Diabetic kidney disease (DKD) accounts for nearly half of all CKD in the United States. The diagnosis of DKD is usually made clinically, on the basis of longstanding (>5 years) diabetes or concurrent microvascular complications, and the presence of albuminuria and/or decreased eGFR. The heterogeneity of DKD and predominant study of blood and urine rather than direct analysis of kidney tissue are both barriers to identifying biomarkers and novel therapeutic targets for progressive disease. Data are lacking regarding the feasibility and safety of obtaining research kidney biopsy cores. To address this issue, we present an interim analysis of the Transforming Research in Diabetic Nephropathy (TRIDENT) study (Clinicaltrials.gov identifier: NCT02986984), a multicenter, longitudinal, observational cohort study of adults with diabetes undergoing clinically indicated kidney biopsies and consenting to an additional research biopsy core (1).

Demographic and biopsy-related data were obtained for all patients who underwent kidney biopsy. A serious biopsy-related complication was defined as hematoma >5 cm, gross hematuria, hospital stay exceeding the center’s standard practice, unplanned blood transfusion, and radiologic or surgical intervention to halt bleeding. Biopsy operator and protocol (postbiopsy imaging and overnight stay) adhered to the existing institutional protocol at each individual TRIDENT site. We present descriptive data without statistical comparison, because of the low number of patients without diabetic glomerulosclerosis and the rare incidence of severe biopsy complications.

Between January 1, 2017 (study inception) and May 1, 2019, 176 patients signed consent for TRIDENT and 160 underwent kidney biopsy. A research biopsy core was successfully obtained in 144 cases (90%). Reasons for a research biopsy core not being obtained were operator choice/all tissue needed for clinical diagnosis (ten of 16, 63%) and bleeding and/or hematoma before collection of research tissue (six of 16, 38%). One patient’s research specimen was placed in incorrect medium. This patient is included in the safety analysis but was not eligible to enroll in TRIDENT because of protocol deviation.

Sociodemographic and clinical characteristics, biopsy descriptors, and adverse events for patients who underwent a kidney biopsy in TRIDENT are presented in Table 1. Cohort characteristics include mean age of 54 (SD 13) years, 43% women, 33% black race, and 33% Hispanic ethnicity. A 16- or 18-gauge needle was used in most cases (54% and 44%, respectively), and the mean number of biopsy passes was 3.6 (SD 1.1). Excessive proteinuria was the most common indication for kidney biopsy (65%), followed by rapid loss of eGFR (24%). The majority of biopsies were performed by nephrology fellows (40%) or attendings (36%). Diabetic glomerulosclerosis (defined using Renal Pathology Society criteria) was present in 82% (117 of 143) of eligible cases.

Table 1.

Demographics, biopsy indications, procedural characteristics, and biopsy complication data for the study

