A 66-year-old woman presented with acute onset of painless gross hematuria and generalized fatigue of 10 days' duration. The patient was asymptomatic until she started to notice blood in her urine, not associated with dysuria or dribbling. Shortly afterward, she developed oliguria, and swelling of both legs, associated with mild exertional dyspnea, diffuse body aches, and fatigue without subjective fevers or chills. She had no oral or nasal ulcers, alopecia, photosensitivity, joint pain, joint swelling or redness, or symptoms of Raynaud's phenomenon. She had mild exertional shortness of breath, but without chest pain or cough. She denied abdominal pain or any change in her bowel habits. Beside the gross hematuria, she denied any other urinary symptom. She denied weight loss, recent travels, or sick contacts. On evaluation in the emergency department, she was afebrile but hypertensive (blood pressure 188/98 mm Hg), which was a new symptom for her. She had periorbital edema as well as bilateral lower extremity pitting edema to the knees. Laboratory evaluation in the emergency room showed normocytic anemia, blood urea nitrogen of 44 mg/dl (normal range 7–18), and a creatinine level of 3.8 mg/dl (normal range 0.6–1.2). Her estimated creatinine clearance was 11.9 ml/minute/1.73 m2 (normal range 80–125) with a fractional excretion of sodium of 2.5%. Her creatinine was checked 3 weeks prior to admission during a routine checkup and found to be elevated at 1.4 mg/dl. She had a potassium level of 5.5 mEq/liter (normal range 3.5–5.1), a phosphorus level of 6.1 mg/dl (normal range 2.8–4.1), and a total calcium level of 8.2 mg/dl (normal range 8.4–10.2). Her urine analysis at presentation revealed 3+ proteins, 3+ blood, gross red blood cells (RBCs), 1–3 white blood cells, and trace leukocyte esterase. A chest radiograph performed on admission showed cardiomegaly with mild congestion. Kidney ultrasound did not show any hydronephrosis, kidney mass, or evidence of nephrolithiasis. She was admitted to the general medical floor with a diagnosis of acute kidney injury and nephrotic syndrome.
Her medical history was remarkable for several chronic medical illnesses, including chronic obstructive pulmonary disease (COPD), type 2 diabetes mellitus, bipolar disorder, hypothyroidism, hypertension, and diastolic heart failure. She was non–oxygen dependent, treated only with intermittent albuterol use, and had multiple hospitalizations for COPD exacerbations/pneumonias requiring mechanical ventilation, but was never taking outpatient steroids. She was receiving insulin therapy, with a most recent glycosylated hemoglobin of 7.2% (normal value <6.5%) checked 3 weeks prior to her admission. She was taking amitriptyline and trazodone for bipolar disorder. Hypothyroidism was treated with levothyroxine. Her primary care doctor held her diuretic and angiotensin-converting enzyme inhibitor 3 weeks prior because of her elevated creatinine level of 1.4 mg/dl. She took no other medications, including over-the-counter medications. She had had 4 full-term pregnancies. She denied any history of spontaneous abortions or thromboembolic diseases. Her only surgery was a cholecystectomy performed approximately 10 years ago.
Social and family history
The patient was born in the US and lived in Rhode Island for many years. She was divorced and lives alone in her own apartment. She stopped smoking approximately 3 years ago. Prior to that, she smoked one pack a day for 40 years. She used to engage in binge drinking of alcohol on a weekly basis, but quit drinking 5 years ago. She denied any illicit drugs. Her family history was remarkable for a mother who died of kidney failure due to “uncontrolled diabetes.”
In the emergency department, the patient appeared pale and tired, but not in acute distress. She was afebrile with an oral temperature of 98.1°F. Vital signs revealed a blood pressure of 188/98 mm Hg, a pulse of 80 beats per minute, a respiratory rate of 18 breaths per minute, and oxygen saturation of 92% on room air. She had no alopecia or nasal or oral ulcers. There was no cervical, axillary, or inguinal lymphadenopathy. She had decreased air entry bilaterally upon lung auscultation. Her heart auscultation showed regular rate and rhythm and normal S1 and S2, without murmurs, rubs, or gallops. Her abdomen was soft, nontender, and slightly distended. There was no organomegaly. She had bilateral pitting lower extremity edema to the knees. Her peripheral pulses were 2+ and symmetric. She had normal range of motion of all of her joints without evidence of synovitis or effusion. She had no skin lesions or rashes. She had normal nails and capillary nailfolds. Her neurologic examination was normal without any focal deficits.
