ENDOUROLOGY AND STONE DISEASES Factors Influencing Complications of Percutaneous Nephrolithotomy: A Single-Center Study Sedat Oner, Muhammed Masuk Okumus, Murat Demirbas,* Efe Onen, Mustafa Murat Aydos, Mehmet Hakan Ustun, Metin Kilic, Sinan Avci Purpose: Percutaneous nephrolithotomy (PNL) is a minimally invasive procedure used for successful treatment of renal calculi. However, it is associated with various complications. We assessed the complications and their potential influencing factors in patients who had undergone PNL. Materials and Methods: In total, 1750 patients who had undergone PNL from November 2003 to June 2011 were evaluated retrospectively. PNL complications and possible contributing risk factors (age, sex, serum creatinine level, previous operations, hydronephrosis, calculi size, localization, opacity, surgeon’s experience, accessed ca- lyxes, number of accesses, and costal entries) were determined. Receiver operating characteristic (ROC) analysis was used to investigate the cutoff values of the data. Ideal cutoff value was determined by Youden's J statistic. All the demographic and clinical variables were examined using backward stepwise logistical regression analysis. Continuous variables were categorized with logistic regression analysis according to the cutoff values. Results: Complications occurred in 396 (24.4%) patients who had undergone PNL. Hemorrhage requiring blood transfusion occurred in 221 (12.6%) patients, hemorrhage requiring arterial embolization occurred in 7 (0.4%) patients, perirenal hematoma occurred in 17 (0.97%) patients, hemo-pneumothorax occurred in 32 (1.8%) patients, and colon perforation occurred in 4 (0.22%) patients. Three patients (0.06%) died of severe urosepsis, and one patient (0.02%) died of severe bleeding. The calculus size, localization, access site, number of accesses, presence of staghorn stones, surgeon’s experience, and duration of the operation significantly affected the complication risk. Conclusion: Our retrospective evaluation of this large patient series reveals that, PNL is a very effective treatment modality for kidney stones. However, although rare, serious complications including death can occur. Keywords: nephrostomy; percutaneous; multivariate analysis; postoperative complications; etiology; retrospec- tive studies; treatment outcome. INTRODUCTION Access to the collecting system of the kidneys was first reported in the 1950s, whereas access to the collecting system for the purpose of kidney stone treatment was first performed in the 1970s and 1980s. (1-3) Nowadays percutaneous nephrolithotomy (PNL) is gold standard treatment modality for upper urinary tract stones.(4) PNL has a success rate of over 95% depending on the kidney anatomy, stone size, stone localization, patient-specific anatomical factors, and surgeon’s ex- perience; however, it is associated with some complica- tions; including death.(5) In this study, we retrospective- ly analyzed the complication rates and the factors that might affect these complications in a large number of patients treated by PNL. MATERIALS AND METHODS Study Population The data of patients who underwent PNL by a single surgeon were examined retrospectively. Detailed phys- ical examinations, blood biochemistry assays, urinal- ysis, and urine culture were performed preoperatively in all the patients. They also underwent preoperative direct urinary tract radiography, urinary ultrasonogra- phy, and intravenous urography, and unenhanced spiral computed tomography, if necessary. Calculated stone surface area (CSA) was evaluated by multiplying the maximum diameter, width, ¼ π of the stone seen on the plain radiography. Appropriate antibiotic therapy was administered to those patients with growth on their preoperative urine cultures. Patients without sterile urinary cultures de- spite appropriate antibiotic therapy, underwent PNL under antibiotic administration. Surgical Procedure A 6 to 7 French (F) ureteral catheter was inserted un- der C-arm fluoroscopy. The patients were placed in the prone position, and the collecting system of the kidney was accessed with needle puncture under fluorosco- py guidance. After Amplatz dilation, 30 F sheath was positioned and only in a small number of children 20 F sheath was used. Stone fragmentation was then per- formed with pneumatic lithotripter. Clearance of the stone fragments was assessed with fluoroscopy. At the end of the procedures, a re-entry nephrostomy catheter Department of