The spectrum of acute renal failure in IgA nephropathy

Abstract

Background: Acute renal failure rarely complicates the course of IgA nephropathy. In this study, we have tried to define the mode of presentation, the spectrum of morphology, and the prognostic factors for renal outcome. Methods: Twenty patients with biopsy-proven IgA nephropathy who developed acute renal failure were identified from 2000 to 2009 at a medical center in Taiwan. The patients' records were retrospectively reviewed with respect to clinical presentation, morphology of renal biopsy, and outcomes. Results: On histology, glomerular crescents were present in 11 patients (55%), acute tubular necrosis was identified in 11 patients (55%), acute interstitial nephritis was seen in 4 patients (20%), and extensive tubular red blood cell casts were present in 4 patients (20%). At the end of follow-up, 2 patients (10%) had died, 11 patients (55%) were in remission, and 7 patients (35%) developed end-stage renal disease. The prognostic factors for renal outcome were peak serum creatinine, dialysis support requirement, morphology (prominent glomerular/tubular injury), percentage of glomeruli affected by crescents, and interstitial infiltration (p = 0.04, <0.001, 0.013, 0.05, 0.02, respectively). Conclusions: Our findings suggested that there were four pathogenic mechanisms involved in IgA nephropathy with acute renal failure including (1) crescentic IgA nephropathy; (2) acute tubular necrosis associated with microhematuria and red blood cell casts occluding tubules; (3) acute tubular necrosis not related to microhematuria; and (4) acute interstitial nephritis, apparently induced by drugs. In general, patients with prominent tubular injury had a much higher remission rate than patients with prominent glomerular injury.

Keywords:

• acute renal failure
• hematuria
• IgA nephropathy
• renal biopsy
• prognosis

INTRODUCTION

IgA nephropathy is the most prevalent glomerular disease throughout the world.¹ In a Chinese study of 13,519 renal biopsies, for example, IgA nephropathy constituted 45% of all cases of primary glomerulonephritis.² The importance of the disease is further highlighted by the fact that 15–40% of the patients eventually progress to end-stage renal disease (ESRD).³ Patients with IgA nephropathy typically present in one of three ways. Approximately 40–50% present with one or recurrent episodes of gross hematuria. It is presumed, although not proven, that the first episode represents the onset of the disease. Another 30–40% have microscopic hematuria and usually mild proteinuria and are incidentally detected on a routine examination. In these patients, the disease is of uncertain duration. Less than 10% present with either nephrotic syndrome. Rarely, acute renal failure may complicate the clinical course of IgA nephropathy. This may be due to crescentic IgA nephropathy or due to heavy glomerular hematuria leading to tubular occlusion by red blood cell casts.^4,5 However, in some instances neither the glomerular involvement nor the extent of tubular obstruction has been thought adequate to provide the complete explanation. An alternative pathogenesis may exist. In this study, we have tried to define the clinical and histological characteristics of acute renal failure in patients with IgA nephropathy.

METHODS

We retrospectively reviewed cases of biopsy-proven IgA nephropathy from July 2000 to June 2009 at a medical center in Taiwan. The diagnosis of IgA nephropathy was based on the demonstration by direct immunofluorescence of IgA as the dominant or co-dominant immunoglobulin in a predominantly mesangial distribution. Patients with clinical or serological evidence for systemic lupus erythematosus or Henoch–Schönlein purpura were excluded. We defined acute renal failure as an increase of serum creatinine of at least 0.5 mg/dL above baseline values. Twenty cases of acute renal failure were identified. For each patient, the following clinical data were compiled from review of written and/or electronic records: age, gender, degree of hematuria, peak serum creatinine (mg/dL), and urinary protein excretion (g/24 hours or protein/creatinine ratio). Scoring of hematuria was on the number of red blood cells per high-power field in sediment and was graded as 0 if 0–5, 1 if 6–20, 2 if 21–50, 3 if 51–100, and 4 if >100. The extent of mesangial proliferation was scored as 1+ if involving <25% of glomeruli, 2+ if involving 25–50%, and 3+ if involving ≥50%. The extent of interstitial infiltration was scored as 1+ if involving <25% of interstitial area, 2+ if involving 25–50%, and 3+ if involving ≥50%. Scoring of the chronic tubulointerstitial damage was based on the percentage of tubular atrophy and interstitial fibrosis and was graded as mild (1+) if involving <25%, moderate (2+) if involving 25–50%, and severe (3+) if involving ≥50%. The intensity of immunofluorescence staining was also graded on a scale of 0 to 3+.

