ORIGINAL ARTICLE Complications of Laparoscopic Splenectomy

Eduardo M. Targarona, MD, PhD; Juan Jose´ Espert, MD; Ernest Bombuy, MD; Oscar Vidal, MD; Gemma Cerda´n, MD; Vicente Artigas, MD, PhD; Manuel Trı´as, MD, PhD

Hypothesis: Analysis of the type and characteristics of Results: One hundred thirteen laparoscopic splenectomies complications after laparoscopic splenectomy may per- were completed (conversion rate, 7.4%). Twenty patients mit the identification of clinical factors with predictive (18%) developed 23 complications. All were Clavien type value for the development of complications. I or II, without mortality. One complication was intraop- erative (diaphragmatic perforation), and 22 were postop- Design: Univariate and multivariate analysis of factors erative: 6 pulmonary (26%), 3 fever (13%), 8 hemorrhagic related to complications in a prospective series of lapa- (35%) (5 episodes of postoperative and 3 abdomi- roscopic splenectomies. nal wall ), and 6 others (26%). Ten (43%) of the 23 were technically related. Univariate analysis showed Setting: A large tertiary referral university–teaching gen- that complications were only related to age (mean±SD, eral hospital. 55±15 vs 39±17 years; PϽ.008) or transfusion (50% vs 11%; PϽ.001). Multivariate analysis showed that the learn- Patients: One hundred twenty-two nonselected con- ing curve (PϽ.005; 95% confidence interval, 2.46), age secutive patients, in whom laparoscopic splenectomy was (PϽ.001; 95% confidence interval, 1.04), spleen weight attempted between February 1993 and July 1999. (PϽ.009; 95% confidence interval, 1.00), and malignant neoplasmdiagnosis(PϽ.007;95%confidenceinterval,3.82) Intervention: Laparoscopic splenectomy. were independent predictors of complications.

Main Outcome Measures: Immediate complica- Conclusions: Laparoscopic splenectomy is feasible, and tions classified according to the Clavien score. Univari- the incidence of severe complications is reduced. How- ate and multivariate analyses were performed of compli- ever, a high proportion of these complications are tech- cations related to age, sex, body mass index, and nique related. Laparoscopic splenectomy requires great malignant nature of the hematologic disease; preopera- technical care but offers major clinical advantages, even tive hematocrit and count; operative time; in less favorable situations, such as in patients with sple- operative position; need of accessory incision; transfu- nomegaly or with malignant neoplasms. sion status; learning curve; and existence of comorbid diseases. Arch Surg. 2000;135:1137-1140

APAROSCOPIC splenectomy RESULTS (LS) has become increas- ingly popular in recent years.1-3 However, it entails One hundred thirteen LSs were com- technical difficulties due to the pleted (conversion, 9 [7.4%] of 122); the Lspleen’s vascular anatomic features and its clinical details are as follows: frail texture. In addition, the specific fea- Clinical Outcome Value* tures of hematologic diseases (low platelet Operative time, min 153 ± 59 (60-240) count, , or systemic disease Transfusion, % 18 []) may increase postoperative Morbidity, % 18 From the Services of General morbidity. The best evaluation of a new pro- Hospital stay, d 4 ± 2 (2-14) and Digestive Surgery, Hospital cedure is via an analysis of its complica- Spleen weight, g 493 ± 588 (60-3200) Accessory spleen, % 12 de Sant Pau, Autonomous tions and the advantages it offers; in the case University of Barcelona Accessory incision, % 36 (Drs Targarona, Cerda´n, of LS, an important argument in its favor is the reduction of the morbidity rate. Thus, *Data are given as mean±SD (range) unless other- Artigas, and Trı´as), and wise indicated. Hospital Clinic (Drs Espert, this study analyzes the type and character- Bombuy, and Vidal), istics of complications related to LS in a con- No conversions were related to Barcelona, Spain. secutive series of 122 patients. intraoperative complications. Twenty