Urology, Sevket Yilmaz Training and Research Hospital, Bursa, Turkey. *Correspondence: Department of Urology, Sevket Yilmaz Training and Research Hospital, Mimar Sinan Mh. Emniyet Cd. Yildirim, Bursa 16310, Turkey. Tel: +90 532 2364838. Fax: +90 224 3660416. E-mail: muratdemirbas@doctor.com Received February 2015 & Accepted September 2015 Vol 12 No 05 September-October 2015 2317 was placed, and antegrade pyelography was performed to check for extravasation and colonic injury. Stone clearance was assessed on a direct urinary tract X-ray. Post-Operative Evaluations Patients with no opacities on the X-ray were considered stone free, those with opacities of under 4 mm were considered to have clinically insignificant residual frag- ments (CIRF) and those with opacities of over 4 mm was defined as failure. Nephrostomies were withdrawn after recovery of hematuria. Fever was considered to be present in patients with a body temperature over 38°C during the postoperative period. Cold compression, an- tipyretics, and antibiotic treatment were administered as necessary. A ureteral double J stent was placed if urine leakage from the nephrostomy tract continued for 72 to 96 hours. Anteroposterior chest radiography was performed to evaluate possible pleural injury in pa- tients in whom supracostal and upper pole access was performed. Patients with mild effusion on chest radio- graphs were followed up, and those with severe effu- sion were treated by insertion of a chest tube. PNL-as- sociated complications such as hemorrhage requiring transfusion, fever, prolonged urinary drainage, severe urosepsis, pleural injury, colon injury, and hemorrhage requiring arterial embolization, perirenal hematoma, and death were classified according to Modified Cla- vien Classification. The patient-related factors including age, sex, serum creatinine level, previous operations, presence of hy- dronephrosis, kidney stone size and opacity, presence of a solitary kidney, horseshoe kidney, staghorn stones and urinary tract infection and the procedure-related factors including the surgeon’s experience, caliceal ac- cesses, number of accesses, duration of the operation, and supracostal punctures that might influence the de- velopment of PNL associated complications were ana- lyzed. Patients were classified according to the presence of hydronephrosis on radiologic assessment. The degree of hydronephrosis was not evaluated as a separate fac- tor. The stones were divided according to location as simple (pelvis, isolated calyx) or complex (multiple cal- ices, staghorn). The presence of a staghorn stone was re-evaluated as a separate factor apart from the stone size. The accessed calyx was evaluated as either isolat- ed or multiple, and the number of accesses were eval- uated as either single or multiple. The cutoff values of the stone size, the surgeon’s experience (assessed ac- cording to the number of patients surgically treated) and the duration of the operation (the time from initial calyx access to the placement of the nephrostomy catheter) for the development of complications was evaluated. Statistical Analysis Receiver operating characteristic (ROC) analysis was used to investigate the cutoff values of the data. Ideal cutoff value was determined by Youden's J statistic. All the demographic and clinical variables were examined using backward stepwise logistical regression analy- sis as to be a risk factor or not. Continuous variables which had significant cutoff values, was categorized with logistic regression analysis according to the cutoff values. Values of P < .05 were considered to be statis- tically significant. The data were entered into an Ex- celtm (Microsoft, Redmond, WA, USA) database and analyzed with Statistical Package for the Social Science (SPSS Inc, Chicago, Illinois, USA) version 20.0. RESULTS A total of 1750 patients underwent PNL from Novem- ber 2003 to June 2011. The patient and operation data are shown in Table 1. Stone free status was obtained in 1485 (84.8%) of the 1750 patients, CIRF were evident in 217 (12.4%), and failure occurred in 48 (2.7%). Com- plications occurred in 396 (24.4%) patients and related data are summarized according to the modified Clavien classification in Table 2. The complications were com- pared with those reported in the literature (Table 3). Urosepsis developed in three (0.17%) patients during the postoperative period, and one patient (0.05%) de- veloped severe bleeding; all four of these patients died. The factors