Complications of Immunosuppression in Glomerular Disease

Abstract

Most glomerular diseases are immunologically mediated disorders of the kidney and are common causes of ESKD. In addition to supportive therapy, a wide range of immunosuppressive agents are used in the management of patients with these conditions. Immunosuppression requires a careful balance of risk and benefits, and many of these agents have a narrow therapeutic window and require close monitoring. This review describes the side effects of immunosuppressive agents used in recent randomized, controlled trials of glomerular disease, and highlights some of the key adverse events that determine the choice and prescription of these medications.

Introduction

Immunosuppressive agents are used in the treatment of glomerular diseases, but are associated with a high risk of adverse effects. This review is intended as a guide for clinicians and focuses on the most commonly prescribed immunosuppressive agents, recognizing that we are in a time of rapid evolution with multiple new therapies being developed. Much of the recent evidence in glomerular disease is drawn from studies of lupus nephritis (LN) and ANCA-associated vasculitis (AAV) because in older studies of primary glomerular disease, the details of adverse events were often not well captured. It should be recognized that each of the medications described below has a wide range of adverse effects beyond the scope of this review, but are described in the package insert supplied with medications or may be reviewed in resources such as Micromedex.

Glucocorticoids

Glucocorticoids (GC) have both anti-inflammatory and immunosuppressive effects, and have a long list of side effects, including infection, bone disease, dysglycemia, obesity, hypertension, psychosis, gastrointestinal bleeding, cataracts, and long-term risks of cardiovascular disease.

Infection Risk

Specific infections that warrant consideration for screening and prophylaxis include Pneumocystis jirovecii pneumonia (PJP), tuberculosis, strongyloidiasis, hepatitis B (HBV) and hepatitis C (HCV), HIV, herpes zoster virus, and candidiasis. In a UK primary care database study (n=275,072), patients taking GC had a two to six fold increased rate of infections compared with those with the same underlying disease not exposed to steroids (1). Risk factors for infection included higher GC doses, increasing age, diabetes, and hypoalbuminemia. In observational studies in patients with rheumatology diseases, low-dose GCs (prednisone <10 mg/d) are associated with small increased risks of bacterial infections, but the risks of more severe opportunistic infections increase at higher doses (prednisone >20 mg/d) (2). Infection risk may be amplified in glomerular disease by urinary loss of Ig and complement, and by the immunodeficiency state of kidney disease.

In primary glomerular disease, steroid monotherapy is widely used in minimal change disease, FSGS, and IgA nephropathy. In three older studies of IgA nephropathy, severe adverse events (SAEs) due to infection were not described (3–5). By contrast, recent studies using high-dose steroids as monotherapy, including the Supportive Versus Immunosuppressive Therapy for the Treatment of Progressive IgA Nephropathy (STOP-IgAN) (6) and Therapeutic Evaluation of Steroids in IgA Nephropathy Global (TESTING) (7) trials (Tables 1 and 2), showed increased infection rates, partly because of more complete adverse event reporting. The TESTING trial was stopped early because of increased adverse events in the oral methylprednisone arm, mostly severe infections, including three patients with PJP (7). Notably, PJP prophylaxis was not used in this study, nor in other recent studies of IgA nephropathy. In a meta-analysis of early LN studies, the risk of major infection did not differ between GCs alone and GCs with the addition of azathioprine or cyclophosphamide (CYC) (8), highlighting the importance of steroid dose in the overall contribution to infection risk. In AAV, a large multicenter study, the Plasma exchange and glucocorticoid dosing in the treatment of anti-neutrophil cytoplasm antibody associated vasculitis (PEXIVAS), will soon report outcomes in patients treated with low- versus standard-dose GCs (Clinicaltrials.gov identifier NCT00987389).

Table 1.

Selected randomized controlled trials of IgA nephropathy

Study	Study Design	Efficacy	Adverse Events
Rauen et al. (64) (STOP-IgA, IgA nephropathy)	Prospective, multicenter, RCT (2008–2011); sample size: 162; intervention: immunosuppression (Pozzi protocol steroids or oral CYC [1.5 mg/kg 3 mo, then oral AZA 1.5 mg/kg 3 yr] versus supportive care only; end point: clinical remission, decline in GFR; follow-up: 3 yr	No significant difference in the annual decline in eGFR or remission rates between the two groups		Immunosuppression (n=82)	Supportive care (n=80)
			SAEs	29 (35%)	21 (26%)
			Serious infections	8 (10%)	3 (4%)
			Total infections	111	174
			Cancer	2 (2%)	0
			Diabetes/impaired glucose	9 (11%)	1 (1%)
Rauen et al. (6) (STOP-IgA, high GFR subgroup)	Prospective, multicenter, RCT (2008–2011); sample size: 162; intervention: immunosuppression corticosteroids (Pozzi protocol) versus supportive care; end point: clinical remission, decline in GFR; follow-up: 3 yr	Increased complete remission with steroid group (20% versus 3%)		Steroids (n=54)	Supportive care (n=55)
			SAEs	12 (22%)	14 (25%)
			Serious infections	4 (7%)	2 (4%)
			Total infections	115	69
			Cancer	0	0
			Diabetes/impaired glucose	9 (17%)	1 (2%)
Lv et al. (7) TESTING, IgA nephropathy)	Prospective, multicenter, RCT (2012–2015); sample size: 262; intervention: oral methylprednisone (0.6–0.8 mg/kg per d, taper after 2 mo) versus placebo; end point: ESKD or 40% decrease in GFR; follow-up: median 2.1 yr	Study terminated early second to excess SAEs in oral methylprednisone group		Oral methylprednisone (n=136)	Placebo (n=126)
			SAE	20 (15%)	4 (3%)
			Serious infection	11 (8%)	0
			Pneumocystis jirovecii	3 (2%)	0
			Gastrointestinal SAE	4 (3%)	1 (1%)
			Osteonecrosis	2 (1%)	0
Hou et al. (65) (IgA nephropathy)	Prospective, multicenter, RCT (2010–2013); sample size: 176; intervention: MMF (1.5 g/d)+prednisone (0.4–0.6 mg/kg per d, taper) versus prednisone only (0.8–1.0 mg/kg per d, taper); end point: complete remission; follow-up: 12 mo	No difference in complete remission rate (48% versus 53%)		MMF+prednisone (n=87)	High-dose prednisone (n=88)
			SAE	5 (6%)	6 (7%)
			Any infection	27 (31%)	20 (23%)
			Serious infection	3 (3%)	4 (5%)
			Diabetes (new)	1 (1%)	12 (14%)
			Gastrointestinal	7 (8%)	10 (11%)
			Osteonecrosis	0	1 (1%)

