Renal biopsy (RB) is useful for diagnosis and therapy guidance of renal diseases but incurs a risk of bleeding complications of variable severity, from transitory haematuria or asymptomatic hematoma to life-threatening hemorrhage. Several risk factors for complications after RB have been identified, including high blood pressure, age, decreased renal function, obesity, anemia, low platelet count and hemostasis disorders. These should be carefully assessed and, whenever possible, corrected before the procedure. The incidence of serious complications has become low with the use of automated biopsy devices and ultrasound guidance, which is currently the “gold standard” procedure for percutaneous RB. An outpatient biopsy may be considered in a carefully selected population with no risk factor for bleeding. However, controversies persist on the duration of observation after biopsy, especially for native kidney biopsy. Transjugular RB and laparoscopic RB represent reliable alternatives to conventional percutaneous biopsy in patients at high risk of bleeding, although some factors limit their use. This aim of this review is to summarize the issues of complications after RB, assessment of hemorrhagic risk factors, optimal biopsy procedure and strategies aimed to minimize the risk of bleeding.

Renal biopsy (RB) is often necessary for diagnosis, prognostic assessment and therapy guidance of various diseases affecting native and transplant kidneys. The final diagnosis differs from the main hypothesis in up to one third of cases[1]. Despite its necessity, RB incurs a potential risk of bleeding complications of variable severity, from transitory hematuria or asymptomatic hematoma to life-threatening hemorrhage[1-3]. Several studies identified risk factors for complications after RB[4-6]. However, controversies persist regarding the optimal assessment and management of bleeding risk. Two surveys, one conducted by the Society of Nephrology in France and another one in United Kingdom paediatric hospitals, highlighted significant variation in RB procedures[7,8]. Therefore, the gold standard for RB practice still remains to be defined. We previously participated to the elaboration of consensual recommendations by the Society of Nephrology in France[8]. Optimizing procedures for RB may improve patient safety and may also provide some logistic benefits and save costs.

This review discusses the issue of complications after RB, optimal biopsy procedure, and strategies aimed to minimize the risk of bleeding. We only address biopsies for the investigation of medical kidney diseases, but not those performed for kidney tumors.

Several large prospective and retrospective studies provide an estimate of the frequency of complications after percutaneous RB[1-3,5,9-12]: (1) Death: < 0.1%; (2) Major bleeding requiring nephrectomy or surgical hemostasis: 0.1% to 0.5%; (3) Arteriovenous fistula requiring invasive intervention: 0.1% to 0.5%; (4) Blood transfusion requirement: 0.3% to 7.4%; (5) Uncomplicated hematoma: 10 to 90%; and (6) Transient macroscopic hematuria : 1% to 10%.

We recently published a series of 312 native kidney biopsies performed at our institution: 15% of patients developed a symptomatic hematoma, 5% macroscopic hematuria, 9% received a red blood cell transfusion and 1% required an angio-intervention[13].

The reported incidence of complications after RB varies in relation to numerous factors, including patient selection, definitions of complications, procedures, and monitoring protocols. Several studies were performed before the implement of ultrasound guidance and automated biopsy devices, which improved the safety and efficiency of RB procedures[4,9]. The rates of complications drawn from these reports may therefore not reflect the risk associated with RB performed nowadays.

Recent studies reported major bleeding and life-threatening complications in less than 0.1% of RB procedures[2,4]. Tøndel et al[12] recently published the largest report of RB complications: 9288 (715 children and 8573 adults) biopsies from the Norwegian kidney biopsy registry, the vast majority of which (99.7%) were guided by ultrasound. In this study, 0.9% of the patients needed blood transfusion, 0.2% required an invasive procedure (surgery or angiointervention), and 1.9% had a macroscopic hematuria[12].

The risk of bleeding complications appears lower for transplant than native kidney biopsie[14,15]. However, major complications can occur after transplant biopsy[16].

An important step before RB is to search for factors increasing the risk of complications, particularly bleeding. Although there are no definitive ways to predict which patients will experience complications, several predisposing factors to bleeding have been identified, at times inconstantly.

High blood pressure, age, a decreased GFR, obesity, anemia, low platelet count and small center size (< 30 biopsies/year) are associated with an increased risk of bleeding[4-6,12,17-19]. Amyloidosis was reported to be associated with bleeding[4], although such association was not found in large study by Tøndel el al[12]. As discussed below, hemostasis disorders, anticoagulant or antiplatelet therapy, and certain anatomic conditions, may also contraindicate or complicate percutaneous RB.

