British Joumal of Dermatology (1979) 100, 113.
Clinical and Laboratory Investigations
Comparison of chlorambucil, azathioprine or cyclophosphamide combined with corticosteroids in the treatment of lupus nephritis M.S.SABBOUR AND LAILA M.OSMAN Department of Medicine and Clinical Pathology, Ain-Shams University, Cairo, Egypt Accepted for publication 28 May 1978
SUMMARY
t63 patients with diffuse lupus glomerulonephritis, proven by renal biopsy, were divided into four therapeutic trial groups; 67 were put on corticosteroids alone, 11 on corticosteroids and azathioprine, 32 on corticosteroids and cyclophosphamide, and 53 on corticosteroids and chlorambucil and were followed up for several years. The addition of azathioprine to corticosteroids did not increase the survival rate, improve the renal function or alter the grim prognosis ofthe patients. Cyclophosphamide appeared to influence favourably the pathological lesion and the renal function when added to corticosteroids, and the disease progressed at a slower rate. The fatal side efFects nearly balanced the therapeutic value of cyclophosphamide. Patients on corticosteroids and chlorambucil had an excellent course. This therapeutic regimen resulted in resolution or regression ofthe renal pathology, marked improvement of the renal function and marked improvement of the survival rate. The authors believe that this therapeutic regimen holds the best chance of becoming the standard treatment for lupus nephritis, particularly since the side effects of chlorambucil were minimal.
Since renal involvement occurs in over three-quarters of patients with systemic lupus erythematosus (SLE) and markedly influences the morbidity and eventual outcome, the treatment of lupus nephropathy is important in the clinical management of patients with SLE. There is considerable evidence that the pathological processes in the nephropathy of SLE are mediated, at least in part, by immunological processes (Koffler et al., 1967; Krisman & Kaplan, 1967; Rothfield & Stollar, 1967). Entrapment of DNA-anti-DNA complexes in glomerular capillary walls is probably a primary event in the initiation ofthe glomerular lesion (Koffler et al., 1967)- It '^ unclear, however, what causes the abnormal immunological response, whether an abnormal antigenic stimulus or abnormal antibody response. The possibility has been raised that human SLE is caused by an immunological response to a viral agent (Christian & Phillips, 1973). Despite the uncertainty of our understanding of the specific immunological determinants that 0007-0963/79/0200-0113502.00 © 1979 British Association of Dermatologists 113
114
M.S.Sabbour and L.M.Osman
modulate the severity of the disease process in individual patients, there is general agreement that the course of lupus nephropathy is largely influenced by the type and severity of the underlying renal lesion. Lupus nephropathy has been classified into 3 varieties (Hecht et al., 1976). In Group I, Uupus glomerulitis' there is a mild proliferation of endothelial cells that is limited to a portion (local) of only some glomeruli (focal). Electron-dense material is not present on ultrastructural examination. Patients with this 'lupus glomerulitis' generally have a benign form of lupus nephropathy (PoIIak, Pirani & Schwartz, 1964; Comerford & Cohen, 1967; Zweiman et al., 1968; Baldwin et al., 1970). The lesion of Group tl 'membranous lupus nephropathy' is characterized by a lack of glomerular hypercellularity and irregular thickening of the glomerular basement membrane. The cpimcmbranous deposition of electron-dense material is indistinguishable from the change found in idiopathic membraneous nephropathy. This type of renal lesion is commonly associated with heavy proteinuria and the nephrotic syndrome but tends to be indolent. By life table analysis, Pollak & Pirani (1969) predicted a greater than 80",, survival 5 years after initial renal biopsy. The more common lesion found in patients with SLE is Group III, 'lupus glomerulonephritis', which is characterized by an irregular endothelial-cell proliferation and the deposition of electron-dense deposits in subendothelial, subepithelial and mesangial sites. By light microscopy, fibrinoid necrosis, exudation, hyaline thrombi and an interstitial infiammatory reaction are frequently present. At present, there is general agreement that patients with glomerulonephritis with 50"., or more of glomeruli affected have a poor outlook despite intensive therapy with corticosteroids (Pollak et aL, 1973). It has become apparent, therefore, that an evaluation ofthe course of patients with lupus nephropathy and the action of therapeutic modalities must take into account the underlying renal lesion. Despite considerable evidence that corticosteroids suppress the systemic manifestations of SLE, only a few reports have suggested that steroid treatment may influence the course of renal disease. Most reports attest to the unfavourable course of patients with moderate and severe glomerulonephritis when steroids are used alone, especially in patients with reduced renal function and/or the nephrotic syndrome. Moreover, because of the side