STOP-IgA, Supportive Versus Immunosuppressive Therapy for the Treatment of Progressive IgA Nephropathy; RCT, randomized controlled trial; CYC, cyclophosphamide; AZA, azathioprine; SAE, serious adverse event; TESTING, Therapeutic Evaluation of Steroids in IgA Nephropathy Global; MMF, mycophenolate mofetil.

Table 2.

Selected randomized controlled trials of primary membranous nephropathy

Study	Study Design Efficacy	Adverse Events
Praga et al. (66) (membranous nephropathy)	Prospective, multicenter, RCT; sample size: 48; intervention: tacrolimus (0.05 mg/kg per d) versus supportive therapy; end point: remission proteinuria; follow-up: 18 mo	Rate of remission at 18 mo higher in tacrolimus group (94%) versus control (35%); 50% relapse rate in tacrolimus group		Tacrolimus (n=25)	Supportive care (n=23)
			Serious infection	0	0
			Diarrhea	2 (8%)	0
			Glucose intolerance	4 (16%)	2 (9%)
			Thrombosis	0	0
Howman et al. (67) (membranous nephropathy)	Prospective, multicenter, RCT (1998–2008); sample size: 108; intervention: prednisone/oral chlorambucil versus oral cyclosporine (5 mg/kg) versus supportive therapy; end point: 20% decline in kidney function; follow-up: 36 mo	20% kidney function decline was lower in prednisone/chlorambucil group (58%)		Prednisone/chlorambucil (n=33)	Cyclosporine (n=37)	Supportive care (n=38)
			SAEs	61%	49%	42%
			Cytopenia	28 (85%)	5 (13%)	3 (8%)
			Infection	3 (9%)	8 (22%)	0
			Gastrointestinal	3 (9%)	3 (8%)	2 (5%)
			Impaired glycemia	8 (24%)	1 (3%)	5 (13%)
Dahan et al. (41) (GEMRITUX, membranous nephropathy)	Prospective, multicenter, RCT; sample size: 75; intervention: rituximab (375 mg/m2×2) versus control; end point: remission proteinuria; follow-up: 6 mo	No difference in remission rates (rituximab 35% versus control 22%)		Rituximab (n=37)	Control (n=38)
			SAEs	6 (16%)	5 (13%)
			Serious infections	1 (3%)	0
			Leukopenia	0	0
			Cardiovascular events	4 (11%)	2 (5%)
			Cancer	0	1 (3%)
			Diarrhea	1 (3%)	0

RCT, randomized controlled trial; SAE, serious adverse event; GEMRITUX, Evaluate Rituximab Treatment for Idiopathic Membranous Nephropathy.

Osteoporosis

GCs are associated with an accelerated bone loss in the first 3–6 months, followed by a persistent decrease in bone formation for the duration of therapy. Fracture risk is related to both the peak and cumulative GC dose, and is increased even at chronic low doses of 5 mg prednisone per day. In patients on long-term steroids, the American College of Rheumatology 2017 guidelines recommend a fracture risk assessment using bone mineral density testing and an online risk assessment tool (FRAX; https://www.shef.ac.uk/FRAX/tool.jsp) for adults over 40 years of age (9). All patients should be treated with oral calcium (1000–1200 mg/d) and oral vitamin D (600–800 U/d). Oral bisphosphonates should be considered for those at moderate to high fracture risk (>10%–20% 10-year fracture risk) (9). Note in the treatment of glomerular disease, the duration of steroid therapy may be short, and in patients with CKD, bone density measurements may be less precise and the safety of bisphosphonate therapy uncertain.

Avascular Necrosis

Avascular necrosis (osteonecrosis), typically affecting the femoral head, is a severe complication of steroid therapy usually attributed to ischemia secondary to abnormalities of lipid metabolism, oxidative stress, and vascular injury. Younger patients with SLE are more prone to avascular necrosis, possibly because of the associated chronic inflammatory and procoagulant milieu. Hip pain typically occurs 2–3 years after the start of GC therapy, but may occur earlier. In contrast to GC-induced osteoporosis (see above), avascular necrosis rarely occurs in patients treated with peak doses of <20–30 mg prednisone per day.

