Safety and Complications of Percutaneous Kidney Biopsies in 715 Children and 8573 Adults in Norway 1988–2010

Summary

Background and objectives

Skepticism about performing renal biopsies is often because of uncertainty regarding risk of complications. The aim of this study was to evaluate safety and relevant complications of renal biopsies in pediatric and adult patients in a large national registry study.

Design, setting, participants, & measurements

Kidney biopsies reported in the Norwegian Kidney Biopsy Registry from 1988 to 2010 were included. Risk factors for major complications (blood transfusion and/or surgical or catheter intervention) were analyzed using logistic regression statistics.

Results

Of the 9288 biopsies included, 715 were from children, and 8573 were from adults (≥18 years). Median age was 49 years (range=2 weeks to 94 years). Gross hematuria appeared after biopsy in 1.9% of the patients; 0.9% of patients needed blood transfusion, and 0.2% of patients needed surgical intervention/catheterization. The frequencies were 1.9%, 0.9%, and 0.2% in adults and 1.7%, 0.1% and 0.1% in children, respectively; 97.9% of the biopsies were without complications. In unadjusted analyses, risk factors for major complications were age>60 years, estimated GFR<60 ml/min per 1.73 m², systolic hypertension, acute renal failure, and smaller clinical center size (<30 biopsies/yr). Adjusted analyses (adjusted for age and/or estimated GFR) showed higher odds ratios (OR) only for smaller clinical center (OR=1.60 [1.02–2.50]) and low estimated GFR (estimated GFR=30–59 ml/min per 1.73 m² [OR=4.90 (1.60–14.00)] and estimated GFR<30 ml/min per 1.73 m² [OR 15.50 (5.60–43.00)]).

Conclusions

Percutaneous renal biopsy is a low-risk procedure in all ages. Reduced estimated GFR and smaller center size are associated with an increased risk of major complications.

Introduction

Renal biopsies have been performed for more than a century (1), and percutaneous biopsy methodology was established in the 1940s (2,3). The technical advances in imaging diagnostics and biopsy procedure have evolved from indirect visualization to real-time ultrasound guidance (4). Most centers use an automated spring-loaded biopsy device constructed in the end of the 1980s (5–7). As a consequence, the general safety and complication rate of the procedure has substantially improved. Kidney morphology is essential in the diagnosis and subsequent evaluation of disease activity (8,9). Indeed, knowledge of the morphologic changes in kidney biopsies has been shown to alter the treatment in a high percentage of the patients (10,11). Postbiopsy bleeding (hematoma and hematuria) is the primary complication of renal biopsies (12–14). However, most of the perirenal hematomas are minor and without clinical significance, and major complications are infrequent in both adults and children when traditional risk factors, like hypertension and bleeding diathesis, are respected (15,16).

In this study, we report the procedure-related complications and safety issues in the period from April of 1988 to November of 2010 from the Norwegian Kidney Biopsy Registry comprising 8573 adult and 715 pediatric cases.

Materials and Methods

Norwegian Kidney Biopsy Registry has registered clinical and histopathologic data for nearly all patients with a native renal biopsy in Norway (current population of 5 million inhabitants) since April of 1988. All included patients or their designees signed informed consent. The local nephrologists (26 different hospitals) report clinical data in a registry notification form, and histopathological data are reported by a limited group of nephropathologists that examines all the biopsies in one center (Haukeland University Hospital). Information regarding biopsy-related complications was lacking in 357 (3.7%) reports, and these cases were excluded from the study. The remaining 9288 cases were included for the statistical analysis. All cases were biopsied with percutaneous techniques: 99.7% were ultrasound-guided, and 0.3% were computer tomography-guided.