(REPRINTED) ARCH SURG/ VOL 135, OCT 2000 WWW.ARCHSURG.COM 1137

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 Table 2. Clavien Classification for Surgical Complications* PATIENTS AND METHODS Clavien Type Description I Mild complications that do not affect the Between February 1993 and July 1999, 122 patients patient’s recovery in the hospital with a wide range of splenic disorders (Table 1) re- IIa Potentially life-threatening quiring splenectomy were approached about under- complications without sequel that going laparoscopy. Clinical data, technical details, and require noninvasive therapy immediate outcome were recorded prospectively. All IIb Potentially life-threatening patients received a preoperative pneumococcal vac- complications without sequel that require interventional therapy cine, and antibiotic prophylaxis was initiated intra- III Complications with residual disability operatively. Laparoscopic splenectomies were per- IV Death formed through an anterior or lateral approach, using 4 techniques previously described. Immediate com- *Data from Clavien et al.5 plications were prospectively recorded and classi- fied according to the Clavien score (Table 2).5 Uni- variate and multivariate analyses were performed of complications related to age, sex, body mass index Table 3. Postoperative Complications After Successful (calculated as weight in kilograms divided by the Laparoscopic Splenectomy square of height in meters), and malignant nature of the hematologic disease; preoperative hematocrit and Complications* Clavien Type Technique Related platelet count; operative time; operative position; need Hemorrhagic of an accessory incision; transfusion status; learning Abdominal wall (3) I Yes curve; and the existence of comorbid diseases. The Subdiaphragmatic hematoma IIa Yes 2 ␹ test was used for comparison of the 2 propor- (3) IIb Yes tions, and the t test was used to compare differences Hemopneumothorax IIb Yes between the 2 series. Multivariate analysis and a model Lung of logistic regression with backward and forward Atelectasis IIa No analysis were used. Pneumonia (2) IIa No Upper airway infection (2) I No Fever and lung tuberculosis II No Septic Wound sepsis I Yes Catheter sepsis II No Table 1. Hematologic Diagnosis in 122 Cases of Attempted Urinary infection I No Laparoscopic Splenectomy* Others Sweet syndrome II No Gout attack I No Hematologic Diagnosis No. of Cases Hypophyseal insufficiency II No Autoimmune low platelet counts 52 Postoperative (2) I No ITP 46 Diaphragmatic perforation II Yes 3 ITP and systemic erythematosus 2 *Numbers in parentheses indicate the number of cases. ITP and 1 HIV related 8 Low platelet counts 5 patients (18%) in the successful LS group developed 23 Hypersplenism 1 complications. All of them were Clavien type I or II, and Splenomegaly and suspicion of malignant neoplasm 2 there was no mortality. Complications and their type are Autoimmune hemolytic 9 described in Table 3. Of the 23 complications, 10 (43%) Spherocytosis 14 Malignant neoplasm 29 were technically related to the procedure, 8 (35%) were Non- 15 hemorrhagic complications, 6 (26%) were pulmonary Myelofibrosis 3 complications, 3 (13%) were infectious complications, Chronic lymphocytic 3 and 6 (26%) were other types of complications. Com- Chronic myeloid leukemia 1 plications were not related to the diagnosis of a malig- Hairy cell lymphoma 4 nant neoplasm (malignant vs benign, 37% vs 21%), spleen Others 3 weight (patients with complications vs those without com- Others 10 Splenomegaly 2 plications, 717±504 vs 474±628 g), or operative time (pa- Splenic tumor tients with complications vs those without complica- Hydatidic cyst 1 tions, 172±75 vs 147±54 minutes; PϽ.06) (data are given Hamartoma 1 as mean±SD unless otherwise indicated). However, pa- Cyst 1 tients with complications were significantly older and re- Angioma 3 quired more blood (53% vs 13%; PϽ.001); complica- Splenic trauma 1 Thrombotic thrombocytopenic 1 tions had a direct impact on postoperative stay (mean±SD, 9±5 vs 3±1 days; PϽ.001) (Table 4). *ITP indicates idiopathic ; HIV, human Logistic regression analysis indicated that factors with immunodeficiency virus. predictive power for morbidity were age, spleen weight,

