Safety and utility of kidney biopsy in patients with estimated glomerular filtration rate < 30 ml/min/1.73 m²

Abstract

Aim

Kidney biopsy (KBx) is the gold standard for evaluation of kidney disease, but is associated with a higher risk of complications in patients with reduced glomerular filtration rate (GFR). We studied the safety and utility of KBx in patients with eGFR <30 ml/min/1.73 m².

Methods

Consecutive adult patients with eGFR <30 ml/min/1.73 m², who were planned for a KBx and consented to participate were prospectively enrolled. Patients with solitary/transplant kidney or acute kidney injury were excluded. Haemoglobin was checked on the day of KBx and repeated 18–24 h later along with a screening ultrasound. Post-KBx complications were noted and their risk-factors analysed. The utility of the KBx was graded as effecting significant, some, or no change to subsequent management.

Results

Of the 126 patients included, 75% were male, 27.7% were diabetic, and the median eGFR was 13.5 ml/min/1.73m². Major complications occurred in 5.6%. Perirenal haematomas were detected in 37.3%, and haematomas ≥2 cm were significantly more frequent in those with eGFR <15 ml/min/1.73 m² (29.2% vs. 13%, p = .032). Dialysis was a risk factor, while pre KBx blood transfusion, diabetes and higher serum albumin were protective against any complication. KBx was more likely to make a significant difference in management in those with eGFR 15–29 ml/min/1.73m² (44.1% vs. 11.1%, p < .001). Increasing age, lower serum creatinine and albumin were independently associated with KBx utility.

Conclusion

KBx is relatively safe in severe kidney disease but its risk to benefit balance needs to be carefully considered when eGFR is <15 ml/min/1.73m².

1. Introduction

Kidney biopsy (KBx) is the gold standard for evaluation of renal parenchymal disease and provides valuable diagnostic information to guide therapeutic decisions.¹ Major complications occur in ~1% of KBxs,^2–4 however in patients with an eGFR <30 ml/min/1.73 m² this rate may be as high as 2%–8.5%.^2,5,6 There is a paucity of prospective studies assessing the risk–benefit ratio of performing a KBx in patients with advanced kidney disease. This study aimed to study the safety and utility of KBx in patients with eGFR <30 ml/min/1.73 m².

2. Methods

Ours is a large tertiary care academic teaching hospital in southern India that routinely performs ~1500 KBxs a year. Over 50% of patients with chronic kidney disease (CKD) present to us in CKD Stage G5,⁷ which is reflective of a country-wide phenomenon.⁸ After a careful consideration of factors such as clinical presentation (pattern of progression, likelihood of a kidney biopsy altering clinical management), feasibility of a KBx (kidney size, cortical thickness), patients’ desire for a concrete diagnosis, and future plan for kidney transplant, they are offered a KBx by their treating nephrologist. In this prospective observational cohort study, consecutive patients with eGFR <30 ml/min/1.73 m² presenting to the in-patient or out-patient service of the Department of Nephrology who had been advised to undergo KBx by the treating nephrologist for their clinical evaluation were invited to participate after informed consent. Patients with traditional risk factors such as solitary kidney, transplant allograft kidney, kidney length < 8 cm on ultrasound imaging, lower pole renal cortical thickness < 1 cm, bleeding diathesis or acute kidney injury were excluded from the study.

