- Email: [email protected]
Post-Percutaneous Nephrolithotomy Extensive Hemorrhage: A Study of Risk Factors
Post-Percutaneous Nephrolithotomy Extensive Hemorrhage: A Study of Risk Factors Ahmed R. El-Nahas, Ahmed A. Shokeir,* Ahmed M. El-Assmy, Tarek Mohsen, Ahmed M. Shoma, Ibrahim Eraky, Mahmoud R. El-Kenawy and Hamdy A. El-Kappany From the Urology and Radiology Departments(TM), Urology and Nephrology Center, Mansoura University, Mansoura, Egypt
Purpose: We identified risk factors predicting severe bleeding due to percutaneous nephrolithotomy. Materials and Methods: Computerized data on 2,909 patients who underwent a total of 3,878 percutaneous nephrolithotomy procedures between January 1995 and December 2005 were retrospectively reviewed. Data on patients who experienced severe bleeding requiring angiographic renal embolization were compared with those on other patients using univariate and multivariate analyses. We tested the characteristics of patients, kidneys and stones together with details of the operative procedure and surgeon experience. Results: Severe bleeding complicated a total of 39 procedures (1%) in 25 males and 14 females with a mean age of 50.7 ⫾ 12.6 years. Associated morbidity included shock in 6 patients and perirenal hematoma in 4. Renal angiography revealed pseudoaneurysm in 20 patients, arteriovenous fistula in 9, the 2 lesions in 8 and arterial laceration in 2. Bleeding could be controlled with superselective embolization in 36 patients (92.3%). Followup was available on 33 patients (mean 21 ⫾ 15 months). Renal function was stable in all patients except 3 who had a post-embolization increase in serum creatinine, of whom all had a solitary kidney and none required renal replacement therapy. Significant risk factors for severe bleeding were upper caliceal puncture, solitary kidney, staghorn stone, multiple punctures and inexperienced surgeon. Conclusions: Percutaneous nephrolithotomy should be performed by an experienced endourologist in patients at risk for severe bleeding, such as those with a solitary kidney or staghorn stones. Key Words: kidney; kidney calculi; nephrolithotomy, percutaneous; hemorrhage; embolization, therapeutic
coagulation profile and liver function tests. Radiological investigations included excretory urography or noncontrast computerized tomography in patients with high serum creatinine. With the patient prone the skin was punctured at the posterior axillary line. Supracostal approach was needed in 531 procedures (13.7%). Percutaneous renal access was established under biplane or multidirectional C-arm fluoroscopic guidance through the posterolateral plane of the kidney. The pelvicaliceal system was entered at the lower posterior calix in patients with renal pelvis or lower caliceal stones. Middle or upper calix punctures were used when stones were present in these calices. One-stage PNL (puncture, dilation and stone retrieval) was performed in 45% of the patients. The tract was dilated using coaxial telescopic dilators, which fit tightly to the 30Fr Amplatz sheath, because a discrepancy between the size of the sheath and the size of the last coaxial dilator can cause renal injury and subsequent bleeding. Small stones were removed with forceps and large ones were disintegrated with pneumatic or ultrasonic lithotrites. A 22Fr nephrostomy tube was placed at the end of the procedure. The tube was removed after 48 hours and the patient was discharged home, provided that there were no complications or residual stones. Mild bleeding was managed conservatively by clamping the nephrostomy tube, adequate hydration, diuretics and hemostatic drugs. Moderate hemorrhage was successfully treated with blood transfusion, in addition to conservative measures. Severe bleeding leading to hemodynamic insta-
ercutaneous nephrolithotomy is an established procedure for large and complex renal calculi.1 Refinement of the technology and increasing experience in the last 2 decades led to increased safety and efficacy. Nevertheless, complications may still occur. Renal hemorrhage is one of the most dangerous complications of PNL. Fortunately conservative measures are adequate to control bleeding in most cases and angiographic embolization is needed in less than 1%.2–5 Although the diagnosis of and treatment for postPNL renal bleeding was previously discussed,6 –9 only a few groups have investigated risk factors 7,10 and one of these studies failed to identify any risk factors.7 Identification of the risk factors of post-PNL severe hemorrhage is of paramount importance for their avoidance. We identified risk factors predicting post-PNL severe bleeding requiring angiographic renal embolization.
P
PATIENTS AND METHODS Computerized data on 2,909 patients who underwent a total of 3,878 PNL procedures between January 1995 and December 2005 at our center were retrospectively reviewed. Preoperative patient evaluation included history, clinical examination, serum creatinine estimation, complete blood count,
Submitted for publication March 28, 2006. * Correspondence: Urology Department, Mansoura Urology and Nephrology Center, Mansoura, Egypt (telephone: ⫹ 2050 226 2222; FAX: ⫹2050 226 3717; e-mail: [email protected]).
