|
|
| |||||||||||||||||||||||||||||||
Background Among patients with the idiopathic nephrotic syndrome who have focal and segmental glomerulosclerosis and undergo renal transplantation, 15 to 55 percent have recurrent nephrotic syndrome. The recurrence may be caused by a plasma factor or factors that increase glomerular permeability, because plasma exchange transiently decreases or abolishes proteinuria in some patients. We studied the effect on proteinuria of the removal of protein (mostly immunoglobulins) by adsorption onto protein A from the plasma of patients with recurrent nephrotic syndrome.
Methods Eight patients were treated with one to three cycles of two to seven 1-day sessions of protein adsorption, and the patients' urinary protein excretion was measured repeatedly. Their immunosuppressive regimens were not changed during the treatment. The adsorbed proteins were eluted from the protein A and injected into rats, and the urinary albumin excretion of the rats was measured.
Results The protein-adsorption treatment consistently decreased urinary protein excretion by an average of 82 percent at the end of a cycle (P<0.001). In one patient proteinuria disappeared, and in another urinary protein excretion remained below 2.5 g per day with repeated cycles of protein adsorption. In all but one patient the effect of adsorption was limited in time, with a return to the preadsorption level of protein excretion within a maximum of two months. The administration to rats of material eluted from the protein A increased urinary albumin excretion 2.9- to 4.6-fold (P<0.001 and P = 0.005, respectively). Although protein A primarily binds immunoglobulins, the active fraction of the eluted proteins had a molecular weight below 100,000, indicating that immunoglobulin was not directly involved.
Conclusions Adsorption of plasma protein decreases urinary protein excretion in patients with recurrence of the nephrotic syndrome after renal transplantation. Studies of the adsorbed proteins should provide information about the mechanism of this disease.
We recently reported that ex vivo adsorption of plasma on protein A Sepharose cartridges decreased proteinuria in a patient who had had a relapse of idiopathic nephrotic syndrome after transplantation25. We have now confirmed the efficacy of this treatment in seven additional patients. In addition, we have demonstrated that material eluted from protein A alters glomerular permeability when injected into normal rats. In this report we provide preliminary data indicating that the active factor (or factors) is not an immunoglobulin.
Methods
Plasma Protein Adsorption
We used a plasma-separation device (plasma filter PF 2000, Gambro, Mechingen, Germany) in seven patients and centrifugation (BT798, Dideco, Mirandola, Italy) in one patient to deliver plasma to two adsorption cartridges (Immuno-absorba, Excorim, Lund, Sweden) at a maximal continuous-flow rate of 35 ml per minute. We used a different pair of cartridges for each session with each patient. The adsorption system (Citem 10, Excorim) monitored the plasma flow, the elution of bound proteins (using 0.13 M sodium citrate, pH 2.2), and the column-washing procedures. These cartridges contained 62.5 ml of protein A covalently linked to Sepharose, with an IgG-binding capacity of 20 to 25 mg per milliliter; they were pyrogen-free, preserved with 0.1 percent merthiolate, and stored at 4 °C26. One adsorption session (a one-day procedure) was used to treat 2.5 volumes of plasma (requiring roughly 2.5 to 4 hours, according to the weight of the patient). Human immunoglobulins (Veinoglobulin, Institut Pasteur Merieux, Lyon, France) were injected intravenously at the end of each cycle. Each cycle consisted of two to seven sessions performed within a 10-day period, because this regimen resulted in efficient depletion of immunoglobulin (>95 percent)27. Further cycles were performed in some patients to confirm the efficacy of the procedure for reducing proteinuria. In addition, three patients (Patients 1, 7, and 8) were treated with plasma exchange, in which 1 to 1.5 plasma volumes were replaced with saline solution supplemented with 4 percent albumin12. In each of these three patients, protein adsorption was performed only when the effect of plasma exchange was no longer evident.
Biochemical Monitoring
Serum concentrations of creatinine and albumin and urinary protein excretion (determined by a colorimetric assay using pyrogallol red) were measured daily during treatment. In addition, the plasma concentrations of complement components (C3 and C4), IgG, IgA, IgM (determined by immunonephelometric assays; BNA, Behring, Rueil Malmaison, France), and fibrinogen were measured and blood-coagulation tests were performed before, during, and at the end of each cycle. An example of the changes in serum albumin and immunoglobulin levels during a cycle of protein adsorption is shown in Figure 1.
|
The sodium citrate eluates (samples containing 2 to 5 µg of protein) of the protein A cartridges were analyzed by sodium dodecyl sulfate-polyacrylamide-gel electrophoresis28 on 5 to 15 percent polyacrylamide gradient gels under reducing (with 2-mercaptoethanol) and nonreducing conditions. The protein bands and molecular-weight standards were stained with silver salts or Coomassie blue.