A recent systematic review and meta-analysis of hemorrhagic complications after percutaneous native kidney biopsy using ultrasound guidance and automated spring-loaded biopsy device reviewed 34 publications and concluded that the predictors of erythrocyte transfusion were: the needle gauge (14 vs 16 or 18), sex (female), serum creatinine (≥ 2 mg/dL), low hemoglobin prior biopsy (≤ 12 g/dL) and acute kidney injury[18].

As it is for any invasive procedure, correction of coagulopathy is mandatory before RB. The platelet count threshold at which a RB can be safely conducted is not clear. Most platelet count thresholds for invasive procedure are based on weak observational evidence. For most major surgery, other than ocular and neurologic, platelet transfusion are considered if the platelet count is below 50000/microL[33]. It is not clear if this can be applied to RB. Many nephrologists consider RB contraindicated if platelet count is <100000/microL, which seems more prudent. Of course, optimal methods for raising platelet count depend on the underlying condition.

In the setting of renal disease, the risk of bleeding can result from dysfunctional platelets resulting from uremia. Indeed, uremic bleeding is a well-known complication of renal failure. The exact underlying mechanisms remain largely unknown, but seem to be multifactorial. The pathophysiology of uremic bleeding and evidence based treatment recommendations were the subject of a review by Hedges et al[17]. Many factors contribute to platelet dysfunction including anemia, dysfunctional von Willebrand factor, platelet membrane abnormalities, uremic toxins inhibiting platelet aggregation, and increased prostacyclin and nitric oxide levels, which are strong anti-platelet aggregating factors[17]. Correction of anemia, deamino-8-D-arginine vasopressin (DDAVP), estrogens and cryoprecipitate have been shown to improve “uremic bleeding”.

Desmopressin (DDAVP) is probably the most common agent used to treat or prevent bleeding in uremic patients. DDAVP improves hemostasis by releasing factor VIII from storage sites. DDAVP can reverse uremic platelet dysfunction rapidly (approximately within one hour of IV injection) for a short period of time (around 24 h)[17].

Several studies demonstrated that recombinant erythropoietin treatment prevents bleeding caused by uremic platelet dysfunction if the hematocrit is increased to more than 30%. Recombinant erythropoietin was shown to improve primary hemostasis in uremia through an increase of hematocrit but also through an effect on platelet function[17,34,35].

Several studies showed that intravenous conjugated estrogens can safely and effectively improve uremic platelet dysfunction and clinical bleeding. Intra-venous conjugated oestrogens improve bleeding time with a maximum effect at 5 to 7 d, lasting from 14 to 21 d[17].

Finally, cryoprecipitate is another therapeutic option in the setting of active uremic bleeding or in patients with high risk of bleeding[17,36]. Cryoprecipitate is prepared from plasma and contains fibrinogen, von Willebrand factor, factor VIII and factor XIII. It has a rapid onset of action (around 1 h) and its effect lasts approximately 24 to 36 h.

The impact of dialysis on uremic bleeding is unsure. Studies are old, and the effect on platelet function and coagulation is inconstant.

In all, the evidence supporting recommendations for the prevention or treatment of uremic bleeding is limited, especially in the context of RB. Despite the absence of robust evidence, it may be prudent to avoid undertaking RB when the hematocrit is lower than 30%, and to consider DDAVP or oestrogens before RB when the glomerular filtration rate is lower than 30 mL/min per 1.73m², as suggested by some authors[37].

Well-trained nephrologists can perform RB as well as radiologists[38,39]. Automated biopsy guns have superseded Tru-cut needles and are probably used in most centers[8]. Several studies suggested that 14-18G needles are appropriate for percutaneous RB[3,15,40]. The use of an automated biopsy gun in combination with real-time ultrasound guidance was reported to provide adequate samples in nearly 99% of cases, with severe hemorrhagic complications occurring in less than 0.1%. This method can be considered the gold standard[2,4]. The use of bedside ultrasound to assess the location and depth of the kidneys was reported as a reliable alternative to real-time guidance[39]. In some instances, especially in obese patients, it may be necessary to perform RB under guidance by CT-scan instead of ultrasound.

Transjugular RB has been reported to be a safe and reliable alternative to conventional percutaneous RB in patients with obesity[41] or those at risk for bleeding, including high-risk patients with coagulopathy and thrombocytopenia[42-44]. In these studies, transjugular RB provided diagnostic yield and safety similar to those of percutaneous approach. However, in most countries, the use of transjugular RB is limited to a few centers because of the necessity of skilled interventional radiologists.

