Recurrent lupus nephritis in renal allograft triggered by pregnancy

Abstract

Herein we report the case of a 37-year-old woman with recurrence of lupus nephritis (LN) in a renal allograft during pregnancy. She had developed end-stage renal disease due to LN and was put on hemodialysis at the age of 26 years. She underwent kidney transplantation at the age 28 years. Maintenance immunosuppressants included methylprednisolone, tacrolimus, and mycophenolate mofetil, which were changed to azathioprine when she desired pregnancy. The renal allograft function remained stable and seemingly disease-free until proteinuria and functional decline occurred during the pregnancy (age: 34 years). The baby was delivered by performing a cesarean section at 33 weeks of gestation. Renal allograft biopsy revealed crescent formation. Light microscopy revealed tuft necrosis and endocapillary proliferation. Immunofluorescence microscopy revealed the deposition of immunoglobulin G and C1q. A recurrence of LN (ISN/RPS class IV-G [A/C]) was diagnosed, and the patient was treated with pulse steroid therapy and azathioprine was replaced with mycophenolate mofetil. This treatment improved acute or active lesions of LN and temporarily benefited the renal allograft function. Unfortunately, there were irreversible chronic changes and a gradual decline in the renal allograft function.

Introduction

Systemic lupus erythematosus (SLE) is an autoimmune disease that primarily affects women of reproductive age. Lupus nephritis (LN) occurs in 22–54% of patients with SLE [1–3], and 14–17% of these patients progress to end-stage renal disease [3, 4]. Kidney transplantation is one of the treatment options for patients with LN-induced end-stage renal disease. Although LN is known to recur in 0–19% of renal allografts [7–18], protocol biopsies show a recurrence rate of 50% [19, 20]. However, kidney transplant recipients with LN have similar graft and patient survival to recipients with other etiologies [5, 7, 8, 14–18]. Its pathologic recurrence rate is high, but its impact on long-term prognosis is low.

Kidney recipients with LN whose allografts are functionally stable (no proteinuria) may safely become pregnant 6 months to 2 years post-transplantation. Pregnant women with LN have a 5–46% risk of experiencing renal flares, depending on LN activity at conception [21–25]. During pregnancy, kidney transplant recipients with LN are at a risk of developing recurrent disease; however, this has not been reported. Herein, we present a kidney transplant recipient who developed proteinuria and deteriorating renal allograft function during pregnancy. This patient was diagnosed with recurrent LN on performing postpartum renal allograft biopsy.

Case report

The patient was a 37-year-old woman who was diagnosed with SLE at the age of 14 years. She presented with facial erythema, arthralgia, and Raynaud’s phenomenon. Blood tests indicated renal impairment and were positive for antinuclear and anti-double stranded DNA (dsDNA) antibodies. She was placed on steroids, cyclophosphamide, cyclosporine, and mizoribine. Her serum creatinine level then stabilized at 0.6–0.7 mg/dl, and proteinuria resolved. However, her renal dysfunction and proteinuria progressed at the age of 22 years. Renal biopsy at that time confirmed LN (International Society of Nephrology [ISN]/Renal Pathology Society [RPS] class IV + V). Despite receiving pulse steroids and cyclophosphamide therapy, she developed end-stage renal disease at the age of 26 years and was put on hemodialysis. Thereafter, SLE activity diminished and quiesced.

She underwent living donor kidney transplantation at the age of 28 years, with the donor being her father. Basiliximab, methylprednisolone, tacrolimus, and mycophenolate mofetil (MMF) were administered as induction immunosuppressive therapy, followed by a maintenance regimen of methylprednisolone at a dose of 4 mg/day, tacrolimus at a dose of 2 mg/day, and MMF at a dose of 1000 mg/day. The trough level of tacrolimus was 3–5 ng/ml. The serum creatinine level was 1.0–1.2 mg/dl, without proteinuria after kidney transplantation.

