Renal haemorrhage risk after extracorporeal shockwave lithotripsy: results from the Japanese Diagnosis Procedure Combination Database


Toru Sugihara, Department of Urology, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. e-mail:


Study Type – Therapy (case series)

Level of Evidence 4

What's known on the subject? and What does the study add?

Renal haemorrhage is a severe adverse event of extracorporeal shock wave lithotripsy with an incidence of about 0.5%. This rarity had made comparative studies among lithotripter models difficult. This study examines a large number of cases and models to reveal risk factors for postoperative renal haemorrhage.


  • • To assess clinical and mechanical risk factors of clinically significant renal haemorrhage after extracorporeal shock wave lithotripsy (ESWL).


  • • Patient data were extracted from the Diagnosis Procedure Combination (DPC) database from 6 months per each year, 2006–2008. The availability of lithotripters in each hospital was identified. We performed logistic regression analysis, which included the generator type (electrohydraulic, electromagnetic or piezoelectric), age, gender, laterality of stones (right, left or uncertain), location of stones (kidney, ureter or uncertain), total number of treatment sessions, anaesthesia and hospital volume (HV), focal size (greater or less than 400 mm3) and F2 angle (greater or less than 70°). Renal haemorrhage events were identified within the database.


  • • Overall, 81 renal haemorrhage events in 26 969 patients (32 476 ESWL sessions) at 482 hospitals with 38 lithotripter models were identified. The incidence of events was 0.50% with renal stones and 0.14% with ureter stones. Specifications of 34 lithotripter models were available. Use of piezoelectric lithotripters (vs electromagnettic, OR 0.13, P= 0.044) and high HV (≥140/year, vs ≤70/year, OR 0.49, P= 0.012) significantly decreased the risk of renal haemorrhage events. Age, gender, focal size and F2 angle did not show statistical significance.


  • • There is a low incidence of renal haemorrhage after ESWL. The less invasive nature of piezoelectric lithotripters and an inverse volume–outcome relationship with ESWL procedures was revealed. Age, focal size and F2 angle do not appear to have a significant impact on renal haemorrhage.

Diagnosis Procedure Combination


extracorporeal shock wave lithotripsy


hospital volume


International Classification of Diseases and Related Health Problems, tenth revision.


Extracorporeal shock wave lithotripsy (ESWL) is one of the most common treatments for urinary tract calculi. Since the development of the first commercially available lithotripter, the Dornier HM-3, in the early 1980s, dozens of modified and improved lithotripters have been introduced into clinical use [1]. There are three types of lithotripters and they are generally classified based on their shock wave generators: electrohydraulic (spark gap), electromagnetic and piezoelectric.

Renal haemorrhage and subcapsular haematoma, although rare, are severe adverse events which can be life-threatening [1,2]. The shock waves themselves or cavitation bubbles surrounding the focus stone can lead to injury [3–5]. The previously reported incidence of clinically significant renal haemorrhage is approximately 0.5% [2,6–8]. This rarity had made comparative studies among lithotripters difficult [3].

In addition, little is known about the impact of hospital volume (HV) on ESWL outcomes. HV is the annual operative caseload and known to be associated with operative outcomes [9–11].

In Japan, more than 900 lithotripters were available for clinical use in 2008 [12]. The present study evaluated the incidence of and risk factors for renal haemorrhage in Japan using the Diagnosis Procedure Combination (DPC) database, a nationwide administrative database.


The DPC database is an inpatient care database in Japan, which includes administrative claims and discharge data [9]. The data comprises i) main diagnoses and comorbidities at admission and complications after admission, recoded by International Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes and the Japanese letters, ii) surgical procedures coded by original Japanese K-codes and iii) discharge status [9]. The database started at 82 teaching hospitals in 2002 and the number of participating hospitals gradually rose to 262 in 2006, 926 in 2007 and 855 in 2008. The number of patients included in the database was 1.08, 2.99 and 2.86 million in 2006, 2007 and 2008, respectively. The number of patients in 2008 represented approximately 40% of all acute care inpatient hospitalizations in Japan.