Injection of Protein A Eluates into Rats
Normal male inbred Wistar rats (175 to 200 g; Charles River, Saint Aubin les Elbeux, France) were placed in metabolic cages, and 24-hour urine samples were collected for five days. After two days, the rats were injected with the eluates (from the samples obtained in the first two sessions) that had been dialyzed extensively against phosphate-buffered saline and concentrated (protein concentration, 7.5 mg per milliliter). Groups of four to six rats were injected both intravenously and intraperitoneally with 2 ml of concentrated eluate twice a day for three consecutive days. Control groups of rats were injected with phosphate-buffered saline, eluates of a plasma sample from a patient with membranous glomerulonephritis treated by protein adsorption (which had no effect on the patient's protein excretion), or eluates of pooled plasma samples from normal subjects prepared in the same way as the samples from the patients. Urinary excretion of total protein and albumin were measured by colorimetry (with pyrogallol red) and a specific rat albumin enzyme-linked immunosorbent assay (ELISA), respectively. The eluates of samples from three patients (Patients 1, 2, and 8) of the five patients tested (Patients 1, 2, 4, 7, and 8) yielded a reproducible effect; only the eluates associated with the greatest excretion of albumin (those from Patients 1 and 2) were used in the subsequent experiments.
Rat Albumin ELISA
Immunoplates containing 96 wells (Maxisorp, Nunc, Strasbourg, France) were coated for two hours at 37 °C with urine samples (dilution, 1:100 or 1:500) in phosphate-buffered saline. Purified rat albumin (Sigma, Vertilliere, France) was used as a standard. The wells were washed extensively with phosphate-buffered saline containing 0.5 percent gelatin and 0.1 percent Tween, followed by coating with the same solution for two hours at 37 °C. Bound albumin was detected by incubation with sheep anti-rat albumin antibody conjugated with horseradish peroxidase (Immunotech, Marseilles, France) for two hours at 37 °C. After washing, the enzyme reaction was activated with 2,2'-azino-di-(3-ethyl-benzothiazolin sulfona 6) (ABTS; Boehringer, Mannheim, Germany), and the plates were analyzed at 405 nm.
Estimation of Molecular Weight of Factors Conducive to Proteinuria
The protein A eluates from Patients 1 and 2 were partially purified by two methods. First, the eluates were passed through a membrane filter with a molecular-weight cutoff of 100,000 (YM 100, Ultramembran, Amicon, Epernon, France). Second, the eluates from Patient 2 were also analyzed by ultracentrifugation as described elsewhere29. The resulting fractions were subjected to dialysis against phosphate-buffered saline and concentrated. The filtered material and ultracentrifuge fractions with a molecular weight less than 150 kd contained no immunoglobulin as determined by analysis of immunoglobulin subclasses and sodium dodecyl sulfate-gel electrophoresis.
Statistical Analysis
The results of the studies of the effect of eluates in rats are expressed as the mean (±SD) albumin excretion of groups of four to six rats during the three-day injection period. The results were analyzed for statistical significance by the two-tailed Student's t-test.
Case Reports
The eight study patients (Table 1) had histologically proved focal and segmental glomerulosclerosis (two patients) or steroid-resistant minimal-change glomerulonephritis (six patients), and all had recurrences of heavy proteinuria within one week after transplantation. The mean (±SD) urinary protein excretion was 10.0 ±5.7 g (range, 2.6 to 20.6) just before the first protein-adsorption cycle. Transplant-biopsy specimens showed minimal-change glomerulonephritis and a typical pattern of foot-process fusion on electron-microscopical examination in five of the eight patients. The study and treatment were approved by the ethics committee of the local university hospital, and all patients gave informed consent.
