Validation of a Difficulty Grading Score in Laparoscopic Splenectomy

Abstract

Introduction:

Laparoscopic splenectomy (LS) is the gold standard for the treatment of many spleen-related disorders. However, in the presence of splenomegaly, the rate of conversion to open surgery can be as high as 33.9% and is associated with longer operative times and higher costs. In an attempt to improve risk stratification and patient selection for LS, a score that includes four preoperative parameters (age, gender, diagnosis, and spleen size) has been developed. The aim was the validation of a difficulty grading score, in predicting conversion and poorer outcomes.

Methods:

Retrospective analysis of 153 consecutive patients subjected to LS from January 2006 through December 2016 was performed. Several parameters were reviewed and correlation with evaluated outcomes was analyzed.

Results:

Conversion to open surgery occurred in 13 (8.50%) patients and was highly associated with serious intraoperative complications. Spearman correlation showed a significant association between the score and risk of conversion, operative time, and postoperative complications, but not with intraoperative bleeding.

Discussion:

Patient and disease features, incorporated in a difficulty grading score, can reliably determine the difficulty of LS and predict risk of conversion, intraoperative, and postoperative complications. This simple and reproducible score improves risk stratification for LS and could be practical in daily clinical activities.

Introduction

Laparoscopic splenectomy (LS) was first described by Delaitre and Maignien in 1992.¹ Owing to the advantages of minimally invasive surgery, namely less postoperative pain, faster recovery, shorter hospital stay, and better cosmetic results,² it has been considered the gold standard for the treatment of benign and malignant spleen-related diseases.³

Splenomegaly was initially considered a relative contraindication for LS, due to inadequate exposure of the upper left quadrant, increased tissue vascularity, and difficulties in specimen retrieval.⁴ In the past two decades, this contraindication has been obviated⁵ due to accumulating experience, innovative instruments,⁶ and hybrid techniques, such as hand-assisted LS.⁷

The rate of conversion of LS to open surgery can be as high as 33.9% which is one of the highest rates among other advanced laparoscopic procedures (colectomy, sleeve gastrectomy, Roux-en-Y bypass, and adrenalectomy), suggesting that LS requires advanced laparoscopic skills.^2,8 Similar to other laparoscopic procedures, conversion to open surgery should not be regarded as a complication but rather as a wise decision to prevent major morbidity or even mortality. When conversion to open surgery is needed, it is associated with increased length of stay and higher costs,⁹ which could be avoided by a correct patient selection for LS.

In the literature, many variables such as indication for splenectomy, individual patient factors, intraoperative findings (namely hemorrhage),⁸ and learning curve have been related to conversion to open surgery.¹⁰ In an attempt to improve risk stratification for surgery, Rodriguez-Otero Luppi et al. developed and validated a predictive model of technical difficulty in LS for nontraumatic diseases.⁹ They evaluated its association with operative time, operative bleeding, and conversion to open surgery. This score with three levels of difficulty was then developed based on four preoperative parameters: age, gender, diagnosis, and spleen size based on final spleen weight.

This study aims to independently evaluate the performance of this scoring tool in the selection of patients for LS.

Materials and Methods

Adult patients (≥18 years old) subjected to LS as a primary elective procedure have been included in an electronic database from January 2006 through December 2016.

This database was retrospectively reviewed. The Ethics Committee of Sao Joao Medical Center approved the study, and all participants signed informed consent. The following variables were analyzed: demographic data, body mass index (BMI), indication for splenectomy, previous abdominal surgeries, spleen size and weight, hematocrit, platelet count, surgeon experience, rate and causes of conversion, intraoperative bleeding, operative time, and intraoperative complications.

The surgical technique has remained unchanged for the duration of the study and is described next. Patients were positioned in right lateral decubitus or dorsal decubitus with axial rotation (30°) to the right and with trunk extension. In general, four trocars were used (15, 10 mm, and two of 5 mm). Splenic mobilization was achieved by sectioning the lienocolic, lienorenal, and lienophrenic ligaments. Short vessels and gastrosplenic ligament were ligated with a harmonic scalpel. The splenic vein and artery were isolated and individually ligated with clips or a stapler. The specimen was retrieved either through the 15 mm port or through a Pfannestiel incision. Operative time was recorded as skin-to-skin time.

