiotensin- converting enzyme inhibitors.

ABSTRACT

Lupus membranous glomerulonephritis (LMG) is sometimes difficult to distinguish from idiopathic membranous glomerulonephritis (IMG) and few is known about its natural history.We performed a retrospective study of 16 patients with LMG, types Va ( "pure" LMG) (8 patients) and Vb (mesangial proliferation) (8 patients). Analitical and clinical characteristics were analyzed in each case at the onset of the disease and during the follow-up.

At the time of diagnosis, 7 patients (43%) did not present any extra-renal manifestation of systemic lupus erythematosus (SLE). Nephrotic syndrome was the most common clinical presentation (81%). 14 patients (87%) showed positive antinuclear antibodies (ANA), but at low titres (<1/300); anti-DNA were positive only in 3 cases (18%) and hypocomplementemia was observed in 2 (12%). The most relevant histological findings were the presence of C1q deposits (83%) and subendothelial and mesangial deposits. Treatment consisted on corticosteroids, adding immunosuppression when no response was obtained in 2-3 months. In recent years, an angiotensin-converting enzyme inhibitor (ACEI) was added before immunosuppression. At the end of the study, 6 patients (42%) developed chronic renal insufficiency and 8 (58%) kept a normal renal function. Actuarial renal survival was 100% at 5 years and 75% at 10 years. The evolution towards a non-nephrotic range proteinuria throughout the follow-up was the only parameter significantly associated with a good prognosis

In conclusion, LMG can present without extra-renal symptoms of SLE and even with very poor expression of its serological markers. Almost half of the patients (42%) developed a chronic renal insufficiency. The evolution towards a non-nephrotic range proteinuria correlated with the maintenance of a normal renal function.

INTRODUCTION

Lupus membranous glomerulonephritis (LMG), type V of the World Health Organization (WHO) classification, is a peculiar entity refering to its clinics and evolution. According to the revised classification of the WHO [1], there are two subtypes of LMG: type Va or "pure" LMG and type Vb with an associated mesangial hypercelularity. Two other subtypes had been classically described: Type Vc with coexisting signs of focal proliferative gomerulonephritis and type Vd with signs of diffuse proliferative glomerulonephritis [2].

LMG is still a not well known entity in different aspects. First of all, and due to its clinical and serological peculiarities, it is sometimes difficult to distinguish from idiopathic membranous glomerulonephritis (IMG). On the other hand, we do not know its natural history, since following diagnosis, treatment is usually iniciated. Therapeutic approaches known to be useful in other types of lupus nephropathies are of doubtful utility in the treatment of LMG because of the lack of prospective studies to address this point. Besides, some of the papers published on this type of nephropathy include histologycally heterogeneous patients, i.e. with and without associated proliferative lesions.

In this study we review our experience with LMG types Va and Vb. We therefore exclude patients with associated proliferative lesions ( the classical types Vc and Vd). We analyze the clinical, serological and histological features that should make us distinguish this entity from IMG. We finally review the evolution of the patients, the effect of the treatments used and factors with a possible prognostic value.

PATIENTS AND METHODS

We present a retrospective study of sixteen patients diagnosed of LMG at the Department of Nephrology of the 12 de Octubre hospital in Madrid (Spain) from 1975 to 1998. Every patient fulfilled the revised criteria of the American Rheumatology Association [3] for the diagnosis of Systemic lupus erythematosus (SLE) sometime during the evolution of the disease. The diagnosis of LMG required histological evidence: diffuse swelling of the capilar wall due to the presence of subepithelial electrondense deposits with an homogeneous distribution. We considered a LMG type Vb when an associated mesangial proliferation existed. Every case with focal or diffuse proliferation was excluded (subtypes Vc and Vd). Each biopsy was reviewed for the present study with optic microscopy in 16 cases (100%), immunofluorescence in 12 (75%) and electron microscopy in 4 (25%).

Analitical and clinical characteristics were analyzed in each case at the onset and during the follow-up of the disease. Patients were periodically revised at the out-patients department except during hospitalization periods. Once the diagnosis of LMG was established, the following parameters were determined in a protocolized way: bioquimics (blood count, autoanalyzer), creatinine clearance, 24 hours urine protein excretion, urine sediment, antinuclear antibodies (ANA), anti-double stranded DNA antibodies and serum levels of components C3 and C4 of complement.

