A retrospective single-center study was conducted to compare the incidence of severe post-kidney biopsy bleeding between adult patients who were administered with desmopressin and patients who were not. Patients who underwent kidney biopsy at our institute from January 1, 2003 to March 31, 2019 were included. The exclusion criteria were as follows: (1) transplanted kidney biopsy (when kidney transplant was performed a day prior to the kidney biopsy or kidney transplant disease codes were available), (2) open kidney biopsy, (3) age <18 years at the time of kidney biopsy, and (4) mass biopsy for cancer diagnosis. A search using our center's Biomedical Research Environment revealed that 6877 patients underwent a kidney biopsy during the period specified above. Among these, 6518 patients were aged 18 years or above. After excluding for transplant kidney biopsy and open kidney biopsy, 4051 patients were identified; of these, 3018 patients were selected as the final study subjects. The patient‘s informed consent was not necessary because the data obtained was collected from clinical practice.

The following data were collected:

• 1.

  Desmopressin administration: Yes/No

• 2.

  Consumption of anticoagulants or antiplatelet drugs before kidney biopsy: Yes/No

• 3.

  Baseline characteristics at the time of kidney biopsy: age, sex, weight, height, and body mass index (BMI)

• 4.

  Comorbidities: diabetes mellitus, hypertension

• 5.

  Blood pressure just before and after the kidney biopsy

  • 1)

    Systolic blood pressure (SBP) (mmHg)

  • 2)

    Diastolic blood pressure (DBP) (mmHg)

  • 3)

    Mean arterial pressure (MAP) (mmHg)

• 6.

  Blood and urine test results obtained just before and after the kidney biopsy

  • 1)

    Hemoglobin (g/dL), platelet (×103/μL), hematocrit (%)

  • 2)

    Prothrombin time (PT international normalized ratio [INR]), activated partial thromboplastin time (sec)

  • 3)

    Serum creatinine (mg/dL), estimated glomerular filtration rate (eGFR) (ml/min/1.73m2), blood urea nitrogen (BUN) (mg/dL)

  • 4)

    Spot urine albumin/creatinine ratio (g/g), spot urine protein/creatinine ratio (g/g)

• 7.

  Kidney biopsy data

  • 1)

    Number of needle passes, number of biopsy segments

  • 2)

    Clinical indication for kidney biopsy

    • (i)

      Hematuria and/or non-nephrotic range proteinuria

    • (ii)

      Azotemia

    • (iii)

      Nephrotic syndrome

  • 3)

    Histological diagnosis

    • (i)

      Glomerulonephritis

    • (ii)

      Nephrosclerosis

    • (iii)

      Tubulo-interstitial nephritis

    • (iv)

      Lupus nephritis

    • (v)

      Diabetic nephropathy

    • (vi)

      Vasculitis

    • (vii)

      Amyloidosis

    • (viii)

      Others

  • 4)

    Department that performed biopsy

    • (i)

      Nephrology

    • (ii)

      Rheumatology

    • (iii)

      Allergy

    • (iv)

      Others

• 8.

  Bleeding events

  • (1)

    Renal artery embolization

  • (2)

    Blood transfusion (limited to red blood cell transfusion)

  • (3)

    Perinephric hematoma on computed tomography (CT) or ultrasound (US), Size of the hematoma

  • (4)

    Nephrectomy due to bleeding

  • (5)

    Length of hospital stay

The eGFR presented in our electronic medical record was estimated using the modification of diet in renal disease (MDRD) equation (MDRD eGFR) and the chronic kidney disease epidemiology collaboration (CKD-EPI) equation (CKD eGFR). In the old data at our institute, it was often expressed as either MDRD eGFR or CKD-EPI eGFR alone. The difference between the two was considered insignificant and the two data were applied together and used to evaluate eGFR statistics. If data for the computation of both were available, then the MDRD equation (which has data from more patients) was applied.