0022-5347/07/1772-0576/0 THE JOURNAL OF UROLOGY® Copyright © 2007 by AMERICAN UROLOGICAL ASSOCIATION
576
Vol. 177, 576-579, February 2007 Printed in U.S.A. DOI:10.1016/j.juro.2006.09.048
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION
577
bility was treated with blood transfusion plus superselective renal angiography to identify the site and type of vascular injury (fig. 1). Embolization of the injured vessels was then performed using platinum microcoils. At the end of the procedure angiography was repeated to ensure occlusion of the offending vessel (fig. 2). The patient remained in bed with vital sign monitoring every 4 to 6 hours. Ultrasonography and complete blood count were performed daily until stabilization of the condition. If bleeding recurred, another embolization session was performed. To identify risk factors for severe bleeding data on patients who needed embolization were compared with those on other patients using chi-square univariate and logistic regression multivariate analyses. We tested all characteristics of patients (age and sex), renal units (side, morphology and whether it was a solitary kidney) and stones (site, burden and recurrence). Stone burden was classified as single, multiple or staghorn stones. The details of the operative procedure (number and site of percutaneous tracts) and surgeon experience in doing PNL were also studied. RESULTS Of 3,878 PNL procedures blood transfusion was needed in 213 (5.5%) and superselective renal angiography was required in 39 (1%). They were 25 males and 14 females with a mean ⫾ SD age of 50.7 ⫾ 12.6 years (range 25 to 80). A total of 29 patients experienced perioperative severe bleeding, while 10 experienced severe hematuria after hospital discharge. Mean time before late hematuria was 6.3 days (range 3 to 22). Associated morbidity included hemorrhagic shock in 6 patients and perirenal hematoma in 4. The median number of blood transfusion units was 3 (range 1 to 6). Renal angiography revealed pseudoaneurysm in 20 patients, AVF in 9 and the 2 lesions in 8 (fig. 1), while arterial laceration was seen in 2. Vascular injuries were related to the upper, middle and lower segmental renal vessels in 9, 6 and 22 patients, respectively, while injury to the upper and lower branches was noted in 1 and injury to a large hilar artery was observed in another.
FIG. 1. Superselective renal angiography of lower segmental branch of left kidney shows pseudoaneurysm (P) and AVF, leading to opacification of renal vein (RV) and inferior vena cava (IVC).
FIG. 2. Superselective renal angiography of lower segmental branch of left kidney after insertion of 3 microcoils demonstrates no perfusion through affected artery.
Bleeding could be controlled with superselective embolization in 36 patients (92.3%) after 1 session in 26 and 2 sessions in 10. The second session was required because of failure to demonstrate vascular injuries during the first session in 2 cases and recurrent hemorrhage in the remaining 8. The early post-embolization course was smooth in 30 patients with cessation of hematuria within 24 hours, while morbidity was observed in 6. Two of these patients had urinary leakage through the nephrostomy tract, which was successfully managed by a double pigtail ureteral stent for 4 weeks. One patient with a perinephric abscess due to perinephric hematoma infection was treated with percutaneous tube drainage. In 3 patients with a solitary kidney serum creatinine increased from preoperative values of 0.8, 1 and 1.3 to 2.5, 3.6 and 6 mg/dl, respectively. Superselective embolization failed to control bleeding in 3 patients, including 2 with a large AVF and 1 with hilar vascular injury. Urgent exploration was performed in the 3 patients. Deep sutures at the site of the nephrostomy tube controlled bleeding in the first patient and nephrectomy was required in the second patient. The third patient died during exploration from profuse internal hemorrhage due to injury to a large hilar artery. Of the 39 patients followup was available on 33 (mean 21 ⫾ 15 months). Renal function was stable in all patients except 3 who had a post-embolization increase in serum creatinine. All of them had a solitary kidney and none required renal replacement therapy. Mean age ⫾ SD in patients with and without severe bleeding was comparable (50.7 ⫾ 12.6 vs 46.9 ⫾ 10.9 years, p ⫽ 0.224). Significant risk factors on univariate analysis were upper calix puncture, solitary kidney, staghorn stone, multiple punctures and inexperienced operator (table 1). On multivariate analysis all of these factors maintained significance as independent risk factors (table 2).
