How Safe Is a Native Kidney Biopsy?

Kidney biopsies are performed to investigate the cause of suspected parenchymal disease. Although these biopsies provide important diagnostic and prognostic information guiding therapeutic decisions, the benefits must be weighed against the potential harms of the biopsy procedure. As the kidneys have an extensive vascular network and receive 20%–25% of the cardiac output, bleeding is the most common postprocedural complication. Since development in the 1950s, modifications of the technique, including the widespread use of real-time ultrasound and automated biopsy devices, have made the procedure safer; however, prebiopsy risk assessment to allow thoughtful decision making with the informed patient is essential.

In this issue of CJASN, Poggio et al. (1) describe a meta-analysis of 87 published studies (January 1983 to March 2018) on the complication rates of native kidney biopsies performed using automated devices under kidney imaging. This study was undertaken to obtain an estimate of percutaneous native kidney biopsy risk to provide subjects in the Kidney Precision Medicine Project (KPMP) with accurate risk information during the informed consent process. The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases–sponsored KPMP is a multicenter biorepository study in which participants undergo protocol kidney biopsies for research purposes to correlate deep molecular phenotyping with patient characteristics, innovative digital pathology, and disease outcomes. A total of 118,064 biopsies were included in this analysis with subjects’ ages ranging between 30–79 years. Complication rates included perinephric hematoma in 11% of the kidney biopsies, macroscopic hematuria in 3.5%, bleeding requiring blood transfusion in 1.6%, and interventions to stop bleeding in 0.3%. Death attributed to native kidney biopsy was rare, occurring in only an estimated 0.06% of all biopsies. Complication rates were higher in hospitalized patients and those with AKI.

In a second biopsy study in this issue of CJASN, Halimi et al. (2) describe their retrospective database study of 52,138 patients who underwent a kidney biopsy in France between 2010 and 2018. Diagnosis and complications were captured using the International Classification of Disease (ICD) codes. Major bleeding (by postprocedure day 8) occurred in 5% of patients, including blood transfusions (5%), angiographic intervention (0.4%), and nephrectomy (0.1%). Death from any cause (by day 30) occurred in 1% but was notably higher in those who had major bleeding (3%) versus no bleeding (1%). The strongest risk factors for postbiopsy bleeding were anemia, AKI, vasculitis, thrombotic microangiopathy, thrombocytopenia, frailty index, and women. Notably, obese subjects and those with diabetic kidney disease appeared to have a lower risk of bleeding, which the authors postulate may be due to a hypercoagulable state. Using multivariable logistic regression, the authors developed a bleeding risk score using data available at the time of kidney biopsy. The bleeding risk increased 75-fold from 0.4% in the lowest score category (zero to four points) to 33% in the highest score category (≥35 points). Risk factors that were not available to study included needle gauge, proceduralist’s specialty (nephrologist versus radiologist), and data on anticoagulants and antiplatelet agents. Notably, there was no decreased bleeding risk in the centers with the largest volume of biopsies, although large centers often perform biopsies on higher-risk patients.

How do we interpret the results of these two studies and incorporate these data into our clinical practice? Although the study by Poggio et al. (1) describes an overall blood transfusion rate of 1.6%, these data were strongly influenced by a single large study of an inpatient cohort (3), which had a much higher bleeding complication rate. It is well recognized that inpatient subjects would be expected to be sicker, often with anemia, thrombocytopenia, AKI, and an expected higher bleeding complication rate. It should also be recognized that the increased transfusion rate in patients with anemia may be due to a lower prebiopsy hemoglobin and not necessarily a higher bleeding rate. Transfusion rates in the outpatient setting would be expected to be much lower. Weaknesses of this study are those common to any meta-analysis of previously published data and, as discussed by the authors, include the mostly small, retrospective, single-center patients series; heterogeneity between studies; lack of uniform end points; and publication bias. Notably, in a prior systemic review, Corapi et al. (4) described a similar transfusion rate of 0.9%.

By contrast, although the study by Halimi et al. (2) described a similar angiography/intervention rate (0.4% versus 0.3%), the overall transfusion rate was much higher at 5%. This study’s strengths include a large sample size and the avoidance of selection and reporting biases by including all biopsies identified in a national database with ICD codes. However, the results of this study must be interpreted cautiously. The seemingly high transfusion rate likely reflects the high-risk baseline characteristics of many subjects in this study: for example, 30% had AKI. These higher-risk characteristics are also reflected by the very high 30-day all-cause mortality rate of 1%. In a similar study using US hospitalization discharge data, Charu et al. (5), described a comparable transfusion rate (5.7%) in hospitalized patients following kidney biopsy.

Considering these important studies, how should we then inform patients about the benefits and harms of a native kidney biopsy? In nephrology practice, we commonly quote a 1% risk of needing a blood transfusion and a 0.5% risk of requiring an angiography procedure. On the surface, the two studies in this issue of CJASN suggest a higher overall risk; however, Halimi et al. (2) nicely illustrate the wide disparity (75-fold) in the bleeding risk between the lowest and highest at-risk subjects. Thus, interpreting their data, outpatients with proteinuric kidney disease and preserved kidney function with normal hematologic parameters have a low risk of major bleeding (<1%). By contrast, inpatient subjects with AKI, anemia, and possible coagulopathy would be at a much higher risk (≥5%). These authors developed a bleeding risk score to quantify the risk of bleeding after kidney biopsy on the basis of baseline risk factors (age, Charlson index, frailty index, sex, hypertension, diabetes mellitus, obesity, dyslipidemia, anemia, thrombocytopenia, cancer, abnormal kidney function, and AKI). The validity of this risk score needs to be confirmed by future studies.

If a biopsy is deemed necessary in those at high risk, measures to mitigate the risk of bleeding, such as preprocedure blood or platelet transfusion, desmopressin (6), use of smaller gauge needles, and limiting the number of biopsy cores, may be required. Alternatives to biopsy should also be considered. Identifying biomarkers, such as antiphospholipase A2 receptor antibodies in membranous nephropathy, may be specific enough for diagnostic purposes in some patients (7). Studies, such as KPMP, hope to identify new biomarkers and molecular patterns that correlate better with diagnosis and clinical course and may obviate kidney biopsies in certain circumstances. By contrast, as we enter the era of precision medicine, kidney biopsies may become more important than ever as we move beyond standard histomorphometric analysis to phenotyping molecular pathways that may facilitate the targeting of specific therapeutic agents to the individual patient.

These two studies are important additions to the literature on the complication rates of native kidney biopsies and highlight the need for careful risk stratification before performing the procedure. So, how safe is a kidney biopsy? In many patients, the risk of major bleeding is low (<1%). In more complicated patients, proceduralists performing kidney biopsies may benefit from adopting a risk stratification tool, like the one proposed by Halimi et al. (2), to identify and potentially modify prebiopsy risk factors and enable informed consent.

Disclosures

All authors have nothing to disclose.

Funding

None.