In the report forms, there were five complication alternatives: hematoma, gross hematuria, blood transfusion, surgery, and a free text field. Angiographic embolization in the free text field was included in the surgery group. The pathologist reported whether the biopsy was representative (i.e., containing sufficient or insufficient material for adequate evaluation). The term complications used in the current study was confined to gross hematuria, blood transfusions, and/or surgery/arterial embolization (i.e., hematoma was not regarded as a significant complication unless associated with one of the other complications listed). The term major complications was defined as blood transfusions and/or surgery/arterial embolization. The following variables were explored: age, estimated GFR (eGFR; calculated with the Modification of Diet in Renal Disease formula in adults [17] and the revised Schwartz formula in children [18]), systolic BP, proteinuria (u-protein≥0.3 g/d or reported proteinuria or nephritic syndrome), hematuria (u-dipstick≥1 or reported hematuria or nephritic syndrome), CKD stages 3–5 (eGFR<60 ml/min per 1.73 m²), nephrotic syndrome (u-protein≥3.0 g/d and s-albumin<35 g/L or clinician-marked nephrotic syndrome), acute renal failure, rapidly progressive glomerulonephritis, amyloidosis, biopsy needle size (gauge), number of needle passes, specialty of the doctor performing the biopsy, and center size (≥650 or <650 biopsies performed during the study period corresponding to ≥30 or <30 biopsies/yr).

The SPSS 17 package (SPSS Inc., Chicago, IL) was used for statistical analysis. Significance testing was performed by t test and Mann–Whitney test. Differences of proportions between patient groups were compared with the chi-squared test (Pearson). A P value of 0.05 was considered statistically significant, and all tests were two-tailed. Logistic regression analysis was performed for the analysis of risk factors, and both unadjusted analyses and analyses adjusted for categories of age (categorized as <18, 18–59, and ≥60 years) and eGFR (categorized as ≥60, 30–59, and <30 ml/min per 1.73 m²) were performed. Age was only adjusted for eGFR, and eGFR was only adjusted for age.

Results

The median age of the 9288 patients was 49 years (range=2 weeks to 94 years). There were 715 children (<18 years of age), with a mean age of 12.0 years (SD=4.9; range=2 weeks to 18 years); 8573 patients were adults (≥18 years), with a mean age of 50.6 years (SD=17.7; range=18–94 years). The clinical characteristics at the time of renal biopsy are shown in Table 1. The most common clinical sign was proteinuria, which was seen in approximately four of five of all the patients, and similar figures were seen in both adults and children (Table 1). Hematuria and nephrotic syndrome were more frequent in children than adults, and chronic renal failure and acute renal failure were more common in adults; 5.9% of the biopsies in the study period were rebiopsies.

Table 1.

Clinical characteristics at the time of renal biopsy given as percentages (total numbers in parentheses) separately for adults and children

Clinical Characteristics	Total (%) (9288)	Adults (%) (8573)	Children (%) (715)	P Value
Proteinuria (>0.3 g/d)a	81.0 (7522)	81.1 (6955)	79.3 (567)	0.23
Hematuriab	68.8 (6393)	68.4 (5862)	74.3 (531)	0.001c
CKD stages 3–5d	61.6 (5726)	64.9 (5568)	22.1 (158)	<0.001c
Nephrotic syndromee	28.7 (2667)	27.8 (2380)	40.1 (287)	<0.001c
Acute renal failure	18.6 (1723)	19.0 (1630)	13.0 (93)	<0.001c
Rapidly progressive glomerulonephritis	3.4 (314)	3.3 (284)	4.2 (30)	0.21

^au-Protein≥0.3 g/d or clinician marked proteinuria or nephritic syndrome.

^bu-Dipstick≥1 or clinician marked hematuria or nephritic syndrome.

^cSignificant difference between adults and children.

^dCKD stages 3–5 indicate eGFR<60 ml/min per 1.73 m².

^eu-Protein≥3.0 g/d and s-albumin<35 g/L or clinician marked nephrotic syndrome.

The biopsy-related complications are shown in Table 2. Serious complications were infrequent in both children and adults (Table 2). The total frequency of blood transfusion or intervention (surgery or angiography) was low (0.9% and 0.2% of all cases, respectively). There was a significantly lower frequency of blood transfusions in children (only one child) than adults. Less than 2% had macroscopic hematuria (Table 2), whereas hematoma was reported in 3.9% of all cases. The frequency of hematoma was significantly higher in children (8.1%) than adults (3.5%; P<0.001). During the last 5 years of the study period, the number of hematomas was higher (7.4% compared with 3.9% in the whole study period), and this change was more evident in children (18.9% compared with 8.1% in the whole study period). The registry did not specify the criteria for the diagnosis of postprocedural hematoma in terms of size or associated clinical symptoms. The total number of hematoma should, therefore, be taken with caution and may reflect an increasing use of systematic safety routines, including postbiopsy ultrasound examination, especially in children. The rates of gross hematuria, blood transfusion, and invasive procedures remained stable throughout the study period (Figure 1).