In order to identify bleeding events, cases where renal artery embolization and blood transfusion were performed within 7 days from the biopsy were investigated. Furthermore, in cases where CT or US was performed within 7 days from the biopsy, the presence of perinephric hematoma was investigated. Hematoma size data were included only when it was described in the official reading. The major bleeding events were defined as cases that underwent renal artery embolization and blood transfusion.

In all kidney biopsies, patients were provided with a detailed explanation of the protocol and informed consent was obtained. Prior to the procedure, the patient's complete blood count (CBC) and coagulation profile were checked. Transfusions were recommended for patients with hemoglobin levels lower than 10g/dL or with platelet counts lower than 100,000/μL. Patients who consumed anticoagulants or antiplatelet drugs prior to kidney biopsy underwent a period of discontinuation. Furthermore, aspirin and clopidogrel consumption was skipped for 7 days, cilostazol for 3 days, and warfarin until the PT INR was normal.

There was no standardized hospital protocol for the administration of desmopressin in this study. Desmopressin was administered to subjects who were deemed to have a high risk of bleeding by the attending physician; these subjects were typically those with a known high risk of bleeding, i.e., patients with impaired renal function and elevated BUN, old age, high blood pressure, low hemoglobin levels, and low platelet counts. Patients in the desmopressin group received a dose of 0.3μg/kg desmopressin (Minirin®, Ferring, Saint-Prex, Switzerland) in 100cc normal saline 30min before the procedure.

A 16- or 18-gauge semiautomated, side-notch disposable biopsy needle was used for the biopsy (STARCUT®, TSK Laboratory, Tochigi, Japan). To reduce the risk of bleeding, 18-gauge needles (9–12cm in length) were used in most cases.10 By using needles of a gauge thinner than that of the usual ones, the risk of bleeding was reduced. Therefore, even in patients with a relatively high bleeding tendency, punctures were repeated several times to obtain sufficient tissue. For native kidney biopsy, we aimed to collect more than 25 glomeruli and obtain an average of about three segments. The biopsies were conducted using real-time ultrasound guidance. After the procedure, a sandbag was placed on the biopsy site for 3h and the patient rested in a supine position. Then, after being transferred to the ward, the blood pressure and pulse rate were measured periodically, and the CBC was checked the next morning. If vital signs became unstable or other bleeding complications, such as gross hematuria or severe abdominal pain, were detected, CT or US was performed and renal artery embolization and blood transfusion were undertaken as required. CT or US was not performed routinely after kidney biopsy in patients without any symptoms.

This study was conducted in accordance with the ethical standards of the Declaration of Helsinki (as revised in Brazil, 2013). This research protocol was approved by the Institutional Review Board of our institute (No. 2019-0427).

When numerical data satisfied normality, they were expressed as mean±standard deviation, and a Student's t-test was performed to compare the means of the two independent groups. The chi-square test and Fisher's exact test were performed to compare categorical data between the two groups. The frequency of occurrence was expressed as a percentage (%). To compare the difference according to whether desmopressin was administered or not, propensity score matching (PS matching) was performed. After performing multiple imputations, the average of PSs estimated from each completed dataset was used. In case of a 2:1 or 3:1 matching, several patients in the desmopressin group were lost; thus, a 1:1 matching was performed. Post-matching outcome comparison was performed using logistic regression, with correlation allowed within the matching pair and a robust estimator. PS matching was performed using the R software, version 3.6.1, while the remaining tests were performed using IBM SPSS Statistics for Windows, version 23 (IBM Corp., Armonk, N.Y., USA). In addition, p<0.05 was considered statistically significant.