Adrenal Suppression

Chronic GC use can suppress the hypothalamic-pituitary-adrenal axis by negative feedback on corticotropin releasing hormone and corticotropin. The hypothalamic-pituitary-adrenal axis may remain suppressed after GC therapy has been reduced or stopped, resulting in adrenal insufficiency. Chronic low cortisol levels may result in nausea and fatigue, but in conditions of acute stress such as surgery, an adrenal crisis may result. Stress-dose steroids may be considered perioperatively for patients at higher risk of adrenal insufficiency. The prevalence of biochemical adrenal insufficiency in observational studies varies widely (14%–63%) (10,11), but clinical insufficiency is uncommon. Some studies have found an association with steroid dose and duration (11), but surprisingly in others, no relationship was found (10). The risk of adrenal insufficiency is minimal if GC duration is <3 weeks. To avoid adrenal insufficiency, GCs are usually prescribed in the morning, alternate day dosing may be used, and with chronic use, although the dose can be tapered relatively rapidly to a physiologic GC level (approximately 7.5 mg/d), a slower taper is needed below this dose.

Mycophenolic Acid

Mycophenolate mofetil (MMF) and enteric-coated mycophenolate sodium are antiproliferative agents that inhibit inosine monophosphate dehydrogenase, the rate-limiting step of purine synthesis in both B and T lymphocytes. Both formulations are highly protein bound (97%) and free mycophenolic acid (MPA) levels may be increased with the hypoalbuminemia of nephrotic syndrome, potentially leading to more adverse events. Notably antacids/phosphate binders (up to 33%) and protein pump inhibitors (25% or more) may inhibit oral absorption and decrease both the efficacy and complication rate. MMF is widely used in glomerular disease, especially LN, but the greatest experience comes from its extensive use in solid organ transplantation. Nausea and diarrhea are the commonest side effects, which may be reduced in some patients with the enteric-coated formulation. MPA salts are strongly contraindicated in pregnancy because of the association with first trimester pregnancy loss and congenital malformations.

Infection Risk

Viral infections (herpes zoster, cytomegalovirus, herpes simplex) are a particular risk with MMF treatment and this risk is enhanced by bone marrow suppression and leucopenia. By contrast, HBV and HCV viral infections may not be greatly increased as MPA may suppress the expression of HBV surface antigen and HBV viral replication (12). MPA may also have a suppressive effect on P. jirovecii. In a review of four kidney transplant studies, PJP occurred in none of the 1068 patients exposed to MMF, compared with ten out of 563 with other immunosuppression (mostly azathioprine) (13).

MMF is widely used as both induction and maintenance therapy for LN (Table 3) (14–16). In the North American Collaborative study, adverse events and severe infections were less common with MMF compared with CYC (16); however, this was not noted in the multicenter Aspreva Lupus Management Study (ALMS) (14). A Cochrane review of six studies (n=683) (17) and a more recent meta-analysis (18) reported no difference in major infections between MMF and CYC.

Table 3.