|
Patient 1 was a 37-year-old woman who presented with idiopathic focal and segmental glomerulosclerosis in 1984, began long-term hemodialysis in March 1989, and received a cadaveric kidney graft in November 1989. (This patient has been described previously25.) She had proteinuria two days after transplantation, and her protein excretion was 6.0 g per day on day 24, at which time her serum creatinine concentration was 1.7 mg per deciliter (150 µmol per liter). A kidney biopsy on day 24 showed only minimal-change glomerulonephritis, with epithelial foot-process fusion on electron microscopy. Six plasma exchanges in 12 days, starting on day 25, resulted in a decrease in protein excretion to 2.3 g per day, but this value increased thereafter, reaching 3.3 g per day on day 70. A second biopsy (on day 60) revealed grade 1 cellular rejection, which was treated successfully with methylprednisolone, with no effect on the proteinuria. Protein A adsorption (six sessions in 10 days) was initiated on day 84 and resulted in a decrease in urinary protein excretion from 3.3 to 0.1 g per day after the third session (Figure 2), but it then increased to 2.6 g per day on day 140. A third kidney biopsy (on day 154) again showed minimal-change glomerulonephritis. A second short cycle of protein adsorption resulted in an almost complete disappearance of proteinuria (urinary protein, <0.3 g per day). Subsequently, this patient was treated again with a single session every three weeks, the approximate interval required for protein excretion to increase to a level above 2.5 g per day. These repeated single sessions were associated with a transient decrease in protein excretion from 3.2 ±0.9 to 1.1 ±0.4 g per day, with a return to the values before the session within three weeks. At 25 months after transplantation, the patient's serum creatinine concentration was 1.4 mg per deciliter (123 µmol per liter) and the urinary protein excretion less than 2.5 g per day.
|
Patient 2 was a 34-year-old man who presented with minimal-change glomerulonephritis in 1978 and had end-stage renal failure two years later. He received a cadaveric kidney graft in 1980 and had an immediate recurrence of proteinuria that led to long-term dialysis five months later. In 1990 he received a second graft. He had proteinuria during the first week after transplantation; protein excretion reached 2.6 g per day on day 42, with a serum creatinine concentration of 3.9 mg per deciliter (350 µmol per liter). Kidney-graft biopsies on days 7 and 36 showed minimal-change glomerulonephritis. Protein adsorption, started on day 53 (with five sessions in eight days), resulted in a decrease in protein excretion from 2.5 to 0.8 g per day and in the serum creatinine concentration to 2.6 mg per deciliter (230 µmol per liter) (Figure 2). Protein excretion then increased within 10 days and reached 4.7 g per day on day 70. A second cycle of protein adsorption, begun on day 73 and again involving five sessions in eight days, led to a decrease in protein excretion to 0.9 g per day. On day 98, the serum creatinine concentration increased and a kidney biopsy showed grade 1 cellular rejection. The patient responded to methylprednisolone with a decrease in serum creatinine, but protein excretion increased from 2.5 to 5.0 g per day. A fourth biopsy (on day 224) again showed minimal-change glomerulonephritis, and a third cycle of protein adsorption resulted in another reduction in protein excretion, to 1.0 g per day. This patient could not receive long-term treatment, and by day 450 after transplantation he had the nephrotic syndrome (10.0 g per day) and poor renal function.
Patient 3
Patient 3 was a 60-year-old man with minimal-change glomerulonephritis that had been diagnosed by biopsy in 1982. Long-term dialysis was started 3 years later, and he received a cadaveric kidney graft 11 months thereafter. He had an immediate recurrence of proteinuria; a biopsy revealed minimal-change glomerulonephritis. His graft was removed in 1989, and a second graft was implanted in 1990. He had proteinuria immediately after transplantation; the urinary protein excretion was 14.0 g per day on day 84 (Figure 2). At this time, a biopsy showed minimal-change glomerulonephritis with vascular lesions thought to be related to preexisting donor lesions. He was treated with protein adsorption starting on day 91 (five sessions in eight days), with a decrease in protein excretion from 14.0 to 5.7 g per day. Another cycle started on day 131 had similar results. A second biopsy on day 194 showed grade 1 cellular rejection, with focal and segmental glomerulosclerosis and chronic vascular lesions. The graft was lost 15 months after transplantation from a recurrence of focal and segmental glomerulosclerosis and chronic rejection.