2.1. Study protocol

As part of the institutional protocol, patients who required dialysis for various indications at the time of study recruitment, received a strict non-heparin dialysis session on the day before proposed KBx and all patients had their blood pressure medications titrated to achieve a target clinic blood pressure < 140/90 mm Hg. Where indicated, packed red cells were transfused to ensure a pre-KBx haemoglobin of ≥8 g/dl. As per study protocol, each subject’s haemoglobin level was checked on the day of KBx at baseline. Percutaneous KBx was performed by nephrologists/trainee nephrologists using a spring-loaded biopsy gun (18 G x 16 cm, BARD® Max Core®, Bard Peripheral Vascular, Inc.) after the radiologist had marked the lower pole and indicated the depth of the kidney below the skin. One core each was obtained for light microscopy and immunofluorescence microscopy, and where indicated, a third core for electron microscopy. Patients were monitored for both major and minor complications for 24 h after the procedure. In addition, each patient was screened 18–24 h after the biopsy, by a radiologist using ultrasound to look for symptomatic or asymptomatic post-biopsy peri-renal hematoma, and a blood sample was obtained to look for any decline in haemoglobin. The final histological confirmed diagnosis was compared with the pre-biopsy predicted clinical diagnosis noted in the patient records. Changes in management plan after availability of the KBx report was noted and the utility of the biopsy was graded by another investigator blinded to the patient’s identity and biopsy.

The primary outcome of the study was the incidence of major complications post-KBx. Secondary outcomes were the incidence of minor complications, any complications (composite of major and minor complications, but excluding asymptomatic haematomas), and incidence of asymptomatic haematomas. Other secondary objectives were to determine potential risk factors for post-KBx complications, to describe the spectrum of histopathology in biopsied patients, to estimate the proportion of cases in which a KBx significantly changed the management of these patients, and to study the factors that could predict when a KBx was likely to change management. The study protocol was approved by the Institutional Review Board and Ethics Committee of our institution. All procedures performed were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments.

2.2. Data sources and measurement

History, treatment details, blood pressure, height, weight, age, gender, co-morbidities, pre-biopsy clinical diagnosis and post-biopsy treatment plan were noted from the clinical records. Haemoglobin, 24-h urine protein, serum creatinine, albumin, kidney size, cortical thickness, final histological report and other investigations were noted from the hospital information system. A standard proforma was used to document complications and post-biopsy ultrasound findings.

2.3. Definitions

Estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI creatinine equation. Prior to KBx, patients were classified as presenting with one of the following clinical syndromes: Nephrotic syndrome, nephritic syndrome, nephritic-nephrotic syndrome, undifferentiated glomerulonephritis, tubulointerstitial disease or unexplained kidney failure. These terms are defined in the Supplementary Material.

In diabetic patients, clinical indications for KBx included one or more of the following as previously described⁹: Nephritic syndrome, unexplained kidney failure with normal sized kidneys, rapidly declining renal function, significant proteinuria occurring abruptly or in the absence of diabetic retinopathy. A KBx was considered to be adequate when it contained at least eight glomeruli and one vessel in the tissue sample. This is an arbitrary cut-off since sample adequacy is not defined in CKD. Samples that did not meet criteria for adequacy but were sufficient for the pathologist to make a diagnosis were termed ‘sub-optimal’.¹⁰ The size of post-biopsy haematomas was defined as the longest diameter measured by ultrasound imaging. A large haematoma was defined as ≥2 cm, based on available literature that found haematomas of this size to be linked to overt bleeding.¹¹ A major complication was defined as requirement for blood transfusion, hypotension (fall in systolic blood pressure by ≥20 mm Hg from baseline), renal artery embolization, nephrectomy, or death directly attributable to KBx.¹² A minor complication was defined as symptomatic hematoma (pain at the haematoma site), bladder clot, or transient gross haematuria not requiring blood transfusion. The term any complication was used for a composite of major and minor complications, but excluded asymptomatic peri-renal haematomas detected on the screening ultrasound. The utility of the KBx was graded on an ordinal scale as follows: Significantly changed management (score 2), changed management to some extent (score 1), did not change management (score 0). A significant change in management was defined as initiation of immunosuppressive therapy for retarding progression of kidney disease based on the KBx report, which would otherwise not have been done. The biopsy was said to have changed management to some extent if it led to discontinuation of immunosuppressive therapy and was useful in prognosticating the patient, which otherwise would not have been done. The biopsy was classified as effecting no change in management if none of the above criteria were met.