effects of high-dose steroid therapy when employed for more than a few months, alternative forms of therapy have been sought. Immunosuppressive drugs have been used to treat systemic lupus and lupus nephritis for over 10 years. The addition of azathioprine (Drinkard et al., 1970; Mahcr & Schreiner, 1970; Shelp et ai, 1971; Sztejnbok et al., 1971; Hayslct et al., 1972; Cade et al., 1973) or cyclophosphamide (Hadidi, 1970; Cameron et ai, 1970; Steinberg et ai, 1971) to corticosteroids in the treatment of lupus nephritis has, in general, been associated with a decrease in proteinuria and lack of deterioration of renal function. The resuhs that have been reported to date do not, however, lead to a confident conclusion that any one form of therapy results in a clearly improved outlook for the patient with lupus glomerulonephritis (Fries, Sharp & McDevitt, 1973; Donadio, Holley & Wagoner, 1974; Decker et ai, 1975). In 1972, Amor et ai reported the use of chlorambucil for the unsuccessful treatment of 12 patients with SLE with a variety of manifestations. On the other hand, Snaith et al. (1973) reported very successful results in treating 6 patients with SLE with chlorambucil; in 5 the decision was taken after failure by corticosteroids to control progressive renal disease. After the introduction of chlorambucil renal function improved, repeat renal biopsy showed quantitative improvement and all patients survived a few years of follow-up, until the report was published. Encouraged by this report and greatly disappointed by the results of therapy of lupus nephritis by corticosteroids, cyclophosphamide, azathioprine, alone or in combination, we decided to put some of our newly diagnosed patients with lupus nephritis on a combined corticosteroid-chlorambucil regimen and to wait for at least 3 years of careful follow-up before reporting our results. The results of this new therapeutic trial were then compared with our experience in the treatment of lupus by either azathioprine or cyclophosphamide.
Treatment of lupus nephritis
.
115
MATERIALS AND METHODS
173 patients with lupus nephritis have been under our care for the last 17 years (between i960 and 1977). AH patients fulfilled the following criteria for diagnosis of lupus nephritis: (1) Clinical features compatible with SLE as defined by Dubois (1966). (2) Positive LE cells in blood on two or more occasions. (3) Positive anti-DNA antibodies in titres greater than 1/16 in patients tested (89 patients; the anti-DNA determinations were not available to us during the early years of the study). (4) Proteinuria more than 05 g/24 h and microscopic haematuria (schistosomiasis, renal stones and other causes of haematuria of course excluded). (5) Renal histology with glomerular abnormality consistent with lupus nephritis on light microscopic studies (all 173 patients) and with the electron microscope (107 patients). The age and sex distribution of the patients is shown in Table i. TABLE r . Age and s"ex distribution of patients Age (years)
0-9
10-19 20-29 30-39 40-49 5(^59 60-69
Total
No. of patients Female
Male
I 25
0
67 53 7
7
4 6
2 0
I 0 0
155
18
Percutaneous needle biopsy of the right kidney was performed using a Franklin-Silvcrman biopsy needle in all 173 patients. Serial biopsies were done in 83 patients; a total of 279 specimens were studied. The morphological types of lupus nephritis, based on the initial biopsy, are shown in Table 2. The majority of the patients (84 8'\,) had lupus glomerulonephritis and 10% of the patients showed a mixed, membrano-proliferative lesion. TABLl! 2. Morphological types of lupus nephritis
Morphological type
No. of patients
Lupus glomerulitis 7 (4%) Membranous lupus nephropathy 3 (l-2%) Lupus giomcruionephritis 146 (84 8%) Mild (less than 50";, of glomeruli affected) 23 Moderately severe (50-75';,, of glomeruli affected) 61 Severe (generalized, diffuse glomerular affection) 62 Mixed membranous and proliferative lesion 17 (10%)
Ii6
M.S.Sabbour and L.M.Osman
Patients with lupus glomcrulitis and with membranous lupus nephropathy were excluded from the clinical comparative therapeutic trial, since they tend to run a rather benign course. Subsequent discussion deals only with 163 patients with an initial histological proof of a renal pathology that most reports have indicated would be expected to progress to renal failure or hypertensive death and persistence of proteinuria. All patients had urinary abnormalities at the time of presentation (Table 3). Combined microscopic haematuria and proteinuria were the commonest urinary abnormality and their incidence and degree increased with the severity ofthe lesion. TABLE 3. Range of urinary abnormalities seen in patients
Morphological type Lupus glomerulonephritis Mild Moderate Severe Mixed membranoproliferative lesion
No. of patients
Hypertension
23 61 62
62
51 62
17
7
7
3
Proteinuria
Miscroscopic haematuria
Nephrotic syndrome
Nitrogen retention
5
18
48
ax
— 27
—
16 41
43
3 39
5
2
Hypertension: BP > 150/95 mmHg. Nephrotic syndrome; proteinuria ' -3 g/24 h and serum albumin - .3-0 g/loo ml. Nitrogen retention: Serum urea --60 mg/icxj ml and serum creatinine - 2 0 mg/ioo ml.