Using the preoperative parameters, scores were assigned based on the difficulty grading score criteria (Table 1), described by Rodriguez-Otero Luppi et al.⁹ The minimum possible score was 2 and the maximum possible score was 10. The degree of surgical difficulty was low for patients with ≤4 points, medium for those with 4.5–5.5 points, and high for those with ≥6 points. The validation of this score was achieved by comparing the three grades of difficulty with operative time, conversion to open surgery, operative bleeding, and postoperative complications.

Table 1.

Difficulty Grading Score

Parameter	Score
Age
≤40 Years	0
40–60 Years	1
≥60 Years	2
Gender
Female	0.5
Male	1
Pathology group
ITP	0.5
Other benign	1
Malignant	2
Spleen weight^a
<400 g	1
400–1000 g	3
>1000 g	5
Difficulty grade
Low	≤4
Medium	4.5–5.5
High	≥6

Minimum possible score, 2; maximum possible score, 10.

Spleen weight formula: width (cm) × length (cm) × height (cm) × 0.6 = splenic weight in grams.

Adapted from Rodriguez-Otero Luppi et al.⁹

ITP, idiopathic trombocitopenic purpura.

Preoperative abdominal ultrasound and/or computed tomography (CT) scan was performed to measure the size of the spleen and to assess the presence of accessory spleens. Splenomegaly was defined as a splenic vertical diameter of >15 cm, according to the European Association for Endoscopic Surgery.⁴ Spleen weight was determined by pathological examination in all cases. As spleen weight was a variable in the difficulty grading score, we confirmed the accuracy of the preoperative weight calculation by a formula previously described¹¹: length × width × thickness of the spleen in centimeters, multiplied by a correction factor (0.6).

Statistical analysis was performed on IBM^® SPSS Statistics Version 20; significant differences were determined by a P value <.05. To compare results between the score groups, χ², Fisher and Kruskal-Wallis tests were used. Association between calculated and observed spleen weight and association between difficulty grading score and the evaluated outcomes were assessed by the Spearman correlation. The discriminatory power of this score refers to its ability to differentiate patients who had intraoperative complications (need for conversion, major intraoperative bleeding—defined as blood loss >500 mL, and prolonged operative time—defined as 1.5 times or more over the median). This power was assessed using receiver operating characteristic (ROC) curve analysis.

Results

Baseline characteristics and outcomes of 153 patients subjected to consecutive LS are described on Table 2. Male-to-female ratio was 0.6, mean age of the patients was 49 years old (range 18–88), and median BMI was 27.6 kg/m². The most common indication for LS was idiopathic trombocitopenic purpura (ITP) in 83 (54.2%) patients, 27 (17.6%) patients presented malignant diseases, such as myeloproliferative disorders or lymphomas, and 43 (28.1%) patients presented other benign diseases, including spherocytosis, autoimmune hemolytic anemia (AIHA), Evans syndrome, splenic artery aneurism, splenic cyst/abscess, among others.

Table 2.

Preoperative, Intraoperative, and Postoperative Results According to Difficulty Grading Score

Parameter	Total	%	Low	%	Medium	%	High	%	P
n	153	100	62	40.5	48	31.4	43	28.1
Age	49 (18–88)		38 (18–59)		60 (18–85)		61 (33–88)		<.001
Gender									.001
Male	57	37.3	18	29.0	13	27.1	26	60.5
Female	96	62.7	44	71.0	35	72.9	17	39.5
BMI (kg/m²)	27.6 (17.6–56.9)		27.8 (17.6–56.9)		26.2 (20.0–43.3)		27.8 (18–43.4)		.543
Pathology									<.001
ITP	83	54.2	60	96.8	23	47.9	0	0
Other benign	43	28.1	0	0	17	35.4	26	60.5
Malignant	27	17.6	2	3.2	8	16.7	17	39.5
Spleen weight (g)	184 (9–1900)		137 (9–360)		177 (67–806)		480 (106–1900)		<.001
Hematocrit (%)	39 (20–52)		40 (27–48)		38 (26–48)		36 (20–52)		.077
Platelet count × 1000/μL	122 (6–547)		84 (10–447)		120 (6–547)		158 (10–353)		.001
Operative time (minutes)	113 (45–287)		100 (45–240)		122 (55–213)		125 (60–287)		.006
Conversion to open surgery	13	8.5	2	3.2	3	6.2	8	18.6	LvsM: .651MvsH: .071LvsH: .015
Bleeding >500 mL	8	5.2	1	1.6	4	8.3	3	7.0	LvsM: .165MvsH: 1.000LvsH: .303
Postoperative complications	20	13.1	2	3.2	8	16.7	10	28.1	.008
Clavien I/II	15	9.8
Clavien III	4	2.6
Clavien IV	2	1.3
Clavien V	1	0.7

Values: median (range).