that might affect the complication rates are shown in Tables 4 and 5 according to categorized and continuous variables, respectively. The cutoff value of stone size for the development of complications was 710 mm2 (sensitivity 41.7%; specificity 69.4%), the Variables Values Mean age, years (range) 45.82 (3-81) Gender, no (%) Male 1055 (60.3) Female 695 (39.7) Stone site, no (%) Left 857 (51) Right 893 (49) Solitary, no (%) 36 (2.05) Previous operation, no (%) Yes 329 (18.8) No 1421 (81.8) Mean serum creatinine (mg/dL) 1.03 Mean stone size, cm2 (range) 7.5 (0.25-95) Presence of hydronephrosis, no (%) Yes 1220 (69.7) No 530 (30.3) Stone opacity, no (%) Semi/Non opaque 35 (2) Radiopaque 1715 (98) Mean operation time, min (range) 56.79 (15-330) Access number, mean (range) 1.40 (1-7) Single 1298 (74.2) Multiple 452 (25.8) Calyx entry, no (%) Supracostal 93 (5.3) Subcostal 1657 (94.7) Mean duration with nephrostomy, days (range) 2.49 (1-12) Mean hospitalization, days (range) 3.08 (1-25 ) Table 1. Clinical characteristics of the study patients and features of operations. Factors Influencing Complications of PNL-Oner et al. Endourology and Stone Diseases 2318 cutoff value of operation time was 67 minutes (sensi- tivity 41.4%; specificity 79.2%) and the cutoff value of the cases for convenient skills to decrease complication rates was 565 (sensitivity 72.3%; specificity 48%). Al- though supracostal access did not increase the risk for general complications, it was a risk factor for pleural injury. Pleural injury was observed in 15 of 95 patients who underwent supracostal access (16.1%), whereas it occurred in 17 of 1657 patients (1.1%) who under- went subcostal access (P < .001). While the presence of staghorn stone, multiple-caliceal accesses, large stone size and the longer duration of operation significantly increased the risk of the most common complication (blood transfusion), the presence of a horseshoe kidney and greater surgical experience were associated with a significantly decreased rate of blood transfusion (Ta- bles 6 and 7). DISCUSSION PNL is considered to be the standard treatment for staghorn renal calculi, large volume renal calculi, up- per tract calculi refractory to other treatment modalities, difficult lower pole stones, cysteine nephrolithiasis, and calculi in anatomically abnormal kidneys.(6) PNL is typ- ically a very safe and well-tolerated procedure, but it is associated with a specific set of complications.(7,8) Michel and colleagues evaluated more than 100 patients and found that the most common complications of PNL were extravasation (7.2%) (Grade 3a), blood transfusion (11.2%-17.5%) (Grade 2), and fever (21.0%-32.1%) (Grade 1).(5) The rarer complications in their study were septicemia (0.3%-4.7%) (Grade 4b), colon inju- ry (0.2%-4.8%) (Grade 4a), and pleural injury (0.3%- 1.0%) (Grade 4a). The complications seen in our study are summarized in Table 2 according to the modified Clavien system. Fever was the most common Grade 1 complication which was treated with cold compression and antipyretics. The most common Grade 3a compli- cation was urinary leakage that exceeded 72 hours in 28 (1.6%) patients. A double J ureteral catheter was placed in these patients. The Grade 3b complications included perirenal hematoma and arteriovenous fistula. Arterial embolization was performed in the patients with arteri- ovenous fistulas, and conservative treatment or place- ment of a double J ureteral catheter was performed in the patients with perirenal hematomas. The most common Grade 4a complications were pleu- ral injury and colon injury. A chest tube was placed in 18 of the patients with pleural injury; the remaining 14 were treated with conservative management. One of the patients with colon injury underwent an operation by a general surgeon perioperatively, and three underwent repair procedures postoperatively. Sepsis developed in three patients (Grade 4b), and these patients died de- spite intense antibiotic and supportive treatment. These findings are consistent with those in the literature.(6,8-10) Hemorrhage is an important morbidity associated with PNL. Kessaris and colleagues reported a 0.8% rate of hemorrhage requiring embolization following PNL.(10) In another large series, the incidence of serious arte- rial bleeding after PNL was reportedly 0.5% to 1.0%. (11) Additionally, Mousavi-Bahar and colleagues report- ed a 0.6% transfusion rate among 671 patients, while El-Nahas and colleagues reported a 16.0% transfusion rate among 241 patients.