Four different parameters were chosen for evaluation ofthe efficacy of any therapeutic regimen. (1) Indices relating to renal function: included serum urea, serum creatinine, creatinine clearance, 24 h urinary protein excretion, serum albumin level and blood pressure. (2) Indices relating to immunological abnormality: included the LE cell test assessed semiquantitatively as —ve to + + +> titre of antinuclear antibody and level of serum gamma globulin. (3) Non-specific indices: included haemoglobin level, total and differential WBC and erythrocyte sedimentation rate (ESR). (4) Histopathological indices: from renal biopsies performed before and 6-12 months after start of treatment (and in a few patients a third or even a fourth renal biopsy has been studied), with detailed, semiquantitative analysis of the various components. All indices for groups 1-3 were repeated at specified intervals for 2 years: monthly for 6 months, then 4 monthly for 2 years, then 6 monthly afterwards (for those who remained alive). Mean values for each index were calculated for each time-point, and comparisons were made with means of pretreatment values for each index; differences were tested for significance by Student's V'-test. The duration of the study extended from entry until the date of death, until the date of commencing haemodialysis therapy, or in the case of survivors until 6 March 1977. Therapeutic programmes given to the patients were as follows: Programme I 'corticosieroids alone' {6j patients)
Fifty-seven patients were put on prednisone or prednisolone 60 mg daily for 6 weeks then the dose was slowly tapered off to the lowest dose that controlled the symptoms. The maintenance dose varied from 5 to 20 mg per day. Seven patients were put on betamethasone, dosage schedule being i/io
Treatment of lupus nephritis
117
that of prednisone; 3 patients were put on triamcinolone starting with 48 mg per day. Patients were kept on the maintenance dose for several years (so long as they survived), with increments of 5-15 mg per day when the disease showed clinical or laboratory evidence of activity. In 7 patients the initial dose was stepped up to 80 mg/day after 2 weeks, when the disease seemed very acute, and then gradually decreased as they improved. This regime was adopted in all the patients between i960 and 1968 (53 patients); corticosteroids constituted the sole therapy in only 14 patients with lupus nephritis within the last 9 years. Programme II 'corticosteroids^azathioprine'' (11 patients) These included 7 patients already on corticosteroids and 4 freshly diagnosed patients. We did not put any patient on azathioprine ourselves (since azathioprine was not available in Egypt until very recently). These patients were our own 7 patients on steroids who flew to Britain for consultation and returned after azathioprine was added to their therapeutic regimen and we continued their followup ; and 4 patients referred to us after being diagnosed and advised treatment in Europe or the United States. They received azathioprine in a dose of 1-2 mg/kg body weight. iOO
V.
c>—o
90
80 70
^
60
I 50 01
\v>.. \# \ ^ \\
1 40 J
T
30
^
20
V
^?\
10
-1—1 ^
4
6
1
1 8
1
i__ 10
Year-;
FIGURE I. Cumulative survival. Mean data for surviving patietits treated for lupus nephritis. (•—• • ) Corticosteroids only; ( • — • ) corticosteroids f azathioprine; ( T . . . . T) corticosteroids \ cyclophosphamide; ( — • — • ) corticosteroids 4- chlorartibucil.