BMI, body mass index; ITP, idiopathic trombocitopenic purpura; LvsH, low versus high; LvsM, low versus medium; MvsH, medium versus high.

The difficulty grading score classified 40.5% of patients as low difficulty, 31.4% as medium difficulty, and 28.1% as high difficulty. Patients' characteristics were compared according to the three score groups, as shown on Table 2. Patients on the high-difficulty group were older (P = .001), predominantly males (P = .001), and with malignant pathology (<.001); spleens were heavier (median 480 g, P < .001). Platelet count was inferior on the lower groups of difficulty (P = .001).

As previously stated, we confirmed the accuracy of the preoperative weight calculation, using a formula from the literature.¹¹ This formula was applied in 92 patients, for whom spleen diameters were recorded. The calculated spleen weights were in accordance with the final weight results obtained in 84.7% of cases.

Median operative time was 113 minutes (45–287) and we registered longer times for higher groups of difficulty (P = .006).

Conversion to open surgery occurred in 13 (8.50%) patients, mostly due to bleeding in 6 cases, adherences in 3 and pneumoperitoneum intolerance in 1 patient. Furthermore, conversion was associated with splenomegaly in 10 cases. In the conversion group, the median weight was 378 g.

Ten patients (6.8%) were subjected to previous abdominal surgeries for other indications, 5 in the upper and 5 in the lower abdomen. Conversion to open splenectomy occurred in 2 patients with previous surgery in the upper abdomen (cholecystectomy and esophagectomy); therefore, conversion rate for patients with previous abdominal surgery was 20%.

There were relevant intraoperative complications in 10 (6.8%) patients: 6 patients had major bleedings, 1 patient a colonic perforation associated with major bleeding, 1 patient gastric perforation, 1 patient diaphragmatic laceration, and 1 patient aorta perforation with the Veress needle. In these 10 patients, 6 (60%) needed conversion to open surgery.

Twenty (13.1%) patients had postoperative complications, including surgical site infections, pneumonia, urinary tract infection, pancreatitis, abdominal collections, and pancreatic fistula. According to Clavien-Dindo classification,¹² morbidity Grades I and II occurred in 15 patients (9.8%). Four patients had Grade IIIa and 2 patients grade IVb complications. In-hospital mortality occurred in 1 patient, a 94-year-old male, admitted for acute heart failure and a concurrent AIHA. His spleen weighted 164 g. During the postoperative period, he presented with nosocomial pneumonia and acute pulmonary edema and died on postoperative day 38.

Complications were present in 7 (53.8%) of the converted patients, were more frequent in medium/high-difficulty groups (P = .008), and were associated with higher conversion rates (P = .07).

We validated the difficulty grading score by comparing the three levels of difficulty with operative time, conversion to open surgery, operative bleeding, and postoperative complications. Spearman correlation showed a significant, although not strong, association between the score and all the evaluated outcomes, except for major bleeding (Table 3). The discriminatory power of this score was assessed using ROC curve analysis. The area under the ROC curve for the score was 0.674 (P = .005, 95% confidence interval 0.567–0.780) (Fig. 1).

FIG. 1.

ROC curve for discriminatory power of difficulty grading score. ROC, receiver operating characteristic.

Table 3.

Score Validation

Score	Low	Medium	High	P	Correlation ^a
Operative time (minutes)	100	122	125	.001	0.259
Conversion	3.2%	6.2%	18.6%	.008	0.214
Bleeding >500 mL	1.6%	8.3%	7.0%	.170	0.112
Complications	3.2%	16.7%	28.1%	.002	0.250

Spearman correlation.