2.4. Sample size calculation

The sample size was calculated using the formula for single proportion—absolute precision. Based on the available literature, the rate of post-biopsy major complications was expected to range from 5% to 8%.^3,12,13 With the proportion of major complications assumed to be 7% and a precision of 4.5% with the desired confidence-interval of 95%, the sample size calculated was 124.

2.5. Statistical analysis

Statistical analysis was done with Statistical Package for the Social Sciences version 11.0 (SPSS Inc.). A p-value <.05 was taken as statistically significant. Univariate analyses were performed using the Pearson’s chi-square and independent t test (or Wilcoxon signed rank test where applicable) and the Multivariate analyses were done by backward conditional method binary logistic regression using clinically relevant covariates and those significant at a p < .2 in the initial model.

3. Results

A total of 980 patients underwent KBx of the native kidneys during the study period. Of these, 140 patients had an eGFR<30 ml/min/1.73 m², out of which 126 patients fulfilled our inclusion criteria and consented to participate in the study (Figure 1). The study population (n = 126) was 75.4% male and had a mean age of 40 ± 13 years, 27.7% were diabetic, 73% were hypertensive, and 5.6% had previously been on anti-platelet agents which were discontinued 7 days prior to KBx. The median creatinine and eGFR of the study population were 4.9 mg/dl and 13.5 ml/min/1.73 m², respectively; 57.1% patients had an eGFR <15 ml/min/1.73 m² at the time of KBx, of which 72.2% were on dialysis. Patients with eGFR <15 ml/min/1.73 m² had a significantly lower haemoglobin (9.1 ± 1.5 g/dl vs. 10 ± 1.5 g/dl, p = .002), and required a pre-biopsy blood transfusion more often than those with eGFR 15–29 ml/min/1.73 m² (27.8% vs. 11%, p = .022, Table 1).

Figure 1. STROBE diagram for the study.

Table 1. Baseline characteristics of the study population as a whole and stratified for estimated glomerular filtration rate (eGFR).

Characteristics	Total (N = 126)	eGFR 15–29 ml/min/1.73 m2 (n1 = 54) n(%)*	eGFR <15 ml/min/1.73 m2 (n2 = 72) n(%)*	p valuea
Age (year)	40 ± 13	42.7 ± 13.3	38.6 ± 13	.089
BMI (kg/m2)	23 ± 4.3	22.7 ± 3.7	23.2 ± 4.7	.598
Male	95(75.4%)	45(83.3%)	50(69.4%)	.073
Diabetes	35(27.8%)	13(24.1%)	22(30.6%)	.421
Hypertension	92(73%)	36(66.7%)	56(77.8%)	.164
Past CAD	8(6.3%)	6(11.1%)	2(2.8%)	.058
Past CVA	4(3.2%)	3(5.6%)	1(1.4%)	.187
Malignancy	2(1.6%)	1(1.9%)	1(1.4%)	.837
Newly diagnosed CKD	36(28.6%)	9(16.7%)	27(37.5%)	.010
Pre-biopsy dialysis	52(41.3%)	0 (0%)	52(72.2%)	<.001
Haemoglobin (g/dl)	9.1 ± 1.5	10 ± 1.5	9.1 ± 1.5	.002
Pre-biopsy BT	26(20.6%)	6(11.1%)	20(27.8%)	.022
>1 BT	5(4%)	1(1.9%)	4(5.6%)	.292
PT (s)	10.8 ± 0.8	10.8 ± 0.8	10.9 ± 0.7	.661
aPTT (s)	30.6 ± 4.5	30.9 ± 5.1	30.2 ± 2.9	.343
Platelet count (lakh/mm3)	2.0 ± 0.75	2.0 ± 0.74	2.0 ± 0.76	.897
Median S. creatinine in mg/dl (IQR)	4.9 (4.8)	3.26 (0.66)	7.43 (4.36)	<.001
Median eGFR in ml/min/1.73 m2 (IQR)	13.5 (12.7)	20.6 (5.05)	8.1 (5.65)	<.001
Median proteinuria in g/day (IQR)	3.4 (4.1)	3.8 (5.37)	3.05 (3.16)	.304
S. albumin (g/dl)	3.3 ± 0.9	3.3 ± 1.0	3.3 ± 0.7	.679
Kidney size (< 9 cm)	21 (16.7%)	10 (18.5%)	11 (15.3%)	.629
Pre-biopsy SBP (mm Hg)	131.5 ± 13.2	130 ± 13.3	132.5 ± 13	.291
Pre-biopsy DBP (mm Hg)	81.9 ± 8.7	81 ± 9.5	82.5 ± 8	.321