Results of the initial laboratory evaluation are shown in Table 1. Her blood smear showed anisocytosis of the RBCs with occasional target cells and basophilic stippling.
Table 1. Laboratory studies*
WBC = white blood cell; fl = femtoliters; RBC = red blood cell; hpf = high-power field; cANCA = cytoplasmic antineutrophil cytoplasmic antibody; pANCA = perinuclear ANCA; GBM = glomerular basement membrane; HBsAg = hepatitis B surface antigen; HBcAg = hepatitis B core antigen; HIV-1 = human immunodeficiency virus type 1.
The patient's hematuria and creatinine increased daily until the creatinine level peaked at 6.0 mg/dl on the fourth day of her hospitalization. The patient underwent multiple serologic and laboratory evaluations, including antinuclear antibody, antineutrophil cytoplasmic antibodies (ANCAs), human immunodeficiency virus serology, hepatitis B and C, anti–glomerular basement membrane (anti-GBM) antibody, cryoglobulins, and complements, which were all negative (Table 1). She continued to have rapidly progressive renal failure. A nephrology consultation was requested upon admission.
The patient, a 66-year-old woman with a medical history significant for hypertension and diabetes mellitus, presented with generalized weakness, body aches, as well as rapidly progressive renal failure in the setting of painless gross hematuria and nephrotic syndrome.
Our 66-year-old female patient met the criteria for the clinical syndrome of rapidly progressive glomerulonephritis (RPGN), defined as loss of renal function of at least 50% in <3 months and evidence of glomerular injury, with hematuria and proteinuria ([1, 2]). The pathologic correlate of the clinical syndrome of RPGN is crescentic glomerulonephritis, in which crescents are formed by extracapillary proliferation of the epithelial cells, affecting >50% of the glomeruli ([3, 4]). The severity of the disease is related to the degree of crescent formation, which is a nonspecific response to severe injury to the glomerular capillary wall (). RPGN is caused by 3 major categories of disease: type 1 or anti-GBM disease (15% of cases), type 2 or immune complex glomerulonephritis (approximately 30%), type 3 or pauci-immune crescentic glomerulonephritis (approximately 60%), or type 4, which includes features of both types 1 and 3 (approximately 5%) ().
The most common cause of RPGN is the pauci-immune necrotizing glomerulonephritides, including ANCA-associated vasculitis (AAV), microscopic polyangiitis, granulomatosis with polyangiitis (Wegener's), and rarely eosinophilic granulomatosis with polyangiitis (Churg-Strauss) ([4, 5]). The majority of patients with renal-limited vasculitis are ANCA positive, with 75–80% having myeloperoxidase (MPO) ANCA, and may have or develop symptoms of underlying vasculitis ([6, 7]). The ANCA tests have a sensitivity of up to 90% and a specificity of up to 98%. In this case, negative ANCA, negative MPO, and negative proteinase 3 (PR3) tests in the absence of any other systemic vasculitis involvement (cutaneous lesions, upper and lower respiratory tracts, mononeuritis multiplex) along with the absence of asthma and peripheral eosinophilia in the presence of renal involvement provide evidence against the diagnosis of AAV, but do not rule it out. ANCA-negative, pauci-immune RPGN is considered part of this spectrum of diseases, and may have comparable clinical features (). The immune complex glomerulonephritides include the hypocomplementemic and normocomplementemic groups. The absence of systemic manifestations and the presence of normal complement levels would argue against systemic lupus erythematosus (SLE), poststreptococcal glomerulonephritis, membranoproliferative glomerulonephritis, endocarditis, visceral abscess, and shunt nephritis seen in glomerulonephritides associated with low C3 levels. The absence of a preceding upper respiratory tract or skin infection would not explain poststreptococcal glomerulonephritis as in this patient's disease. SLE is most often manifested in young women, but can also occur in the elderly and in men. This elderly woman had none of the classic features of SLE and her serologies remained negative. Similarly, there was no clinical evidence of endocarditis or visceral abscess or shunt nephritis. She had normal C4 levels in conjunction with the absence of cryoglobulins, which would argue against cryoglobulinemic glomerulonephritis. Normocomplementemic glomerulonephritides include IgA nephropathy, IgA vasculitis (Henoch-Schönlein) (IgAV), and fibrillary glomerulonephritis. This elderly patient had none of the classic features of IgAV such as palpable purpura and gastrointestinal manifestations, including abdominal pain, bleeding, infarction, and intussusceptions (). IgA nephropathy is the most common cause of glomerulonephritis, and symptoms can range from asymptomatic hematuria to RPGN (). Although our patient is a woman in her 60s with a history of heavy smoking, the absence of pulmonary symptoms and negative antibodies to NC1 of type IV collagen in the GBM excludes Goodpasture's syndrome as the cause of her renal disease ([11, 12]).