578
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION
TABLE 1. Univariate analysis of risk factors for post-PNL severe renal vascular injuries requiring superselective angiography
Factor
Overall
No. sex: M F Side (No. kidneys): R L Renal morphology (No. kidneys): No hydronephrosis Hydronephrosis Pyelonephritis Congenital anomalies (No. kidneys): No Yes Renal unit (No. kidneys): Not solitary Solitary Stone burden (No. kidneys): Single Multiple Staghorn Recurrence (No. kidneys): No Yes Stone site (No. kidneys): Pelvicaliceal Caliceal Pelvic Procedure (No. kidneys): 1-Stage 2-Stage Operator (No. kidneys): Experienced endourologist General urologist or trainee Tracts (No. kidneys): Single Multiple Punctured calix (No. punctures): Lower Middle Upper Skin puncture (No. punctures): Subcostal Supracostal
2,909 1,861 1,048 3,878 1,876 2,002
No. Severe Vascular Injury (%)
p Value
25 (1.3) 14 (1.3)
0.990
16 (0.9) 23 (1.1)
0.356
1,705 1,715 458
12 (0.7) 18 (1) 9 (2)
0.054
3,683 195
35 (1) 4 (2.1)
0.133
3,517 361
29 (0.8) 10 (2.8)
⬍0.001
1,296 2,353 229
12 (0.9) 18 (0.8) 9 (3.9)
⬍0.001
2,582 1,296
28 (1.1) 11 (0.8)
0.488
1,969 903 1,006
26 (1.3) 4 (0.4) 9 (0.9)
0.084
1,750 2,128
22 (1.3) 17 (0.8)
0.155
2,617 1,261
17 (0.6) 22 (1.7)
0.001
3,497 381 4,348 3,695 434 219 4,348 3,817 531
28 (0.8) 11 (2.9)
⬍0.001
24 (0.6) 5 (1.2) 10 (4.6)
⬍0.001
34 (0.9) 5 (0.9)
0.990
DISCUSSION PNL is a safe and reliable technique for renal stones. It has replaced open surgery as the treatment of choice for large, multiple and staghorn renal calculi.1 However, it is an invasive procedure with reported complication rates of 3% to 18% according to different investigators.2– 4 One of the most serious complications is renal hemorrhage. Blood loss is a normal feature of PNL because some bleeding can occur during renal puncture, tract dilation, nephroscopy and stone disintegration. It is considered a complication only when blood transfusion is required. A transfusion rate of 3% to
23% has been reported.2–7,9 –12 Fortunately in most cases bleeding can be controlled with conservative measures, such as clamping the nephrostomy, hydration and diuretics, hemostatic medications and Kaye balloon tamponade.13,14 Therefore, the necessity of renal embolization to control severe bleeding is low (range 0.3% to 1.4%).2–7,9 –12 The transfusion and embolization rates in our series were comparable to these ranges (5.5% and 1%, respectively). In PNL venous bleeding can be controlled with conservative measures in most cases because the intrarenal venous system is quite resilient. On the other hand, arterial injuries result in severe bleeding that requires embolization. Therefore, the most commonly detected vascular lesions in this and other studies were AVF or arterial pseudoaneurysm. Blood passage from the high pressure of the injured artery to the injured adjacent vein results in AVF and blood passage to the parenchyma leads to pseudoaneurysm.10 Delayed bleeding may result from pseudoaneurysm rupture.15 Arterial laceration is rare but it can lead to severe intraoperative bleeding. The only death in our series was due to inadvertent laceration of a main branch of the renal artery because of over dilation. The success rate of superselective embolization for controlling post-PNL severe bleeding in the current study was 92.3%, which is comparable to rates in previous studies.6,7,10 The risk of blood transfusion after PNL is influenced by many factors, including operative technique, patient status and stone complexity. The most commonly incriminated operative risk factor was too medial a port of entry to the kidney, leading to direct injury to the posterior segmental arterial branch. Perforation of the pelvicaliceal system during dilation or stone disintegration and forceful manipulation of the rigid nephroscope to access stones in different calices may damage adjacent blood vessels.11 Lam et al reported that improved skills and flexible nephroscopy decreased transfusion rates.16 In our study operator experience was a significant predictor of the need for embolization because experienced endourologists had the skills needed to perform proper puncture, judicious dilation and gentle manipulations using rigid and flexible nephroscopes. The number of punctures has also been incriminated in vascular injuries in some series,6,7 while other groups did not find it to be a significant risk factor.9,10 Our results support its incrimination. It would seem logical that the chances of damage to renal vasculature would be greater as the number of punctures increases. The incidence of severe PNL associated vascular injuries was directly related to stone complexity. Meta-analysis of publications of percutaneous removal of staghorn stones demonstrated higher transfusion rates.1 Kessaris et al ob-
TABLE 2. Multivariate analysis of risk factors for post-PNL severe renal vascular injuries requiring superselective angiography Variable Punctured calix lower (referent) vs upper Stone burden single (referent) vs staghorn Renal unit not solitary (referent) vs solitary No. punctures single (referent) vs multiple Operator experience endourologist (referent) vs general urologist or trainee
B Regression Coefficient
SE
Exp. B*
95% CI
p Value
1.790 1.017 0.980 0.940 0.879
0.414 0.491 0.393 0.404 0.332
5.991 2.765 2.665 2.559 2.408
2.661–13.492 1.056–7.238 1.235–5.755 1.160–5.648 1.256–4.618
⬍0.001 0.038 0.013 0.020 0.008
* RR, Relative risk when the category of a certain variable was compared to the referent.
