Complications of Laparoscopic Splenectomy

Hypothesis  Analysis of the type and characteristics of complications after laparoscopic splenectomy may permit the identification of clinical factors with predictive value for the development of complications.

Design  Univariate and multivariate analysis of factors related to complications in a prospective series of laparoscopic splenectomies.

Setting  A large tertiary referral university–teaching general hospital.

Patients  One hundred twenty-two nonselected consecutive patients, in whom laparoscopic splenectomy was attempted between February 1993 and July 1999.

Intervention  Laparoscopic splenectomy.

Main Outcome Measures  Immediate complications classified according to the Clavien score. Univariate and multivariate analyses were performed of complications related to age, sex, body mass index, and malignant nature of the hematologic disease; preoperative hematocrit and platelet count; operative time; operative position; need of accessory incision; transfusion status; learning curve; and existence of comorbid diseases.

Results  One hundred thirteen laparoscopic splenectomies were completed (conversion rate, 7.4%). Twenty patients (18%) developed 23 complications. All were Clavien type I or II, without mortality. One complication was intraoperative (diaphragmatic perforation), and 22 were postoperative: 6 pulmonary (26%), 3 fever (13%), 8 hemorrhagic (35%) (5 episodes of postoperative bleeding and 3 abdominal wall hematomas), and 6 others (26%). Ten (43%) of the 23 were technically related. Univariate analysis showed that complications were only related to age (mean ± SD, 55 ± 15 vs 39 ± 17 years; P<.008) or transfusion (50% vs 11%; P<.001). Multivariate analysis showed that the learning curve (P<.005; 95% confidence interval, 2.46), age (P<.001; 95% confidence interval, 1.04), spleen weight (P<.009; 95% confidence interval, 1.00), and malignant neoplasm diagnosis (P<.007; 95% confidence interval, 3.82) were independent predictors of complications.

Conclusions  Laparoscopic splenectomy is feasible, and the incidence of severe complications is reduced. However, a high proportion of these complications are technique related. Laparoscopic splenectomy requires great technical care but offers major clinical advantages, even in less favorable situations, such as in patients with splenomegaly or with malignant neoplasms.

LAPAROSCOPIC splenectomy (LS) has become increasingly popular in recent years.^1-3 However, it entails technical difficulties due to the spleen's vascular anatomic features and its frail texture. In addition, the specific features of hematologic diseases (low platelet count, splenomegaly, or systemic disease [lymphoma]) may increase postoperative morbidity. The best evaluation of a new procedure is via an analysis of its complications and the advantages it offers; in the case of LS, an important argument in its favor is the reduction of the morbidity rate. Thus, this study analyzes the type and characteristics of complications related to LS in a consecutive series of 122 patients.

Between February 1993 and July 1999, 122 patients with a wide range of splenic disorders (Table 1) requiring splenectomy were approached about undergoing laparoscopy. Clinical data, technical details, and immediate outcome were recorded prospectively. All patients received a preoperative pneumococcal vaccine, and antibiotic prophylaxis was initiated intraoperatively. Laparoscopic splenectomies were performed through an anterior or lateral approach, using techniques previously described.⁴ Immediate complications were prospectively recorded and classified according to the Clavien score (Table 2).⁵ Univariate and multivariate analyses were performed of complications related to age, sex, body mass index (calculated as weight in kilograms divided by the square of height in meters), and malignant nature of the hematologic disease; preoperative hematocrit and platelet count; operative time; operative position; need of an accessory incision; transfusion status; learning curve; and the existence of comorbid diseases. The χ² test was used for comparison of the 2 proportions, and the t test was used to compare differences between the 2 series. Multivariate analysis and a model of logistic regression with backward and forward analysis were used.

One hundred thirteen LSs were completed (conversion, 9 [7.4%] of 122); the clinical details are as follows:

No conversions were related to intraoperative complications. Twenty patients (18%) in the successful LS group developed 23 complications. All of them were Clavien type I or II, and there was no mortality. Complications and their type are described in Table 3. Of the 23 complications, 10 (43%) were technically related to the procedure, 8 (35%) were hemorrhagic complications, 6 (26%) were pulmonary complications, 3 (13%) were infectious complications, and 6 (26%) were other types of complications. Complications were not related to the diagnosis of a malignant neoplasm (malignant vs benign, 37% vs 21%), spleen weight (patients with complications vs those without complications, 717 ± 504 vs 474 ± 628 g), or operative time (patients with complications vs those without complications, 172 ± 75 vs 147 ± 54 minutes; P<.06) (data are given as mean ± SD unless otherwise indicated). However, patients with complications were significantly older and required more blood (53% vs 13%; P<.001); complications had a direct impact on postoperative stay (mean ± SD, 9 ± 5 vs 3 ± 1 days; P<.001) (Table 4).

Logistic regression analysis indicated that factors with predictive power for morbidity were age, spleen weight, malignant nature of the disease, and procedures performed early in our experience (learning curve) (Table 5).