Selected randomized controlled trials of lupus nephritis

Study	Study Design	Efficacy	Adverse Events
Houssiau et al. (72) (Eurolupus, lupus nephritis)	Prospective, multicenter RCT; sample size: 90; intervention: high-dose IV CYC (eight doses 0.5–1 g/m2) versus low-dose IV CYC (6 doses 500 mg); end point: treatment failure; follow-up: 41 mo	High-dose and low-dose IV CYC showed equivalent efficacy		High-dose CYC (n=45)	Low-dose CYC (n=45)
			Severe infection	10 (22%)	5 (11%)
			Other infection	7 (16%)	5 (11%)
			Leucopenia	5 (11%)	5 (11%)
			Amenorrhea	3 (7%)	3 (7%)
Houssiau et al. (32) (Eurolupus, 10-yr follow-up)	Prospective, multicenter RCT; sample size: 90; intervention: high-dose IV CYC (eight doses 0.5–1 g/m2) versus low-dose IV CYC (six doses 500 mg); end point: treatment failure; follow-up: 10 yr	Long-term results showed no difference in efficacy		High-dose CYC (n=45)	Low-dose CYC (n=45)
			Cardiovascular events	4 (9%)	3 (7%)
			Cancers	1 (2%)	6 (14%)
			Successful pregnancy	9 (20%)	10 (22%)
Contreras et al. (73) (lupus nephritis)	Prospective, single-center RCT (1996–2003); sample size: 59; intervention: IV CYC (0.5–1 g 3 monthly) versus oral AZA (1–3 mg/kg per d) versus oral MMF (1–3 g/d); end point: patient and kidney survival; follow-up: 72 mo	Patient and kidney survival was lower in the IV CYC group	Ratesa	CYC (n=20)	AZA (n=19)	MMF (n=20)
			Infections	77%	25%	32%
			Major infections	25%	2%	2%
			Leucopenia	10%	6%	2%
			Amenorrhea	32%	8%	6%
			Vomiting	55%	7%	14%
			Diarrhea	12%	9%	12%
			Death	4 (25%)	0	1 (5%)
Ginzler et al. (16) (lupus nephritis)	Prospective, multicenter RCT; sample size: 140; intervention: oral MMF (3 g/d) versus IV CYC (0.5–1 g monthly), both arms high-dose glucocorticoids; end point: remission in proteinuria; follow-up: 6 mo	MMF was superior in rate of remission compared with CYC (52.1% versus 30.4%)			MMF (n=71)		CYC (n=69)
			Serious infections		1 (1%)		6 (9%)
			Other infection		3 (4%)		5 (7%)
			Upper GI symptoms		23 (32%)		25 (36%)
			Diarrhea		15 (21%)		2 (3%)
			Lymphopenia		18 (25%)		28 (40%)
			Alopecia		0		8 (12%)
			Amenorrhea		0		2 (3%)
Appel et al. (14) (ALMS, lupus nephritis)	Prospective, multicenter RCT; sample size: 370; intervention: oral MMF (3 g/d) versus IV CYC (0.5–1 g monthly), both arms high-dose glucocorticoids; end point: decrease in proteinuria or serum creatinine; follow-up: 6 mo	No difference in response rates (56.2% versus 53%)			MMF (n=184)		CYC (n=180)
			SAE		51 (28%)		41 (23%)
			Severe infection		22 (12%)		18 (10%)
			Diarrhea		52 (28%)		23 (13%)
			Withdrawal		24 (13%)		13 (7%)
			Death		9 (5%)		5 (3%)
Dooley et al. (15) (ALMS maintenance, lupus nephritis)	Prospective, multicenter RCT; sample size: 227; intervention: oral MMF (2 g) versus oral azathioprine (2 mg/kg per d); end point: treatment failure (death, ESKD, doubling creatinine, kidney flare, rescue therapy); follow-up: 36 mo	MMF was superior to azathioprine (16.4% versus 32.4%)			MMF (n=115)		AZA (n=111)
			SAE		27 (23%)		37 (33%)
			Infection		91 (79%)		87 (78%)
			Serious infection		11 (10%)		13 (12%)
			Leucopenia		0		4 (4%)
			Withdrawal		29 (25%)		44 (40%)
			Death		0		1 (1%)
Houssiau et al. (74) (MAINTAIN, lupus nephritis)	Prospective, multicenter RCT (2002–2006); sample size: 105; intervention: oral AZA (2 mg/kg) versus oral MMF (2 g/d); end point: time to relapse; follow-up: 3 yr	No difference in time to relapse, or relapse rate			AZA (n=52)		MMF (n=53)
			Any infection		14 (27%)		21 (40%)
			Leucopenia		11 (21%)		2 (4%)
			Nausea/diarrhea		8 (15%)		8 (15%)
Rovin et al. (76) (LUNAR, induction – lupus nephritis)	Prospective, multicenter RCT (2006–2008); sample size: 144; intervention: RTX (500 mg ×2, repeat 6 mo) versus placebo, background prednisone and MMF (3 g/d); end point: kidney response; follow-up: 18 mo	No difference in kidney response between the groups			RTX/MMF (n=72)		MMF (n=72)
			SAE		29 (40%)		24 (33%)
			Severe infection		14 (19%)		14 (19%)
			Infusion reaction		12 (17%)		6 (8%)
			Neutropenia		2 (3%)		1 (1%)
			Death		2 (3%)		0
ACCESS trial group (76) (lupus nephritis)	Prospective multicenter RCT; sample size: 134; intervention: IV abatacept (500–1000 mg monthly) versus placebo, on background Eurolupus regime of CYC+AZA; end point: complete kidney response; follow-up: 12 mo	No difference between the groups (33% versus 31%)			Abatacept+CYC/AZA (n=66)		Placebo+CYC/AZA (n=68)
			SAE		19 (29%)		20 (29%)
			Severe infection		8 (12%)		5 (7%)
			Any infection		31 (47%)		32 (47%)
			Death		0		1 (1%)
Liu et al. (39) (lupus nephritis)	Prospective multicenter RCT; sample size: 362; intervention: MMF (0.5 mg twice a day)+TAC (2 mg twice a day) versus IV CYC (0.75 mg/m2 monthly ×6); end point: complete remission; follow-up: 6 mo	More patients in MMF/TAC group achieved complete remission (46% versus 26%)			MMF+TAC (n=181)		IV CYC (n=181)
			SAE		27 (15%)		10 (6%)
			Severe infection		24 (13%)		8 (4%)
			Any infection		91 (50%)		95 (52%)
			Leukopenia		1 (1%)		12 (7%)
			Gastrointestinal		25 (14%)		43 (24%)
			Menstrual abnormality		2 (1%)		7 (5%)
			Hyperglycemia		5 (3%)		4 (3%)

RCT, randomized controlled trial; IV, intravenous; CYC, cyclophosphamide; AZA, azathioprine; MMF, mycophenolate mofetil; GI, gastrointestinal; ALMS, Aspreva Lupus Management Study; SAE, serious adverse event; MAINTAIN, Azathioprine versus mycophenolate mofetil for long-term immunosuppression in lupus nephritis; LUNAR, Lupus Nephritis Assessment with Rituximab; RTX, rituximab; ACCESS, Abatacept and Cyclophosphamide Combination Efficacy and Safety; TAC, tacrolimus.

^a% rates defined as 100×number of events/number of patient-years follow-up.

Small studies have suggested that therapeutic drug monitoring may improve the efficacy and decrease the incidence of side effects in glomerular disease (19), although the use of MPA levels as a guide to therapy has not been widely adopted.

Cyclophosphamide

CYC is an inactive prodrug that is metabolized to an active metabolite, phosphoramide mustard, and the bladder-toxic metabolite acrolein, by cytochrome P450 enzymes. Functional variants in cytochrome P450 enzyme genes have been associated with increased leukopenia (but also increased efficacy) in AAV (20). CYC is associated with multiple serious adverse effects, many of which occur early (bone marrow suppression, infection, hemorrhagic cystitis, infertility), but others may occur more than 10 years after therapy has stopped (malignancy).