For outcome analysis, complete remission of acute renal failure was defined as normalization of serum creatinine to baseline levels or to ≤1.2 mg/dL for those patients in whom baseline levels of creatinine were unavailable; and partial remission was defined by elevation of serum creatinine above baseline levels or follow-up creatinine >1.2 mg/dL for those patients in whom baseline levels of creatinine were unavailable. ESRD was defined as requiring maintenance dialysis therapy.

Statistical analyses were performed using SPSS version 12 (SPSS Inc., Chicago, IL, USA) for Windows and produced by simple nonparametric test (Mann–Whitney U-test and Fisher's exact test). A p-value <0.05 was considered significant.

RESULTS

There were 15 males and 5 females, with a mean age of 52 ± 23 years (range 9–80 years). All patients were submitted to renal biopsy due to acute renal failure, with mean peak serum creatinine of 6.4 ± 4.1 mg/dL (range 1.4–17.5 mg/dL). Nine out of the 20 patients (45%) required dialysis support during the disease course. The mean 24-hour urinary protein excretion was 3.1 ± 4.7 g (range 0.1–20.5 g) and eight patients (40%) had nephrotic range proteinuria. Eleven patients (55%) had an intercurrent disease, mostly infection (Table 1).

TABLE 1.  Intercurrent diseases in cases of IgA nephropathy with acute renal failure

Bacteremia (n = 7)

Methicillin-sensitive Staphylococcus aureus (osteomyelitis of spine)

Methicillin-sensitive Staphylococcus aureus (septic arthritis of hip)

Methicillin-resistant Staphylococcus aureus (central line infection)

Escherichia coli (peritonitis)

Escherichia coli (liver abscess)

Klebsiella pneumoniae (septic arthritis of knee)

Acinetobacter baumannii (undetermined origin)

Pneumonia (n = 2)

Methicillin-resistant Staphylococcus aureus

Pseudomonas aeruginosa

Malignancy (n = 1)

Small cell carcinoma of lung

Autoimmune hemolysis (n = 1)

On histology, glomerular crescents were present in 11 patients (55%) (affected ≥30% of viable glomeruli in 9 patients, affected ≥50% of viable glomeruli in 6 patients). Acute tubular necrosis with luminal ectasia, epithelial flattening and simplification, and loss of brush border was identified in 14 patients (70%). Acute interstitial nephritis with extensive mixed inflammatory cells infiltration with eosinophils, and even tubulitis, was seen in four patients (20%). Red blood cell casts in tubules were present in four patients (20%).

Except 2 patients who died of sepsis (1 pneumonia, 1 central line infection) within 3 months after diagnosis, the remaining 18 patients were followed for a mean of 27 ± 23 months. Eleven patients (55%) entered remission (complete and partial) and seven patients (35%) with ESRD had to be submitted to chronic dialysis. By univariate analysis, the prognostic factors for renal outcome were peak serum creatinine, dialysis support requirement, morphology (prominent glomerular/tubular injury), percentage of glomeruli affected by crescents, and interstitial infiltration (p = 0.04, <0.001, 0.013, 0.05, 0.02, respectively) (Table 2).