Urinary tract infection was very unlikely at this point despite the presence of bacteriuria; pyuria was also absent, as were urinary symptoms except for painless gross hematuria (). The latter was later confirmed by a negative urine culture. Bacteriuria could be explained by vaginal flora contamination (). Basophilic stipplings of the RBCs found in our patient were believed to be a nonspecific finding, since they might occur in a small percentage of the normal population and are found in a variety of rheumatic diseases (). Furthermore, our patient did not have any risk of lead exposure.
Confronted with a negative serologic assay result, painless gross hematuria, and nephrotic syndrome in the setting of rapidly progressive renal failure, our patient was empirically treated with glucocorticoids and underwent a renal biopsy. The patient was started on a pulse dosage of intravenous methylprednisolone (1,000 mg/day) on day 3 for a total of 3 days, and a kidney biopsy was performed on day 4. Figure 1 shows a representative slide of the kidney biopsy demonstrating focal segmental necrotizing and crescentic glomerulosclerosis. Weak mesangial staining for IgA and C3 was also noted on immunochemistry. An electron microscopy study demonstrated subendothelial edema, interstitial mononuclear inflammatory cells, and the absence of immune complex deposits (Figure 2). A computed tomography scan of the chest and sinuses was negative for granulomas. The patient's ANCA serologies were repeated during her hospitalization and remained negative. Therefore, a diagnosis of ANCA-negative pauci-immune crescentic glomerulonephritis was made. As a result of rapidly progressive renal failure, she was started on hemodialysis every other day, transitioned to oral prednisone (1 mg/kg), and was started on oral cyclophosphamide (CYC) along with Pneumocystis jiroveci prophylaxis. In addition, the patient received a total of 7 sessions of plasmapheresis alternating with days of hemodialysis. The gross hematuria subsided after 4 weeks of immunosuppressive therapy, and a repeat urine analysis revealed a decrease in the number of RBCs to 11–30 per high-power field.
The majority of cases of pauci-immune crescentic glomerulonephritis are attributed to ANCA-associated systemic vasculitis, and a small number of these cases involve renal-limited vasculitis. ANCA plays a major pathogenic role in ANCA-positive pauci-immune glomerulonephritis. However, in a subgroup of patients with pauci-immune crescentic glomerulonephritis, the serologic hallmark of ANCA remains persistently negative.