Cytopenia and Infection

Early studies of LN and AAV used CYC for both induction and maintenance therapy, resulting in a high cumulative dose. Recent protocols seek to minimize CYC exposure by limiting therapy to induction, using lower doses and pulse intravenous therapy rather than daily oral administration and by switching early to rituximab (RTX) after one or two doses (21,22). Infection rates in recent studies of LN and AAV are described in Table 3. Lymphopenia (<1×10⁹/L) resulting from CYC therapy correlates with an increase in opportunistic infections (23).

Cancer Risk

CYC is associated with an increase in malignances (mostly skin cancer, myeloid malignancies, and bladder cancer), because of direct DNA damage and reduced immune surveillance. Early studies, with typically high cumulative doses of CYC, showed a 2–2.4-fold increase in cancer rates, but more recent studies show a lower risk (24,25). In a Danish registry study with 10-year follow-up, patients with AAV treated with CYC (n=293) had a 27% incidence of malignancy with a standardized incidence ratio (SIR) of 1.9 compared with national cancer incidence rates, driven mostly by non-myeloma skin cancer (NMSC) (24). Notably, in the subgroup with a lower cumulative CYC exposure (<36 g), there was no increased cancer risk. In a review of four studies of AAV from the European Vasculitis Study Group (n=535), the use of CYC was associated with an increased SIR of 1.58, driven entirely by NMSC, although the median follow-up was short (5 years) (25). In a single-center series of patients with AAV (n=119) with a median 5.6 year follow-up, the SIR was 3.1, again by NMSC only, without an increase in bladder cancer or myeloid malignancy (26). Of note, this study did not find any increase in the cancer rate in patients treated with RTX, an alternative agent in the treatment of AAV. Bladder cancer is a specific concern after CYC use and may be related to the bladder toxic metabolite acrolein. A high frequency (4.8%–14%) was described in older studies using prolonged courses of CYC (50–100 g cumulative dose), but no increased risk was noted in two recent studies with lower cumulative drug exposure (<20–36 g) (24,27). The value of mesna for bladder protection is uncertain at the doses of CYC typically used for glomerular disease (28).

To put the cumulative dose thresholds into context, a patient weighing 70 kg, treated with a variety of common CYC protocols, would receive a cumulative dose much lower than 20 g: (modified Ponticelli regime for membranous nephropathy [13 g]; 6 month course for LN [6 g]; Eurolupus protocol for LN [3 g]; 3 month course of oral CYC, 2 mg/kg for minimal change disease, LN, or AAV [13 g]). Note in many of these studies, the role of the underlying disease, and the contribution from other immunosuppression is difficult to determine. For example, azathioprine is commonly used as maintenance therapy in AAV, and itself is associated with a three-fold increase in skin cancer risk.

Hemorrhagic Cystitis

This complication occurred commonly (12%–41%) with prolonged courses of high-dose oral CYC (50–100 g), and but is rare with intravenous CYC and with lower dose oral CYC (28). Mesna binds acrolein in the urine and has been used to prevent hemorrhagic cystitis, however, recent studies question its efficacy (28,29).

Fertility

CYC therapy is associated with decreased fertility in both men and women and is associated with cumulative drug exposure. In a review of the National Institutes of Health studies of LN, chronic amenorrhea occurred more commonly with higher dose (approximately 10–15 g) versus lower dose (approximately 5–6 g) CYC (36% versus 12%) (30). The rate also increased with age: <25 years (12%), 26–30 years (27%), and >31 years (62%). A second study of 84 women with AAV or LN treated with intravenous CYC (mean exposure 12 g) showed chronic amenorrhea in 70% of women over 35 years, but in 0% of those aged <26 years (31). The risk of amenorrhea seems to increase significantly after 10–15 g exposure, although no safe threshold has been demonstrated (30). In the Eurolupus study, much lower doses (3 g) were used and fertility rates were relatively preserved in this study (32). Gonadotropin-releasing hormone agonists (e.g., leuprolide) for ovarian suppression are often used for protection, but efficacy is debated (33), and oocyte or embryo cryopreservation should be considered if higher dose CYC is required.

In men, transient azoospermia is common, with recovery related to the cumulative dose. In patients treated for sarcoma, exposure to a cumulative dose <7.5 g/m² (approximately 14 g for a 70 kg man) resulted in a 72% recovery, whereas only 11% recovered with doses >7.5 g/m² (34). The risk of permanent azoospermia may be low for cumulative doses <12 g (equivalent to 2 mg/kg for 12 weeks) (35). Notably, gonadotropin-releasing hormone analogs have shown disappointing results for fertility protection in men and semen cryopreservation should be considered.

Calcineurin Inhibitors

The calcineurin inhibitors (CNIs) cyclosporine and tacrolimus are extensively used in the treatment of membranous nephropathy, minimal change disease, and FSGS, and more recently in the treatment of LN as part of multitarget therapy. CNIs inhibit T cell activation, but also have direct effects on the podocyte, stabilizing the actin cytoskeleton (36).