	All Patients Who Underwent Biopsy, n=160	Research Tissue Not Obtained, n=16	Research Tissue Obtained
			Total, n=144a	Diabetic Glomerulosclerosis Present, n=117	Diabetic Glomerulosclerosis Absent, n=26
Demographics
Women, n (%)	68 (43%)	3 (19%)	65 (45%)	49 (42%)	15 (58%)
Age, mean (SD), yr	54 (13)	53 (12)	54 (13)	53 (12)	61 (13)
Black race, n (%)	52 (33%)	8 (50%)	44 (30%)	37 (32%)	7 (27%)
Hispanic ethnicity, n (%)	53 (33%)	3 (19%)	50 (35%)	41 (35%)	9 (35%)
Indication for biopsy
Urinary findings	14 (9%)	1 (6%)	13 (9%)	2 (2%)	1 (4%)
Excessive proteinuria	104 (65%)	11 (69%)	93 (65%)	71 (61%)	21 (81%)
Rapid eGFR loss	39 (24%)	4 (25%)	35 (24%)	32 (27%)	3 (12%)
Other indicators of nondiabetic kidney disease	2 (1%)	0 (0%)	2 (1%)	12 (10%)	0 (0%)
Not available	1 (0.7%)	0 (0%)	1 (0.6%)	0 (0%)	1 (4%)
Procedural characteristics
Needle gauge, n (%)
16 G	86 (54%)	12 (75%)	74 (51%)	58 (50%)	16 (62%)
17 G	2 (1%)	0 (0%)	2 (1%)	1 (0.9%)	0 (0%)
18 G	70 (44%)	4 (25%)	66 (46%)	56 (48%)	10 (38%)
Not available	2 (1%)	0 (0%)	2 (1%)	2 (2%)	0 (0%)
Mean (SD) number of passes	3.6 (1.1)	3.3 (1.5)	3.6 (1.0)	3.6 (1.0)	3.4 (0.8)
Biopsy operator, n (%)
Nephrology Fellow	64 (40%)	9 (56%)	55 (38%)	46 (39%)	9 (35%)
Nephrology Attending	58 (36%)	5 (31%)	53 (37%)	43 (37%)	10 (38%)
Nephrology Fellow with Attending takeover	3 (2%)	0 (0%)	3 (2%)	3 (3%)	0 (0%)
Interventional radiologist	32 (20%)	2 (13%)	30 (21%)	23 (20%)	6 (23%)
Physician assistant	3 (2%)	0 (0%)	3 (2%)	2 (2%)	1 (4%)
Biopsy complications
Hematoma >5 cm, n (%)	7 (4%)	1 (6%)	6 (4%)	6 (5%)	0 (0%)
Gross hematuria, n (%)	3 (2%)	0 (0%)	3 (2%)	3 (3%)	0 (0%)
Hospital stay exceeding the center’s standard practice, n (%)	6 (4%)	1 (6%)	5 (3%)	3 (3%)	2 (8%)
Blood transfusion, n (%)	3 (2%)	0 (0%)	3 (2%)	2 (2%)	1 (4%)
Surgical/radiological intervention, n (%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)

Open in a separate window

^aOne patient had a research core obtained that was placed in formalin and is included in the safety analysis (n=144) but was not eligible to enroll in the study (n=143) because of this protocol deviation.

In total, 11 of 160 patients (7%) experienced 19 biopsy-related complications. Three patients (2%) required blood transfusion, seven patients (4%) had a hematoma >5 cm, three patients (2%) experienced gross hematuria, and six patients (4%) required a prolonged hospital stay or readmission. Of the 144 patients who had a research core successfully obtained, nine patients (6%) experienced 17 biopsy-related complications. Of the three patients who required blood transfusion, two had diabetic glomerulosclerosis on pathology. No patient (zero of 160) required surgical/radiologic intervention.

A 2012 systematic review and meta-analysis of 34 studies inclusive of 9474 adult patients found a 0.9% (95% confidence interval, 0.4% to 1.5%) rate of blood transfusion and 0.6% (95% confidence interval, 0.4% to 0.8%) rate of need for angiographic intervention after clinically indicated kidney biopsies (2). Transfusion rates were higher in studies where the mean creatinine level was >2.0 mg/dl, >50% of patients were women, higher proportions of biopsies were in the setting of AKI, and prebiopsy hemoglobin levels were <12 g/dl. Higher complication rates after biopsy have been described in large, single-center case series at major academic centers (3–5). A recent study of hospitalized patients with AKI found 8% (95% confidence interval, 5% to 15%) required a blood transfusion and 2% (95% confidence interval, 1% to 5%) required intervention (5). One recent, large, single-center study found that patients with diabetic nephropathy on kidney biopsy had a 3% transfusion rate, with no surgical/radiologic interventions (4).

To our knowledge, this interim analysis of the TRIDENT study is the first report describing the feasibility and safety of obtaining research tissue during clinically indicated kidney biopsies. We demonstrate that obtaining research tissue in people with diabetes undergoing clinically indicated kidney biopsy is feasible, as we successfully obtained research cores in 89% of participants. There was also no clear evidence of a negative impact of obtaining a research core, which is reassuring for other studies obtaining research kidney biopsy cores. We acknowledge our study has limitations, as this is a descriptive and interim analysis of TRIDENT. A complete analysis of risk factors for biopsy complications and biopsy practice at individual TRIDENT sites is planned for when TRIDENT has been fully recruited. These data will help potentiate the safety of obtaining kidney tissue for research, ultimately improving care for patients with DKD.