In order to evaluate renal manifestations, we have used the following definitions: Non nephrotic proteinuria: urine protein excretion lower than 3.5 g/day; Nephrotic syndrome: urine protein excretion greater than 3.5 g/day with hypoalbuminemia and hypoproteinemia. We considered arterial hypertension if blood pressure levels were greater than 140/90 mmHg in a sustained way. Renal insufficiency was defined as a serum creatinine value greater than 1.4 mg/dl and/or a creatinine clearance lower than 60 ml/min.

Once the diagnosis of LMG was established, the usual therapy consisted on oral corticosteroids. When after a follow-up period of about 2-3 months no significant improvement in renal manifestations was seen, immunosuppressive therapy was added: cyclophosphamide (oral or intravenous bolus) or azathioprine. Since a long period of time has been analyzed (1975-1998), therapeutic protocols have been changing, but this same scheme has been basically kept. In the early period, azathioprine was the immunosuppressive drug used in those cases which did not improve with corticosteroids; lately, is was cyclophosphamide the selected drug, which was administered p.o. or in periodical intravenous bolus more recently. In the last years we have addopted the policy of using angiotensin-converting enzyme inhibitors (ACEI) along with corticosteroids in cases with significant proteinuria (greater than 1 g/day), before introducing immunosuppressive therapy.

In order to analyze those factors with a possible prognostic value, we have studied the following parameters: age, gender, serum creatinine, proteinuria and blood pressure at the time of diagnosis; treatment used (corticosteroids alone, corticosteroids plus azathioprine, corticosteroids plus cyclophosphamide) and remission of proteinuria towards a non nephrotic range during follow-up. These parameters were related to the fact of developing a chronic renal failure versus mantaining a normal renal function during the study. Data have been expressed as media ± standard deviation and comparison between groups has been performed using Fisher test for qualitative variables and non-parametrical Wilcoxon and Mann-Whitney tests for quantitative variables.

RESULTS

Our series consists on sixteen patients ( 5 males and 11 females), with a medium age of 41.9 ± 12 years (range 27- 64 years). The average time between set-up of clinics and renal biopsy was 32.4 ± 16 months.

CLINICAL PRESENTATION

Renal disease showed as nephrotic syndrome in 13 patients (81%) with an initial proteinuria of 5.96 ± 3.2 g/day while the rest of the patients showed non nephrotic proteinuria. 6 patients (37%) had renal insufficiency at the onset of the process with a serum creatinine value of 1.7 ± 0.2 mg/dl, but none of them required dialysis and all presented an associated nephrotic syndrome. 7 patients (43%) showed arterial hypertension on set-up. Before or during renal manifestations, patients suffered from different signs and symptoms of SLE (Table 1), although at the moment renal biopsy was performed 7 patients (43%) showed non extra-renal clinics of SLE.

Table 1: Extra-renal manifestations of SLE , making the difference among those present at the moment renal biopsy was performed and those that appeared later on.

SEROLOGICAL MARKERS

Serological manifestations of SLE were poor in our series. 14 patients (87%) were positive for ANA, but most of them at low titres: below 1/100 in 7 patients and higher in 4, although lower than 1/300 in these 4 cases. Moreover, 2 patients (12.5%) were negative for ANA. Consistently, anti-DNA were positive only in 3 patients (18%) and just in one of them at a high titre. Presence of hypocomplementemia, as descent on serum C3 and C4 levels, was found only in 2 cases (12.5%).

PATHOLOGY

Renal biopsy was performed in every patient showing membranous glomerulonephritis on optic microscopy. 8 patients (50%) were WHO type Va and the rest Vb (with associated mesangial proliferation). Immunofluorescence was available in 12 cases (75%), with IgG deposits in 11 (91%), IgM in 1 (8%), IgA in 3 (25%), C3 in 10 (83%), C4 in 2 (16%) and obvious deposit of C1q in 10 (83%). In 4 patients (25%) electron microscopy was available; in all of them, along with electrondense subepithelial deposits, such deposits were found also in other locations such as mesangium or subendothelium.