Limited prevalence data are available on ANCA-negative pauci-immune crescentic glomerulonephritis. A limited number of case series and population-based studies have focused on the clinical spectrum of this disorder ([8, 15-20]). In a 10-year retrospective study, the annual incidence rate was 3.9 patients per 1 million individuals and approximately 27% of affected patients with pauci-immune crescentic glomerulonephritis were ANCA negative (). In a Taiwanese and a Chinese case series, 38% and 33% of patients with pauci-immune crescentic glomerulonephritis were ANCA negative, respectively ([19, 20]). However, these studies lacked a unified definition for “crescentic” and “pauci-immune” glomerulonephritis, which limits the comparability of these results. Similarly, the small number of patients in these 2 case series provided selection bias. Evaluation of larger cohorts from both the European Vasculitis Study Group and the Chapel Hill Group implied that approximately 10–30% of patients with pauci-immune crescentic glomerulonephritis lack positive ANCAs ([21, 22]). In both the Taiwanese and the Chinese case series, ANCA-negative pauci-immune crescentic glomerulonephritis presented at a much younger age than those with ANCA-positive disease ([19, 20]). For example, ANCA-negative patients were usually much younger than the ANCA-positive patients (mean ± SD age 45.1 ± 13.5 versus 59.1 ± 17.6 years &lsqbr;&rsqbr; and 39 ± 17 versus 56 ± 14 years &lsqbr;&rsqbr;). However, in a study in the UK, no major differences in age at presentation were found between these 2 groups (). These differences may reflect variations in disease susceptibilities observed in 2 different ethnic populations. ANCA-negative patients had fewer constitutional symptoms (including weight loss, fevers, myalgias, and arthralgias) and fewer extrarenal manifestations (including respiratory, eye, and ear/nose/throat symptoms) than those with positive ANCAs ([19, 20]). However, ANCA-negative patients had a greater degree of proteinuria (mean ± SD 5.47 ± 3.32 gm/24 hours versus 2.23 ± 2.27 gm/24 hours; P < 0.001) and a higher prevalence of nephrotic syndrome (46% versus 8.8%; P < 0.001) than ANCA-positive patients ().
The criteria of pauci-immune crescentic glomerulonephritis are defined as the intensity of glomerular immunoglobulin staining by direct immunofluorescence assay in renal sections on a scale of 0 to ≥4 ([8, 23]). A new pathologic and prognostic classification of ANCA-associated glomerulonephritis was developed in 2010 by an international working group of renal pathologists and validated in both studies of the European Vasculitis Study Group: CYCAZAREM (CYC or azathioprine as a remission therapy for vasculitis) and MEPEX (methylprednisolone versus plasma exchange as adjunctive therapy for severe renal vasculitis) trials (). The classification system was composed of a total of 4 categories () (Table 2). Patients with focal and/or crescentic ANCA-associated glomerulonephritis presented with highly active renal disease, but had a favorable renal outcome (). Patients with a mixed phenotype had an intermediate renal outcome (). However, sclerotic ANCA-associated glomerulonephritis ran the poorest prognosis, with an increased risk of death in the first year of diagnosis (). This classification takes into account patients with ANCA-negative pauci-immune crescentic glomerulonephritis. ANCA-negative patients have more severe glomerular lesions () and severe chronic lesions, including interstitial fibrosis and glomerulosclerosis ([8, 19]), than their ANCA-positive counterparts. These findings have been attributed to a delay in diagnosis by the lack of systemic symptoms and negative ANCAs seen in these patients ([9, 20]). Our patient had focal and crescentic phenotypes; however, the presence of the glomerulosclerosis pattern precludes a poor renal prognosis. Patients with crescentic, mixed, and sclerotic biopsies had a worse estimated glomerular filtration rate at 1 year postbiopsy compared to those with the focal type (). A recent retrospective cohort study evaluating 129 patients from a single center in China reevaluated the new classification of ANCA-associated glomerulonephritis for its predictive power with respect to renal outcome as well as its response to treatment (). The likelihood of progressing to end-stage renal disease was increased and predicted with the ascending sequence of focal, mixed, crescentic, and sclerotic glomerulonephritis (). This differed from the sequence previously posted in the study by Berden et al () (Table 2). This difference was attributed to the fact that in the Chinese cohort, the patients with ANCA-associated crescentic glomerulonephritis had more chronic lesions, and MPO ANCA was more common than PR3 ANCA, which limited the response to the immunosuppressive therapy (). This study confirmed again that chronic lesions are more abundant in patients with MPO ANCA than PR3 ANCA ([26, 27]). Chronic lesions such as interstitial fibrosis and glomerulosclerosis are particularly severe in ANCA-negative patients (), leading to poorer renal outcomes and survival than their ANCA-positive counterparts ([19, 20]).
Table 2. Classification schema for antineutrophil cytoplasmic antibody–associated glomerulonephritis
aPauci-immune staining pattern on immunofluorescence microscopy and >1 glomerulus with necrotizing or crescentic glomerulonephritis on light microscopy are required to meet inclusion criteria in all 4 categories.