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION served staghorn stones in 8 of 17 patients (47%) who required embolization7 and Srivastava et al concluded that stone size significantly predicted severe vascular lesions after PNL.10 In this study staghorn stones were a significant independent risk factor because during PNL for such complex stones multiple tracts and excessive manipulation were needed. Identification of risk factors influencing the incidence of severe vascular injuries is of the utmost importance for decreasing the rate of this serious and sometimes fatal complication. In the current study we addressed this issue in a large number of patients using univariate and multivariate statistical analyses of preoperative and operative risk factors, which were lacking in previous reports. However, because this was a retrospective study, the influence of some factors, such as body mass index and distance of the skin puncture from the midline, could not be tested. Moreover, followup of differential renal function using radioisotope renal scan was not available. Therefore, other studies are invited to overcome these shortcomings. In this series we identified new, significant risk factors for severe bleeding, such as upper calix puncture and solitary kidney. The reason for the increased incidence of vascular injury with upper caliceal puncture may be related to the oblique and longer length of the tract. In many cases the upper calix was superior to the site of skin puncture even if a supracostal approach was used. Therefore, an oblique and longer tract was needed. Moreover, changing the direction of the tract to reach the renal pelvis may result in injury to the adjacent parenchyma with its vascular supply. Therefore, we advise entering the upper calix through a straight and direct tract when possible and avoiding excessive changes in the direction of the upper caliceal tract using another lower caliceal puncture or flexible nephroscope to remove migrating stone fragments. Solitary kidney was a significant risk factor because compensatory hypertrophy is a normal physiological response as the thickness of renal parenchyma increases with the increase in kidney size. It is speculated that puncture and dilation through thick renal parenchyma may increase the possibility of bleeding damage to more renal tissue and its vascular supply.
Abbreviations and Acronyms AVF ⫽ arteriovenous fistula PNL ⫽ percutaneous nephrolithotomy REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
CONCLUSIONS Based on the results of this study we can divide patients with renal stones amenable for PNL into 2 categories. The first group includes patients at low risk for vascular complications, in whom PNL could be performed by general urologists and endourological trainees. The second group includes patients at risk for severe bleeding, such as those with staghorn stones or solitary kidney, in whom PNL should be performed by an experienced endourologist.
579
14. 15.
16.
Preminger GM, Assimos DG, Lingeman JE, Nakada SY, Pearle MS and Wolf JS Jr (AUA Nephrolithiasis Guideline Panel): Chapter 1: AUA guideline on management of staghorn calculi: diagnosis and treatment recommendations. J Urol 2005; 173: 1991. Segura JW, Patterson DE, Le Roy AJ, Williams HJ Jr, Barrett DM, Benson RC Jr et al: Percutaneous removal of kidney stones: review of 1,000 cases. J Urol 1985; 134: 1077. El-Kenawy MR, El-Kappany HA, El-Diasty TA and Ghoneim MA: Percutaneous nephrolithotomy for renal stones in over 1000 patients. Br J Urol 1992; 69: 470. Lee WJ, Smith AD, Cubelli V, Badlani GH, Lewin B, Vernace F et al: Complications of percutaneous nephrolithotomy. AJR Am J Roentgenol 1987; 148: 177. Sacha K, Szewczyk W and Bar K: Massive haemorrhage presenting as a complication after percutaneous nephrolithotomy (PCNL). Int Urol Nephrol 1996; 28: 315. Patterson DE, Segura JW, LeRoy AJ, Benson RC and May GR: The etiology and treatment of delayed bleeding following percutaneous lithotripsy. J Urol 1985; 133: 447. Kessaris DN, Bellman GC, Pardalidis NP and Smith AG: Management of hemorrhage after percutaneous renal surgery. J Urol 1995; 153: 604. Gupta M, Bellman GC and Smith AD: Massive hemorrhage from renal vein injury during percutaneous renal surgery: endourological management. J Urol 1997; 157: 795. Martin X, Murat FJ, Feitosa LC, Rouviere O, Lyonnet D, Gelet A et al: Severe bleeding after nephrolithotomy: results of hyperselective embolization. Eur Urol 2000; 37: 136. Srivastava A, Singh KJ, Suri A, Dubey D, Kumar A, Kapoor R et al: Vascular complications after Percutaneous nephrolithotomy: are there any predictive factors? Urology 2005; 66: 38. Stoller ML, Wolf JS Jr and St. Lezin MA: Estimated blood loss and transfusion rates associated with percutaneous nephrolithotomy. J Urol 1994; 152: 1977. Reddy PK, Hulbert JC, Lange PH, Clayman RV, Marcuzzi A, Lapointe S et al: Percutaneous removal of renal and ureteral calculi: experience with 400 cases. J Urol 1985; 134: 662. Galek L, Darewicz B, Werel T and Darewicz J: Hemorrhagic complications of percutaneous lithotripsy: original methods of treatment. Int Urol Nephrol 2000; 32: 231. Kaye KW and Clayman RV: Tamponade nephrostomy catheter for percutaneous nephrolithotomy. Urology 1986; 27: 441. Gavant ML, Gold RE and Church JC: Delayed rupture of renal pseudoaneurysms: complication of percutaneous nephrostomy. AJR Am J Roentgenol 1982; 138: 948. Lam HS, Lingeman JE, Mosbaugh PG, Steele RE, Knapp PM, Scott JW et al: Evolution of the technique of combination therapy for staghorn calculi: a decreasing role for extracorporeal shock wave lithotripsy. J Urol 1992; 148: 1058.