(12,13) In our study, hemorrhage requiring transfusion occurred in 221 (12.6%) patients, while hemorrhage requiring arterial embolization oc- curred in 7 (0.4%). Sampaio reported a 67% vessel injury rate and a 17% arterial (interlobar) injury rate in percutaneous inter- Table 2. Complication rates according to Modified Clavien Classifica- tion. Modified Clavien Classification Values Grade 1 (Fever) 80 (4.5) Grade 2 (Blood transfusion) 221 (12.6) Grade 3a (Extravasation) 28 (1.6) Grade 3b Perirenal hematoma 7 (0.4) Arteriovenous fistula 17 (0.97) Grade 4a Colon injury 4 (0.22) Pleural injury 32 (1.82) Grade 4b (Sepsis) 3 (0.17) Grade 5 (Death) 4 (0.22) Data are presented as no. (%). Complications Mousavi-Bahar(12) (n = 671) Rosetta(6) (n = 5803) Lee(9) (n = 582) Our Experience (n = 1750) Transfusion 0.6 5.7 11.2 12.6 Hemorrhage requiring intervention 0.15 NA 1.2 0.4 Fever NA 10.5 22.4 4.5 Sepsis 0 NA 0.8 0.17 Colon injury NA NA 0.2 0.22 Pleural injury 0.7 1.8 3.1 1.82 Extravasation/Urine leakage 5.2 3.4 7.2 1.6 Death 0.3 0.3 0.3 0.22 Abbreviation: NA, not available. Data are presented as %. Table 3. Comparison of our complications in percutaneous nephrolithotomy with the literature. Factors Influencing Complications of PNL-Oner et al. Vol 12 No 05 September-October 2015 2319 ventions performed on upper calices.(14) Considering this anatomic feature of the kidney, lower-calyx access is considered to be safest with respect to complications. However, Kukreja and colleagues showed that the loca- tion of the calyx did not affect the development of com- plications.(15) Although this is a controversial finding, the complication rates associated with upper calyx ac- cess in our study were lower than those associated with lower calyx access, although not statistically significant (19.2% vs. 21.5%, respectively) (Table 4). Upper calyx Table 4. The factors that might affect the complication rates (categorized variables). Factors Patients, no Complications, no (%) P Value Gender Male 1055 198 (18.8) _____ Female 695 198 (28.5) Localization Simple 772 149 (19.3) _____ Complex 978 247 (25.3) Access site Lower calyx 619 133 (21.5) < .001 Middle calyx 592 113 (19.1) Upper calyx 172 33 (19.2) Multiple calices 367 117 (31.9) Number of accesses Single Access 1298 251 (19.3) < .001 Multiple Accesses 452 145 (32.1) Costal entry Subcostal 1657 374 (22.6) _____ Supracostal 93 22 (23.7) Staghorn stone Yes 164 59 (36) < .001 No 1586 337 (21.2) Solitary kidney Yes 36 12 (33.3) _____ No 1714 384 (22.4) Horseshoe kidney Yes 39 5 (12.8) _____ No 1711 391 (22.8) Opacity Semi/non opaque 35 6 (17.1) _____ Radiopaque 1715 390 (22.7) Previous operation Yes 329 72 (21.9) _____ No 1421 324 (22.8) Hydronephrosis Yes 1220 262 (21.5) _____ No 530 134 (25.3) Preoperative infection Yes 165 45 (27.3) _____ No 1585 351 (22.1) Factors Influencing Complications of PNL-Oner et al. Endourology and Stone Diseases 2320 access increases the risk of pleural injury; however, es- tablishing access from this site may be easier than lower and middle calyx access, because the guide can be more easily placed in the ureter when it is inserted from the upper calyx. Amplatz dilatation performed through a guide in the ureter can be more easily and rapidly ac- complished. Moreover, manipulations related the renal pelvis and other calices can be performed more com- fortably with upper-pole than with lower-pole access. Placement of a guide in the renal pelvis and ureter is sometimes difficult with lower calyx access; this can re- sult in problems, particularly bleeding during dilatation. In the present series, most blood transfusions occurred in association with access of multiple calyxes, followed by lower calyx access (22.1% and 12.6%, respective- ly) (Table 5). This rate was 8.3% in the middle-calyx group, and lowest requirement for blood transfusion oc- curred in association with isolated upper calyx access. The blood transfusion rates in patients with isolated lower calyx access were significantly higher than those in patients with isolated middle and isolated upper ca- lyx access (P = .014 and P = .009, respectively). No significant difference in the rate of blood transfusion was found between middle and upper calyx access (P = .884). Lower calyx access is generally recommended in the literature. According to our own experience, howev- er, upper calyx access seems to be more convenient and is associated with lower hemorrhage rates. Therefore, we do not