At the age of 31 years, given her desire to become pregnant, MMF was replaced by azathioprine (AZA) at a dose of 50 mg/day. This decision was reversed a year later as the serum creatinine level increased (1.2–1.4 mg/dl). A subsequent allograft biopsy did not contained glomeruli in light microscopy. No interstitial and vascular lesions caused by calcineurin inhibitor toxicity were observed. Immunofluorescence microscopy proved negative for immunoglobulins or complement components in the glomeruli. Electron microscopy revealed that there were no electron-dense deposits in the glomerular basement membrane to substantiate the recurrence of LN. Furthermore, she had no hypocomplementemia, and anti-dsDNA antibody testing was negative. The cause of deterioration of the renal allograft function was unclear, but since then, the renal allograft function was stable without proteinuria. MMF was again replaced with AZA at the age of 33 years. She became pregnant after intrauterine insemination, but had a miscarriage at 7 weeks of gestation. Tests for lupus anticoagulant, anticardiolipin antibodies, and anti-β₂glycoprotein-1 were negative, ruling out antiphospholipid syndrome.

A second intrauterine insemination resulted in pregnancy at the age of 34 years and a viable birth at the age of 35 years. Prior to this gestation, the serum creatinine level was 1.3–1.4 mg/dl and the urine protein–creatinine ratio was 0.1–0.2 g/g creatinine. Urinary protein excretion increased at 14 weeks of gestation, plateauing at 2–3 g/g creatinine. The patient was hospitalized for abdominal pain and elevated blood pressure at 28 weeks of gestation. Her blood pressure, which was around 120/70 mmHg before admission, was 150/100 mmHg at the time of admission. The cause of abdominal pain was uncertain, but it resolved spontaneously, and her blood pressure normalized in response to antihypertensive drugs. Her serum creatinine level rose to 1.6–1.7 mg/dl at 33 weeks of gestation, despite unremarkable ultrasound imaging of the renal allograft. She was diagnosed with superimposed preeclampsia. The baby was delivered by a cesarean section, with Apgar scores of 6/7/8 and a birth weight of 1493 g.

After delivery, the serum creatinine level was 1.6–1.8 mg/dl, with the urine protein–creatinine ratio at 2–3 g/g creatinine. She had no hypocomplementemia, and anti-dsDNA antibodies were negative. There were no signs of extrarenal SLE involvement. She underwent renal allograft biopsy at 3 months postpartum. The tissue contained 49 glomeruli. Routine histologic preparations showed crescent formation and tuft necrosis in 2 glomeruli, endocapillary proliferation in 10 glomeruli, and double contouring of glomerular basement membrane in 3 glomeruli. There were no wire loop lesions. Overall, 65% of the glomeruli appeared sclerotic, with 30% of cortex exhibiting interstitial fibrosis and tubular atrophy. Arteriolar hyalinosis was also detected. There was no evidence of graft rejection. Immunofluorescence microscopy revealed immunoglobulin G deposits in the glomerular basement membrane and C1q deposits in the glomerular basement membrane and mesangium (Fig. 1). These findings led to the diagnosis of LN (ISN/RPS class IV-G [A/C]) and arteriolopathy due to calcineurin inhibitor toxicity. AZA was replaced with MMF at a dose of 1250 mg/day, and pulse methylprednisolone therapy (500 mg, intravenously daily for 3 days) was initiated. Prednisolone (40 mg/day) was then administered, with gradual dose reduction to 5 mg/day over 6 months. The patient developed cytomegalovirus viremia on several occasions, without overt disease manifestations. Viremia disappeared after immunosuppressant adjustment and valganciclovir administration. Following treatment, the serum creatinine level was 1.5–1.6 mg/dl and the urine protein–creatinine ratio was 1–2 g/g creatinine. Another allograft biopsy performed at 11 months postpartum showed segmental endocapillary proliferation with no global endocapillary proliferation, crescentic formation, or tuft necrosis. Overall, 50% of the glomeruli appeared sclerotic, with 50% of the cortices marked by interstitial fibrosis and tubular atrophy. These findings led to the diagnosis of LN (ISN/RPS class IV-S [A/C]). The treatment improved the acute or active lesions of LN as well as the renal allograft function. Unfortunately, there were irreversible chronic changes and a gradual rise in the serum creatinine level to 1.8–1.9 mg/dl (Fig. 2).

Fig. 1.