Table 2.

Frequency of different complications given as percentages (total numbers in parentheses) separately for adults and children

Type of Complication	All Patients (%) (9288)	Adults (%) (8573)	Children (%) (715)	P Value
Gross hematuria	1.9 (178)	1.9 (167)	1.5 (11)	0.44
Blood transfusion	0.9 (79)	0.9 (78)	0.1 (1)	0.03a
Surgery/arterial embolization	0.2 (18)	0.2 (17)	0.1 (1)	0.73
Complicationsb	2.6 (215)	2.7 (233)	1.8 (13)	0.15
Major complicationsc	0.9 (88)	1.0 (86)	0.3 (2)	0.06

^aSignificant difference between adults and children.

^bComplications were gross hematuria, blood transfusions, and/or surgery/arterial embolization.

^cMajor complications were blood transfusions and/or surgery/arterial embolization.

Figure 1.

Kidney biopsy complications.

In the free text field, one patient was reported with arteriovenous fistula, one patient was reported with urosepsis and one patient was reported with ascites leakage. No patient was reported with death. A total of 97.9% of the patients (9092 patients) had no complications.

Figure 2 shows the change in biopsy needle size used in the study period. There was an overall shift from 14 to 16 gauge (G) size, and in the majority of the biopsies (82%), the needle size was 16 or 18 G (Table 3). A significantly higher rate of macroscopic hematuria was reported after use of 18 G compared with 16 G. Major complications (Table 3) were not related to needle size. The percentage of biopsies containing representative tissue was similar across all needle types (94.3%). The median number of glomeruli in the tissue samples per subject was 10 (range=0–86), and the number was lowest when an 18-G needle was used (9, range=0–76). In 3.0% of the biopsies, no glomeruli were found.

Figure 2.

Kidney biopsy needle sizes.

Table 3.

Comparison of complications and number of glomeruli for different needle diameter

Needle Thickness	14 Gauge	16 Gauge	18 Gauge	Unknown	Total
N total	858	2146	4998	744	8746
Percent total	9.8	24.5	57.1	8.5	100
Percent (N) complicationsa	2.6 (22)	2.0 (43)b	3.0 (151)	2.3 (17)	2.7 (233)
Percent (N) major complicationsc	0.6 (5)	1.2 (25)	0.9 (46)	0.8 (6)	0.9 (82)
Percent (N) gross hematuria	2.2 (19)	1.0 (22)d	2.4 (118)	1.6 (12)	2.0 (171)
Percent (N) blood transfusions	0.2 (2)	1.1 (23)	0.9 (43)	0.7 (5)	0.8 (73)
Percent (N) surgery/angiographic embolization	0.4 (3)	0.2 (5)	0.1 (7)	0.3 (2)	0.2 (17)
Percent representative tissue	94.6	93.5	94.5	94.9	94.3
Median number of glomeruli (range)	12 (0–86)d	12 (0–61)d	9 (0–76)	11 (0–69)d	10 (0–86)

^aComplications were gross hematuria, blood transfusions, and/or surgery/arterial embolization.

^bP<0.05 (P value compared with 18-gauge needle).

^cMajor complications were blood transfusions and/or surgery/arterial embolization.

^dP<0.001 (P value compared with 18-gauge needle).

Analysis of the number of needle passes per biopsy procedure showed that one or two passes were performed in the majority of the patients (47.5% and 37.9%, respectively). The biopsy was performed by a nephrologist in 33.4% of the cases, and this number decreased during the study period (data not shown). Radiologists performed the biopsy in 53.5%, and in 13.1% of the cases, the nephrologist and the radiologist handled the procedure together.

Risk factors for major complications (blood transfusions, surgery, or angiographic embolization) are analyzed in Table 4. Older age, lower eGFR, higher systolic BP, acute renal failure, and smaller center sizes were all highly significant risk factors in the unadjusted analyses. The unadjusted relative risk of major complications was significantly lower in children; when corrected by eGFR level, the risk was similar in children and adults. The needle size, number of passes, and physician specialty did not influence the rate of major complications. The diagnosis of rapidly progressive glomerulonephritis or amyloidosis did not carry any increased procedural risk. A lower eGFR (CKD stages 3–5) and smaller center size (<30 biopsies/yr) were the only significant risk factors after adjustment for age and/or eGFR.