believe that the surgeon should be insisted on lower calyx access; if the middle and upper calyxes seem to facilitate stone removal, it would be wise to utilize access through these calyxes. Table 5. The factors that might affect the complication rates (continuous variables). Variables Values Age Cutoff value _____ AUC (SE) 0.478 (0.017) P value .188 Duration of operation (min) Cutoff value 67 AUC (SE) 0.637 (0.016) P value < .001 Surgeon’s experience (no.) Cutoff value 565 AUC (SE) 0.383 (0.017) P value < .001 Stone size (cm2) Cutoff value 7.10 AUC (SE) 0.560 (0.016) P value < .001 Abbreviations: AUC, area under the curve; SE, standard error. Factors Patients (no.) Blood Transfusion no. (%) P Value Access site Lower calyx 619 78 (12.6) < .001 Middle calyx 592 49 (8.3) Upper calyx 172 13 (7.6) Multiple calices 367 81 (22.1) Number of access Single 1298 127 (9.8) < .001 Multiple 452 94 (20.8) Staghorn stone Yes 164 41 (25.0) < .001 No 1586 180 (11.3) Horseshoe kidney Yes 39 0 (0) - No 1711 221 (12.9) Table 6: Factors affecting blood transfusion (categorized variables). Factors Influencing Complications of PNL-Oner et al. Vol 12 No 05 September-October 2015 2321 PNL is associated with lower success rates and higher risks of complications in the treatment of staghorn and complex stones than in the treatment of simple stones. Generally, more than one working channel is needed to clear these stones. Stoller and colleagues found that the formation of multiple working channels increased the hemorrhage rate.(16) According to the kidney stone guidelines of the American Urological Association (AUA), the complication rate associated with staghorn stones is 7% to 27%, and the transfusion rate is about 18%.(7) The presence of staghorn kidney stones and the formation of multiple working channels are independ- ent factors associated with the development of hem- orrhage.(16,17) In a retrospective study of factors affect- ing hemorrhage in 193 patients who underwent PNL, multiple accesses increased bleeding.(18) Additionally, Akman and colleagues reported that the bleeding rates were higher in association with multiple calyx than iso- lated calyx access.(19) In a study of 619 individuals, Akman and colleagues found that stone size was an enhancing factor for trans- fusion rates.(19) In the present study, we found that a larger stone was an enhancing factor for both complica- tions and blood transfusions (P = < .001 and P = .001, respectively; cutoff value, 593 mm2). In a study of the vascular structure of horseshoe kidneys, Jenetschek and Kunzel reported that a significant portion of the blood flow in the kidneys occurs on the medial surface; therefore, posterior access would cause fewer vascular injuries in horseshoe kidneys.(20) The present study sup- ports this finding. The transfusion rate in our series was 12.6% in general, and no blood transfusion was needed following PNL in 39 patients with horseshoe kidneys. Similarly Darabi and colleagues and Ghoneimy and colleagues reported no blood transfusion in their series of PNL in horseshoe kidney.(21,22) A prolonged operation time is another factor that en- hances the rates of complications and blood transfu- sions according to the general literature. Akman and colleagues reported that the cutoff operation time for blood transfusion was 58 minutes; operation times ex- ceeding this value increased the need for blood trans- fusions 2.82-fold.(19) In the present study, the cutoff time was found to be 67 minutes, and the complication frequency increased in patients who underwent opera- tions exceeding this duration. The surgeon’s experience is also a potential factor influencing the complications and the hemorrhage risk. Previous studies have shown that there is a negative correlation between surgical experience and bleeding risk.(17) Allen and colleagues suggested that 60 PNL cases are needed for surgical ad- equacy, while 115 are needed for surgical excellence. (23) In the present study, the blood transfusion rate was 31.0% for the first 100 cases and 9.2% for the 500th to 1750th case. The cutoff surgeon experience was found to be 565 cases for decreasing complications and 647 cases for decreasing blood transfusion rates. The pos- sible explanation for the high cutoff values of surgical experience in our study was considered to be the effect of increasing the number of difficult and complex cases as surgeon experience increases. CONCLUSIONS Although PNL is generally a safe treatment modality for kidney stones, the surgeon must remember that se- rious complications such as death may occur. The com- plication rates in the present study were consistent with those in the general literature. Larger stones, complex stone, multiple-calyx access, an increased number of accesses, presence of staghorn stones, lower surgical experience, and prolonged operation times increased the complication rates. CONFLICT OF INTEREST None declared. REFERENCES 1. Ogg CS, Saxton HM, Cameron JS. Percutaneous needle nephrostomy. Br Med J. 1969;4:657-60. 