TREATMENT

8 patients (50%) out of the 16 included in the study were treated with prednisone, 3 (18.7%) were given prednisone plus oral cyclophosphamide (2 mg/kg/day),1 patient (6%) received prednisone plus intravenous cyclophosphamide (6 monthly bolus of 1 g/m2) and 2 (12.5%) were treated with prednisone plus azathioprine. 2 patients (12.5%) did not receive specific treatment. During the last years, we have introduced the policy of sistematically using ACEI even in normotensive patients soon after iniciating treatment with corticosteroids. In the present series 6 patients (37.5%) were treated with ACEI as referred above.

FOLLOW- UP AND PROGNOSTIC FACTORS

2 cases were lost for follow-up after the diagnosis of LMG was established. The remaining 14 cases have a medium follow-up of 60.6 ± 47 months. 2 patients died during the study; one of them 30 months after diagnosis due to a stroke and the other patient 3 months after renal biopsy in the context of a thrombotic thrombocytopenic purpura. At the end of the follow-up, 6 patients (42%) have developed a chronic renal failure, leading to dialysis in two of them. 8 patients (57.1%) keep a normal renal function. Actuarial renal survival is 100% at 5 years and 75% at 10 years.

We have found no statistically significant differences in terms of age or gender between the group of patients with normal renal function and the one with chronic renal failure at the end of the study. Initial serum creatinine tends to be higher in the last group but without an statistical importance. Neither statistically significant differences have been found between both groups according to initial proteinuria, presence of hypertension or therapeutic protocols. Only the evolution to or the maintenance of a non nephrotic proteinuria was more frequent in the group of patients that keeps a normal renal function at the end of the study, being this difference statistically significant (p=0.02) (Table 2).

Table 2: Comparison of variables between the group of patients developing chronic renal failure and the group of patients keeping a normal renal function at the end of the follow-up.

M:Males, F:Females, Cort:Corticosteroids, IS:Immunosuppressive drugs,NS:Non statistically significant

In this way, the 6 patients developing a chronic renal failure showed a nephrotic syndrome at the onset of the disease and in 5 of them (83%), proteinuria was in the nephrotic range along the entire follow-up. On the other hand, among the 8 patients keeping normal renal function, 7 (87%) kept with or evolutioned towards a non nephrotic proteinuria during the study, being 5 out of these 7 patients (71%) treated with ACEI following the above described protocol (Figure 1).

Improvement of proteinuria with ACEI, without other immunosuppressive therapy but corticosteroids was significant, especially in 3 cases that showed a certainly spectacular evolution (Figure 2).

Given serological and clinical peculiarities of LMG, it is sometimes difficult to make the differential diagnosis with IMG. A brief report of two patients is described to ilustrate this point. Both of them had no extra-renal symptoms of SLE and serology was poor, so an incorrect diagnosis of IMG was made. After a period free of immunosuppressive therapy, an acute and severe lupus outbreak showed up. In one of them evolution of membranous glomerulonephritis towards a diffuse proliferative form occurred.

A 30 years old woman showed a 2.5 g/day proteinuria with normal values of both blood pressure and renal function. Immunology was negative, including ANA and anti-DNA with normal C3 and C4 levels. Renal biopsy showed a diffuse sweling of the basal glomerular membrane with spikes and no hypercelularity; on immunofluorescence, deposits of IgG, C3 and C1q were found. Electron microscopy was not available. As a diagnosis of IMG was made, no immunosuppressive treatment was iniciated. During 3 years, the patient kept a normal renal function with a proteinuria of 3 g/day. 2 years later, the patient was admitted at the emergency room with anasarca and progressive renal failure, reaching a serum creatinine value of 4 mg/dl and a proteinuria of 8 g/day. Severe lung and neurological manifestations were also present. ANA titres were 1/1280 and anti-DNA 1/128; serum values of C3 and C4 were 40 and 9 mg/dl respectively. A second renal biopsy was performed, this time showing, along with signs of membranous glomerulonephritis, capillary walls with a "wire-loop" appearance, hialine thrombi, necrotic areas and celular proliferation, all these signs being compatible with a diffuse proliferative glomerulonephritis. On immunofluorescence, IgG, IgA, IgM, C3 and C1q deposits were found. Electron microscopy showed electrondense subepithelial, subendothelial and mesangial deposits. Treatment was iniciated with methilprednisolone bolus at a high dosage, followed by oral prednisone and azathioprine. An spectacular improvement of the patient was observed, and extra-renal symptoms progressively remitted. Renal function evolutioned towards normality and proteinuria decreased. After 10 years of follow-up, serum creatinine value is 1.4 mg/dl and proteinuria 2 g/day.