^ap value—for comparison between the eGFR <15 and 15–29 groups.

^*Values given in mean ± SD or n(%) or median (IQR) as applicable.

Abbreviations: aPTT, active partial thromboplastin time; BMI, Body mass index; BT, blood transfusion; CAD, coronary artery disease; CKD, chronic kidney disease; CVA, cerebrovascular disease; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate by CKD EPI formula; PT, prothrombin time; SBP, systolic blood pressure.

The common clinical syndromes at the time of KBx were undifferentiated glomerulonephritis (31%), unexplained renal failure (21%) followed by nephritic syndrome (18%), nephritic-nephrotic presentation (13%), nephrotic syndrome (9%) and possible tubulointerstitial disease (8%).

3.1. Major and minor complications post-KBx

Major complications and any complications were seen in 5.6% and 12.7% of the study population respectively. Both types of complications were more common in patients with eGFR <15 ml/min/1.73 m² (8.3% vs. 1.9%; 18% vs. 5.6%) though this was not statistically significant (Table 2). Post-KBx hematomas were seen in 37.3% patients, 78.7% of which were asymptomatic. About 22.2% of patients had a haematoma ≥2 cm in size, occurring significantly more common in those with eGFR <15 ml/min/1.73 m² (29.2% vs. 13%, p = .032). All major complications were detected within 12 h post-KBx except the single case of page kidney which presented 3 days later with acute pain and accelerated hypertension and was confirmed on computed tomography (Figure S1). There was no instance of nephrectomy, massive transfusion requirement or death.

Table 2. Comparison of various complications between subjects with eGFR 15–29 ml/min/1.73 m² and eGFR<15 ml/min/1.73 m².

Complications	Total N = 126(%)	eGFR 15–29 ml/min/1.73m2 (n1 = 54), n(%)	eGFR <15 ml/min/1.73m2 (n2 = 72), n(%)	ap value
Any complication	16 (12.7)	3 (5.6)	13 (18)	.053
Major complication	7 (5.6)	1 (1.9)	6 (8.3)	.232
Any haematoma	47 (37.3)	17(31.5)	30 (41.7)	.242
Asymptomatic haematoma	37 (29.4)	15 (27.8)	22 (30.6)	.731
Haematoma ≥2 cm	28 (22.2)	7 (13)	21(29.2)	.032
Symptomatic haematoma	10 (7.9)	2 (3.7)	8 (11.1)	.181
Transient gross haematuria	7 (5.6)	1 (1.9)	6 (8.3)	.232
Bladder clot	1 (0.8)	0 (0)	1 (1.4)	1.000
Severe haematuria requiring BT	2 (1.6)	0 (0)	2 (2.8)	.501
Blood transfusion post-biopsy	5 (4)	1 (1.9)	4 (5.6)	.391
Hypotension	2 (1.6)	0 (0)	2 (2.8)	.501
Infection	6 (4.8)	2 (3.7)	4 (5.6)	.702
Intervention (DSA guided embolisation)	2 (1.6)	0 (0)	2 (2.8)	.501
Page kidney	1(0.8)	0 (0)	1 (1.4)	1.000
Prolonged hospitalization >24 h	15 (11.9)	2 (3.7)	13 (18)	.023

^ap value—for comparison between the eGFR <15 and 15–29 groups.