(REPRINTED) ARCH SURG/ VOL 135, OCT 2000 WWW.ARCHSURG.COM 1138

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 malignant nature of the disease, and procedures per- formed early in our experience (learning curve) (Table 5). Table 4. Results of Univariate Analysis of Clinical Factors The reasons for the 9 conversions were as follows: Related to Complications After Laparoscopic Splenectomy diffuse oozing (n=2), spleen size (n=6), and adhesions Patients Patients (n=1). The characteristics of the 9 patients are as follows: With Without Characteristic Value Complications Complications Age, mean±SD 55±20 Clinical Factor (n = 20) (n = 93) P Male-female ratio 5:4 Sex, M/F 9/11 25/68 Ͼ.18 Body mass index, mean±SD 24±3 Age, mean ± SD, y 55 ± 15 39 ± 17 Ͻ.001 Spleen weight, mean±SD (range), g 1843±1323 (200-3500) Body mass index, 25.0 ± 5.0 25.0 ± 4.5 Ͼ.99 Maligant-benign disease ratio 4:5 mean ± SD* Complications, No. (%) 2 (22%) Preoperative hematocrit 0.37 ± 0.05 0.35 ± 0.07 Ͻ.22 Transfusions 7 Preoperative platelet count, 111 ± 114 104 ± 93 Ͻ.78 Postoperative stay, mean±SD, d 6±1 mean ± SD, ϫ 109/L Operative time, mean±SD, min 170±46 Operative time, mean ± SD, 172±75 147±54 Ͻ.10 min Spleen weight, mean ± SD, g 717 ± 504 474 ± 628 Ͻ.12 COMMENT Malignant-benign neoplasm 7:13 (35) 19:74 (20) Ͻ.25 ratio† Lateral-supine approach 17:3 (85) 87:6 (93) Ͻ.86 Open splenectomy is not usually technically demand- ratio† ing, except for the treatment of a massively enlarged Accessory incision− 9:11 (45) 24:69 (26) Ͻ.09 spleen.6 However, it requires a wide laparotomy to morcellation† gain access to the left hypochondriac region fossa. The Transfusion−no transfusion† 10:10 (50) 11:82 (12) Ͻ.001 Comorbid diseases, yes-no 4:16 (20) 19:74 (20) Ͻ.84 spleen’s rich vascularization and its intimate anatomic ratio† relations with intra-abdominal organs, along with the traction and maneuvers necessary for exposure, mean *Calculated as weight in kilograms divided by the square of height in meters. that the procedure is associated with complications in †Data in parentheses are percentages (first part of the ratio/total of that 5% to 60% of patients.7-10 The complications are column). mainly secondary to pancreatic tail injury or bleeding, or are pulmonary in origin. Splenectomy for malignant diseases or enlarged spleen is associated with a mor- Table 5. Results of Multivariate Analysis of Clinical Factors bidity of 40% to 60%.7-10 Related to Complications After Laparoscopic Splenectomy Potential advantages of laparoscopy are a decrease in the complication rate due to the less aggressive na- 95% ture of the approach and the great visual detail ob- Clinical Factor P Confidence Interval tained; also, it avoids manipulation of the left side of the Learning curve .005 2.46 diaphragm. Large series of LSs available report a signifi- Age Ͻ.001 1.04 Spleen weight (measured in grams) .009 1.00 cant decrease in complications compared with open sur- Malignant vs benign neoplasm .007 3.82 gery, although most of these studies are retrospective and randomized trials are lacking.11-13 In our series, we observed a morbidity rate of 18% in an unselected group of 122 LSs. Forty-three percent proach facilitates dissection and mobilization of the spleen, were clearly technique related, and most were of hem- reduces operative time, and avoids the manipulation of orrhagic origin. In univariate analysis, the only factor that the spleen and tears or capsule fractures. In this series, showed a significant difference comparing complicated there was no massive intraoperative bleeding, nor in- with noncomplicated cases was age. However, multivar- jury to great splenic vessels, which in most patients were iate analysis showed that the learning curve, age, malig- controlled by an endostapler.16 An important reason for nant neoplasm, and spleen weight were predictive fac- this was the use of the lateral approach, which permits tors for complications. mobilization of the spleen and the placement of the sta- We began to use LS in early 1993, and we have in- pler without tension. Two cases were converted early in cluded all our patients in this series; we thus include our our experience due to bleeding originating in epiploon early learning curve, before we mastered the technique. vessels. Technical results have also been improved by the The demonstration that the learning curve is an indepen- development of instruments that facilitate the dissec- dent factor for complication should be highlighted. How- tion maneuvers during LS, such as the harmonic scal- ever, at present, LS is a well-systematized procedure that pel.17 The use of this scalpel does not require clips, which can be taught and learned by trainees or inexperienced sur- may become dislodged, and simplifies the control of low- geons, thus reducing morbidity due to the learning curve. caliber epiploon vessels. Several researchers1,14,15 have shown that the procedure can One severe hemorrhagic complication is postopera- be performed in a teaching environment, with most of the tive bleeding, which occurred 4 times (3%): 3 times af- procedures being performed by residents. ter successful LS, and a fourth after conversion. A po- In our experience, another factor has favored the re- tential hazard when using mechanical devices to control duction of the incidence of complications: the switch from splenic vessels is the risk of bleeding due to a high- the anterior to the lateral approach.4 The lateral ap- pressure flux on a row of staples. We try to interrupt the