No controlled, prospective studies have been conducted on the treatment of patients with ANCA-negative pauci-immune crescentic glomerulonephritis. Treatment protocols are, therefore, based on those for patients who are ANCA positive following the principle of combined remission induction and maintenance strategy based on the stage and activity of AAV for each patient. Randomized controlled trials have been conducted in order to obtain evidence-based recommendations ([28-30]). Daily oral CYC plus corticosteroids remains the gold standard therapy for this group of patients (). Studies have found remission rates between 70% and 100% and early mortality rates of <20% with increased treatment-related morbidity (). Our patient presented with a serum creatinine level <500 μmoles/liter (3.8 mg/dl) and based on these recommendations was treated with prednisone and oral CYC ([30, 31]). Oral and pulse CYC has been shown to be equivalent in terms of clinical response (). In a randomized controlled trial comparing the efficacy of oral CYC (2 mg/kg/day) with intravenous pulse CYC (15 mg/kg every 2 weeks for the first 3 pulses, then every 3 weeks) with the same corticosteroid regimen in both arms (CYCLOPS trial [randomized trial of daily oral versus pulse CYC as therapy for ANCA-associated systemic vasculitis]) (), CYC pulses were found to be associated with fewer episodes of leukopenia and the cumulative dose of the drug was inferior to that found in patients treated with oral CYC, but there was a tendency toward more frequent relapses in the patients treated with pulse CYC ([32-34]). Our patient was placed on daily oral CYC along with corticosteroids, since she had an increased risk of relapse due to her age and rapidly worsening kidney function, as well as the presence of the sclerosis pattern in the tissue specimen. Harper et al evaluated the long-term outcomes of the 148 patients previously enrolled in the CYCLOPS study (). At a median duration of 4.3 years, patient survival was similar between the 2 treatment arms, despite an increased risk of relapse in patients treated with pulse CYC (). In addition, our patient's serum creatinine rapidly worsened during her hospitalization, with a serum creatinine level <500 μmoles/liter (5.7 mg/dl), suggestive of severe and rapidly progressive renal vasculitis, immediately life-threatening disease, and poor renal outcome at a 1-year interval (); therefore, she was treated with a course of plasma exchange (7 sessions in 2 weeks) in conjunction with the prednisone and oral CYC. Plasma exchange has been shown to be beneficial in patients with segmental necrotizing glomerulonephritis and severe renal disease ([36, 37]). The randomized MEPEX trial enrolled 137 patients with a new diagnosis of pauci-immune glomerulonephritis and the results supported the findings that plasma exchange was more effective than methylprednisolone pulses in renal function recovery (). Therefore, plasma exchange is currently the best complementary treatment to immunomodulation in advanced renal disease ([29-31]). Nevertheless, the potential combination of plasma exchange and methylprednisolone in patients with ANCA-negative pauci-immune glomerulonephritis and its use in less severe renal disease remains unclear and requires further investigation.
Recently, rituximab (RTX) has emerged as an alternative for CYC in the remission induction of patients with generalized and severe AAV. RTX, a chimeric monoclonal antibody to CD20, is recommended as an alternative for standard remission induction in newly diagnosed AAV patients. RTX is as effective as CYC in inducing remission in AAV and should be preferred to CYC, at least in younger patients, in order to preserve fertility ([38-40]), or in patients at risk for malignancies (). Furthermore, it may be considered particularly in relapsing patients ([38, 42, 43]). In a subgroup analysis, RTX at 6 months was found to be superior to CYC in inducing remission in those who relapsed than CYC in patients enrolled in the Rituximab in ANCA-Associated Vasculitis (RAVE) trial (). Evaluation over a longer period of time of patients in the RAVE trial with severe AAV found that a single course of RTX was as effective as continuous conventional immunosuppressive therapy for induction and maintenance of remission over an 18-month period (). RTX also has been used to maintain remission in noncontrolled trials, with promising results in an ongoing randomized controlled trial ([40, 42]). Larger prospective studies enrolling ANCA-negative patients with a longer-term followup are required to confirm the potential benefits of RTX to determine its efficacy and safety, and whether patients with ANCA-negative pauci-immune crescentic glomerulonephritis have ANCAs of as yet undiscovered or undetected specificity.