Purpose: We identified risk factors predicting severe bleeding due to percutaneous nephrolithotomy. Materials and Methods: Computerized data on 2,909 patients who underwent a total of 3,878 percutaneous nephrolithotomy procedures between January 1995 and December 2005 were retrospectively reviewed. Data on patients who experienced severe bleeding requiring angiographic renal embolization were compared with those on other patients using univariate and multivariate analyses. We tested the characteristics of patients, kidneys and stones together with details of the operative procedure and surgeon experience. Results: Severe bleeding complicated a total of 39 procedures (1%) in 25 males and 14 females with a mean age of 50.7 ⫾ 12.6 years. Associated morbidity included shock in 6 patients and perirenal hematoma in 4. Renal angiography revealed pseudoaneurysm in 20 patients, arteriovenous fistula in 9, the 2 lesions in 8 and arterial laceration in 2. Bleeding could be controlled with superselective embolization in 36 patients (92.3%). Followup was available on 33 patients (mean 21 ⫾ 15 months). Renal function was stable in all patients except 3 who had a post-embolization increase in serum creatinine, of whom all had a solitary kidney and none required renal replacement therapy. Significant risk factors for severe bleeding were upper caliceal puncture, solitary kidney, staghorn stone, multiple punctures and inexperienced surgeon. Conclusions: Percutaneous nephrolithotomy should be performed by an experienced endourologist in patients at risk for severe bleeding, such as those with a solitary kidney or staghorn stones. Key Words: kidney; kidney calculi; nephrolithotomy, percutaneous; hemorrhage; embolization, therapeutic
coagulation profile and liver function tests. Radiological investigations included excretory urography or noncontrast computerized tomography in patients with high serum creatinine. With the patient prone the skin was punctured at the posterior axillary line. Supracostal approach was needed in 531 procedures (13.7%). Percutaneous renal access was established under biplane or multidirectional C-arm fluoroscopic guidance through the posterolateral plane of the kidney. The pelvicaliceal system was entered at the lower posterior calix in patients with renal pelvis or lower caliceal stones. Middle or upper calix punctures were used when stones were present in these calices. One-stage PNL (puncture, dilation and stone retrieval) was performed in 45% of the patients. The tract was dilated using coaxial telescopic dilators, which fit tightly to the 30Fr Amplatz sheath, because a discrepancy between the size of the sheath and the size of the last coaxial dilator can cause renal injury and subsequent bleeding. Small stones were removed with forceps and large ones were disintegrated with pneumatic or ultrasonic lithotrites. A 22Fr nephrostomy tube was placed at the end of the procedure. The tube was removed after 48 hours and the patient was discharged home, provided that there were no complications or residual stones. Mild bleeding was managed conservatively by clamping the nephrostomy tube, adequate hydration, diuretics and hemostatic drugs. Moderate hemorrhage was successfully treated with blood transfusion, in addition to conservative measures. Severe bleeding leading to hemodynamic insta-
ercutaneous nephrolithotomy is an established procedure for large and complex renal calculi.1 Refinement of the technology and increasing experience in the last 2 decades led to increased safety and efficacy. Nevertheless, complications may still occur. Renal hemorrhage is one of the most dangerous complications of PNL. Fortunately conservative measures are adequate to control bleeding in most cases and angiographic embolization is needed in less than 1%.2–5 Although the diagnosis of and treatment for postPNL renal bleeding was previously discussed,6 –9 only a few groups have investigated risk factors 7,10 and one of these studies failed to identify any risk factors.7 Identification of the risk factors of post-PNL severe hemorrhage is of paramount importance for their avoidance. We identified risk factors predicting post-PNL severe bleeding requiring angiographic renal embolization.
P
PATIENTS AND METHODS Computerized data on 2,909 patients who underwent a total of 3,878 PNL procedures between January 1995 and December 2005 at our center were retrospectively reviewed. Preoperative patient evaluation included history, clinical examination, serum creatinine estimation, complete blood count,
Submitted for publication March 28, 2006. * Correspondence: Urology Department, Mansoura Urology and Nephrology Center, Mansoura, Egypt (telephone: ⫹ 2050 226 2222; FAX: ⫹2050 226 3717; e-mail: [email protected]).