Discussion

LS has been extensively validated as a safe, feasible, and effective technique in the treatment of various benign and malignant spleen-related diseases. The laparoscopic approach is the gold standard in the majority of indications for elective splenectomy, with numerous advantages, namely less complications and shorter recovery period. For the second-line treatment of ITP (the most common indication), LS is associated with short- and long-term response rates of 94% and 91%, respectively, with no mortality and 6% of major postoperative complications.¹³

Conversion of LS to open surgery ranges from 2.4% to 33.9%^2,8 and is associated with higher morbidity (14.6%) and hemorrhage (9.3%) than in patients initially subjected to open splenectomy.⁸ In our series, we observed a conversion rate of 8.5% that was associated with serious intraoperative complications, namely bleeding, arguably one of the main reasons for conversion.¹⁴ An effort should, therefore, be made to adequately select the patients for LS or for open splenectomy.

Several studies have analyzed preoperative factors to predict risk of conversion to open surgery and postoperative complications. Most series refer to spleen weight when making considerations about indication for laparoscopic approach. Matharoo et al.⁸ published results of 37,006 splenectomies, 6898 of which were laparoscopic, and successful in only 4938 patients. Morbidity (7.4% versus 10.4%) and mortality (1.3% versus 2.5%) were significantly lower in the laparoscopic group than in the open group. For heavier spleens, other strategies should be considered, such as open splenectomy, hand-assisted LS, or preoperative splenic artery embolization.⁴

The patient's age should not preclude LS,⁴ but, as expected, patient comorbidities are associated with longer hospital stay and a higher rate of postoperative complications.⁴ Several cardiovascular and respiratory diseases could be associated with pneumoperitoneum intolerance in laparoscopic surgery and raise the number of conversions, which only occurred in 1 of our patients.

Some authors^8,15,16 stated, in a series ranging from 676 to 1344 patients, that conditions such as AIHA and splenic lymphoma presented a higher risk for conversion and higher operative time, possibly due to morphological features.

Taking these factors into account, and in an attempt to improve risk stratification to LS, Rodriguez-Otero Luppi et al. developed a difficulty grading score.⁹ This score is simple, easily reproducible, and only incorporates preoperative factors. The authors created the score through the analysis of patients from their Surgical Department in Hospital de la Santa Creu I Sant Pau, Barcelona, and performed external validation on patients from Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City. We were encouraged by their results and considered that the difficulty grading score would be a useful tool in surgical decision.

The score showed a statistical correlation with the evaluated outcomes. The correlation with the score was 0.259 (P = .001) for operative time, 0.214 (P = .008) for risk of conversion, 0.250 (P = .002) for postoperative complications, and 0.112 (P = .170) for bleeding, which was the only factor not statistically significant in our results. The discriminatory power of this score was equivalent to the one obtained by the initial validation.

In our opinion, this score should not restrict patients from being selected to the laparoscopic approach, but instead should be used as a tool to stratify patients to high- or low-risk groups of surgery. This information is of great value, in the era of personalized medicine, as patients receive more precise and adequate information about the risks of their surgery before informed consent. As spleen weight, calculated by a formula using the diameters registered on CT scans, was one of the parameters of the validated score, we consider that all patients proposed to LS should have a CT scan requested, with spleen measurements. Also, with the use of this score, the surgical and anesthetic teams are able to anticipate and prepare for greater difficulties during the perioperative period. For high-risk patients, the surgical team should be more experienced and prepared for a possible conversion or hand-assisted approach, with the required material readily available.

Several authors^17,18 suggest that 20 LSs might be regarded as a threshold to acquire sufficient experience. After performing 20 LSs, a surgeon can observe significant reduction in the operative time, intraoperative blood loss, and hospital costs, but no difference in the rate of conversion. Matharoo et al.⁸ also propose that to decrease conversion rate, LS should be performed by fellowship-trained or advanced laparoscopic surgeons.

There were some limitations to this study regarding its retrospective nature. Other limitations included being performed by a single surgical team and only routinely assessed variables were analyzed.

In conclusion, according to these results, patient and disease features, incorporated in the presented grading score, can reliably predict the risk of conversion and of perioperative complications. This simple and reproducible score improves risk stratification for LS, so the patient and the surgical team must be aware of the expected difficulties. High-risk patients should be managed by experienced teams.

Footnotes

Disclosure Statement

No competing financial interests exist.