Table 4.

Odds ratios for major complications (blood transfusion, surgery, and/or arterial embolization)

	N Total	N (%) with Major Complications	Unadjusted Analyses		Adjusted Analysesa
			Odds Ratio (CI)	P Value	Odds Ratio (CI)	P Value
Age (years)
<18	715	2 (0.3)	0.40 (0.10–1.70)		0.46 (0.06–3.40)
18–59	5609	39 (0.7)	1.00 (ref)	<0.001	1.00 (ref)	0.34
≥60	2964	47 (1.6)	2.30 (1.50–3.50)		1.20 (0.80–2.00)
eGFR (ml/min per 1.73 m2)
≥60	3307	4 (0.1)	1.00 (ref)		1.00 (ref)
30–59	2307	14 (0.6)	5.00 (1.70–15.00)	<0.001	4.90 (1.60–14.00)	<0.001
<30	3309	66 (2.0)	16.80 (6.10–46.00)		15.50 (5.60–43.00)
Systolic BP
<140	4141	27 (0.7)	1.00 (ref)		1.00 (ref)
140–159	2476	24 (1.0)	1.50 (0.86–2.60)	0.001	1.10 (0.61–1.90)	0.29
≥160	2127	32 (1.5)	2.30 (1.40–3.90)		1.30 (0.77–2.30)
Proteinuria (g/d)
0.3–0.99	2435	18 (0.7)	1.00 (ref)		1.00 (ref)
1–2.99	1933	23 (1.2)	1.60 (0.87–3.00)	0.28	1.60 (0.84–3.00)	0.14
≥3.0	3154	33 (1.0)	1.40 (0.80–2.50)		1.60 (0.89–2.90)
Acute renal failure
Yes	1723	30 (1.7)	2.29 (1.50–3.60)	<0.001	1.10 (0.69–1.80)	0.48
No	7576	58 (0.8)	1.00 (ref)		1.00 (ref)
Rapidly progressive glomerulonephritis
Yes	314	3 (1.0)	1.00 (0.31–3.20)	0.99	1.70 (0.54–5.50)	0.41
No	8974	85 (1.0)	1.00 (ref)		1.00 (ref)
Amyloidosis
Yes	255	3 (1.2)	1.21 (0.38–3.90)	0.75	0.92 (0.28–3.00)	0.88
No	6256	61 (1.0)	1.00 (ref)		1.00 (ref)
Needle size (gauge)
14	858	5 (0.6)	0.63 (0.25–1.60)		0.79 (0.31–2.00)
16	2146	25 (1.2)	1.30 (0.78–2.10)	0.55	1.40 (0.85–2.30)	0.92
18	4998	46 (0.9)	1.00 (ref)		1.00 (ref)
Unknown	744	6 (0.8)	0.88 (0.37–2.10)		0.93 (0.39–2.20)
Number of needle passes
1	698	6 (0.9)	1.60 (0.40–6.40)		1.80 (0.44–7.10)
2	557	3 (0.5)	1.00 (ref)	0.18	1.00 (ref)	0.35
3	158	4 (2.5)	4.80 (1.10–22.00)		4.00 (0.87–18.00)
≥4	56	1 (1.8)	3.40 (0.34–33.00)		2.80 (0.28–28.00)
Specialty
Nephrologist	2975	22 (0.7)	1.00 (ref)		1.00 (ref)
Radiologist	4766	50 (1.0)	1.40 (0.86–2.40)	0.17	1.10 (0.65–1.80)	0.66
Both	1163	30 (2.6)	1.50 (0.76–3.00)		1.20 (0.58–2.30)
Center size (biopsies in study period)
≥650	4873	32 (0.7)	1.00 (ref)		1.00 (ref)
<650	4415	56 (1.3)	1.90 (1.30–3.00)	0.003	1.60 (1.02–2.50)	0.04

CI, confidence interval; eGFR, estimated GFR.

^aCorrected for categories of age and eGFR. Age is only corrected for eGFR, and eGFR is only corrected for age.