Abbreviations: BT, blood transfusion; DSA, digital subtraction angiography; eGFR, estimated glomerular filtration rate by CKD EPI formula.

Hospitalization was prolonged beyond the 24 h in 93.7% (15/16) patients with any complication and 100% (7/7) of those who had a major complication, and was significantly more common in those with eGFR <15 ml/min/1.73 m² (18% vs. 3.7%, p = .023).

Median fall in haemoglobin from baseline to 18–24 h post biopsy was significantly greater in the any complication group compared with the no complication group (0.6 g/dl, IQR 1.95 vs. 0.1 g/dl, IQR 0.5, p = .011) and in the major complication group compared with the no complication group (1.9 g/dl, IQR 1.7 vs. 0.1 g/dl, IQR 0.5, p = .002). Overall, post-biopsy haemoglobin drop >1 g/dl from baseline was noted in 37.5% of those with any complication and 71.4% of those with a major complication.

3.2. Risk factors for complications

By multivariate analysis, the advanced kidney disease that required dialysis before KBx was an independent risk factor predicting any complication (HR 10.11, 95% CI 2.43–42.00). However, diabetes (HR 0.16, 95% CI 0.03–0.94), higher serum albumin (HR 0.42, 95% CI 0.19–0.92), and pre-biopsy blood transfusion (HR 0.06, 95% CI 0.01–0.57) were protective (Table 3). For major complications, increasing age was the only factor found to be protective (HR 0.87, 95% CI 0.78–0.97, analysis not shown). Variables such as clinical syndrome at presentation, gender, platelet count, kidney length, systolic and diastolic blood pressure, eGFR, number of biopsy needle passes and whether the biopsy was done by a nephrologist or a trainee, were not risk factors for post-KBx complications.

Table 3. Multivariate analysis of independent risk factors for any complication.

Characteristic	p-value	Exp(B)	95.0% CI for Exp(B)
			Lower bound	Upper bound
Nephritic syndrome	.058	3.773	0.979	14.542
Higher albumin	.031	0.425	0.196	0.921
Pre-biopsy blood transfusion	.012	0.059	0.006	0.576
Diabetes mellitus	.047	0.161	0.027	0.945
Pre-biopsy dialysis	.001	10.119	2.435	42.050

Note: Variables entered on step 1: Age, creatinine, albumin, platelet count, haemoglobin at baseline, blood transfusion prior to kidney biopsy, HD done prior to biopsy, number of biopsy passes, pre-biopsy systolic blood pressure, diabetes, nephritic syndrome, unexplained renal failure.

3.3. The spectrum of histopathological findings

The tissue sample for KBx was adequate in 81 (64.3%) cases with a median of nine glomeruli (IQR 6) per sample. For samples reported as inadequate, insufficient glomeruli was the reason in all cases. A pathological diagnosis could be made in 93.3% of inadequate samples (hence more appropriately termed sub-optimal), or 97.6% of the whole cohort.

The most common histopathological diagnosis was a glomerular disease in 56.3% (IgA nephropathy in 27%), followed by diabetic nephropathy (DN) either alone or with a non-diabetic kidney disease—NDKD (22.2%), and chronic interstitial nephritis (12.6%).

Among the subgroup of patients with diabetes (n = 35), 60% (n = 21) had DN, 20% (n = 7) had DN with a non-diabetic kidney disease (NDKD) and 20% (n = 7) had a NDKD alone. Patients with DN, DN with NDKD and NDKD alone had a median duration of diabetes 10 years (IQR 14), 16 years (IQR 11) and 6 years (IQR 11) respectively (p = .150), with presence of diabetic retinopathy in 52.4%, 71.4% and 28.5% (p < .001), respectively.

Post-KBx diagnosis matched the pre-biopsy clinical diagnosis in 53.2% while in 44.4% KBx changed the clinical diagnosis (Table S1) and in 2.4% the KBx results were inconclusive.