Programme III'corticosteroids-\-cyclophosphamide^ (32 patients) Some patients already on corticosteroids before 1968 were given cyclophosphamide in addition to their maintenance dose of steroids (17 patients). The oral dose was 150 mg daily for i month, 100 mg daily for i month and 50 mg daily for 1-6 months. The remaining 15 patients, diagnosed after 1968, were given 60 mg prednisone-rcyclophosphamide from the start. In 4 patients cyclophosphamide was given intravenously 200 mg every other day for 10 doses, then maintenance was by
M.S.Sabbour and L.M.Osman TABLE 4. Survival rates of patients under different therapeutic regimes "„ Survival rate after Therapeutic regime Corticosteroids only Corticosteroids- azathioprine Corticosteroids -t cyclophosphamide Corticosteroid + chlorambucil
I year
2 years
3 years
4 years
8r
51 64 80
36 52 73 96
27 31 67 96
81 90 IOO
IOO
an oral dose of 50 mg per day. The other 11 patients were given the previously described oral dose schedule. Programme IV ^corticosteroidsi chlorambucil' ( j j patients) Since 1973, most of our newly diagnosed patients with lupus nephritis have been started on 60 mg prednisolone and 10 mg chlorambucil for 2-4 weeks. The dose of the steroid was usually gradually reduced to 5-15 mg/day and chlorambucil to 2-5 mg/day and patients are kept on this maintenance dose. We also added chlorambucil to the treatment of some patients who were on steroids only. We did not institute it in place of azathioprine or cyclophosphamide if the patient had already been taking either of them. The initial mean serum urea and creatinine clearance were comparable in all four groups. There were no significant dififerences in severity of the disease process histologically among the patients included in the four programmes. RESULTS
Cumulative survival data Using the method of Cutler & Ederer (1958) we have calculated the survivors in each therapeutic trial group following the establishment of the diagnosis by renal biopsy and the institution of treatment (Fig. i). The addition of azathioprine to corticosteroids did not improve the survival rate, while the addition of cyclophosphamide and especially chlorambucil resulted in a definite improvement of the survival rate. Table 4 compares the i-, 2-, 3- and 4-year survival rates in patients under different therapeutic regimens. Among the corticosteroid-chlorambucil-treated group only one patient died during the whole follow-up period. She was a 25-year-old girl who had cerebral as well as renal lupus involvement and she died in status epilepticus. Causes of death Of our original 163 patients 95 have died. The causes of death were analysed and were grouped as follows: (1) Deaths due to progressive renal disease. (2) Non-renal lupus deaths. (3) Deaths due to other causes. Table 5 summarizes these results. Examples of 'lupus non-renal deaths' are: status epilepticus due to systemic lupus with cerebral
Treatment of lupus nephritis
119
TABLE 5. Analysis of causes of death of patients lander different therapeutic regimes Renal death
Therapeutic regime Corticosteroids only (67 patients) Corticosteroids (azathioprine (11 patients) Corticosteroids -f- cyclophosphamide (32 patients) Corticosteroids -)- chlorambucil (53 patienrs) Total
Non-renal lupus death
Other deaths
Total
43
9
7
59
5
2
3
zo
15
3
7
as
0
X
.0'
t
63
15
m
involvement (6 patients), massive cerebral thrombosis (2 patients), cerebral haemorrhage (i patient, she was also hypertensive), multiple pulmonary infarctions (possibly due to local pulmonary vaseulitis—3 patients), acute myocarditis and acute congestive cardiac failure (3 patients, all were hypertensive). Under 'other deaths' arc included conditions as varied as bone marrow aplasia (3 patients), agranulocytosis (i patient), generalized bleeding tendency (i patient), severe haemorrhagic gastroenteritis (i patient), haemorrhagic cystitis (i patient), septicaemia (4 patients), fulminating pulmonary infection {3 patients), acute psychosis (2 patients), fulminating hepatitis (i patient). All other 'renal deaths' were patients with nitrogen retention. Most of them were hypertensive and nephrotic. Most of them had one or several episodes of urinary tract infection, in addition to their iupus nephritis, during the course of their illness. Progress of the renal disease under therapy
The means and standard deviations for serum urea, serum creatinine and creatinine clearance in all four groups showed no significant difference at the initial assessment. However, when data for surviving patients were analysed at specific points, differences could easily be shown. There was no significant difference between patients treated with corticosteroids only and patients treated with a combination of corticostroids and azathioprine at any assessment point during the 24 months (Table 6). TABLE 6. Serum urea (mg/roo ml) in patients on corticosteroids only and in patients on steroids + azathioprine CorticosteroJds only Before treatment After 6 months After 12 months After 24 months
8s 121 163 237
(±23-1) (±19-2) (±17-3) (±12)
Corticosteroids -}- azathioprine 81 117 154 219
(J:lI-7) (±9-3) (±5-7) (only one survivor)