(REPRINTED) ARCH SURG/ VOL 135, OCT 2000 WWW.ARCHSURG.COM 1139

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 arterial flux with a clip to reduce the spleen size. This Presented as a poster at the 1999 Scientific Meeting of maneuver also has a certain autotransfusion effect. Sages, San Antonio, Tex, April 25-27, 1999; and as a free One dangerous technical complication is diaphrag- paper at the European Association of Endoscopic Surgery matic perforation, which is rare during open surgery. A Annual Meeting, Linz, Austria, June 23-26, 1999. small cautery injury of the diaphragm dome, mainly in We thank M. T. Puig, MD, and J. Gich, MD, for their the muscled part, may enlarge as a consequence of the help with the statistical analysis. pneumoperitoneum and develop into a hole. This has been Reprints: Manuel Trı´as, MD, PhD, Service of General described by other researchers16; it can usually be con- and Digestive Surgery, Hospital de Sant Pau, Avda P Claret trolled by laparoscopy. 167, 08025 Barcelona, Spain (e-mail: [email protected]). Age, the malignant nature of the hematologic dis- ease, and increased spleen weight are well-known fac- REFERENCES tors related to complications after open splenectomy, and 6-10 all 3 usually are interrelated. Several medical compli- 1. Kathkouda N, Hurtwitz MB, Rivera RT, et al. Laparoscopic splenectomy: out- cations occurred in our patients, which were unrelated come and efficacy in 103 consecutive cases. Ann Surg. 1998;228:568-578. to the laparoscopic technique (Sweet syndrome, gout at- 2. Gigot JF, Lengele B, Gianello P, Etienne J, Claeys N. Present status of laparo- tack, and hypophysial insufficiency). Central line infec- scopic splenectomy for hematological diseases: certitudes and unresolved is- tion and urinary sepsis were also observed in this sub- sues. Semin Laparosc Surg. 1998;5:147-167. 3. Klinger PJ, Tsiotos GG, Glaser KS, Hinder RA. Laparoscopic splenectomy and group of patients, who usually underwent more current status. Surg Laparosc Endosc. 1999;9:1-8. manipulation than the easier patients (ie, those who were 4. Trias M, Targarona EM, Espert JJ, Balague´ C. Laparoscopic surgery for splenic dis- younger and healthier). Laparoscopic management of sple- orders: lessons learned from a series of 64 cases. Surg Endosc. 1998;12:66-72. nomegaly entails a more demanding technique, and ma- 5. Clavien PA, Sanabria JR, Strasberg JM. Proposed classification of complica- tions of surgery with example of utility in cholecystectomy. Surgery. 1992;111: nipulation of a large organ and extraction of the viscera 518-526. 18,19 may be more difficult. Splenomegaly has been clas- 6. Hiatt JR, Allins A, Kong LR. Open splenectomy. In: Hiatt JR, Phillips EH, Mor- sically associated with increased morbidity after open sple- genstern L, eds. Surgical Diseases of the Spleen. Berlin, Germany: Springer; 1997: nectomy.6-10 However, in our experience, LS for spleno- 197-210. megaly, when successful (LS success rates of 85% for 7. McRae HM, Yakimets WW, Reynolds T. Perioperative complications of splenec- tomy for hematologic disease. Can J Surg. 1992;35:432-436. patients with spleens Ͼ400 g and 75% for those with 8. Aksnes