3.4. The proportion of cases in which KBx significantly changed management

According to the grading scale, KBx significantly changed management (graded 2) in 25.4%, changed management to some extent (graded 1) in 41.3%, and did not change management (graded 0) in 33.3% cases respectively. A KBx was more likely to make a significant difference in management when the eGFR was 15–29 ml/min/1.73 m² (44.4% vs. 11.1%, p < .001, Figure 2).

Figure 2. Utility of the kidney biopsy stratified by eGFR <15 ml/min/1.73m² and 15–29 ml/min/1.73m², and graded as major difference, some difference and no difference.

Table S2 describes the immunosuppressive treatment prescribed to patients in whom KBx made a significant difference in management. One patient was diagnosed to have hereditary nephropathy consistent with Alport’s syndrome which helped in kidney transplant donor selection.

3.5. Factors predicting when a kidney biopsy is likely to change management (Utility grade 1 and 2)

By multivariate analysis, increasing age (HR 1.05, 95% CI 1.02–1.1), lower creatinine (HR 0.78, 95% CI 0.67–0.90) and lower serum albumin (HR 0.31, 95% CI 0.16–0.61) were independent factors associated with the KBx making any difference in the management (Table 4).

Table 4. Multivariate analysis for factors that predicted whether biopsy made any difference.

Characteristic	p-value	Exp(B)	95.0% CI for Exp(B) lower upper bound
Age (year)	.004	1.058	1.019	1.099
S. Creatinine	.001	0.781	0.671	0.901
S. Albumin	.001	0.312	0.160	0.609

Note: Variables entered on step 1: Age, BMI, creatinine, serum albumin, ultrasound kidney size, platelet count, dialysis done prior to kidney biopsy, diabetes, 24-h urine protein, nephritic syndrome, undifferentiated glomerulonephritis.

4. Discussion

Our study aimed to assess the safety and utility of a KBx in patients with eGFR <30 ml/min/1.73 m² in terms of the incidence of major and minor complications, the risk factors for major or any complications, the correlation between histological and pre-biopsy clinical diagnosis, the main types of renal pathology encountered in this population, the utility of the KBx in changing clinical management and predictors for the utility of a KBx in this setting.

4.1. Post-KBx complications: Incidence and timing

The major complication rate of 5.6% is comparable to that reported by contemporaneous studies addressing the safety of kidney biopsies in patients with eGFR <30 ml/min/1.73 m² (2%–5.5%) that have used a similar definition.^2,5,6,14 The proportion of patients with any complication (composite of major and minor complications) was 12.7%, which is lower than that reported from large biopsy registries not stratified for eGFR (13%–14.7%).^13,15 Perirenal haematomas were found in 37.3%, of which 78.7% were asymptomatic. Other investigators have reported that perirenal haematomas may be found in between 29% and 86% of KBxs,^11,16 and though none have reported the proportion of asymptomatic haematomas, the presence of a peri-renal haematoma does not reliably predict a major post-KBx bleed unless it measures ≥2 cm.¹¹ We found that around one-fifth of all patients, and more than one-fourth of those with an eGFR <15 ml/min/1.73 m² had a peri-renal haematoma ≥2 cm in size. The latter subgroup showed a trend towards a higher frequency of major or any complication, though this did not reach statistical significance. The increased risk for post-KBx bleed with declining eGFR is almost universally acknowledged,^2,6,13,15,17 and has been attributed to uraemic platelet dysfunction, anaemia,⁶ and tissue fibrosis that increases arterial stiffness and impedes vessel contraction after injury.^18,19

Most major complications were detected within 12 h of the KBx, so an observation period of 24 h in this high-risk group is justified. Whittier et al have shown that 89% of post-KBx complications present within 24 h, and an observation period <8 h would miss 33% of complications.¹⁵ It is important to note that 55.5% of those with large asymptomatic haematomas in our cohort were on haemodialysis, and inadvertent exposure to heparin during dialysis may result in an increase in haematoma size and delayed presentation of a major complication. This was observed in the patient who presented with a page kidney 72 h after KBx.