However, patients on cyclophosphamide together with corticosteroids showed a significant difference from those on corticosteroids alone at 12 and 24 months (P-^o-os), and patients on corticosteroids and chlorambucil showed a very significant difference {P
Clinical and Laboratory Investigations
Comparison of chlorambucil, azathioprine or cyclophosphamide combined with corticosteroids in the treatment of lupus nephritis M.S.SABBOUR AND LAILA M.OSMAN Department of Medicine and Clinical Pathology, Ain-Shams University, Cairo, Egypt Accepted for publication 28 May 1978
SUMMARY
t63 patients with diffuse lupus glomerulonephritis, proven by renal biopsy, were divided into four therapeutic trial groups; 67 were put on corticosteroids alone, 11 on corticosteroids and azathioprine, 32 on corticosteroids and cyclophosphamide, and 53 on corticosteroids and chlorambucil and were followed up for several years. The addition of azathioprine to corticosteroids did not increase the survival rate, improve the renal function or alter the grim prognosis ofthe patients. Cyclophosphamide appeared to influence favourably the pathological lesion and the renal function when added to corticosteroids, and the disease progressed at a slower rate. The fatal side efFects nearly balanced the therapeutic value of cyclophosphamide. Patients on corticosteroids and chlorambucil had an excellent course. This therapeutic regimen resulted in resolution or regression ofthe renal pathology, marked improvement of the renal function and marked improvement of the survival rate. The authors believe that this therapeutic regimen holds the best chance of becoming the standard treatment for lupus nephritis, particularly since the side effects of chlorambucil were minimal.
Since renal involvement occurs in over three-quarters of patients with systemic lupus erythematosus (SLE) and markedly influences the morbidity and eventual outcome, the treatment of lupus nephropathy is important in the clinical management of patients with SLE. There is considerable evidence that the pathological processes in the nephropathy of SLE are mediated, at least in part, by immunological processes (Koffler et al., 1967; Krisman & Kaplan, 1967; Rothfield & Stollar, 1967). Entrapment of DNA-anti-DNA complexes in glomerular capillary walls is probably a primary event in the initiation ofthe glomerular lesion (Koffler et al., 1967)- It '^ unclear, however, what causes the abnormal immunological response, whether an abnormal antigenic stimulus or abnormal antibody response. The possibility has been raised that human SLE is caused by an immunological response to a viral agent (Christian & Phillips, 1973). Despite the uncertainty of our understanding of the specific immunological determinants that 0007-0963/79/0200-0113502.00 © 1979 British Association of Dermatologists 113
114
M.S.Sabbour and L.M.Osman
modulate the severity of the disease process in individual patients, there is general agreement that the course of lupus nephropathy is largely influenced by the type and severity of the underlying renal lesion. Lupus nephropathy has been classified into 3 varieties (Hecht et al., 1976). In Group I, Uupus glomerulitis' there is a mild proliferation of endothelial cells that is limited to a portion (local) of only some glomeruli (focal). Electron-dense material is not present on ultrastructural examination. Patients with this 'lupus glomerulitis' generally have a benign form of lupus nephropathy (PoIIak, Pirani & Schwartz, 1964; Comerford & Cohen, 1967; Zweiman et al., 1968; Baldwin et al., 1970). The lesion of Group tl 'membranous lupus nephropathy' is characterized by a lack of glomerular hypercellularity and irregular thickening of the glomerular basement membrane. The cpimcmbranous deposition of electron-dense material is indistinguishable from the change found in idiopathic membraneous nephropathy. This type of renal lesion is commonly associated with heavy proteinuria and the nephrotic syndrome but tends to be indolent. By life table analysis, Pollak & Pirani (1969) predicted a greater than 80",, survival 5 years after initial renal biopsy. The more common lesion found in patients with SLE is Group III, 'lupus glomerulonephritis', which is characterized by an irregular endothelial-cell proliferation and the deposition of electron-dense deposits in subendothelial, subepithelial and mesangial sites. By light microscopy, fibrinoid necrosis, exudation, hyaline thrombi and an interstitial infiammatory reaction are frequently present. At present, there is general agreement that patients with glomerulonephritis with 50"., or more of glomeruli affected have a poor outlook despite intensive therapy with corticosteroids (Pollak et aL, 1973). It has become apparent, therefore, that an evaluation ofthe course of patients with lupus nephropathy and the action of therapeutic modalities must take into account the underlying renal lesion. Despite considerable evidence that corticosteroids suppress the systemic manifestations of SLE, only a few reports have suggested that steroid treatment may influence the course of renal disease. Most reports attest to the unfavourable course of patients with moderate and severe glomerulonephritis when steroids are used alone, especially in patients with reduced renal