Data were collected from July 1 to December 31 2006–2008 (6 months per each year). The patients included in the present study were those who were diagnosed with ‘hydronephrosis with renal and ureteral calculous obstruction’ (ICD-10 code; N13.2), ‘calculus of kidney’ (N20.0), ‘calculus of ureter’ (N20.1) and ‘calculus of kidney with calculus of ureter’ (N20.2), and those who underwent ‘ESWL for urolithiasis’ (assigned K-code K768). Patients were excluded from the study if they were under the age of 15 years. Given the anonymous nature of the data collection process, informed consent was not required. Study approval was obtained from the Institutional Review Board in the University of Occupational and Environmental Health, Japan.

The following information was extracted for each patient: gender, age, laterality of stones (right, left or uncertain), location of stones (kidney, ureter or uncertain), total number of treatment sessions (one, two or more) and anaesthesia use. Laterality and location were determined from the diagnosis reports, which were recorded in Japanese.

In the present study, HV for ESWL in each hospital was determined by using the unique identifier of each hospital. HV was categorized into three groups (i.e. low, medium and high volume) to ensure that there was an equivalent number of patients included in each group. After the data extraction process, levels of low, medium and high volume were defined as ≤70 (n= 8992), 71–139 (n= 9025) and ≥140 ESWLs/year (n= 8925), respectively.

The types of lithotripters used in each hospital were identified from a Japanese medical device magazine, ‘Shin-Iryo’, which presented about 900 lithotripter locations [12] or an online search (e.g. hospital websites). Data on focal size and angle of the second geometrical focus (F2 angle) were also collected. Focal size is defined as the volume within which the shock wave pressure remains greater than 50% of the maximum pressure. F2 angle was calculated from the aperture of the shock wave reflector and the depth of F2. Focal sizes and F2 angles were converted into dichotomous variables (i.e. smaller or larger) to ensure that patient numbers in each group were equivalent. The specifications assessed in the present study are the necessary data for lithotripter approval and clinical use in Japan.

Renal haemorrhage was the major outcome in the present study. The events were detected with diagnoses which contain Japanese terms ‘renal haemorrhage’ or ‘renal haematoma’, since there are no appropriate ICD-10 codes which directly point out the renal haemorrhage or haematoma. Validity of the extracted disease names was also manually verified. Furthermore, not only detected admissions for ESWL, but also readmissions within 30 days after discharge for ESWL hospitalization were assessed.

Univariate comparisons of each variable were performed using a chi-squared test. Next, we performed logistic regression analysis on renal haemorrhage events adjusting for the concurrent effects of age, gender, HV, generator type (electrohydraulic, electromagnetic or piezoelectric type), laterality, location of stones, total treatment sessions, anaesthesia use, focal size and F2 angle. The threshold for significance was a P < 0.05. All P values and 95% CIs were calculated using PASW version 18.0 (SPSS Inc. Chicago, IL, USA).


A total of 32 476 ESWL sessions and 26 969 patients from 482 hospitals were identified using the DPC database and lithotripter matching at each hospital. There were 15 electrohydraulic lithotripter models in 124 hospitals, 17 electromagnetic lithotripter models in 318 hospitals and six piezoelectric lithotripter models in 40 hospitals. There were 81 renal haemorrhage events detected and four of those events were readmission cases.

Patient demographics, distribution of lithotripters and renal haemorrhage events are presented in Table 1. The median age (interquartile range) was 56 (44–65) years. The ratio of males to females was 2.34 (P < 0.001). Among the shock wave generator types, there was a significant difference in age, stone location, total treatment sessions, anaesthesia use and HV. In the piezoelectric generator group, anaesthesia was used less frequently than with the other generator types.