J, Abdelnoor M, Mathisen O. Risk factors associated with mortality and spleens Ͼ1000 g), presented the same advantages as LS morbidity after elective splenectomy. Eur J Surg. 1995;161:253-258. for smaller spleens, and yielded better results than open 9. Horowitz J, Smith JL, Weber TK, Rodriguez-Bigas MA, Petrelli NJ. Postopera- splenectomy. tive complications after splenectomy for hematologic malignancies. Ann Surg. 1996;223:290-296. The laparoscopic approach has intrinsic potential 10. Coon W. Surgical aspects of splenic disease and lymphoma. Curr Probl Surg. complications due to the type of access (trocars and 1998;35:543-646. pneumoperitoneum).20 An unusual bleeding complica- 11. Friedman RL, Hiatt JR, Korman JL, Facklis K, Cymerman J, Phillips EH. Laparo- tion during open surgery is abdominal wall hematoma. scopic or open splenectomy for hematological disease: which approach is su- Trocar placement is performed through several muscu- perior? J Am Coll Surg. 1997;185:49-54. 12. Targarona EM, Espert JJ. Effect of spleen size on splenectomy outcome: a com- lar planes when the lateral approach is used, and subse- parison of open and laparoscopic surgery. Surg Endosc. 1999;13:559-562. quent pneumoperitoneum may dissect these anatomic 13. Park A, Marcaccio M, Sternbach M, Witzke D, Fitzgerald P. Laparoscopic versus planes. Despite careful suture of portholes, we observed open splenectomy. Arch Surg. 1999;134:1263-1269. a postoperative decrease of hematocrit and blood accu- 14. Poulin E, Mamamzza J. Laparoscopic splenectomy: lessons from the learning mulation in the abdominal wall in 3 patients. All of them curve. Can J Surg. 1998;41:28-36. 15. Rege RV, Joehl RJ. A learning curve for laparoscopic splenectomy at an aca- had a low platelet count and a slow elevation of platelet demic institution. J Surg Res. 1999;81:27-32. count. 16. Navarro RA, Korman JE, Phillips EH. Complications of laparoscopic splenec- Laparoscopic splenectomy is feasible, and the inci- tomy. Semin Laparosc Surg. 1997;4:182-189. dence of severe complications is reduced. However, a high 17. Gossot D, Fritsch S, Celerier M. Laparoscopic splenectomy: optimal vascular con- trol using the lateral approach and ultrasonic dissection. Surg Endosc. 1999;13: proportion of these complications are technique re- 21-25. lated. Laparoscopic splenectomy requires great techni- 18. Terrossu G, Donini A, Baccarini U, et al. Laparoscopic versus open splenectomy cal care but offers major clinical advantages, even in less in the management of splenomegaly: our preliminary experience. Surgery. 1998; favorable situations, such as in patients with spleno- 124:839-843. megaly or with malignant neoplasms. 19. Targarona EM, Espert JJ, Balague´ C, Piulachs J, Artigas V, Trias M. Spleno- megaly should not be considered a contraindication for laparoscopic spleno- megaly. Ann Surg. 1998;228:35-39. This study was supported by grant 97/970 from the Fondo 20. McMahon MJ. Complications of laparoscopic access. Semin Laparosc Surg. 1997; de Investigaciones Sanitarias, Madrid, Spain. 4:139-146.

(REPRINTED) ARCH SURG/ VOL 135, OCT 2000 WWW.ARCHSURG.COM 1140

©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021