function and/or the nephrotic syndrome. Moreover, because of the side effects of high-dose steroid therapy when employed for more than a few months, alternative forms of therapy have been sought. Immunosuppressive drugs have been used to treat systemic lupus and lupus nephritis for over 10 years. The addition of azathioprine (Drinkard et al., 1970; Mahcr & Schreiner, 1970; Shelp et ai, 1971; Sztejnbok et al., 1971; Hayslct et al., 1972; Cade et al., 1973) or cyclophosphamide (Hadidi, 1970; Cameron et ai, 1970; Steinberg et ai, 1971) to corticosteroids in the treatment of lupus nephritis has, in general, been associated with a decrease in proteinuria and lack of deterioration of renal function. The resuhs that have been reported to date do not, however, lead to a confident conclusion that any one form of therapy results in a clearly improved outlook for the patient with lupus glomerulonephritis (Fries, Sharp & McDevitt, 1973; Donadio, Holley & Wagoner, 1974; Decker et ai, 1975). In 1972, Amor et ai reported the use of chlorambucil for the unsuccessful treatment of 12 patients with SLE with a variety of manifestations. On the other hand, Snaith et al. (1973) reported very successful results in treating 6 patients with SLE with chlorambucil; in 5 the decision was taken after failure by corticosteroids to control progressive renal disease. After the introduction of chlorambucil renal function improved, repeat renal biopsy showed quantitative improvement and all patients survived a few years of follow-up, until the report was published. Encouraged by this report and greatly disappointed by the results of therapy of lupus nephritis by corticosteroids, cyclophosphamide, azathioprine, alone or in combination, we decided to put some of our newly diagnosed patients with lupus nephritis on a combined corticosteroid-chlorambucil regimen and to wait for at least 3 years of careful follow-up before reporting our results. The results of this new therapeutic trial were then compared with our experience in the treatment of lupus by either azathioprine or cyclophosphamide.
Treatment of lupus nephritis
.
115
MATERIALS AND METHODS
173 patients with lupus nephritis have been under our care for the last 17 years (between i960 and 1977). AH patients fulfilled the following criteria for diagnosis of lupus nephritis: (1) Clinical features compatible with SLE as defined by Dubois (1966). (2) Positive LE cells in blood on two or more occasions. (3) Positive anti-DNA antibodies in titres greater than 1/16 in patients tested (89 patients; the anti-DNA determinations were not available to us during the early years of the study). (4) Proteinuria more than 05 g/24 h and microscopic haematuria (schistosomiasis, renal stones and other causes of haematuria of course excluded). (5) Renal histology with glomerular abnormality consistent with lupus nephritis on light microscopic studies (all 173 patients) and with the electron microscope (107 patients). The age and sex distribution of the patients is shown in Table i. TABLE r . Age and s"ex distribution of patients Age (years)
0-9
10-19 20-29 30-39 40-49 5(^59 60-69
Total
No. of patients Female
Male
I 25
0
67 53 7
7
4 6
2 0
I 0 0
155
18
Percutaneous needle biopsy of the right kidney was performed using a Franklin-Silvcrman biopsy needle in all 173 patients. Serial biopsies were done in 83 patients; a total of 279 specimens were studied. The morphological types of lupus nephritis, based on the initial biopsy, are shown in Table 2. The majority of the patients (84 8'\,) had lupus glomerulonephritis and 10% of the patients showed a mixed, membrano-proliferative lesion. TABLl! 2. Morphological types of lupus nephritis
Morphological type
No. of patients
Lupus glomerulitis 7 (4%) Membranous lupus nephropathy 3 (l-2%) Lupus giomcruionephritis 146 (84 8%) Mild (less than 50";, of glomeruli affected) 23 Moderately severe (50-75';,, of glomeruli affected) 61 Severe (generalized, diffuse glomerular affection) 62 Mixed membranous and proliferative lesion 17 (10%)
Ii6
M.S.Sabbour and L.M.Osman
Patients with lupus glomcrulitis and with membranous lupus nephropathy were excluded from the clinical comparative therapeutic trial, since they tend to run a rather benign course. Subsequent discussion deals only with 163 patients with an initial histological proof of a renal pathology that most reports have indicated would be expected to progress to renal failure or hypertensive death and persistence of proteinuria. All patients had urinary abnormalities at the time of presentation (Table 3). Combined microscopic haematuria and proteinuria were the commonest urinary abnormality and their incidence and degree increased with the severity ofthe lesion. TABLE 3. Range of urinary abnormalities seen in patients
Morphological type Lupus glomerulonephritis Mild Moderate Severe Mixed membranoproliferative lesion
No. of patients
Hypertension
23 61 62
62
51 62
17
7
7
3
Proteinuria
Miscroscopic haematuria
Nephrotic syndrome
Nitrogen retention
5
18
48
ax
— 27
—
16 41
43
3 39
5
2
Hypertension: BP > 150/95 mmHg. Nephrotic syndrome; proteinuria ' -3 g/24 h and serum albumin - .3-0 g/loo ml. Nitrogen retention: Serum urea --60 mg/icxj ml and serum creatinine - 2 0 mg/ioo ml.