A 64 years old female patient, with chronic renal failure (serum creatinine 2.2 mg/dl, creatinine clearance 43 ml/min), showed also a 4.5 g/day proteinuria and hypertension. Except ANA titres of 1/80, immunological study was uneventful. On renal biopsy (optic microcopy exclusively available) data of membranous glomerulonephritis with no hypercelularity was observed. On diagnosis of IMG, no specific treatment was given to the patient who developed, two years later, a process consisting on fever, arthralgias, arthritis and cutaneous lesions. Renal function was rapidly impaired, reaching a serum creatinine value of 4.2 mg/dl and a creatinine clearance of 22 ml/min. Simultaneously, 24 hours proteinuria reached 10 grams. Anti-DNA kept negative but ANA turned remarkably positive at 1/1280. C3 levels mantained on normal values, but C4 descended to 7 mg/dl. The rest of the immunological study was negative. Due to the uncommon evolution, a second biopsy was performed, with the following findings: homogeneous widening of the glomerular basement membrane with no hypercelularity on optic microscopy; deposits of IgG, C3 and C1q on immunofluorescence and electrondense subepithelial, subendothelial and mesangial deposits on electron microscopy. Changing the diagnosis to LMG, corticosteroids were given with initial improvement (creatinine 2.4 mg/dl and descending proteinuria), but the patient died 6 months later because of a stroke.

DISCUSSION

LMG may have a peculiar way of presentation, sometimes making it difficult to differentiate from IMG. In a patient with a previous diagnosis of SLE and renal implication, the histological finding of membranous glomerulonephritis is easy to relate with the first diagnosis, so being a LMG. When no other data of SLE are present, which is otherwise rather frequent in clinical practice, it may be a really difficult task to make the diagnosis of LMG: clinical, serological and pathological data are the clues.

Refering to renal disease, in "pure" LMG proteinuria is the most relevant finding; we detected it in 81% of the patients in our series, which is consistent with other series in the literature [4, 5, 6]. However, the most relevant datum is the lack of extra-renal clinics on set-up, which is also consistent with other series [6, 7, 8, 9]. In fact, 7 out of 16 patients in our study (43%), i.e., almost half of our patients, showed no extra-renal clinics suggestive of SLE when renal biopsy was performed. Therefore, we conclude that LMG can present with poor clinics, making it difficult to detect a latent SLE; the last process should later show up, sometimes violently, as in the clinical reports showed above.

Toghether with lack of clinics, poor serological data are also characteristic of this type of lupus nephropathy. 14 patients (7%) were initially positive for ANA, but most of them at a low titre and ANA were negative in 2 cases (12.5%). Moreover, anti-DNA were positive only in 3 patients (18%), and hypocomplementemia was found just in 2 cases (12.5%). Poor serological manifestations have also been described by other authors [6, 7], while no detected by others. In this way, refering to positivity for ANA in LMN, Jennette [8] finds it in 100% of his patients, Donadio [6] in 78% and Appel [4] in 50%. However, no specific reference is made to this positivity, and all four types of LMG classically described by the WHO are included. Pascuali [10], who does make a difference between associated or not prolifferative lesions, describes, unlike we do, positive ANA at moderate or high titres in all his patients with "pure" LMG. On the other hand, it is relevant in the same series the high frequency of hypocomplementemia when comparing to ours. Jennete [8] describes hypocomplementemia in 65% of his patientes, Appel [4] in 75% and Donadio [6] in most of his patients. Perhaps frequent hypocomplementemia is due again to the fact that pathologically heterogeneous patients are included, i.e. with and without associated proliferative lesions. This explanation is based upon Pascuali`s observation; he classifys types Va+Vb versus Vc+Vd when driving out conclusions and finds hypocomplementemia only in 11.5% of the patients in the first group, so with similar results to ours.