There were no significant differences in the weight, height, and BMI between the two groups; however, the age was significantly higher in the desmopressin group than in the no-desmopressin group (p<0.001). Furthermore, the SBP, DBP, and MAP were also significantly higher in the desmopressin group than in the no-desmopressin group (p<0.001). The hemoglobin, platelet, and hematocrit levels were significantly lower in the desmopressin group than in the no-desmopressin group (p<0.001, p=0.001, and p<0.001, respectively). The PT (p<0.001) and aPTT (p=0.025) were significantly higher in the desmopressin group than in the no-desmopressin group. Serum creatinine and BUN were significantly higher (p<0.001), while eGFR was significantly lower (p<0.001) in the desmopressin group than in the no-desmopressin group. The urine protein/creatinine ratio (p=0.007) was also significantly higher in the desmopressin group than in the no-desmopressin group. The use of anticoagulants or antiplatelet drugs before renal biopsy was significantly more in the desmopressin group as compared to in the no-desmopressin group (p<0.001) (Table 2).

The incidence of clinical indications for kidney biopsy were significantly different between the desmopressin and no-desmopressin groups (p<0.001). Hematuria and/or non-nephrotic range proteinuria (54.3% in the desmopressin group, 67.2% in the no-desmopressin group) were the most common indications. The incidence of azotemia as an indication was higher in the desmopressin group than in the no-desmopressin group (36.3% vs. 13.4%), while the incidence of nephrotic syndrome as an indication was higher in the no-desmopressin group than in the desmopressin group (19.4% vs. 9.4%). The histological diagnosis also differed between the two groups (p<0.001): Glomerulonephritis was the most common (66.8% in the desmopressin group, 58.3% in the no-desmopressin group), while lupus nephritis was more common in the no-desmopressin group (14.9% vs. 6.1%). The departments where the biopsy was performed differed significantly between the two groups (p<0.001); however, the nephrology department was the most common in both groups (87.0% in the desmopressin group, 76.8% in the no-desmopressin group) (Table 2).

The incidence of renal artery embolization did not differ significantly between the two groups (p=0.077); however, blood transfusions occurred significantly more frequently in the desmopressin group than in the no-desmopressin group (p<0.001). The frequency of perinephric hematoma (detected by CT and US) was significantly higher in the desmopressin group than in the no-desmopressin group (p<0.001). In both groups, there was no case in which nephrectomy was performed due to complications from renal biopsy. After biopsy, hemoglobin levels were significantly lower in the desmopressin group than in the no-desmopressin group (p<0.001); moreover, hemoglobin reduction was significantly higher in the desmopressin group (p<0.001). The size of the hematoma was significantly larger in the desmopressin group as compared to in the no-desmopressin group (p=0.016). Post-biopsy SBP (p<0.001) and MAP (p=0.002) were higher in the desmopressin group than in the no-desmopressin group; moreover, the reduction in SBP was significantly greater in the desmopressin group (p=0.007). Hospital stay was significantly longer in the desmopressin group than in the no-desmopressin group (p<0.001) (Table 3).

The age, sex, BMI, SBP, MBP, hemoglobin, platelet count, hematocrit, PT, aPTT, serum creatinine, eGFR, and BUN were chosen as variables that could influence bleeding events. Following PS matching for these variables, there were no differences in the variables between the two groups, as the standardized mean difference (SMD) was less than 0.1 (Table 4).

Comparison of the two adjusted groups revealed no intergroup differences in the incidence of renal artery embolization (p=0.341), blood transfusion (p=0.579), and total major bleeding events (p=0.442). Furthermore, total bleeding events (including perinephric hematoma on CT/US) did not differ between the two groups (p=0.239) (Table 5).

The degree of serum sodium reduction was significantly greater in the desmopressin group than in the no-desmopressin group by an average of 1.92 (p<0.001). Furthermore, the incidence of significant hyponatremia of less than 125mmol/L was significantly higher in the desmopressin group than in the no-desmopressin group (p=0.008) (Table 5).

Comparison of the two groups based on a creatinine level of 1.8mg/dL revealed that desmopressin did not significantly change the major bleeding risk in both groups (p=0.808 for serum creatinine less than 1.8mg/dL, p=0.482 for more than 1.8mg/dL, and p=0.870 indicating the difference between the two groups) (Table 6).