Table 1.  Patient demographics, distribution of lithotripters and renal haemorrhage events
Generator typeElectrohydraulicElectromagneticPiezoelectric P valueTotal
n (%) n (%) n (%) n (%)
  • *

    P < 0.05,

  • P < 0.01, P value calculated using chi-squared test among each subcategory.

Lithotripter models15176 38
Hospitals12431840 482
Treatment sessions791321 9282635 32 476
Patients654118 3942034 26 969
 Male4650 (71.1)12 833 (69.8)1407 (69.2)0.09018 890 (70.0)
 Female1891 (28.9)5 561 (30.2)627 (30.8)8 079 (30.0)
Age (years)     
 15–29248 (3.8)698 (3.8)74 (3.6)0.033*1 020 (3.8)
 30–492102 (32.1)5 676 (30.9)631 (31.1)8 409 (31.2)
 50–693163 (48.4)8 874 (48.2)948 (46.6)12 985 (48.1)
 ≥701028 (15.7)3 146 (17.1)381 (18.7)4 555 (16.9)
 Right1793 (27.4)5 065 (27.5)554 (27.2)0.6387 412 (27.5)
 Left2243 (34.3)6 395 (34.8)733 (36.1)9 371 (34.7)
 Uncertain2505 (38.3)6 934 (37.7)747 (36.7)10 186 (37.8)
Stone location     
 Kidney1649 (25.2)4 826 (26.2)541 (26.6)<0.0017 016 (26.1)
 Ureter4141 (63.3)10 801 (58.7)954 (46.9)15 896 (58.9)
 Uncertain751 (11.5)2 767 (15.1)539 (26.5)4 057 (15.0)
Total treatment sessions     
 15658 (86.5)15 972 (86.8)1719 (84.5)0.014*23 349 (86.6)
 ≥2883 (13.5)2 422 (13.2)315 (15.5)3 620 (13.4)
Anaesthesia use     
 No5421 (82.9)15 674 (85.2)1894 (93.1)<0.00122 989 (85.2)
 Yes1120 (17.1)2 720 (14.8)140 (6.9)3 980 (14.8)
Hospital volume     
 Low (≤70/year)2337 (35.7)5 864 (31.9)791 (38.9)<0.0018 992 (33.3)
 Medium (71–139/year)1705 (26.1)6 640 (36.1)680 (33.4)9 025 (33.5)
 High (≥140/year)2499 (38.2)5 890 (32.0)563 (27.7)8 952 (33.2)
Renal haemorrhage18621 81

Lithotripter specifications and the prevalence of renal haemorrhage events are listed in Table 2. No lithotripter had a high incidence of renal haemorrhage (P= 0.820). Specification data were available on 34 models, which covered 97.3% (n= 26 243) of all patients. Focal sizes were divided into two categories: >400 mm3 (n= 13 286) or ≤400 mm3 (n= 12 957). Also, F2 angles were divided into two categories: >70° (n= 14 268) or ≤70° (n= 11 975).

Table 2.  Lithotripter specifications and the prevalence of renal haemorrhage events
 PatientsRenal haemorrhageFocal sizeF2 angle
n (%)*W × L (mm)Volume (mm3)>400 mm3Aperture (cm)F2 depth (cm)Angle>70°
  • *

    P= 0.820 (among the lithotripter models by chi-squared test).

  • Volume = W/2 × W/2 × L ×π/3.

  • F2 Angle = 2 × Arctan (Aperture/2/F2 depth). NA,not available.