Patient 4
Patient 4 was a 32-year-old woman with minimal-change glomerulonephritis proved by biopsy in 1982. Her proteinuria improved transiently with corticosteroid therapy, but hemodialysis was required in 1985. She received a cadaveric kidney graft in March 1987 but had rapid graft failure caused by steroid-resistant rejection. A second cadaveric graft was implanted in 1991. Renal function improved, but her urinary protein excretion increased immediately, from 6.0 to 11.0 g per day during the first two weeks after transplantation. A kidney biopsy on day 12 showed minimal-change glomerulonephritis with epithelial foot-process fusion. On day 42 she had severe acute rejection resistant to methylprednisolone, but she was treated with rabbit antithymocyte globulin (Merieux, Lyons, France) for 10 days with a decrease in the serum creatinine concentration but no change in protein excretion. A first cycle of protein adsorption (five sessions in nine days) started on day 112 (Figure 2) resulted in a decrease in protein excretion from 6.6 to 2.2 g per day, with a subsequent increase in two weeks. After a second cycle starting on day 139, protein excretion decreased from 5.4 to 0.4 g per day. Two additional sessions on days 168 and 172 resulted in a decrease in protein excretion from 3.0 to 0.4 g per day. On day 357, the patient still had no proteinuria (protein excretion, <0.1 g per day), and her serum creatinine concentration was 1.1 mg per deciliter (100 µmol per liter).
Patient 5
Patient 5 was a 27-year-old woman with biopsy-proved focal and segmental glomerulosclerosis who had a nephrotic syndrome resistant to therapy with corticosteroids and cyclosporine. End-stage renal failure developed that required hemodialysis in 1989. She received a cadaveric kidney graft in 1991. One week later, her urinary protein excretion was 9.2 g per day, and it reached 20.6 g per day on day 18. A biopsy of the transplanted kidney showed minimal-change glomerulonephritis. Protein adsorption was started on day 18, with one session every day for five days, resulting in decreases in protein excretion (to 4.5 g per day after the fifth session) and in the serum creatinine concentration (from 3.2 mg per deciliter [286 µmol per liter] to 1.1 mg per deciliter [95 µmol per liter]) (Figure 2). However, protein excretion increased to 9.3 g per day shortly thereafter. Three months after transplantation, her renal function was stable and urinary protein excretion was 15.5 g per day. Another cycle of protein adsorption (four sessions in four days) resulted in a decrease in protein excretion to 2.7 g per day. This value increased to 9.0 g per day within two weeks and remained elevated for six months thereafter.
Patient 6
Patient 6 was a 16-year-old girl with a steroid-resistant nephrotic syndrome who had minimal-change nephropathy on the initial renal biopsy but had end-stage renal failure within two years. She received a cadaveric kidney graft after four years of hemodialysis. By day 10 after transplantation, her urinary protein excretion was 11.0 g per day (Figure 2). Five weeks after transplantation, severe cytomegalovirus pneumonitis developed. At three months a renal biopsy showed focal and segmental glomerulosclerosis. Protein adsorption (three sessions on alternate days) resulted in a decrease in urinary protein excretion from 8.5 to 1.2 g per day, but two weeks later urinary protein excretion was 10.0 g per day.
Patient 7
Patient 7 was a 23-year-old man who had an idiopathic nephrotic syndrome resistant to corticosteroids and cyclosporine. After 7 years, hemodialysis was required; he received a cadaveric kidney graft 18 months later. He had elevated protein excretion immediately after transplantation (10.0 g per day) that decreased slowly to 1.0 g per day on day 45 but then increased to 37.0 g per day on day 55. A biopsy at this time showed minimal-change glomerulonephritis. On day 61 plasma exchange was started, with eight exchanges in 22 days; urinary protein excretion decreased slightly, and the serum creatinine concentration remained stable (1.9 mg per deciliter [165 µmol per liter]) until 18 months after grafting. Protein adsorption (four sessions in seven days), started after a second biopsy showed minimal-change glomerulonephritis, resulted in a decrease in urinary protein excretion from 9.0 to 1.0 g per day, after which this value increased (Figure 2).