0022-5347/07/1772-0576/0 THE JOURNAL OF UROLOGY® Copyright © 2007 by AMERICAN UROLOGICAL ASSOCIATION
576
Vol. 177, 576-579, February 2007 Printed in U.S.A. DOI:10.1016/j.juro.2006.09.048
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION
577
bility was treated with blood transfusion plus superselective renal angiography to identify the site and type of vascular injury (fig. 1). Embolization of the injured vessels was then performed using platinum microcoils. At the end of the procedure angiography was repeated to ensure occlusion of the offending vessel (fig. 2). The patient remained in bed with vital sign monitoring every 4 to 6 hours. Ultrasonography and complete blood count were performed daily until stabilization of the condition. If bleeding recurred, another embolization session was performed. To identify risk factors for severe bleeding data on patients who needed embolization were compared with those on other patients using chi-square univariate and logistic regression multivariate analyses. We tested all characteristics of patients (age and sex), renal units (side, morphology and whether it was a solitary kidney) and stones (site, burden and recurrence). Stone burden was classified as single, multiple or staghorn stones. The details of the operative procedure (number and site of percutaneous tracts) and surgeon experience in doing PNL were also studied. RESULTS Of 3,878 PNL procedures blood transfusion was needed in 213 (5.5%) and superselective renal angiography was required in 39 (1%). They were 25 males and 14 females with a mean ⫾ SD age of 50.7 ⫾ 12.6 years (range 25 to 80). A total of 29 patients experienced perioperative severe bleeding, while 10 experienced severe hematuria after hospital discharge. Mean time before late hematuria was 6.3 days (range 3 to 22). Associated morbidity included hemorrhagic shock in 6 patients and perirenal hematoma in 4. The median number of blood transfusion units was 3 (range 1 to 6). Renal angiography revealed pseudoaneurysm in 20 patients, AVF in 9 and the 2 lesions in 8 (fig. 1), while arterial laceration was seen in 2. Vascular injuries were related to the upper, middle and lower segmental renal vessels in 9, 6 and 22 patients, respectively, while injury to the upper and lower branches was noted in 1 and injury to a large hilar artery was observed in another.
FIG. 1. Superselective renal angiography of lower segmental branch of left kidney shows pseudoaneurysm (P) and AVF, leading to opacification of renal vein (RV) and inferior vena cava (IVC).
FIG. 2. Superselective renal angiography of lower segmental branch of left kidney after insertion of 3 microcoils demonstrates no perfusion through affected artery.
Bleeding could be controlled with superselective embolization in 36 patients (92.3%) after 1 session in 26 and 2 sessions in 10. The second session was required because of failure to demonstrate vascular injuries during the first session in 2 cases and recurrent hemorrhage in the remaining 8. The early post-embolization course was smooth in 30 patients with cessation of hematuria within 24 hours, while morbidity was observed in 6. Two of these patients had urinary leakage through the nephrostomy tract, which was successfully managed by a double pigtail ureteral stent for 4 weeks. One patient with a perinephric abscess due to perinephric hematoma infection was treated with percutaneous tube drainage. In 3 patients with a solitary kidney serum creatinine increased from preoperative values of 0.8, 1 and 1.3 to 2.5, 3.6 and 6 mg/dl, respectively. Superselective embolization failed to control bleeding in 3 patients, including 2 with a large AVF and 1 with hilar vascular injury. Urgent exploration was performed in the 3 patients. Deep sutures at the site of the nephrostomy tube controlled bleeding in the first patient and nephrectomy was required in the second patient. The third patient died during exploration from profuse internal hemorrhage due to injury to a large hilar artery. Of the 39 patients followup was available on 33 (mean 21 ⫾ 15 months). Renal function was stable in all patients except 3 who had a post-embolization increase in serum creatinine. All of them had a solitary kidney and none required renal replacement therapy. Mean age ⫾ SD in patients with and without severe bleeding was comparable (50.7 ⫾ 12.6 vs 46.9 ⫾ 10.9 years, p ⫽ 0.224). Significant risk factors on univariate analysis were upper calix puncture, solitary kidney, staghorn stone, multiple punctures and inexperienced operator (table 1). On multivariate analysis all of these factors maintained significance as independent risk factors (table 2).
