Abstract
Background
Resection is rarely indicated in giant hepatic hemangiomas (HHs) that are symptomatic. Enucleation (EN), compared with anatomical resection (AR), is considered the better technique to resect them as EN has been reported to have lower morbidity while conserving the normal liver tissue. But no study has yet clearly established the superiority of EN over AR. In addition, the independent predictors of postoperative morbidity have not been established.
Methods
All consecutive patients operated for HH at two specialized hepatobiliary centers were reviewed. Patient demographics, operative variables, and postoperative outcomes were analyzed and compared between two techniques. Postoperative complications were graded as per Clavien-Dindo classification of surgical complications. The aims of this study were to compare two techniques of HH resection with respect to postoperative outcomes and to identify the risk factors for 90-day major postoperative morbidity and mortality.
Results
A total of 64 patients, including 41 who underwent AR, 22 who underwent EN, and 1 who underwent liver transplantation, were operated for hemangiomas during the study period. Ten patients (9 who were operated for hemangiomas of size ≤4 cm and 1 who underwent transplantation) were excluded. Fifty-four patients, the majority being women (85%), with a median age of 48 years, were operated for giant HH. These patients were classified into two groups based on the technique of resection, namely, EN (22 patients) and AR (32 patients). Both groups were comparable in all aspects except that the number of liver segments resected was significantly more with AR. Postoperative outcomes were similar in both groups. Independent predictors of 90-day major complications including mortality were the use of total vascular exclusion (relative risk [RR]: 2.3, p = 0.028) and duration of surgery >4.5 h (RR: 2.3, p = 0.025).
Conclusion
Both techniques yield similar results with respect to 90-day postoperative morbidity and mortality. The choice of technique should be based on the location of tumor and simplicity of liver resection.
Keywords: benign liver tumor, cavernous hemangioma, hepatectomy, vascular tumor of the liver
Abbreviations: AR, anatomical resection; EN, enucleation; HH, giant hepatic hemangioma; POD, postoperative day; RBC, red blood cell; RR, relative risk; TVE, total vascular exclusion
Cavernous hemangiomas are vascular congenital malformations that grow slowly by ectasia and are the most common benign liver tumors, being found in 5–7% % of autopsies.1,2 They are usually diagnosed incidentally on imaging and very rarely require surgery as they are very small in size. Resection is proposed when the patient becomes symptomatic, when the patient develops complications owing to the size of hemangiomas, and/or when there is a diagnostic dilemma.1, 2, 3, 4, 5, 6, 7, 8 Hemangiomas >4 cm are called giant hepatic hemangiomas (HHs).8 Various aspects of surgical management from the indication for surgery to the technique of resection remain controversial.4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 Enucleation (EN) described by Baer et al9 in 1992 is considered the standard technique of resection as it conserves normal liver tissue and has been reported to have lower postoperative complication rates than anatomical resection (AR).9,17, 18, 19, 20 However, no study has clearly demonstrated the advantages of EN over AR. In this study, we aim to compare these two techniques with respect to postoperative outcomes and determine independent risk factors for 90-day major postoperative complication and mortality.
Patients and methods
All consecutive patients who underwent resection for HH at two specialized hepatobiliary centers over two decades were included in the study. Patients who were operated for either hemangiomas ≤4 cm or transplanted were excluded from the study. Patient demographics including clinical presentation, indication for surgery, operative variables including technique of resection, intraoperative blood loss, transfusion, and postoperative outcomes including length of hospitalization, complications, and mortality were reviewed retrospectively and analyzed. The primary end point of the study was the 90-day major postoperative complication and mortality for the two techniques of HH resection, namely, EN and AR.