Electrohydraulic type         
  HM3560 (0.00%)10 × 25653Yes15.61362°No
  MFL50001 0192 (0.20%)5.4 × 38289No171366°No
  MPL90001570 (0.00%)4.2 × 34157No20.51281°Yes
  Sonolith 3000410 (0.00%) NA   NA 
  Sonolith Praktis1 6639 (0.54%)3 × 2866No21.91380°Yes
  Sonolith VISION1 1252 (0.18%)3 × 2866No21.91380°Yes
  Tripter X-1680 (0.00%)5 × 17111No1813.567°No
  New Tripter Nova4613 (0.65%)5 × 17111No1813.567°No
  Nova Ultima4650 (0.00%)5 × 17111No181465°No
  Duet1080 (0.00%)5 × 17111No181465°No
  Integra240 (0.00%)6 × 50471Yes201471°Yes
  Econolith 20006692 (0.30%)13 × 582566Yes1713.564°No
  Litho Diagnost M1070 (0.00%) NA   NA 
  Litho Diagnost ME2030 (0.00%) NA   NA 
 Health Tronics         
  Litho Tron3750 (0.00%) NA   NA 
Electromagnetic type         
  Compact S5870 (0.00%)7 × 801027Yes12.51352°No
  Lithotripter U506431 (0.16%)3 × 57135No2214.574°Yes
  Compact Delta3 5129 (0.28%)7.5 × 781149Yes221573°Yes
  Lithotripter S1 7575 (0.26%)3 × 71166No141550°No
  Compact Sigma4140 (0.00%)7.5 × 781149Yes141550°No
  Lithotripter S II1960 (0.00%)3 × 2763No141550°No
  Compact Delta II2 0065 (0.25%)8 × 1071793Yes221573°Yes
  Lithostar 2810 (0.00%)9 × 601272Yes12.51255°No
  Lithostar 2 Plus1330 (0.00%)3.5 × 95305No121837°No
  Lithostar Multiline4 09821 (0.51%)4.5 × 95502Yes14.51262°No
  Lithostar Modularis C3792 (0.53%)8 × 961608Yes12.51255°No
  Lithostar Modularis C plus1701 (0.59%)6 × 80754Yes141260°No
  Lithoskop1 2905 (0.39%)8 × 1402346Yes15.81653°No
 Karl Storz         
  Modulith SL-203041 (0.33%)6 × 28264No3016.584°Yes
  Modulith SLX-UX4091 (0.24%)6 × 28264No3016.584°Yes
  Modulith SLX-MX1 5618 (0.51%)6 × 28264No3016.584°Yes
 Modulith SLX-F28543 (0.35%)6 × 28264No301880°Yes
9 × 28
Piezoelectric type         
  LT02610 (0.00%)4 × 37155No301494°Yes
  LT02X1950 (0.00%)4 × 37155No301494°Yes
  Piezolith 230060 (0.00%)4 × 1146No5015118°Yes
  Piezolith 25007800 (0.00%)4 × 1146No5015118°Yes
  Piezolith 30007901 (0.13%)4 × 25105No261582°Yes
  ESL-500A2020 (0.00%)3 × 2047No5014121°Yes
Total26 96981 (0.30%)       

Table 3 shows the results of the chi-squared tests and the logistic regression analysis for the incidence of renal haemorrhage after ESWL. Using the piezoelectric lithotripter (odds ratio (OR) 0.13, 95% CI 0.02, 0.95, P= 0.044, compared with electromagnetic type), having ureter stones (OR 0.29, 95% CI 0.17, 0.49, P < 0.001, compared with kidney stones) and being treated in a high volume hospital (OR 0.49, 95% CI 0.28, 0.85, P= 0.012, compared with low volume) significantly reduced the risk of renal haemorrhage. Neither focal size nor F2 angle had a significant effect on the incidence of renal haemorrhage events.

Table 3.  Univariate and multivariate analysis of renal haemorrhage events after extracorporeal shock wave lithotripsy (ESWL)
 PatientsRenal haemorrhageUnivariate analysisMultivariate analysis
n (%) P valueOR (95% CI) P value
  • *

    P < 0.05,

  • P < 0.01.