578
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION
TABLE 1. Univariate analysis of risk factors for post-PNL severe renal vascular injuries requiring superselective angiography
Factor
Overall
No. sex: M F Side (No. kidneys): R L Renal morphology (No. kidneys): No hydronephrosis Hydronephrosis Pyelonephritis Congenital anomalies (No. kidneys): No Yes Renal unit (No. kidneys): Not solitary Solitary Stone burden (No. kidneys): Single Multiple Staghorn Recurrence (No. kidneys): No Yes Stone site (No. kidneys): Pelvicaliceal Caliceal Pelvic Procedure (No. kidneys): 1-Stage 2-Stage Operator (No. kidneys): Experienced endourologist General urologist or trainee Tracts (No. kidneys): Single Multiple Punctured calix (No. punctures): Lower Middle Upper Skin puncture (No. punctures): Subcostal Supracostal
2,909 1,861 1,048 3,878 1,876 2,002
No. Severe Vascular Injury (%)
p Value
25 (1.3) 14 (1.3)
0.990
16 (0.9) 23 (1.1)
0.356
1,705 1,715 458
12 (0.7) 18 (1) 9 (2)
0.054
3,683 195
35 (1) 4 (2.1)
0.133
3,517 361
29 (0.8) 10 (2.8)
⬍0.001
1,296 2,353 229
12 (0.9) 18 (0.8) 9 (3.9)
⬍0.001
2,582 1,296
28 (1.1) 11 (0.8)
0.488
1,969 903 1,006
26 (1.3) 4 (0.4) 9 (0.9)
0.084
1,750 2,128
22 (1.3) 17 (0.8)
0.155
2,617 1,261
17 (0.6) 22 (1.7)
0.001
3,497 381 4,348 3,695 434 219 4,348 3,817 531
28 (0.8) 11 (2.9)
⬍0.001
24 (0.6) 5 (1.2) 10 (4.6)
⬍0.001
34 (0.9) 5 (0.9)
0.990
DISCUSSION PNL is a safe and reliable technique for renal stones. It has replaced open surgery as the treatment of choice for large, multiple and staghorn renal calculi.1 However, it is an invasive procedure with reported complication rates of 3% to 18% according to different investigators.2– 4 One of the most serious complications is renal hemorrhage. Blood loss is a normal feature of PNL because some bleeding can occur during renal puncture, tract dilation, nephroscopy and stone disintegration. It is considered a complication only when blood transfusion is required. A transfusion rate of 3% to
23% has been reported.2–7,9 –12 Fortunately in most cases bleeding can be controlled with conservative measures, such as clamping the nephrostomy, hydration and diuretics, hemostatic medications and Kaye balloon tamponade.13,14 Therefore, the necessity of renal embolization to control severe bleeding is low (range 0.3% to 1.4%).2–7,9 –12 The transfusion and embolization rates in our series were comparable to these ranges (5.5% and 1%, respectively). In PNL venous bleeding can be controlled with conservative measures in most cases because the intrarenal venous system is quite resilient. On the other hand, arterial injuries result in severe bleeding that requires embolization. Therefore, the most commonly detected vascular lesions in this and other studies were AVF or arterial pseudoaneurysm. Blood passage from the high pressure of the injured artery to the injured adjacent vein results in AVF and blood passage to the parenchyma leads to pseudoaneurysm.10 Delayed bleeding may result from pseudoaneurysm rupture.15 Arterial laceration is rare but it can lead to severe intraoperative bleeding. The only death in our series was due to inadvertent laceration of a main branch of the renal artery because of over dilation. The success rate of superselective embolization for controlling post-PNL severe bleeding in the current study was 92.3%, which is comparable to rates in previous studies.6,7,10 The risk of blood transfusion after PNL is influenced by many factors, including operative technique, patient status and stone complexity. The most commonly incriminated operative risk factor was too medial a port of entry to the kidney, leading to direct injury to the posterior segmental arterial branch. Perforation of the pelvicaliceal system during dilation or stone disintegration and forceful manipulation of the rigid nephroscope to access stones in different calices may damage adjacent blood vessels.11 Lam et al reported that improved skills and flexible nephroscopy decreased transfusion rates.16 In our study operator experience was a significant predictor of the need for embolization because experienced endourologists had the skills needed to perform proper puncture, judicious dilation and gentle manipulations using rigid and flexible nephroscopes. The number of punctures has also been incriminated in vascular injuries in some series,6,7 while other groups did not find it to be a significant risk factor.9,10 Our results support its incrimination. It would seem logical that the chances of damage to renal vasculature would be greater as the number of punctures increases. The incidence of severe PNL associated vascular injuries was directly related to stone complexity. Meta-analysis of publications of percutaneous removal of staghorn stones demonstrated higher transfusion rates.1 Kessaris et al ob-
TABLE 2. Multivariate analysis of risk factors for post-PNL severe renal vascular injuries requiring superselective angiography Variable Punctured calix lower (referent) vs upper Stone burden single (referent) vs staghorn Renal unit not solitary (referent) vs solitary No. punctures single (referent) vs multiple Operator experience endourologist (referent) vs general urologist or trainee
B Regression Coefficient
SE
Exp. B*
95% CI
p Value
1.790 1.017 0.980 0.940 0.879
0.414 0.491 0.393 0.404 0.332
5.991 2.765 2.665 2.559 2.408
2.661–13.492 1.056–7.238 1.235–5.755 1.160–5.648 1.256–4.618
⬍0.001 0.038 0.013 0.020 0.008
* RR, Relative risk when the category of a certain variable was compared to the referent.