Patient 8
Patient 8 was a 26-year-old woman with idiopathic nephrotic syndrome who required hemodialysis five years after diagnosis. Eleven months later she received a cadaveric kidney graft, but it was lost within five years because of a recurrence of the initial disease and chronic rejection. She then received a second graft. On day 15 her serum creatinine concentration was 1.1 mg per deciliter (97 µmol per liter). Her urinary protein excretion was 1.8 g per day on day 48 and increased progressively to 5.0 g per day by day 190. Two renal-graft biopsies (on days 135 and 480) showed minimal-change glomerulonephritis with foot-process cell fusion. Starting on day 482, she underwent one cycle of three sessions of protein adsorption, which led to a decrease in protein excretion from 3.5 to 0.2 g per day (Figure 2), while the serum creatinine concentration remained stable at 1.1 mg per deciliter. Her urinary protein excretion increased thereafter, reaching the pretreatment level by day 550. It then remained stable until day 632, when three plasma exchanges were carried out over a 5-day period, resulting in a decrease in protein excretion from 3.2 to 0.8 g per day that lasted for 40 days.
Results
Changes in Urinary Protein Excretion during Protein-Adsorption Treatment
The mean changes in urinary protein excretion in all eight patients during and after the 15 cycles of protein-adsorption treatment are shown in Figure 3. The mean (±SD) excretion decreased by 82 ±11 percent at the end of the cycle, after which it increased. The three patients who were treated with plasma exchanges had decreases of 76, 62, and 13 percent in protein excretion (mean, 49 percent) after the exchanges.
|
The electrophoretic patterns of the protein A Sepharose eluates of plasma from the eight patients after protein-adsorption treatment, as well as those from a control patient with membranous glomerulonephritis and from normal subjects, are shown in Figure 4. Under nonreducing conditions, most of the material migrated with proteins of molecular weight greater than 100,000, a pattern compatible with immunoglobulin. Under reducing conditions, there were broad and intense bands around 25, 50, and 75 kd, corresponding to the light and heavy chains of IgG, IgA, and IgM, as confirmed by immunoblotting (data not shown). A few additional bands were detected, but most of them were also present in the control eluates, and none of these additional bands were found consistently in the eluates of plasma from the eight patients.
|
Urinary protein excretion, as measured by the pyrogallol-red method, did not differ between rats injected with control eluates and rats injected with eluates derived from patients, and electron microscopy showed only inconsistent foot-process cell fusion. All subsequent experiments were performed with the ELISA specific for rat albumin. The urinary albumin excretion in rats injected only with phosphate-buffered saline ranged from 20 to 482 µg per day (base-line value before injection, <100 µg per day). Injections of eluates derived from Patients 1 and 2 resulted in significant increases in albumin excretion (Table 2). The injection of eluates obtained from a patient with membranous glomerulonephritis and from normal subjects did not result in substantial albuminuria as compared with the injection of phosphate-buffered saline.
|
100,000) fractions from Patient 2 caused a borderline increase in albumin excretion (P = 0.04). Three pools of differing molecular weights, prepared by ultracentrifugation of the eluate from Patient 2, were also tested for their ability to induce albuminuria (Table 2). Only rats injected with fractions corresponding to a molecular weight of less than 150,000 (a protein concentration of 0.5 mg per milliliter) had significantly higher urinary albumin excretion than rats injected with similarly prepared fractions from a patient with membranous glomerulonephritis or rats injected with phosphate-buffered saline. Discussion
Idiopathic nephrotic syndrome is frequently characterized by recurrences5. Some therapeutic regimens, including high doses of cyclosporine,31 angiotensin-converting-enzyme inhibitors,32 or nonsteroidal antiinflammatory drugs,33 have been reported to decrease proteinuria in some patients. Because the effect of plasma exchanges14,15,16,17,18,19,20 suggested that the proteinuria could be caused by a circulating factor or factors, we tested a more specific procedure for plasma modification.
In eight patients with recurrence of the idiopathic nephrotic syndrome, protein adsorption decreased proteinuria consistently and profoundly. The three patients (Patients 1, 7, and 8) who were also treated by plasma exchange responded to that treatment, but somewhat less so than to protein adsorption. Furthermore, when plasma exchange was effective in decreasing proteinuria in other, similar patients, the average decrease was less pronounced than that occurring after protein adsorption14. One patient (Patient 4) still had no proteinuria six months after the final cycle of protein adsorption, and in another patient (Patient 2) an average of two sessions per month maintained urinary protein excretion below 2.5 g per day, whereas delaying treatment led to an increase to more than 4.0 g per day.