Laparoscopic RB has also been reported as an alternative for patients in whom percutaneous approach was not feasible or was contraindicated, because of obesity, solitary kidney, anticoagulation or coagulopathy, or failed percutaneous biopsy[45,46]. However the number of patients included in these studies was limited and no study has compared the safety of percutaneous, transjugular and laparoscopic RB in patients at high risk for bleeding. In addition, when considering these procedures, one should carefully contemplate the risk of general anesthesia, perioperative risk and recovery time.

After RB, patients have to be monitored closely for the occurrence of complicatons such as gross hematuria, flank pain, hypotension and acute renal obstruction.

The standard practice after RB has traditionally been to observe the patient overnight, as suggested by early studies[47]. In our French survey, almost all nephrologists observed patients for at least 24 h after a native kidney biopsy[8]. However, controversies have emerged regarding the optimal duration of observation after RB and it has been proposed that patients be discharged after 6-8 h of observation[48,49]. Performing RB as an outpatient procedure offers several advantages but raises the concern of missing late complications. Whittier et al[22] reported a large series of 750 native kidney biopsies in adults. In this study, 13% patients developed biopsy-related complications; minor complications occurred in 6.6% and major complications (most requiring a blood transfusion) occurred in 6.4% patients. Around 30% of the patients had a biopsy performed using a manual biopsy device. The analysis of the timing of complications showed that 89% of complications were identified within 24 h after RB, and that an observation period less than 8 hours missed 33% of complications. On the contrary, several smaller studies suggested that outpatient observation time of 6 to 8 h is safe (Table 1)[13,19,49-57]. Most of outpatients in these studies were selected as low risk. Considering this, an outpatient biopsy may be an option in a carefully selected population with no risk factor.

Renal transplant biopsies are routinely performed as an outpatient procedure in some centers. In our survey in France, approximately 25% of nephrologists performed transplant biopsies with observation times limited to 4-8 h[8]. In a multicentric study by Furness et al[58] on 2127 protocol transplant biopsies, only 9 (0.42%) severe complications occurred, all presenting within four hours after biopsy. In another study, no severe complications were observed after 251 protocol transplant biopsies[59]. Therefore, an observation time of 4-8 h after a transplant biopsy appears to be a relatively safe practice, at least in patients without risk factors for bleeding.

Some protocols use a routine renal ultrasound or measurement of hemoglobin or hematocrit control before discharge, in addition to clinical monitoring. Systematic ultrasound reveals perirenal hematoma in 40%-90% of procedures[11,60]. Arteriovenous fistula may be detected in 10% of RB, but they usually disappear spontaneously after a few months[61,62]. In biopsies that are otherwise uncomplicated with an asymptomatic course, hematomas are usually small (< 3 cm)[48,63]. These hematomas are almost always asymptomatic, and such a finding usually occurs without therapeutic consequence. In a study that evaluated the use of renal ultrasound one hour post-RB, the presence of a hematoma was poorly predictive of complications[63]. The absence of a hematoma was predictive of an uncomplicated course in after RB[63]. However, a period of observation is required after RB, even in the absence of hematoma right after the biopsy. Early routine repeat imaging is therefore of limited usefulness and is not necessary in patients otherwise asymptomatic.

The use of a hemoglobin or hematocrit measurement after a RB as a predictor of bleeding is controversial. Systematic hemoglobin monitoring was shown to be of little value in detecting complications after RB in one study[22], although in another study, a direct relationship was found between the change of hematocrit at 6 h and the hematocrit at 24 h following a RB, suggesting that the absence of fall at 6 h makes a significant fall of hematocrit at 24 h unlikely[64].

The RB is an indispensable tool to establish the diagnosis and management of kidney diseases. Although the overall incidence of serious complications is low, risk factors for bleeding must be carefully assessed and, whenever possible, corrected before the procedure. If contraindications, especially high blood pressure and hemostasis abnormalities, are respected, percutaneous RB with an automated biopsy device and ultrasound guidance is safe for the vast majority of patients. Some controversies remain regarding the optimal duration of observation and the possibility to perform RB as an outpatient procedure. To address these issues, further studies are warranted to improve our ability to predict and stratify the risk of bleeding.

We are grateful to Dr. Stephan Troyanov for his assistance in improving the quality of writing.