POST-PERCUTANEOUS NEPHROLITHOTOMY ANGIOEMBOLIZATION served staghorn stones in 8 of 17 patients (47%) who required embolization7 and Srivastava et al concluded that stone size significantly predicted severe vascular lesions after PNL.10 In this study staghorn stones were a significant independent risk factor because during PNL for such complex stones multiple tracts and excessive manipulation were needed. Identification of risk factors influencing the incidence of severe vascular injuries is of the utmost importance for decreasing the rate of this serious and sometimes fatal complication. In the current study we addressed this issue in a large number of patients using univariate and multivariate statistical analyses of preoperative and operative risk factors, which were lacking in previous reports. However, because this was a retrospective study, the influence of some factors, such as body mass index and distance of the skin puncture from the midline, could not be tested. Moreover, followup of differential renal function using radioisotope renal scan was not available. Therefore, other studies are invited to overcome these shortcomings. In this series we identified new, significant risk factors for severe bleeding, such as upper calix puncture and solitary kidney. The reason for the increased incidence of vascular injury with upper caliceal puncture may be related to the oblique and longer length of the tract. In many cases the upper calix was superior to the site of skin puncture even if a supracostal approach was used. Therefore, an oblique and longer tract was needed. Moreover, changing the direction of the tract to reach the renal pelvis may result in injury to the adjacent parenchyma with its vascular supply. Therefore, we advise entering the upper calix through a straight and direct tract when possible and avoiding excessive changes in the direction of the upper caliceal tract using another lower caliceal puncture or flexible nephroscope to remove migrating stone fragments. Solitary kidney was a significant risk factor because compensatory hypertrophy is a normal physiological response as the thickness of renal parenchyma increases with the increase in kidney size. It is speculated that puncture and dilation through thick renal parenchyma may increase the possibility of bleeding damage to more renal tissue and its vascular supply.
Abbreviations and Acronyms AVF ⫽ arteriovenous fistula PNL ⫽ percutaneous nephrolithotomy REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
CONCLUSIONS Based on the results of this study we can divide patients with renal stones amenable for PNL into 2 categories. The first group includes patients at low risk for vascular complications, in whom PNL could be performed by general urologists and endourological trainees. The second group includes patients at risk for severe bleeding, such as those with staghorn stones or solitary kidney, in whom PNL should be performed by an experienced endourologist.
579
14. 15.
16.
Preminger GM, Assimos DG, Lingeman JE, Nakada SY, Pearle MS and Wolf JS Jr (AUA Nephrolithiasis Guideline Panel): Chapter 1: AUA guideline on management of staghorn calculi: diagnosis and treatment recommendations. J Urol 2005; 173: 1991. Segura JW, Patterson DE, Le Roy AJ, Williams HJ Jr, Barrett DM, Benson RC Jr et al: Percutaneous removal of kidney stones: review of 1,000 cases. J Urol 1985; 134: 1077. El-Kenawy MR, El-Kappany HA, El-Diasty TA and Ghoneim MA: Percutaneous nephrolithotomy for renal stones in over 1000 patients. Br J Urol 1992; 69: 470. Lee WJ, Smith AD, Cubelli V, Badlani GH, Lewin B, Vernace F et al: Complications of percutaneous nephrolithotomy. AJR Am J Roentgenol 1987; 148: 177. Sacha K, Szewczyk W and Bar K: Massive haemorrhage presenting as a complication after percutaneous nephrolithotomy (PCNL). Int Urol Nephrol 1996; 28: 315. Patterson DE, Segura JW, LeRoy AJ, Benson RC and May GR: The etiology and treatment of delayed bleeding following percutaneous lithotripsy. J Urol 1985; 133: 447. Kessaris DN, Bellman GC, Pardalidis NP and Smith AG: Management of hemorrhage after percutaneous renal surgery. J Urol 1995; 153: 604. Gupta M, Bellman GC and Smith AD: Massive hemorrhage from renal vein injury during percutaneous renal surgery: endourological management. J Urol 1997; 157: 795. Martin X, Murat FJ, Feitosa LC, Rouviere O, Lyonnet D, Gelet A et al: Severe bleeding after nephrolithotomy: results of hyperselective embolization. Eur Urol 2000; 37: 136. Srivastava A, Singh KJ, Suri A, Dubey D, Kumar A, Kapoor R et al: Vascular complications after Percutaneous nephrolithotomy: are there any predictive factors? Urology 2005; 66: 38. Stoller ML, Wolf JS Jr and St. Lezin MA: Estimated blood loss and transfusion rates associated with percutaneous nephrolithotomy. J Urol 1994; 152: 1977. Reddy PK, Hulbert JC, Lange PH, Clayman RV, Marcuzzi A, Lapointe S et al: Percutaneous removal of renal and ureteral calculi: experience with 400 cases. J Urol 1985; 134: 662. Galek L, Darewicz B, Werel T and Darewicz J: Hemorrhagic complications of percutaneous lithotripsy: original methods of treatment. Int Urol Nephrol 2000; 32: 231. Kaye KW and Clayman RV: Tamponade nephrostomy catheter for percutaneous nephrolithotomy. Urology 1986; 27: 441. Gavant ML, Gold RE and Church JC: Delayed rupture of renal pseudoaneurysms: complication of percutaneous nephrostomy. AJR Am J Roentgenol 1982; 138: 948. Lam HS, Lingeman JE, Mosbaugh PG, Steele RE, Knapp PM, Scott JW et al: Evolution of the technique of combination therapy for staghorn calculi: a decreasing role for extracorporeal shock wave lithotripsy. J Urol 1992; 148: 1058.