2. Fernström I, Johansson B. Percutaneous pyelolithotomy. A new extraction technique. Scand J Urol Nephrol. 1976;10:257-9. 3. Badlani G, Eshghi M, Smith AD. Percutaneous surgery for ureteropelvic junction obstruction Table 7. Factors affecting blood transfusion (continuous variables). Age Cutoff Value Values AUC (SE) 0.496 (0.021) P value .858 Duration of operation (min) Cutoff value 67 AUC (SE) 0.679 (0.020) P value < .001 Surgeons experience (no.) Cutoff value 647 AUC (SE) 0.361 (0.021) P value < .001 Stone size (cm2) Cutoff value 5.93 AUC (SE) 0.588 (0.022) P value < .001 Abbreviations: AUC, area under the curve; SE, standard error. Factors Influencing Complications of PNL-Oner et al. Endourology and Stone Diseases 2322 (endopyelotomy): technique and early results. J Urol. 1986;135:26-8. 4. Basiri A, Tabibi A, Nouralizadeh A, et al. Comparison of safety and efficacy of laparoscopic pyelolithotomy versus percutaneous nephrolithotomy in patients with renal pelvic stones: a randomized clinical trial. Urol J. 2014;11:1932-7. 5. Michel MS, Trojan L, Rassweiler JJ. Complications in Percutaneous Nephrolithotomy. Eur Urol. 2007;51:899-906. 6. de la Rosette J, Assimos D, Desai M, et al. The Clinical Research Office of the Endourological Society Percutaneous Nephrolithotomy Global Study: indications, complications, and outcomes in 5803 patients. J Endourol. 2011;25:11-7. 7. Preminger GM, Assimos DG, Lingeman JE, Nakada SY, Pearle MS, 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- 2000. 8. Rudnick DM, Stoller ML. Complications of percutaneous nephrostolithotomy. Can J Urol. 1999;6:872-5. 9. Lee WJ, Smith AD, Cubelli V, et al. Complications of percutaneous nephrolithotomy. AJR Am J Roentgeno. 1987;148:177-80. 10. Kessaris DN, Bellman GC, Pardalidis NP, Smith AG. Management of hemorrhage after percutaneous renal surgery. J Urol. 1995;153:604-8. 11. Patterson DE, Segura JW, LeRoy AJ, Benson RC Jr, May G. The etiology and treatment of delayed bleeding following percutaneous lithotripsy. J Urol. 1985;133:447-51. 12. Mousavi-Bahar SH, Mehrabi S, Moslemi MK. Percutaneous nephrolithotomy complications in 671 consecutive patients: a single-center experience. Urol J. 2011;8:271-6. 13. el-Nahas AR, Eraky I, Shokeir AA, et al. Factors affecting stone-free rate and complications of percutaneous nephrolithotomy for treatment of staghorn Stone. Urology. 2012;79:1236-41. 14. Sampaio FJR. How to place a nephrostomy safely. Contemp Urol. 1994;6:41-6. 15. Kukreja R, Desai M, Patel S, Bapat S, Desai M. Factors affecting blood loss during percutaneous nephrolithotomy: Prospective study. J Endourol. 2004;18:715-22. 16. Stoller ML, Wolf JS Jr., St Lezin MA. Estimated blood loss and transfusion rates associated with percutaneous nephrolithotomy. J Urol. 1994;152:1977-81. 17. El-Nahas AR, Shokeir AA, El-Assmy AM, et al. Post-percutaneous nephrolithotomy extensive hemorrhage: a study of risk factors. J Urol. 2007;177:576-9. 18. Turna B, Nazli O, Demiryoguran S, Mamadov R, Cal C. Percutaneous nephrolithotomy: variables that influence hemorrhage. Urology. 2007;69:603-7. 19. Akman T, Binbay M, Sari E, et al. Factors Affecting Bleeding During Percutaneous Nephrolithotomy: Single Surgeon Experience. J Endourol. 2011;25:327-33. 20. Janetschek G and Kunzel KH. Percutaneous nephrolithotomy in horseshoe kidneys. Br J Urol. 1988;62:117-22. 21. Darabi Mahboub MR, Zolfaghari M, Ahanian A. Percutaneous nephrolithotomy of kidney calculi in horseshoe kidney. Urol J. 2007;4:147-50. 22. El Ghoneimy MN, Kodera AS, Emran AM, Orban TZ, Shaban AM, El Gammal MM. Percutaneous nephrolithotomy in horseshoe kidneys: is rigid nephroscopy sufficient tool for complete clearance? A case series study. BMC Urol. 2009;16;9:17. 23. Allen D, O’Brien T, Tiptaft R, Glass J. Defining the learning curve for percutaneous nephrolithotomy. J Endourol. 2005;19:279- 82. Factors Influencing Complications of PNL-Oner et al. Vol 12 No 05 September-October 2015 2323