The side-effect profiles of these two agents are similar, although some adverse events are more common with cyclosporin (hirsutism, gingival hyperplasia, hyperuricemia, hypertension, hyperlipidemia) and others with tacrolimus (glucose intolerance, neurotoxicity). One of the major concerns of CNIs is nephrotoxicity, both as an early hemodynamic reversible acute injury and as a chronic progressive nephropathy. Tacrolimus is also associated with hyperkalemic metabolic acidosis and hypertension mediated by upregulation of the sodium chloride transporter in the distal convoluted tubule (37). Both CNIs have a narrow therapeutic window with marked variability in bioavailability and metabolism, and therapeutic drug monitoring is required. Genetic variation in genes encoding CYP3A4, CYP3A5, and P-glycoprotein contribute to the variable pharmacokinetics and have been associated with an risk of nephrotoxicity (38). Anemia and hypoalbuminemia may increase the unbound (active) fraction of CNI and increase side effects for a given whole-blood level.

In LN, recent trials have shown a benefit from multitarget therapy, adding a CNI to lower dose prednisone and mycophenolate. Liu et al. (39), showed that the MMF/tacrolimus group was superior to the intravenous CYC group in achieving a complete kidney response at 6 months (46% versus 26%); however, adverse events were increased, with more serious infections (Table 3). A novel CNI, voclosporin, has been studied in a phase IIb trial of LN (AURA-LV) (ClinicalTrials.gov: NCT02141672). The addition of voclosporin to prednisone/MMF increased remission rates at 48 weeks, but serious adverse events and mortality were increased in the voclosporin groups. The ongoing phase 3 study will provide further information on the adverse drug reactions with this combination.

CNIs continue to be widely used in the treatment of glomerular disease with drug interactions remaining one of the most challenging management issues. New extended release formulations of tacrolimus, and novel CNIs (e.g., voclosporin) may dramatically change the use of these agents.

Rituximab

This chimeric mAb to CD20 leads to depletion of circulating and tissue resident B cells, but not plasma cells (which lack the CD20 antigen). In the United States, it is approved by the US Food and Drug Administration for the treatment of AAV, but is widely used off-label in the treatment of other glomerular diseases. RTX has a prolonged duration of action. The drug can be found in the circulation for 3–6 months, and B cell depletion may be sustained for 6–12 months, and sometimes longer.

Infusion Reactions

Immediate hypersensitivity reactions are relatively common and were described in 28% of patients with idiopathic membranous nephropathy (40), but in a more recent study using appropriate prophylaxis, no severe infusion reactions were noted (41). These reactions are mediated by a cytokine release syndrome, not anaphylaxis, and manifest by severe flu-like symptoms (fever, chills, myalgias, shortness of breath). Pretreatment with acetaminophen, diphenhydramine, and methylprednisone, and starting with a slow infusion rate are commonly used to prevent this.

Infectious Complications

Infection rates after RTX therapy vary according to the indication for treatment, age, comorbidities, and the cumulative effect of other immunosuppressive agents. Hypogammaglobulinemia and late-onset neutropenia are additional risk factors. In the Rituximab in ANCA-associated vasculitis (RAVE) and Rituximab versus cyclophosphamide in ANCA-associated vasculitis (RITUXVAS) trials of AAV, severe infection was described in 7% and 18%, respectively, and somewhat surprisingly, did not differ from the CYC arm (22,42) (Table 4). By contrast, in a retrospective study (n=370) of autoimmune disorders (mostly rheumatoid arthritis, LN, and AAV), serious infection events occurred in only 3.7% (43). A recent study (n=98) of primary and secondary glomerular disease described an overall infection rate of 21.6 per 100 patient years (44). In other glomerular diseases, lower infection rates were described. No infections were described in cohorts of membranous nephropathy (n=100) (40), minimal change disease (n=17) (45), and only a single case in another cohort of membranous nephropathy (n=15) (46).

Table 4.