Four different parameters were chosen for evaluation ofthe efficacy of any therapeutic regimen. (1) Indices relating to renal function: included serum urea, serum creatinine, creatinine clearance, 24 h urinary protein excretion, serum albumin level and blood pressure. (2) Indices relating to immunological abnormality: included the LE cell test assessed semiquantitatively as —ve to + + +> titre of antinuclear antibody and level of serum gamma globulin. (3) Non-specific indices: included haemoglobin level, total and differential WBC and erythrocyte sedimentation rate (ESR). (4) Histopathological indices: from renal biopsies performed before and 6-12 months after start of treatment (and in a few patients a third or even a fourth renal biopsy has been studied), with detailed, semiquantitative analysis of the various components. All indices for groups 1-3 were repeated at specified intervals for 2 years: monthly for 6 months, then 4 monthly for 2 years, then 6 monthly afterwards (for those who remained alive). Mean values for each index were calculated for each time-point, and comparisons were made with means of pretreatment values for each index; differences were tested for significance by Student's V'-test. The duration of the study extended from entry until the date of death, until the date of commencing haemodialysis therapy, or in the case of survivors until 6 March 1977. Therapeutic programmes given to the patients were as follows: Programme I 'corticosieroids alone' {6j patients)
Fifty-seven patients were put on prednisone or prednisolone 60 mg daily for 6 weeks then the dose was slowly tapered off to the lowest dose that controlled the symptoms. The maintenance dose varied from 5 to 20 mg per day. Seven patients were put on betamethasone, dosage schedule being i/io
Treatment of lupus nephritis
117
that of prednisone; 3 patients were put on triamcinolone starting with 48 mg per day. Patients were kept on the maintenance dose for several years (so long as they survived), with increments of 5-15 mg per day when the disease showed clinical or laboratory evidence of activity. In 7 patients the initial dose was stepped up to 80 mg/day after 2 weeks, when the disease seemed very acute, and then gradually decreased as they improved. This regime was adopted in all the patients between i960 and 1968 (53 patients); corticosteroids constituted the sole therapy in only 14 patients with lupus nephritis within the last 9 years. Programme II 'corticosteroids^azathioprine'' (11 patients) These included 7 patients already on corticosteroids and 4 freshly diagnosed patients. We did not put any patient on azathioprine ourselves (since azathioprine was not available in Egypt until very recently). These patients were our own 7 patients on steroids who flew to Britain for consultation and returned after azathioprine was added to their therapeutic regimen and we continued their followup ; and 4 patients referred to us after being diagnosed and advised treatment in Europe or the United States. They received azathioprine in a dose of 1-2 mg/kg body weight. iOO
V.
c>—o
90
80 70
^
60
I 50 01
\v>.. \# \ ^ \\
1 40 J
T
30
^
20
V
^?\
10
-1—1 ^
4
6
1
1 8
1
i__ 10
Year-;
FIGURE I. Cumulative survival. Mean data for surviving patietits treated for lupus nephritis. (•—• • ) Corticosteroids only; ( • — • ) corticosteroids f azathioprine; ( T . . . . T) corticosteroids \ cyclophosphamide; ( — • — • ) corticosteroids 4- chlorartibucil.