TABLE 2. Clinical and histological predictors of remission: univariate analysis

Variables	Remission (n = 11) [mean ± SD or no. (%)]	No remission (n = 7) [mean ± SD or no. (%)]	p-Value
Age (years)^a	46.1 ± 23.3	53.7 ± 23.4	0.497
Gender (male/female)^b	9/2	4/3	0.326
Proteinuria (g/24 hours)^a	1.2 ± 1.3	5.7 ± 7.1	0.135
Peak serum creatinine (mg/dL)^a	4.0 ± 2.4	9.9 ± 4.2	0.004
Dialysis support requirement^b	0 (0)	7 (100)	<0.001
Morphology (glomerular/tubular injury)^b	2/9	6/1	0.013
% Glomerular obsolete^a	8.6 ± 9.2	23.1 ± 16.8	0.063
% Glomeruli affected by crescents^a	9.2 ± 14.0	63.0 ± 34.9	0.005
Interstitial infiltration (2+ or 3+)^b	0 (0)	5 (71.4)	0.002

^{Notes: a}Data were analyzed by Mann–Whitney U-test.

^bData were analyzed by Fisher's exact test.

*p-Value < 0.05.

We subdivided the 20 patients into two groups. Group 1 consisted of nine patients with prominent glomerular injury presenting as crescentic IgA nephropathy (defined as ≥30% of glomeruli affected by crescent). In this group, six patients had diffuse crescentic glomerulonephritis (generally defined as ≥50% of glomeruli affected by crescent). One patient also had tubulointerstitial nephritis. In this group, three patients had intercurrent infection including one osteomyelitis of spine (methicillin-sensitive Staphylococcus aureus), one septic arthritis of hip (methicillin-sensitive S. aureus), and one pneumonia (methicillin-resistant S. aureus). Hypocomplementemia was present in two patients. In this group, induction therapy consisted of steroids and cyclophosphamide in five patients (three with intravenous cyclophosphamide, two with oral cyclophosphamide) and steroids alone in two patients. In addition to immunosuppressive therapy, one patient underwent plasmapheresis. The remaining two patients did not receive immunosuppressive therapy because of concurrent sepsis. Six patients developed ESRD, two patients achieved remission, and one patient (the patient with pneumonia) died of sepsis.

Group 2 consisted of 11 patients with prominent tubular injury and could be further subdivided into three subgroups. Subgroup 1 consisted of four patients with acute tubular necrosis associated with microhematuria and red blood cell casts occluding tubules. The mean duration of microhematuria was 2.3 ± 1.3 weeks (range 1–4 weeks). In this subgroup, two patients had intercurrent infection including one liver abscess (Escherichia coli) and one bacteremia with undetermined origin (Acinetobacter baumannii). Treatment consisted of hydration and oral dipyridamole, and antimicrobial agents were administrated in the two patients with concurrent infection. All patients got complete recovery of renal function. Subgroup 2 consisted of four patients with acute tubular necrosis not related to microhematuria. In this subgroup, one patient had concurrent autoimmune hemolysis and three patients had intercurrent infection including one central line infection (methicillin-resistant S. aureus), one peritonitis (E. coli), and one septic arthritis of knee (Klebsiella pneumoniae). Hypocomplementemia was noted in one patient. Treatment consisted of steroids in the patient with autoimmune hemolysis, and antimicrobial agents were administrated in the three patients with concurrent infection. Three patients achieved complete remission and one patient (the patient with central line infection) died of sepsis. Subgroup 3 consisted of three patients with acute interstitial nephritis. In this subgroup, one patient had concurrent pneumonia (Pseudomonas aeruginosa) and received antibiotic treatment, and one patient with intercurrent diesease of small cell lung cancer ever took nonsteroidal anti-inflammatory drug. Two patients recovered from renal function and one patient (the patient with small cell lung cancer) developed ESRD.