Generator type     
 EH5 79718 (0.31) 0.933 (0.514, 1.692)0.819
 EM18 33262 (0.34) Reference 
 Piezoelectric2 0331 (0.05)0.0850.126 (0.017, 0.946)0.044*
 Male18 31759 (0.32) Reference0.374
 Female7 84522 (0.28)0.5790.798 (0.485, 1.313)
Age (years)     
 15–299943 (0.30) Reference 
 30–498 13920 (0.25) 0.793 (0.234, 2.678)0.708
 50–6912 61848 (0.38) 1.136 (0.352, 3.666)0.831
 ≥704 41110 (0.23)0.2530.678 (0.185, 2.483)0.557
 Right7 19819 (0.26) Reference 
 Left9 09726 (0.28) 1.052 (0.580, 1.902)0.870
 Uncertain9 86736 (0.36)0.4451.355 (0.776, 2.367)0.286
Stone location     
 Kidney6 82535 (0.51) Reference 
 Ureter15 35523 (0.15) 0.290 (0.171, 0.493)<0.001**
 Uncertain3 98223 (0.57)<0.0011.203 (0.705, 2.052)0.498
Total treatment sessions     
 122 64064 (0.28) Reference0.058
 ≥23 52217 (0.48)0.048*1.703 (0.982, 2.954)
Anaesthesia use     
 No22 56372 (0.32) Reference0.336
 Yes3 5999 (0.25)0.4900.709 (0.352, 1.429)
Hospital volume     
 Low (≤70/year)8 84737 (0.42) Reference 
 Middle (71–139/year)8 71124 (0.27) 0.680 (0.402, 1.151)0.152
 High (≥140/year)8 60420 (0.23)0.0690.488 (0.279, 0.852)0.012*
Focal size     
 ≤400 mm312 92136 (0.28) Reference0.942
 >400 mm313 24145 (0.34)0.3741.019 (0.610, 1.704)
F2 angle     
 ≤70°11 93436 (0.34) Reference0.603
 >70°14 22839 (0.28)0.3670.885 (0.560, 1.400)


In the present study, the incidence of renal haemorrhage events was 0.50% with renal stones and 0.14% with ureter stones.

Corroborating our findings, the previously reported incidences (per patient) of clinically significant renal haemorrhage were around 0.5%; i.e. 0.66% (21/3208) by Knapp et al. [8], 0.77% (5/645) by Newman et al. [7], 0.08% (18/19 962) by Mobley et al. [2], 0.28% (31/10 953) by Collado Serra et al. [5] and 0.72% (3/415) by Dhar et al. [6] The detection rate of renal haemorrhage was reported to be 4.1% when assessed by sonography [6], 19% by computerized tomography [13] and 29% by magnetic resonance imaging [14].

We also confirmed the effect of HV on the renal haemorrhages after ESWL. These inverse volume–outcome relationships are recognized in numerous medical fields, including urology [9–11]. Our result suggests that the so-called ‘practice makes perfect’ theory is also applicable to ESWL procedures. In fact, there are some recommendations for best practices to reduce tissue damage [15], such as keeping shock wave rate under 1.0 Hz [4], ramping the shock wave power in a stepwise fashion [2], inserting a brief pause between the ramping steps [16] and using an appropriate gel for fitting the therapy head to the patient [17].

Of the three different shock wave generators available, we established that the piezoelectric type was associated with a significantly lower incidence of renal haemorrhage than the other types (adjusted OR 0.12). The piezoelectric generator consists of small polarized polycrystalline ceramic elements, and owing to the limited power of each element, a large spherical dish (large F2 angle) is required to produce a sufficient shock wave pressure. The advantages of using this generator are that low energy levels are delivered through the skin and renal parenchyma and that it requires less anaesthesia [3]. This characteristic of requiring less anaesthesia with piezoelectric generator use was confirmed in the present study. The proportion of anaesthesia use among piezoelectric users was 6.9% which was significantly lower than 17.1% for electrohydraulic users and 14.8% for electromagnetic users (Table 1).