The reasons for the 9 conversions were as follows: diffuse oozing (n = 2), spleen size (n = 6), and adhesions (n = 1). The characteristics of the 9 patients are as follows:

Open splenectomy is not usually technically demanding, except for the treatment of a massively enlarged spleen.⁶ However, it requires a wide laparotomy to gain access to the left hypochondriac region fossa. The spleen's rich vascularization and its intimate anatomic relations with intra-abdominal organs, along with the traction and maneuvers necessary for exposure, mean that the procedure is associated with complications in 5% to 60% of patients.^7-10 The complications are mainly secondary to pancreatic tail injury or bleeding, or are pulmonary in origin. Splenectomy for malignant diseases or enlarged spleen is associated with a morbidity of 40% to 60%.^7-10

Potential advantages of laparoscopy are a decrease in the complication rate due to the less aggressive nature of the approach and the great visual detail obtained; also, it avoids manipulation of the left side of the diaphragm. Large series of LSs available report a significant decrease in complications compared with open surgery, 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 were clearly technique related, and most were of hemorrhagic origin. In univariate analysis, the only factor that showed a significant difference comparing complicated with noncomplicated cases was age. However, multivariate analysis showed that the learning curve, age, malignant neoplasm, and spleen weight were predictive factors for complications.

We began to use LS in early 1993, and we have included all our patients in this series; we thus include our early learning curve, before we mastered the technique. The demonstration that the learning curve is an independent factor for complication should be highlighted. However, at present, LS is a well-systematized procedure that can be taught and learned by trainees or inexperienced surgeons, thus reducing morbidity due to the learning curve. Several researchers^1,14,15 have shown that the procedure can be performed in a teaching environment, with most of the procedures being performed by residents.

In our experience, another factor has favored the reduction of the incidence of complications: the switch from the anterior to the lateral approach.⁴ The lateral approach facilitates dissection and mobilization of the spleen, reduces operative time, and avoids the manipulation of the spleen and tears or capsule fractures. In this series, there was no massive intraoperative bleeding, nor injury to great splenic vessels, which in most patients were controlled by an endostapler.¹⁶ An important reason for this was the use of the lateral approach, which permits mobilization of the spleen and the placement of the stapler without tension. Two cases were converted early in our experience due to bleeding originating in epiploon vessels. Technical results have also been improved by the development of instruments that facilitate the dissection maneuvers during LS, such as the harmonic scalpel.¹⁷ The use of this scalpel does not require clips, which may become dislodged, and simplifies the control of low-caliber epiploon vessels.

One severe hemorrhagic complication is postoperative bleeding, which occurred 4 times (3%): 3 times after successful LS, and a fourth after conversion. A potential hazard when using mechanical devices to control splenic vessels is the risk of bleeding due to a high-pressure flux on a row of staples. We try to interrupt the arterial flux with a clip to reduce the spleen size. This maneuver also has a certain autotransfusion effect.

One dangerous technical complication is diaphragmatic perforation, which is rare during open surgery. A small cautery injury of the diaphragm dome, mainly in the muscled part, may enlarge as a consequence of the pneumoperitoneum and develop into a hole. This has been described by other researchers¹⁶; it can usually be controlled by laparoscopy.

Age, the malignant nature of the hematologic disease, and increased spleen weight are well-known factors related to complications after open splenectomy, and all 3 usually are interrelated.^6-10 Several medical complications occurred in our patients, which were unrelated to the laparoscopic technique (Sweet syndrome, gout attack, and hypophysial insufficiency). Central line infection and urinary sepsis were also observed in this subgroup of patients, who usually underwent more manipulation than the easier patients (ie, those who were younger and healthier). Laparoscopic management of splenomegaly entails a more demanding technique, and manipulation of a large organ and extraction of the viscera may be more difficult.^18,19 Splenomegaly has been classically associated with increased morbidity after open splenectomy.^6-10 However, in our experience, LS for splenomegaly, when successful (LS success rates of 85% for patients with spleens >400 g and 75% for those with spleens >1000 g), presented the same advantages as LS for smaller spleens, and yielded better results than open splenectomy.

The laparoscopic approach has intrinsic potential complications due to the type of access (trocars and pneumoperitoneum).²⁰ An unusual bleeding complication during open surgery is abdominal wall hematoma. Trocar placement is performed through several muscular planes when the lateral approach is used, and subsequent pneumoperitoneum may dissect these anatomic planes. Despite careful suture of portholes, we observed a postoperative decrease of hematocrit and blood accumulation in the abdominal wall in 3 patients. All of them had a low platelet count and a slow elevation of platelet count.

Laparoscopic splenectomy is feasible, and the incidence of severe complications is reduced. However, a high proportion of these complications are technique related. Laparoscopic splenectomy requires great technical care but offers major clinical advantages, even in less favorable situations, such as in patients with splenomegaly or with malignant neoplasms.

This study was supported by grant 97/970 from the Fondo de Investigaciones Sanitarias, Madrid, Spain.

Presented as a poster at the 1999 Scientific Meeting of Sages, San Antonio, Tex, April 25-27, 1999; and as a free paper at the European Association of Endoscopic Surgery Annual Meeting, Linz, Austria, June 23-26, 1999.

We thank M. T. Puig, MD, and J. Gich, MD, for their help with the statistical analysis.

Reprints: Manuel Trías, MD, PhD, Service of General and Digestive Surgery, Hospital de Sant Pau, Avda P Claret 167, 08025 Barcelona, Spain (e-mail: 7366@hsp.santpau.es).