References

Delaitre

, Maignien

, Icard

. Laparoscopic splenectomy. Br J Surg, 1992; 79:1334.

Winslow

, Brunt

. Perioperative outcomes of laparoscopic versus open splenectomy: A meta-analysis with an emphasis on complications. Surgery, 2003; 134:647–653.

Knauer

, Ailawadi

, Yahanda

, Obermeyer

, Millie

, Ojeda

, Mulholland

, Colletti

, Sweeney

. 101 Laparoscopic splenectomies for the treatment of benign and malignant hematologic disorders. Am J Surg, 2003; 186:500–504.

Habermalz

, Sauerland

, Decker

, Delaitre

, Gigot

, Leandros

, Lechner

, Rhodes

, Silecchia

, Szold

, Targarona

, Torelli

, Neugebauer

. Laparoscopic splenectomy: The clinical practice guidelines of the European Association for Endoscopic Surgery (EAES). Surg Endosc, 2008; 22:821–848.

Poulin

, Thibault

. Laparoscopic splenectomy for massive splenomegaly: Operative technique and case report. Can J Surg, 1995; 38:69–72.

Al-Mulhim

. Laparoscopic splenectomy for massive splenomegaly in benign hematological diseases. Surg Endosc, 2012; 26:3186–3189.

Qian

, He

, Hua

, Gong

, Lin

, Song

. Hand-assisted versus conventional laparoscopic splenectomy: A systematic review and meta-analysis. ANZ J Surg, 2014; 84:915–920.

Matharoo

, Afthinos

, Gibbs

. Trends in splenectomy: Where does laparoscopy stand?. JSLS, 2014; 18.

Rodriguez-Otero Luppi

, Targarona Soler

, Balague Ponz

, Pantoja Millán

, Turrado Rodriguez

, Pallares Segura

, Bollo Rodriguez

, Trias Folch

. Clinical, anatomical, and pathological grading score to predict technical difficulty in laparoscopic splenectomy for non-traumatic diseases. World J Surg, 2016; 41:439–448.

10.

Nyilas

, Paszt

, Simonka

, Abraham

, Borda

, Man

, Lazar

. Laparoscopic splenectomy is a safe method in cases of extremely large spleens. J Laparoendosc Adv Surg Tech A, 2015; 25:212–216.

11.

Kluhs

, Teichgraber

, Schneider

, Ludwig

, Dorken

, Benter

. Accuracy of the sonographic determination of the splenic weight in comparison with the weight at autopsy [in German]. Rofo, 2003; 175:532–535.

12.

Dindo

, Demartines

, Clavien

. Classification of surgical complications: A new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg, 2004; 240:205–213.

13.

Costa-Pinho

, Fernandes

, Bessa-Melo

, Aral

, Graça

, Costa-Maia

. Laparoscopic splenectomy for second-line treatment of immune thrombocytopenia—Analysis of 53 patients and current perspectives. Eur Surg, 2016; 48:1–5.

14.

Fan

, Liu

, Wang

, Li

, Kang

, Zhao

, Zhang

. Study on the efficacies of splenic pedicle transection by using manual manipulation and Endo-GIA procedure for laparoscopic splenectomy. Int J Clin Exp Med, 2015; 8:19430–19435.

15.

Casaccia

, Torelli

, Pasa

, Sormani

, Rossi

, Centers

. Putative predictive parameters for the outcome of laparoscopic splenectomy: A multicenter analysis performed on the Italian Registry of Laparoscopic Surgery of the Spleen. Ann Surg, 2010; 251:287–291.

16.

Bagrodia

, Button

, Spanheimer

, Belding-Schmitt

, Rosenstein

, Mezhir

. Morbidity and mortality following elective splenectomy for benign and malignant hematologic conditions: Analysis of the American College of Surgeons National Surgical Quality Improvement Program data. JAMA Surg, 2014; 149:1022–1029.

17.

Cusick

, Waldhausen

. The learning curve associated with pediatric laparoscopic splenectomy. Am J Surg, 2001; 181:393–397.

18.

Peters

Jr , Camacho

, Ojeda

, Reichenbach

, Knauer

, Yahanda

, Cooper

, Sweeney

. Defining the learning curve for laparoscopic splenectomy for immune thrombocytopenia purpura. Am J Surg, 2014; 188:522–525.