Microscopic features of renal allograft biopsy: a crescent formation, tuft necrosis, endocapillary proliferation, and double contouring of glomerular basement membrane in periodic acid-Schiff (PAS) stain, 400×; b hyalinosis caused by calcineurin inhibitor toxicity in the arteriole (arrow) and interstitial infiltrates only around atrophic tubules were observed via periodic acid-Schiff (PAS) stain, 100×; c deposition of immunoglobulin G at glomerular basement membrane on immunofluorescence stain, 400×; d deposition of C1q at glomerular basement membrane and mesangium on immunofluorescence stain, 400×

Fig. 2.

Clinical course of patient. AZA, azathioprine; MMF, mycophenolate mofetil

Discussion

We presented a case of pregnancy-induced recurrent LN in a renal allograft. Until now, only rates of recurrent LN in renal allografts (0–19%) [5–18] or renal flares in pregnant women with LN (5–46%) [21–25] have been established; however, there have been no reports of pregnancy-induced recurrent LN in renal allografts. The mechanism by which SLE exacerbates during pregnancy is unclear. In murine models, an elevated level of estrogen was found to be associated with increased SLE activity, promoting the survival and activation of high affinity autoreactive B cells [26, 27]. Reported risk factors for recurring LN in renal allografts are as follows: non-Hispanic black race (odds ratio [OR] = 1.88, 95% confidence interval [CI] 1.37–2.57), female sex (OR = 1.70, 95% CI 1.05–2.76), and age < 33 years (OR = 1.69, 95% CI 1.23–2.31) [10]. Transplant recipients differ significantly according to recurrence status (present vs. absent) with regard to time from the diagnosis of LN to end-stage renal disease (52 ± 37 months vs. 149 ± 54 months; p = 0.016), duration of dialysis (21 ± 37 months vs. 58 ± 50 months; p = 0.045), and time from the diagnosis of LN to kidney transplantation (6 ± 3 years vs. 16 ± 4 years). Recurrent LN is also associated with significantly lowered antithymocyte globulin induction therapy (p = 0.004) [20]. Risk factors for renal flare in pregnant women with LN are treatment with azathioprine (hazard ratio [HR] of 9.1, p = 0.010) and with prednisolone (HR of 7.3, p = 0.018) [23]. Patients with proteinuria > 1 g or glomerular filtration rates < 60 ml/min/1.73 m² have a nine-fold higher risk of renal flares, and patients in partial (vs. complete) remission show a three-fold greater risk [22]. Serum levels of C3 complement fraction and anti-DNA antibodies likewise correlate significantly with risk of renal flare [25]. The patient in the present case displayed some of these risk factors; however, we considered the risk of recurrence low owing to no hypocomplementemia, negative anti-dsDNA antibody status, low-level proteinuria (urine protein–creatinine ratio, 0.1–0.2 g/g creatinine), and no evidence of LN recurrence on renal allograft biopsy prior to pregnancy. The renal allograft biopsy that demonstrated at 3 months postpartum already showed irreversible chronic lesions; therefore, it may have been better to have a little earlier. The worsening proteinuria seen at 14 weeks of gestation signaled a need for allograft biopsy. Renal biopsies may be safely performed in this setting before mid-term (median 16 weeks; range 9–27 weeks), without biopsy-related complications [28].

Our patient had consistently shown no hypocomplementemia and anti-dsDNA antibody negativity since the initiation of dialysis. During pregnancy, the serum levels of complement components typically increase as hepatic protein synthesis mounts [28, 29]. Consequently, they do not reliably reflect SLE disease activity. In this respect, however, the status of the current patient remained unchanged after delivery. Although transplant recipients with recurrent LN have shown detectable anti-dsDNA antibodies and hypocomplementemia at the rates of 15–33% and 40–67%, respectively [9, 11, 19], serologic parameters are not always useful in assessing disease activity or predicting recurrence of post-transplantation LN.

In conclusion, the differential diagnosis of renal allograft dysfunction in patients with SLE during pregnancy is comparatively broad, and even if hypocomplementemia and anti-dsDNA antibodies are lacking, the possibility of recurrent LN must be considered.

Declarations

Conflict of interest

All the authors have declared no competing interest.

Human and animal rights

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee at which the studies were conducted (IRB approval number RIN A19-097) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Written informed consent obtained from the patient.