Raeman et al. [18] found that a piezoelectric lithotripter is less destructive to murine kidney tissues than an electrohydraulic lithotripter. Graber et al. [19] compared kidney trauma between the electrohydraulic lithotripter ‘HM3’ and the electromagnetic lithotripter ‘Lithostar Plus’ by measuring urinary N-acetyl-β-glucosaminidase and β-galactosidase and concluded that there were no significant differences.

Dhar et al. [6] reported that the probability of haematoma increases 1.67 times for every 10 year increase in age. However, in the present study, age did not significantly affect the outcome. According to the analysis conducted by Dhar et al. [6], haemorrhage events that were detected by routine screening with sonography after ESWL were included, and as a result, the incidence of events reached 4.1% (17 patients). However, of these 17 patients, only three events were symptomatic (0.7%). Thus, differences in the definition of haematoma events can influence the interpretation of the results.

Several comorbidities are known to be risk factors for renal haemorrhage. Knapp et al. [8] reported that pre-existing hypertension, especially if poorly controlled, increased the risk of post-ESWL haemorrhage. The rates were 0.66% overall, 2.5% in hypertensive patients and 3.8% in poorly-controlled hypertensive patients. Atherosclerosis caused by hypertension may make vessels vulnerable to shock waves [7,20]. On the other hand, Dhar et al. [6] could not find a relationship between mean arterial blood pressure and renal haematoma. In addition, coagulation disorders [5], diabetes mellitus, coronary artery disease, obesity [7] and generalized atherosclerosis [21] were suggested as risk factors for post-ESWL haemorrhage in several retrospective case series. Unfortunately, in the present study, less than 5% of patients were assigned ICD-10 codes associated with such conditions as hypertension, diabetes mellitus or anticoagulant/antiplatelet medication. Because of this, we elected not to add these comorbidities to the explanatory variables.

Newer generation lithotripters tend to have a smaller focus and a high peak pressure. Some investigators suggest that a narrow focal size is associated with higher re-treatment rates and a greater incidence of adverse events, including renal haematoma [3,22,23]. We analyzed the influence of two mechanical factors, focal size and F2 angle, among three models, and did not observe a significant difference. It should be mentioned that certain specifications of each lithotripter were not taken into consideration because of their complexity and/or unavailability of information regarding the parameters used in the clinical setting, such as the number of shock waves, voltage and frequency.

There were other limitations in the present study. Because our study was based on an administrative claims database and the definition of renal haemorrhage depended completely on the recording physician, we could not evaluate the clinical severity and our ability to determine the validity of coded diagnoses was limited. Furthermore, the present study focused on inpatient cases only, and consequently there may be a sample bias towards patients who are more seriously ill. However, according to the Survey of Medical Care Activities in Public Health Insurance in 2008 [24], the ratio of inpatient to outpatient ESWL cases was approximately 1.2. This indicates that, in Japan, inpatient ESWL therapies are as common as outpatient ones. Also, since the DPC database collects data from July through to December, the total numbers of treatment sessions of each patient were underestimated. According to a questionnaire survey by the Japanese Society of Endourology and ESWL [25], the average number of treatment sessions with ESWL in 406 hospitals was estimated to be around 1.5–1.9 sessions, while in the present study, the average number of sessions was 1.2. Lastly, some important clinical parameters, such as stone size, component, lithotripter setting parameters, equipped stone localizers (fluoroscopy or sonography), were not recorded in the database. Despite these limitations, using the nationwide database enabled us to analyze how frequently renal haemorrhage events happened after ESWL based on large, real-world clinical data.

In conclusion, we revealed that the incidence of renal haemorrhage was 0.50% with renal stones and 0.14% with ureter stones. Piezoelectric lithotripters and a high HV (≥140/year) significantly decreased the risk of renal haemorrhage. Multiple ESWL sessions nearly reached the significance level. There was no significant impact of age, focal size and F2 angle on the outcome.


None declared. Source of funding: this study was funded by a Grant-in-Aid for Research on Policy Planning and Evaluation from the Ministry of Health, Labour and Welfare, Japan (Grant number: H19-Policy-001).