Discussion

The main finding in the present study is that a renal biopsy should be regarded as a safe procedure (both in adults and children) provided that general contraindications are respected, and the overall rate of major complications was less than 1% (Table 2). Of note, the relative risk was even lower in children, and only 2 of 715 children (0.3%) were affected. Importantly, the lowest complication rates were seen in centers performing more than 30 biopsies/yr, indicating the best clinical routines occur when biopsies are done approximately on a weekly basis.

To our knowledge, the current study is the only nationwide registry-based and largest report of renal biopsy complications, and our findings confirm the general impression of low complication rates and improved safety of modern renal biopsy procedures (19). Similar low complication rates have been reported in recent studies (16,19,20) in contrast to earlier studies (reporting 5%–7% complication rates) (21–25). The crucial safety evaluation of kidney biopsies is characterized by the overall risk and frequency of major complications, which in our study, are defined as need of blood transfusion and/or surgery or catheter-based intervention (embolization) secondary to bleeding, and they should not be jeopardized by procedural events of minor clinical relevance. Therefore, the description of kidney biopsy complications should be restrained to clinically meaningful complications, especially because there often is an unfounded skepticism toward biopsies because of many reports of frequent occurrence of clinically less relevant events, like microscopic hematuria, transient gross hematuria, and asymptomatic minor perirenal hematoma. The latter should, in fact, be regarded as an epiphenomenon because of the high incidence (reported in up to 60%–90% of cases after examination prospectively with ultrasonography or computertomography). Of note, several authors found that a hematoma alone is not a reliable predictor of a serious adverse outcome (7,16,26,27). Perirenal hematomas should not be regarded as a clinical significant complication in the absence of significant bleeding or persisting pain necessitating delayed hospitalization. The number of hematomas in the present study was significantly higher in children, probably because of a more systematic use of routine ultrasonographic examination, and this finding must, therefore, be interpreted with caution. Such a view is supported by several studies. The work by Kersnik Levart et al. (16) found mostly small perinephritic hematoma in 63.2% of 87 children, and 10 of the hematomas (11.5%) were more than 2 cm in diameter. Other authors report that the majority of hematomas are less than 2 cm, and about 1%–2% are described as symptomatic (15,28,29).

In the present large study, gross hematuria was seen in less than 2%, and the benign course supports the view that such events should not be considered serious complications unless bleeding is of a magnitude causing a significant decline in hemoglobin concentration or delayed hospitalization. The great majority of the patients (97.9%) had neither macroscopic hematuria nor blood transfusion or surgery/embolization. These findings are in line with other studies (27,30–32).

In the current study, the risk factors for major complications were older age, low eGFR, systolic hypertension, acute renal failure, and smaller center size (<30 biopsies/yr) in the unadjusted analyses. Only low eGFR and smaller center size were associated with increased risk in the adjusted analyses. In fact, the relative risk was 16 times higher when eGFR was below 30 ml/min per 1.73 m² (Table 4). Although elevated risk has been shown previously, the increased risk incurred by kidney failure may be underscored in many reports (14,21,33,34), and the work by Whittier and Korbet (25) found that serious complications were about two times as common in patients with serum creatinine above 5.0 mg/dl. It is conceivable that this observation is associated with the general increased bleeding tendency seen in kidney failure patients (33). Similarly, acute renal failure more than doubled the risk of major complications in our study, and the reason for this finding is probably also related to increased bleeding tendency (Table 4). As a consequence, careful recognition of reduced kidney function before the procedure is important. To minimize biopsy risk, meticulous control of clinical routines (to avoid inappropriate use of anticoagulant medication 1 week before as well as after the biopsy procedure) is mandatory, especially in patients with reduced eGFR.

An overview of studies from many centers shows the wide variation in complication rates (27). In the current nationwide registry study, we found the lowest complication rates in the most active and experienced centers (e.g., university hospitals performing biopsies, on average, on a weekly basis or >30 biopsies/yr). This finding is a reassuring observation that strengthens the importance of introducing adequate training programs (27) and supports the view that modern biopsy procedures are safe in experienced hands. Furthermore, the general low complication rate indicates similar adequate clinical practice among our centers. This finding is in contrast to two recent nationwide surveys from the United Kingdom (19) and France (35) showing relatively great variability in clinical practice and complication rates among centers and highlighting the importance of establishing safe procedural standards.