The effect of protein adsorption cannot be mistaken for a spontaneous remission of the disease, which is an extremely rare event,3,34,35 as evidenced by the decreases in proteinuria during treatment and the increases thereafter (in most patients). Moreover, the decrease in proteinuria cannot be related either to a hemodilution effect or to albumin depletion, since the patients' serum albumin concentrations remained within 90 percent of the initial values. All the patients received polyclonal immunoglobulin at the end of each treatment cycle. Immunoglobulin therapy is beneficial in some autoimmune diseases,36 but since the immunoglobulin was given after urinary protein excretion had decreased, it is unlikely to have caused the decrease. Furthermore, in a recent study, polyclonal immunoglobulin therapy had no effect in patients with idiopathic nephrotic syndrome37.
Four patients were also treated with high doses of methylprednisolone for acute rejection crises without any decrease in urinary protein excretion, indicating the resistance of the proteinuria to steroid treatment. Furthermore, although cyclosporine treatment has been associated with a decrease in protein excretion in some children with recurrent idiopathic nephrotic syndrome,31,38,39 there was no such change in our patients after this drug was introduced.
Protein A, which is extracted from a strain of Staphylococcus aureus, specifically binds the constant domains of immunoglobulins40 and possibly also fibronectin26,41. The serum concentrations of all IgG isotypes (except IgG3) are decreased by procedures involving protein A adsorption, whereas concentrations of other immunoglobulin isotypes are less altered27. Therefore, the decrease in urinary protein excretion after protein adsorption suggests the possible role of an immunoglobulin (for the most part, IgG) in the proteinuria, although there is no other evidence of the involvement of immunoglobulin in this disease. We found that protein A eluates contained not only immunoglobulin but also other proteins. Using an assay specific for rat albumin, we attempted to characterize the biochemical properties of this factor or factors, focusing first on its possible immunoglobulin nature, because of the capacity of protein A to bind these molecules26. In an assessment using two methods of separation by size, we found that the capacity to alter glomerular permeability in rats was present in molecular-weight fractions below 100,000 and 150,000, sizes incompatible with intact immunoglobulin. However, some activity was also found in the large-molecular-weight fractions prepared by membrane filtration of the eluate derived from Patient 2. Therefore, if we cannot totally exclude the possibility that some activity is related to a factor larger than 100 kd, there is clear evidence that the majority of the effect was not due to an immunoglobulin. In addition, Western blotting of the eluates did not identify immunoglobulin fragments smaller than 100 kd in nonreducing conditions (data not shown). These results do not favor the possibility that a fragment of immunoglobulin is involved.
The precise identity of the active factor or factors bound to protein A remains speculative. One can, however, hypothesize two possibilities that would reconcile the specificity of protein A for immunoglobulin and our preliminary data: first, the presence of immune complexes in which the factors conducive to albuminuria were active only when they were dissociated from immunoglobulin; and second, the possibility that the factors could be attached to immunoglobulin through binding to the constant heavy-chain part of the molecule. However, because the size-separation methods used should not have dissociated complexes of this type, a unique factor conducive to albuminuria is more likely. Ultimate biochemical characterization will require time, largely because no in vitro test clearly predictive of the in vivo effect is yet available and because our test produces a low albuminuric signal (measured in milligrams at the most), probably owing to the small amount of active factor or factors transferred.
We are indebted to Professor R. Sibley (Department of Pathology, Stanford University) for discussion and criticism of the manuscript, to Dr. F. Buzelin for examination of the kidney-biopsy specimens, and to Ms. Aline Bertho for assistance in the preparation of the manuscript.
Source Information
From the Service de Nephrologie-Immunologie Clinique (J.D., A.T., J.P.S.), and INSERM U.211 (Unite de Recherche sur les Effecteurs Lymphocytaires T) (J.D., E.B., W.B., Y.J., J.P.S.), Centre hospitalier regional et universitaire (C.H.R.U.), Nantes; the Service de Nephrologie, Hopital de Bicetre, Le Kremlin Bicetre (F.K., B.C.); the Service de Nephrologie Pediatrique, Hopital Necker Enfants Malades, Paris (P.N.); and the Service de Nephrologie, C.H.R.U. Clemenceau, Caen (B.H.L.) -- all in France.
Address reprint requests to Dr. Dantal at the Service de Nephrologie-Immunologie Clinique, C.H.R.U., Pl. Alexis Ricordeau, 44035 Nantes CEDEX, France.
References
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