As a conclusion made out of above, LMG may be found with no clinical data of LES and even with negativity for the classical serological markers. This makes it imperative to consider some pathological data to distinguish LMG from IMG. Jennette [8] separately analyzes some of these data. Consistent with our findings, it is remarkable the importance that he gives to C1q deposits in immunofluorescence and to the presence of mesangial and subendothelial electrondense deposits in electron microscopy. Other authors have also described mesangial and subendothelial electrondense deposits as more frequent in LMG and seldom found in non lupus membranous glomerulonephritis [4, 6, 11, 12, 13]. In this way, 10 out of 12 patients (83%) with available immunofluorescence in our series had C1q deposits and in every 4 (100%) with electron microscopy, there were mesangial and subendothelial electrondense deposits, adding to the expected subepithelial ones.

We still do not know the natural history of LMG because the experience driven from other types of lupus nephropathy makes early specific treatment been iniciated. However, there is an important issue to be analyzed consisting on some clinical data with relevant prognostic value according to renal function evolution. In our experience, age, gender, initial proteinuria, hypertension and treatment protocols do not behave as prognostic markers; in other words, they are similar between both groups of patients, those keeping a normal renal function an those who progress to chronic renal insufficiency. Renal failure at set-up is neither a negative prognostic marker according to renal function evolution. Some authors consider initial renal failure as a negative prognostic marker, although refering to lupus nephropathy as a whole [14,15,16,17]. Sloan [18] makes a difference with treated LMG and finds a negative prognostic value in initial serum creatinine, but just in type Vd. Unlike the mentioned authors and consistent with us, Appel [19] found that the presence of initial renal insufficiency has not a negative prognostic value in patients with lupus nephropathy, although pointing out that patients included in his study could have had more potentially reversible lesions comparing with other series.

The only factor we have showed to have a important prognostic value is the remission of proteinuria towards a non nephrotic range. This event is clearly more frequent in patients keeping normal renal function during follow-up, being this difference statiscally significant. Prognostic value of proteinuria in lupus nephropathy was previously described by Appel [19]. In his study, nephrotic syndrome at set-up was found to render a higher probability of renal failure. However, this probability diminishes in those patients in whom the nephrotic syndrome dissapears. Wallace [14] and Ginzler [15] also show that severe proteinuria and/or nephrotic syndrome in lupus nephropathy render a worse prognosis. None the less, other authors have not obseved a worse evolution in those patients with lupus nephropathy who show a heavy proteinuria and/or a nephrotic syndrome [16, 20, 21].

LMG has showed to have a worse prognosis than expected is series with long term follow-up. In this way, Pollak [22] found that none of seven patients with LMG evolutioned towards renal failure or death, although follow-up was short, with a medium of 3.6 years. After 3.9 years of follow-up, Donadio [6] found 8 out of 28 (29%) patients who died or developed chronic renal failure. Appel, in his study from 1978 [4] describes this unfavourable evolution in exclusively 5 out of 10 patients, but after a longer follow-up period [19], the number of patients with a bad evolution raised up to 7. Baldwin [23] and Adler [24], also with longer follow-up periods of 5.5 and 6 years respectively, show an evolution towards death or chronic renal failure in almost half of the patients with LMG. Our experience with LMG is rather similar to that of the authors described above. With a medium follow-up period of 5 years, 8 out of 16 patients (50%) have either died or developed a chronic renal failure. It seems like renal survival in LMG is good on short term evolution getting worse with longer follow-up periods.

Our series is too smal and therapeuthic protocols too heterogeneous to drive out conclusions according to the most accurate treatment in this kind of lupus nephropathy. Prospective and multicentric studies are necessary to clarify this issue. However, we would like to point out the utility of ACEI in this entity. Antiproteinuric effect of ACEI has been described in different glomerular processes, but it has not been proved its usefulness in lupus nephropathy [25, 26]. None the less, our results with ACEI are rather satisfactory according to reduction of proteinuria, which is, as remarked above, clearly related to a better prognosis of renal function. Once corticosteroid treatment is initiated and no favourable response is detected, it would be advisable to associate ACEI before adding an immunosuppressive drug; some cases show such an evident response that immunosuppressive treatment can be totally avoided, so avoiding long-term adverse events of this therapy.

In conclusion, LMG can present without extra-renal symptoms of SLE and even with very poor expression of its serological markers. Almost half of the patients (42%) developed a chronic renal insufficiency. The evolution towards a non-nephrotic range proteinuria correlated with the maintenance of a normal renal function.

BIBLIOGRAPHY

DISCUSSION BOARD PANEL DE DISCUSION