Once the disease is in remission (usually within 3–6 months), azathioprine is as effective as CYC and more effective than mycophenolate mofetil to maintain remission ([28, 29, 31]). Bosch et al reported that azathioprine and daily prednisone can maintain remission in patients who test negative for ANCAs, whereas patients positive for PR3 ANCA should be closely monitored because of a higher probability of relapse (). Methotrexate has been shown to be equivalent to azathioprine, but it should not be used in patients with renal impairment (). How long to continue maintenance therapy is not known, but at least 1 year since remission is recommended.
The pathogenesis of ANCA-negative pauci-immune crescentic glomerulonephritis remains unclear; most of the understanding of the pathogenic mechanism of pauci-immune crescentic glomerulonephritis has been mainly focused on patients with positive ANCAs (). Substantial in vitro and in vivo evidence suggests that ANCAs have an important role in the pathogenesis of vasculitis in AAV, and are not simply a serologic marker of the disease (). Neutrophils are likely to have a major role in ANCA-negative pauci-immune crescentic glomerulonephritis ([8, 46]). Both immunohistochemical and degranulation studies demonstrated more prominent neutrophil activation and degranulation in ANCA-negative patients than in those with positive ANCAs ([47, 48]). Studies are currently being carried out to determine the factors that lead to neutrophil activation. Neutrophils could be activated in 2 ways: first, via antibodies other than ANCAs such as anti–endothelial cell antibodies (AECAs), autoantibodies to human lysosome-associated membrane protein 2 (LAMP-2), or others. Approximately 50% of patients with ANCA-negative pauci-immune crescentic glomerulonephritis had AECAs in their serum, although the prevalence was lower than in patients with ANCAs (). Similarly, LAMP-2 is found in >90% of individuals with active pauci-immune crescentic glomerulonephritis and causes neutrophil activation and endothelial cell apoptosis ([50, 51]). LAMP-2 has strong homology with FimH, an adhesin commonly found in gram-negative bacteria, suggesting that bacterial infection in a susceptible host may induce autoantibodies that result in pauci-immune crescentic glomerulonephritis (). Presently, whether patients without ANCAs have antibodies to LAMP-2 is unclear (). Second, cell-mediated immunity could lead to activation of neutrophils via interleukin-8 (IL-8) through IL-17 producing T helper cell lineage ([53, 54]). Further studies are needed to determine if similar pathogenic mechanism(s) are involved in ANCA-negative pauci-immune crescentic glomerulonephritis as in AAV.
The proportions of patients with ANCA-negative pauci-immune crescentic glomerulonephritis are not considerable, and the existence of these individuals has both important clinical and prognostic implications. ANCA-negative pauci-immune crescentic glomerulonephritis presents with a different clinical spectrum than those with ANCA-positive disease, with fewer constitutional symptoms and extrarenal manifestations delaying the diagnosis. The absence of ANCAs may be the reason for these differences. Negative ANCAs seem to be associated with increased severity of glomerular lesions and poor renal outcomes; therefore, evaluation of patients with RPGN with negative ANCAs is a medical emergency where a renal biopsy should be rapidly performed and the institution of immunosuppressive therapy and plasma exchange should not be delayed.
Although both the RITUXVAS (RTX versus cyclophosphamide in ANCA-associated renal vasculitis) and RAVE trials have provided evidence for targeted therapy to manage ANCA-associated diseases, larger prospective studies enrolling ANCA-negative patients with longer-term followup are needed to confirm the potential benefits of RTX translate into improvements in efficacy and safety, and whether patients with ANCA-negative vasculitis have ANCAs of as yet or undiscovered specificity.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Najem had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Najem, Sfeir, Estrada, Mbuyi, Valicenti, Reginato.
Acquisition of data. Najem, Sfeir, Estrada, Mbuyi, Valicenti, Reginato.
Analysis and interpretation of data. Najem, Sfeir, Reginato.
The photographs of the kidney biopsy are courtesy of M. Barry Stokes, MD, Department of Pathology, Columbia University, New York.