Selected randomized controlled trials of ANCA-associated vasculitis

Study	Study Design	Efficacy	Adverse Events
Jayne et al. (68) (CYCAZAREM, maintenance AAV)	Prospective, multicenter, RCT; sample size: 155; intervention: oral CYC (1.5 mg/kg) versus oral AZA (2 mg/kg); end point: relapse rate; follow-up: 18 mo	No difference in relapse rate		AZA (n=71)	CYC (n=73)
			SAEs	12 (28%)	15 (21%)
			Any infection	13 (18%)	13 (14%)
			Severe infection	4 (6%)	3 (4%)
			Leukopenia	22 (31%)	35 (48%)
			Cystitis	1 (1%)	3 (4%)
			Amenorrhea	1 (1%)	2 (3%)
de Groot et al. (69) (CYCLOPS, AAV)	Prospective, multicenter, RCT; sample size: 147; intervention: pulse IV CYC (15 mg/kg q2–3 weekly) versus oral CYC (2 mg/kg); end point: time to remission; follow-up: 18 mo	No difference in time to remission		Pulse IV CYC (n=76)	Oral CYC (n=71)
			SAE	19 (25%)	31 (44%)
			Serious infections	7 (9%)	10 (14%)
			Other infection	20 (26%)	21 (30%)
			Leucopenia	20 (26%)	33 (46%)
			Alopecia	0	2 (3%)
			Amenorrhea	1 (1%)	0
			New/worse diabetes	8 (11%)	8 11%)
Harper et al. (70) (CYCLOPS, AAV)	Follow-up: median 43 mo	Relapse rate increased in pulse IV CYC group (39.5% versus 20.8%)		Pulse IV CYC (n=76)	Oral CYC (n=71)
			Severe infection	19 (25%)	15 (21%)
			Cancer	8 (11%)	6 (8%)
			Cardiovascular event	6 (8%)	3 (4%)
			New onset diabetes	8 (11%)	5 (7%)
			Fracture	6 (8%)	3 (4&)
Hiemstra et al. (71) (IMPROVE, maintenance AAV)	Prospective, multicenter RCT (2002–2009); sample size: 156; intervention: oral AZA (2 mg/kg) versus oral MMF (2 g/d); end point: relapse-free survival; follow-up: 39 mo	AZA was superior to MMF (relapse rate 38% versus 55%)		AZA (n=80)	MMF (n=76)
			SAE	13 (16%)	8 (11%)
			Any infection	17 (21%)	12 (16%)
			Severe infection	8 (10%)	3 (4%)
			Leucopenia	7 (9%)	4 (5%)
			Cancer	3 (4%)	1 (1%)
			Gastrointestinal	8 (10%)	8 (11%)
Stone et al. (22) (RAVE, induction AAV)	Prospective, multicenter RCT (2004–2008); sample size: 197; intervention: RTX (375 mg/m2×4) versus oral CYC (2 mg/kg); end point: remission rate; follow-up: 6 mo	No difference in primary end point; RTX superior in relapsing disease		RTX (n=99)	CYC (n=98)
			SAE	11 (11%)	5 (5%)
			Severe infection	7 (7%)	7 (7%)
			Leucopenia	3 (3%)	10 10%)
			Cancer	1 (1%)	1 (1%)
			Gastrointestinal	8 (8%)	8 (8%)
			Hemorrhagic cystitis	1 (1%)	1 (1%)
			Venous thrombosis	6 (6%)	9 (9%)
			Death	1 (1%)	2 (2%)
Jones et al. (42) (RITUXVAS, induction AAV)	Prospective, multicenter RCT (2006–2007); sample size: 44; intervention: RTX (375 mg/m2×4) versus IV CYC (3–6 m), then maintenance AZA (2 mg/kg); end point: sustained remission rate; follow-up: 12 mo	No difference in sustained remission rate		RTX (n=33)	CYC/AZA (n=11)
			SAE	14 (42%)	4 (36%)
			Any infection	12 (36%)	7 (64%)
			Severe infection	6 (18%)	3 (27%)
			Hypogammaglobulinemia	1 (3%)	0
			Neutropenia	2 (6%)	1 (9%)
			Infusion reaction	2 (6%)	0
			Cancer	2 (6%)	0
			Death	3 (9%)	1 (9%)
Guillevin et al. (21) (MAINRITSAN, maintenance AAV)	Prospective, multicenter RCT (2008–2010); sample size: 115; intervention: RTX (500 mg×2, then 6 monthly) versus AZA 2 mg/kg, then 1.5 mg/kg); end point: major relapse; follow-up: 28 mo	RTX was superior to AZA in preventing relapse (5% versus 29%)		RTX (n=57)	AZA (n=58)
			SAE	25 (44%)	25 (43%)
			Severe Infection	11 (19%)	8 (14%)
			Pneumocystis jirovecii	1 (2%)	0
			Leukopenia	0	6 (10%)
			Cancer	0	2 (3%)
			Death	0	2 (3%)

CYCAZAREM, Cyclophosphamide versus Azathioprine for Early Remission Phase of Vasculitis; AAV, ANCA-associated vasculitis; RCT, randomized controlled trial; CYC, cyclophosphamide; AZA, azathioprine; SAE, serious adverse event; CYCLOPS, Pulse versus daily oral cyclophosphamide for induction of remission in ANCA-associated vasculitis; IV, intravenous; IMPROVE, International Mycophenolate Mofetil Protocol to Reduce Outbreaks of Vasculitides; MMF, mycophenolate mofetil; RAVE, Rituximab in ANCA-Associated; RTX, rituximab; RITUXVAS, Rituximab versus cyclophosphamide in ANCA-associated vasculitis; MAINRITSAN, Rituximab versus azathioprine for maintenance in ANCA-associated vasculitis.

Certain opportunistic infections are of concern with RTX therapy. Progressive multifocal leukoencephalopathy (PML) is a rare, but often fatal, neurologic disease caused by reactivation of the John Cunningham polyomavirus, described in patients taking RTX, and a “black box” warning is included in the product label information in the United States. The large majority of cases are described in patients treated for lymphoma, who are confounded by other risk factors, including concomitant chemotherapy and the underlying hematologic malignancy (47). Multiple case reports of PML have also been described in patients with LN (47,48). In a database study of hospital discharges, PML occurred in 4 per 100,000 patients with SLE (49). The overall intensity of immunosuppression is critical and many cases of PML were described in patients with LN treated with alkylating agents, and not RTX.

PJP is more commonly associated with impaired T cell immunity (e.g., CYC, HIV infection), but is rarely described with RTX. Concomitant high corticosteroid use increases the risk. In AAV, one patient (n=57) developed PJP in the RTX arm of the MAINRITSAN study (Table 4) (21). In a cohort of patients with AAV (n=53), one fatal case of PJP was described 2 months after discontinuation of prophylaxis (50). Similarly, two cases of PJP were described in a cohort of 80 patients with AAV (51). In LN, the incidence in hospitalized patients may be similar, and is estimated at 1% (52).

Reactivation of HBV is listed as a product black box warning by the FDA, and screening for hepatitis B surface antigen and anti-hepatitis B core should be performed before initiation of treatment. If positive, antiviral prophylaxis with entecavir or tenofovir should be started and continued for a least 1 year after the last RTX dose, in view of late immune reconstitution.