4.2. Post-KBx complications: Risk factors

Dialysis was an independent risk factor for any complication after KBx in our study. While the requirement for dialysis indicates more severe kidney damage and possibly either tissue fibrosis or inflammation, which are risk factors for post-KBx bleeding, its independent effect on platelet function is controversial. Dialysis has been shown to remove dialyzable uraemic middle molecules that impair platelet function,^20,21 and may improve platelet count by stimulating the release of immature platelet fractions and thrombopoietin from the bone marrow.²² However, even in the absence of heparin, exposure of blood to the dialyzer and dialysis tubing may result in platelet degranulation and loss of glycoprotein receptors, which may impede the haemostatic function of platelets.^23–25 This effect may last for up to 24 h after dialysis.

Pre-KBx blood transfusion was protective against any complication. Anaemia is a risk factor for post-KBx bleeding,^6,17,26 and in vitro²⁷ and in vivo²⁸ studies have demonstrated that red cell transfusions improve platelet reactivity, aggregation and adhesion via ADP mediated pathways. Diabetes was found to be protective, as reported by others,^26,29 a possible explanation being platelet hyper-aggregation seen in diabetes.³⁰ We found higher serum albumin to be protective, as reported by other authors¹⁹ though the biological basis for this is not known. Older age was found to be protective against major complications, which has been variably attributed to the higher vascularity of younger kidneys and greater propensity to be physically active post biopsy in this age group.^3,31 However, we cannot rule out confounding as an explanation for the above findings. Since the decision to offer biopsies was determined by the treating physician in each case, there is likely to be a bias by indication, with various factors such as age, eGFR, perceived likelihood of the biopsy yielding a clinically meaningful result and the patient’s willingness to undergo the procedure, playing a part in determining which patients are offered a KBx, and by extension, the risk factors identified.

4.3. Biopsy adequacy, histological diagnosis versus clinical diagnosis

With an arbitrary cut-off of eight glomeruli and one vessel, only 64.3% of biopsy specimens were adequate. This may be due to the smaller (18G) gauge of the biopsy gun used. A randomized controlled trial demonstrated higher sample adequacy without an increase in major complications with the use of 16G needles compared with 18G needles.³² However, it is important to note that in the vast majority, a diagnosis could still be made, which suggests that biopsy adequacy in CKD may need to be re-defined.

Our results confirm that patients with eGFR <30 ml/min/1.73 m² who are offered a KBx may have a wide spectrum of kidney diseases—over 50% were found to have a glomerular disease in our cohort, IgA nephropathy being the most common. These findings are similar to those from the Triveneto Register of Renal biopsies from north-eastern Italy.¹⁴ NDKD either alone, or in combination with DN was seen in 40% of the diabetic patients selected for kidney biopsy. Although those with a pure NDKD tended to have a shorter duration of diabetes, the absence of diabetic retinopathy was more likely to predict the presence of a pure NDKD compared with the duration of diabetes. The classical clinical predictors of NDKD when applied in the absence of diabetic retinopathy, have been shown to have a positive predictive value of 87% for NDKD,⁹ and this may be a useful guide for nephrologists when considering the feasibility and utility of kidney biopsy in diabetics with advanced kidney disease.