Programme III'corticosteroids-\-cyclophosphamide^ (32 patients) Some patients already on corticosteroids before 1968 were given cyclophosphamide in addition to their maintenance dose of steroids (17 patients). The oral dose was 150 mg daily for i month, 100 mg daily for i month and 50 mg daily for 1-6 months. The remaining 15 patients, diagnosed after 1968, were given 60 mg prednisone-rcyclophosphamide from the start. In 4 patients cyclophosphamide was given intravenously 200 mg every other day for 10 doses, then maintenance was by
M.S.Sabbour and L.M.Osman TABLE 4. Survival rates of patients under different therapeutic regimes "„ Survival rate after Therapeutic regime Corticosteroids only Corticosteroids- azathioprine Corticosteroids -t cyclophosphamide Corticosteroid + chlorambucil
I year
2 years
3 years
4 years
8r
51 64 80
36 52 73 96
27 31 67 96
81 90 IOO
IOO
an oral dose of 50 mg per day. The other 11 patients were given the previously described oral dose schedule. Programme IV ^corticosteroidsi chlorambucil' ( j j patients) Since 1973, most of our newly diagnosed patients with lupus nephritis have been started on 60 mg prednisolone and 10 mg chlorambucil for 2-4 weeks. The dose of the steroid was usually gradually reduced to 5-15 mg/day and chlorambucil to 2-5 mg/day and patients are kept on this maintenance dose. We also added chlorambucil to the treatment of some patients who were on steroids only. We did not institute it in place of azathioprine or cyclophosphamide if the patient had already been taking either of them. The initial mean serum urea and creatinine clearance were comparable in all four groups. There were no significant dififerences in severity of the disease process histologically among the patients included in the four programmes. RESULTS
Cumulative survival data Using the method of Cutler & Ederer (1958) we have calculated the survivors in each therapeutic trial group following the establishment of the diagnosis by renal biopsy and the institution of treatment (Fig. i). The addition of azathioprine to corticosteroids did not improve the survival rate, while the addition of cyclophosphamide and especially chlorambucil resulted in a definite improvement of the survival rate. Table 4 compares the i-, 2-, 3- and 4-year survival rates in patients under different therapeutic regimens. Among the corticosteroid-chlorambucil-treated group only one patient died during the whole follow-up period. She was a 25-year-old girl who had cerebral as well as renal lupus involvement and she died in status epilepticus. Causes of death Of our original 163 patients 95 have died. The causes of death were analysed and were grouped as follows: (1) Deaths due to progressive renal disease. (2) Non-renal lupus deaths. (3) Deaths due to other causes. Table 5 summarizes these results. Examples of 'lupus non-renal deaths' are: status epilepticus due to systemic lupus with cerebral
Treatment of lupus nephritis
119
TABLE 5. Analysis of causes of death of patients lander different therapeutic regimes Renal death
Therapeutic regime Corticosteroids only (67 patients) Corticosteroids (azathioprine (11 patients) Corticosteroids -f- cyclophosphamide (32 patients) Corticosteroids -)- chlorambucil (53 patienrs) Total
Non-renal lupus death
Other deaths
Total
43
9
7
59
5
2
3
zo
15
3
7
as
0
X
.0'
t
63
15
m
involvement (6 patients), massive cerebral thrombosis (2 patients), cerebral haemorrhage (i patient, she was also hypertensive), multiple pulmonary infarctions (possibly due to local pulmonary vaseulitis—3 patients), acute myocarditis and acute congestive cardiac failure (3 patients, all were hypertensive). Under 'other deaths' arc included conditions as varied as bone marrow aplasia (3 patients), agranulocytosis (i patient), generalized bleeding tendency (i patient), severe haemorrhagic gastroenteritis (i patient), haemorrhagic cystitis (i patient), septicaemia (4 patients), fulminating pulmonary infection {3 patients), acute psychosis (2 patients), fulminating hepatitis (i patient). All other 'renal deaths' were patients with nitrogen retention. Most of them were hypertensive and nephrotic. Most of them had one or several episodes of urinary tract infection, in addition to their iupus nephritis, during the course of their illness. Progress of the renal disease under therapy
The means and standard deviations for serum urea, serum creatinine and creatinine clearance in all four groups showed no significant difference at the initial assessment. However, when data for surviving patients were analysed at specific points, differences could easily be shown. There was no significant difference between patients treated with corticosteroids only and patients treated with a combination of corticostroids and azathioprine at any assessment point during the 24 months (Table 6). TABLE 6. Serum urea (mg/roo ml) in patients on corticosteroids only and in patients on steroids + azathioprine CorticosteroJds only Before treatment After 6 months After 12 months After 24 months
8s 121 163 237
(±23-1) (±19-2) (±17-3) (±12)
Corticosteroids -}- azathioprine 81 117 154 219
(J:lI-7) (±9-3) (±5-7) (only one survivor)
However, patients on cyclophosphamide together with corticosteroids showed a significant difference from those on corticosteroids alone at 12 and 24 months (P-^o-os), and patients on corticosteroids and chlorambucil showed a very significant difference {P