Although there has been a change in practice (fewer patients with isolated microscopic hematuria and more elderly patients undergo a biopsy), the total frequency of renal biopsies in Norway has remained stable for the last 20 years and was 10.9/100,000 in 2009 (13/100,000 in adults and 3.8/100,000 in children). Other reports have shown a range of biopsy frequencies between 3.3 and 23 per 100,000 inhabitants (36,37).

In many centers, the numbers of kidney biopsies are relatively low in children because of a traditional view of many nephrologists and pediatricians that the procedure is associated with an unacceptable high complication risk. General skepticism to the routine use of anesthesia is probably also contributing to the restraint of kidney biopsies. In our cohort, kidney biopsies in children below the age of 15 years were usually performed with short-lasting general anesthesia (sedation) without intubation. Our experience supports the findings in the work by Mauer and Drummond (38) that general anesthesia often can be avoided in patients above the age of 12 years (9). No complications caused by general anesthesia were reported in our study. Hence, the present encouraging findings should contribute to attenuating unfounded skeptic attitudes against kidney biopsies in children as long as general clinical safety routines are respected (14).

Valid biomarkers are generally lacking in many renal diseases, and several important studies have shown the role of kidney morphology as an important prognostic and therapeutic guide in common renal diseases (e.g., IgA nephritis [39], lupus nephritis [40], and diabetes nephropathy [41]); potential prognostic capacity has been shown in orphan diseases like Fabry disease (42). Furthermore, modern therapies of kidney diseases include potent and potentially toxic drug intervention in many common as well as a rapidly increasing number of rare diseases identified by genetic molecular techniques. The latter is mirrored by an increasing general focus from biotechnology companies on development of orphan drugs (43). Importantly, the Food and Drug Administration and the European Medical Association have recently presented guidelines that highlight the importance of meticulous safety and follow-up control of tolerance and side effects of new medical treatment in children (44). Hence, this fundamental change of attitude should also be reflected in the use of evidence-based risk evaluation when it comes to indications of renal biopsy.

It has been suggested that the complication rate is lower with the use of thin biopsy needles (27,45). However, valid comparative studies on needle size and complications in native kidneys are lacking. In the current study, the shift of needle size from 14 to 16 G over time (Figure 2) and the dominance of 16- and 18-G needles (Table 3) may be seen as a result of higher focus, in general, on minimizing risk factors and the acceptance of less tissue per needle pass. The median number of glomeruli per subject in our study (10, range=0–86) was comparable with the median number in other studies (7) and significantly higher by use of both 14- and 16-G needles compared with 18-G needles (P<0.001). However, the percentage of biopsies characterized as representative tissue was in the same level irrespective of needle size. Surprisingly, there was a significantly higher rate of gross hematuria by use of 18 compared with 16 G. This finding has not been reported previously and may be because of the tendency of thinner needles to deviate from the lower pole sector in a proximal direction (where vessel density is higher), with subsequent bleeding into the pelvis. Of note, serious complications, like surgery/embolization and blood transfusions, were of the same order in all needle sizes (Tables 3 and 4).

Interestingly, we did not observe any influence of number of needle passes on the rate of serious complications. The reason for this finding is not known, but it may be related to restraint with multiple needle insertions in high-risk patients and children. In our study, the majority of the biopsy cases (85.4%) had one or two needle passes, and the percentage of patients with three or more passes was low (14.6%). This finding is in contrast to the recent British audit, where the standard target of >80% needing three or less passes was achieved in 86.4% (19). No difference in complication rate was seen between different medical specialists. In our study, most biopsies were performed by a radiologist, whereas 33.4% were done by a nephrologist; the latter percentage has been decreasing over the last decade as reported previously, and it represents an educational challenge in nephrology (27).

Limitations of this study are potential underreporting, reporting bias, and lack of detailed information about other potential complications not prespecified in the registry report form. All GFR analyses were based on eGFR, although imprecision of eGFR is well known in subgroups of patients.

In conclusion, we have shown that kidney biopsies are safe procedures with low complication rates in children and adults when contraindications are respected. Lower eGFR and biopsies performed in smaller centers are risk factors.

Disclosures

None.