4.4. The utility of KBx in changing diagnosis and management

The KBx report changed the pre-biopsy clinical diagnosis in 44.4% and helped change the management approach to some extent in 65% of patients. Important therapeutic decisions to treat with immuno-suppression were taken in over 25% of the patients and was graded as a significant change in management. Previous estimates of the proportion of biopsies that have led to initiation of immunosuppression in patients with severe kidney failure have varied from 25% in the United Kingdom³³ to as high as 40% in Italy¹⁴ and may be reflective of variations in the histopathological spectrum and severity of disease as well as clinical practice. At our centre, given that most patients present to us for the first time with advanced kidney disease, a trial of immunosuppression is offered to patients with a nephritic-nephrotic presentation and a histopathological diagnosis of primary or secondary glomerular disease in a bid to delay initiation of renal replacement therapy, which is a devastating out-of-pocket expenditure.³⁴ This practice, although unconventional, has been shown to improve short-term outcome in diseases such as advanced IgA nephropathy in high-risk ethnic groups.³⁵ In our study, there was a significant difference in utility in favour of patients with an eGFR >15 ml/min/1.73 m², which is in line with evidence that has correlated the utility of KBx with the degree of histological chronicity and duration of CKD.¹⁸ Higher age, lower albumin and lower creatinine were independent predictors of ‘whether KBx made any difference’. Lower albumin is indicative of a glomerular disease, which may warrant immunosuppression; lower creatinine is suggestive of less extensive histological damage, which is more likely to encourage physicians to offer definitive treatment.¹⁸ The lower incidence of major complications and the higher utility in the elderly suggests that KBx should be offered to this age group where clinically indicated.

4.5. Strengths and limitations

This prospectively study evaluates the safety and utility of KBxs in patients with eGFR <30 ml/min/1.73 m². We used robust inclusion criteria, recruited and followed up the pre-specified sample size for the study, subjected all patients to a screening ultrasound and repeat haemoglobin 18–24 h post-KBx to document clinically significant bleeding, and were thus able to offer an accurate estimate of major and minor complications and asymptomatic significant haematomas. We were also able to determine the clinical utility of the KBx in each case based on a careful review of the histopathological report and subsequent management plan.

The study has certain limitations. The primary outcome (major complications) was lower in our study cohort than what was used to calculate the sample size and the study may, therefore, be underpowered to detect significant differences between the those with an eGFR above and below 15 ml/min/1.73 m². We did not study the association of histological parameters of chronicity with the complication rate or clinical utility of the biopsy, however, we aimed to determine pre-biopsy parameters that could predict these outcomes. Being an observational cohort study, confounding in the assessment of risk factors for post-KBx bleeding cannot be ruled out. Our assessment of KBx utility may be biased and specific to clinical practice in this region. This being a single centre study with considerable renal biopsy experience (both in procedure and reporting), study results may not be generalizable to centres that perform fewer biopsies, since complication rates are heavily operator dependent.² Biopsies were performed blindly following ultrasound marking, and it is possible that real-time ultrasound guidance may have resulted in fewer major complications.³⁶ Lastly, a high proportion of biopsy samples were reported as inadequate based on pre-specified criteria. While biopsy adequacy in CKD has not been defined, it is likely that the use of a 16G biopsy needle,³² real-time ultrasound guidance³⁶ and on-site biopsy adequacy assessment¹⁰ may have resulted in higher biopsy adequacy.

5. Conclusions

KBx in patients with eGFR <30 ml/min/1.73 m² is associated with a major complication in 5.6%. While severe kidney dysfunction requiring dialysis is a risk factor, pre biopsy blood transfusion to correct anaemia, presence of diabetes and a higher serum albumin are protective for any complication. KBx is likely to alter clinical management in 65% of cases, especially in the elderly, those with lower serum albumin and lower degree of kidney dysfunction. In light of the association between eGFR <15 ml/min/1.73 m² and significant post-KBx haematomas as well as the limited utility of KBx in significantly changing clinical management, its risk–benefit ratio needs to be carefully considered in this group.

Summary at a Glance.

The utility and safety of kidney biopsies in patients with significantly reduced kidney function is a common clinical conundrum. This paper reports a single centre case series of kidney biopsies in patients with an eGFR < 30 ml/min/1.73 m², where clinical suspicion of an alternative diagnosis is high. The authors demonstrate that in a high volume centre, biopsies in advanced kidney disease are relatively safe and can alter management decisions.