Hypogammaglobulinemia

Hypogammaglobulinemia is typically a late complication of RTX therapy. In patients with AAV, severe hypogammaglobulinemia (<3–4 g/L) was described in 4%–10% of patients (53–55). In these patients, infection risk is increased, especially pulmonary infections, and intravenous Ig replacement may be required. This complication appears more commonly in AAV, where risk factors include prolonged therapy with RTX, prior treatment with CYC, and low IgG levels before treatment. IgG levels should be checked in high-risk patients before treatment with RTX, and in those who develop infectious complications.

Late-Onset Neutropenia

This is defined as an absolute neutrophil count <1.5×10⁹/L, occurring more than 1 month (usually 2–6 months) after the last RTX infusion. It occurs in patients with AAV, but rarely in other forms of primary glomerular disease. The incidence ranges from 6.5% to 11.9% in patients with AAV treated with RTX (56–58). The exact mechanism is unclear, but bone marrow biopsy studies have documented myeloid maturation arrest (59). It is often asymptomatic and recovers spontaneously, but granulocyte colony-stimulating factor may be considered when associated with infection. Notably, after recovery, re-exposure to RTX rarely leads to recurrent neutropenia (56,57).

Eculizumab

Eculizumab is a humanized IgGκ mAb that blocks the cleavage of complement C5, inhibiting formation of the membrane attack complex (C5b-9) and release of the anaphylatoxin C5a. It is used in complement-mediated disorders such as C3 glomerulopathy and thrombotic microangiopathy. Notably, multiple new complement inhibitors are currently in development, with many acting more proximally in the complement cascade, and these will likely have very different adverse event profiles.

Infectious Complications

The clearance of encapsulated organisms is heavily dependent on C5b-9. Infection with Streptococcus pneumoniae or Haemophilus influenza is less common, possibly related to opsonization, but the risk of infection with Neisseria meningitidis is markedly increased, even in those who have been vaccinated (60). Starting at least 2 weeks before therapy, patients should receive two doses (2 months apart) of the quadrivalent meningococcal conjugate vaccine (MenACWY), with a booster dose every 5 years. Simultaneous immunization with a recombinant meningococcal B vaccine (MenB) is also recommended. Chemoprophylaxis with penicillin V or ciprofloxacin should be used for the first 4 weeks of therapy. Disseminated gonococcal infection has also been described and sexually active patients should be counseled. Upper respiratory tracts infections may also be more common (60).

Other Complications

Eculizumab is generally well tolerated and infusion reactions are rare. The development of low titer neutralizing antibodies to eculizumab is described (approximately 2%), but does not seem to interfere with clinical response.

Recommendations for Screening and Prophylaxis

Screening for Latent Infection

Before starting therapy with immunosuppression, screening for HBV, HCV, HIV, latent tuberculosis, and strongyloidiasis should be considered (Table 5). The patient should be assessed for evidence of mucosal candidiasis.

Table 5.

Screening and prophylaxis with immunosuppression in glomerular diseases

Assessment	Recommendation
Infection screening
Hepatitis B	HBsAg, anti-HBc antibody
Hepatitis C	Anti-HCV antibody
HIV	Anti-HIV antibody±HIV p24 antigen
Latent tuberculosis	Chest x-ray, tuberculin skin test or IFN-γ release assay
Strongyloidiasis	Anti-strongyloides antibody
Vaccination	Per CDC guidelines (62) (avoid live vaccines during and within 4 wk of starting immunosuppression)
Infection prophylaxis
Pneumocystis jirovecii	Cotrimoxazole or dapsone, atovaquone, inhaled pentamadine
Candida	Clotrimazole troche or weekly fluconazole
Other prophylaxis
Osteoporosis	Calcium and vitamin D; ±bisphosphonates
Gastroprophylaxis	Histamine (H2) antagonist, proton pump inhibitor

HBsAg, hepatitis B surface antigen; HBc, hepatitis B core; HCV, hepatitis C virus; CDC, Centers for Disease Control.

Vaccination

The optimal time for vaccination is at least 4 weeks before starting treatment with immunosuppression, but this is not often possible in glomerular disease. Notably, after RTX or other anti-B cell therapies, the response to vaccination may be markedly attenuated. Vaccination should be performed according to latest Centers for Disease Control guidelines for immunocompromised patients (61). In general, patients should be vaccinated according to the standard Centers for Disease Control annual schedule, including inactivated pneumococcal and influenza vaccines, but avoiding live vaccines such as varicella. A new inactivated recombinant zoster vaccine became available in the United States in 2017.

Infection Prophylaxis

The question of when to use cotrimoxazole or other PJP chemoprophylaxis is difficult for clinicians caring for patients with glomerular disease. PJP has been described with GC therapy at doses as low as 16 mg/d for 8 weeks (62), but is more common at higher doses, and when GCs are used in combination with other immunosuppression. Impaired kidney function and hypoalbuminemia have been described as risk factors for PJP infection in patients with IgA nephropathy (63). Prophylaxis is typically used in AAV, and often in LN in the setting of high-dose steroids with additional immunosuppressive agents. In minimal change disease/FSGS/IgA nephropathy when high-dose GC monotherapy is used, prophylaxis should also be considered. Candida prophylaxis with clotrimazole troche or weekly oral fluconazole should also be considered in patients taking high-dose GCs (Table 5).

Other Prophylaxis

Recommendations for osteoporosis prophylaxis are described above. Patients treated with high-dose GCs are at risk for gastritis and peptic ulcer disease, and prophylaxis with histamine antagonists or proton pump inhibitors may be considered.

Disclosures

None.