Comparison of the clinical and histological parameters between the two groups was listed in Table 3. In comparison of clinical parameters, the glomerular injury group had a greater urinary protein excretion (5.7 ± 6.0 vs. 0.9 ± 0.9 g/24 h, p = 0.014), a higher level of peak serum creatinine (8.8 ± 4.4 vs. 4.5 ± 2.8 mg/dL, p = 0.025), and more frequently had requirement for dialysis support (77.8% vs. 18.2%, p = 0.022) than the tubular injury group. In comparison of histological parameters, the glomerular injury group had a higher degree of mesangial hypercellularity (score 2.3 ± 0.5 vs. 1.5 ± 0.5, p = 0.009) and significantly more percentage of glomeruli affected by crescents (64.3 ± 25.4% vs. 3.2 ± 7.6%, p < 0.001). Regarding renal outcome, the tubular injury group had a much higher remission rate than the glomerular injury group (81.8% vs. 22.2%, p = 0.022).

TABLE 3. Comparison of clinical and histological parameters between patients with prominent glomerular injury and patients with prominent tubular injury

Variables	Glomerular injury group (n = 9) [mean ± SD or no. (%)]	Tubular injury group (n = 11) [mean ± SD or no. (%)]	p-Value
Clinical parameters
Age (years)^a	51.3 ± 19.9	52.2 ± 26.7	0.849
Gender (male/female)^b	6/3	9/2	0.617
Hematuria (score)^a	3.6 ± 0.9	3.3 ± 1.6	1.000
Proteinuria (g/24 hours)^a	5.7 ± 6.0	0.9 ± 0.9	0.014
Peak serum creatinine (mg/dL)^a	8.8 ± 4.4	4.5 ± 2.8	0.025
Dialysis support requirement^b	7 (77.8)	2 (18.2)	0.022
Intercurrent disease^b	3 (33.3)	8 (72.7)	0.175
Histological parameters
Globally sclerotic glomeruli (%)^a	15.8 ± 12.2	14.5 ± 15.4	0.704
Mesangial cellularity (score)^a	2.3 ± 0.5	1.5 ± 0.5	0.009
% Glomeruli affected by crescents^a	64.3 ± 25.4	3.2 ± 7.6	<0.001
Acute tubular injury^b	5 (55.6)	11 (100)	0.026
Red blood cell casts in tubules^b	0 (0)	4 (36.4)	0.094
Interstitial infiltration (2+ or 3+)^b	4 (44.4)	1 (9.1)	0.127
Outcome
Death^b	1 (11.1)	1 (9.1)	1.000
Remission^b	2 (22.2)	9 (81.8)	0.022
End-stage renal disease^b	6 (66.7)	1 (9.1)	0.017

^{Notes: a}Data were analyzed by Mann–Whitney U-test.

^bData were analyzed by Fisher's exact test.

*p-Value < 0.05.

DISCUSSION

Our findings suggested that there might be four pathogenic pathways leading to acute renal failure in IgA nephropathy. The first mechanism was immune-mediated glomerular injury associated with extensive crescent formation. Several studies have shown a positive correlation between the percentage of glomeruli affected by crescent formation and the severity of initial and subsequent renal failure. In our series, this group had a higher degree of proteinuria and severity of renal failure. Several patients also had tubular red blood cell casts, but the extent was mild (involving <25% of tubular area). Six patients with ≥50% of glomeruli affected by crescent were generally classified as diffuse crescentic glomerulonephritis. These patients had the worst prognosis even though most patients had received aggressive immunosuppressive therapy.

The second mechanism was acute tubular necrosis associated with microhematuria and red blood cell casts occluding tubules. Previous studies reported the very high incidence of glomerular crescents seen in biopsies taken at the time of macroscopic hematuria.^6,7 However, as these authors have subsequently pointed out, the glomerular changes seen in the renal biopsy were not sufficient to explain themselves the renal function impairment. An alternative mechanism proposed that glomerular hematuria with red blood cells in renal tubules may cause intratubular obstruction and subsequent tubular damage. Acute renal failure from gross hematuria in patients with glomerulonephritis was first reported by Kincaid-Smith et al., who demonstrated that the renal function impairment was due to tubular necrosis which came along with intratubular red blood cell casts, phagocytosis of erythrocytes by tubular epithelium, and extensive interstitial edema, rather due to the glomerular changes.^8,9 Subsequently, Praga et al. prospectively studied 29 episodes of gross hematuria in 21 patients with IgA nephropathy.⁷ They found that acute renal failure was significantly correlated with a longer during of gross hematuria, higher amounts of intratubular red blood cell casts, and a more extensive tubular necrosis. Fogazzi et al. found no back diffusion of Tamm-Horsfall protein into the glomerular Bowman's space, so they suggested that acute renal failure due to gross hematuria might be not associated with intratubular obstruction.⁵ Furthermore, the correlation between duration of microhematuria and final renal function has been pointed out in a recent study.¹⁰ Because the tubulointerstitial changes were found particularly in areas with the most extensive red blood cells accumulation, one may assume that it was an expression of the toxic effects of hemoglobin or some other, yet unidentified, substances released from red blood cell casts. The severity and the duration of acute renal failure varied from patient to patient, some of whom required temporary dialysis, but in all cases spontaneous recovery of renal function was seen. In our series, 4 out of the 20 biopsies (20%) showed diffuse tubules filled by red blood cells. Among them, two patients had a small percentage of glomeruli affected by crescent (24 and 11%). The renal prognosis was excellent in this group and all of the four patients entered complete remission.

However, Lupo et al. found only 2 out of 14 episodes in their series were accompanied by intratubular red blood cell casts formation sufficient to explain the development of acute renal failure.¹¹ Packham et al. also pointed out that red blood cell casts were present in only a minority of tubules, which was thought inadequate to explain the severe deterioration in renal function.¹² This is consistent with our experience. In our series, there were four biopsies presenting with acute tubular necrosis in which neither glomerular crescent formation nor intratubular red blood cell casts could be found. Moreover, acute interstitial nephritis was present in three biopsies. In these cases, the tubular injury was most likely secondary to infection and/or exposure of nephrotoxic agents such as antibiotics and nonsteroidal anti-inflammatory drugs and the IgA nephropathy might be coincident. However, the mean age of patients subdivided into prominent tubular injury group was 52 years. Since idiopathic IgA nephropathy is less common to present at this age and these patients had no pre-existing renal manifestations such as proteinuria and hematuria, we suppose that the IgA nephropathy was associated with infection, which was increasingly identified in the recent years.^13–16It has been speculated that enterotoxins produced by methicillin-resistant S. aureus may serve as superantigens which contribute to this type of glomerulonephritis.¹⁷ However, whether similar mechanism exists in other bacterial infection is not known. None of our patients had subepithelial hump-shaped deposits. According to previous reports, IgA-dominant postinfectious glomerulonephritis was difficult to differentiate from idiopathic IgA nephropathy by histology alone and electron microscopy often could not show characteristic subepithelial hump-shaped deposits.^14,18–21However, these patients had clinical features that favor postinfectious glomerulonephritis over idiopathic IgA nephropathy include older age, documented infection, and presentation of acute renal failure which is unusual in idiopathic IgA nephropathy without extensive crescent formation.^14,15 Furthermore, our findings suggested that acute tubular necrosis was no more simply caused by infection per se in this setting and could be caused by the consequence of IgA nephropathy (glomerular hematuria leading to tubular obstruction by red blood cell casts). These findings might provide some therapeutic implications.

In conclusion, our finding suggested that there were four pathogenic mechanisms involved in IgA nephropathy with acute renal failure including (1) crescentic IgA nephropathy; (2) acute tubular necrosis associated with microhematuria and red blood cell casts occluding tubules; (3) acute tubular necrosis not related to microhematuria; and (4) acute interstitial nephritis, apparently induced by drugs. Moreover, the severity of renal failure and prognosis largely depended on the histological presentation. In general, patients with prominent tubular injury had a much higher remission rate than patients with prominent glomerular injury. This also highlighted the importance of renal biopsy in IgA nephropathy patients presenting with acute renal failure.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.