Operative technique
Laparotomy was performed through right subcostal incision with an extension to the left or Makuuchi incision. Intraoperative ultrasonography was routinely carried out to define the lesion and its relation to intrahepatic portal pedicles and hepatic veins. After mobilization of the liver, the hepatic pedicles and hepatic veins were looped and controlled using a vessel loop for vascular control during resection. The aim of the surgery was to resect the largest and symptomatic lesion while conserving normal liver tissue. The senior surgeon decided the choice of operative technique, EN or AR, based on the size and location of HH, its relation to major vascular structures, and the ease of resection. For example, a tumor located in the peripheral segments of the liver was enucleated, and a centrally located or a large tumor replacing the entire lobe underwent AR. Inflow control such as the intermittent Pringle maneuver or selective hepatic arterial clamping was usually performed to reduce the size of HH and to reduce the risk of bleeding. Total vascular exclusion (TVE) and extracorporeal circulation was performed for very large tumors with a very high risk of hemorrhage. EN was performed by developing the plane between the tumor and normal hepatic parenchyma using CUSA™ after incising Glisson's capsule by cautery. The tumor was progressively separated from the parenchyma using an ultrasonic dissector. Small bridging vessels and bile ducts were either ligated using absorbable sutures or cauterized. AR was performed using standard liver resection techniques. Once the extent of resection was decided, the Glissonean pedicle to the part of the liver to be resected was ligated using ligatures, and transection was performed using CUSA™ and Ligasure™ along the line of demarcation. Hepatic veins were also clamped at the point of entry into the inferior vena cava when there was significant venous bleeding during transection. After the hemangioma was removed, hemostasis and biliostasis was ensured on the cut surface by bipolar cautery, suture ligatures, and/or argon beam coagulation.
Definitions
-
1.
Giant HHs were those with a diameter of >4 cm.8
-
2.
Major hepatectomy was defined as the resection of >2 Couinaud segments of the liver.
-
3.
High-risk hepatectomy was one in which the cut surface exposed major Glisson's sheath and/or included the hepatic hilum, for example, central hepatectomy and caudate lobectomy.
-
4.
Clavien-Dindo classification was used to grade the complications in the 90-day postoperative period, and complications with grade >2 were defined as major complications.21
Statistical analysis
Discrete variables are shown as counts (percentage), and parametric variables are shown as medians (range). Fischer's exact test and the Mann-Whitney U test were used to compare the two groups: EN and AR. Univariate and multivariate logistic regressions were used for determining the risk factors associated with 90-day major complications and mortality. Differences were considered statistically significant when the P value <0.05. SPSS software, version 20.0, was used for all statistical calculations (SPSS Inc, Chicago, IL).
Results
A total of 64 patients, including 41 who underwent AR, 22 who underwent EN, and 1 who underwent liver transplantation, were operated for hemangiomas during the study period. Nine patients who underwent resection for hemangiomas ≤4 cm and one who underwent transplantation were excluded from further analysis. Fifty-four patients, 46 women (85%) with a median age of 48 years, operated for HH were included in the further analysis. All patients except one were diagnosed as having HH preoperatively by abdominal ultrasonography, computed tomography, or magnetic resonance imaging. Table 1 summarizes patient demographics, HH characteristics, various operative parameters, and postoperative outcomes. Abdominal pain (83%) and presence of complications (33%) were the most common indications for surgery. The median number and size of HHs resected was one and 14 cm, respectively. The median (range) number of segments resected was two (1–5). EN was performed in 22 patients (41%). Thirty-two patients (59%) underwent AR including left lateral segmentectomy (12 patients), right hepatectomy (9), right extended hepatectomy (3), left extended hepatectomy (3), left hepatectomy (2), right bisegmentectomy (2), and unisegmentectomy (1). High-risk hepatectomy was performed in 8 patients (15%). The majority of patients (87%) had vascular control during transection. The median (range) blood loss during surgery was 375 (100–11300) mL, and 15 patients (28%) required intraoperative packed red blood cell (RBC) transfusion, with seven of them (13%) requiring ≥4 units of transfusion.
Table 1.
Patient Demographics, Intraoperative Parameters, and Postoperative Outcomes.
| Variables | Study population, N = 54 (100%)a |
|---|---|
| Age (years), median (range) | 48 (31–70) |
| Gender, female | 46 (85%) |
| Hemangioma characteristics | |
| Size (cm), median (range) | 14 (5–32) |
| Number, median (range) | 1 (1–10) |
| Liver resection | |
| Anatomical resection | 32 (59%) |
| Enucleation | 22 (41%) |
| High-risk hepatectomy | 8 (24%) |
| Couinaud segments resected, median (range) | 2 (1–5) |
| Intraoperative parameters | |
| Inflow control | |
| None | 7 (13%) |
| Intermittent Pringle maneuver | 29 (54%) |
| Selective hepatic artery clamping | 6 (11%) |
| Total hepatic vascular exclusion | 12 (22%) |
| Duration of inflow control (min), median (range) | 40 (12–130) |
| Blood loss (mL), median (range) | 375 (100–11300) |
| Packed red blood cell transfusion of ≥4 units | 7 (13%) |
| Duration of surgery (min), median (range) | 222 (90–600) |
| Postoperative outcomes | |
| Duration of hospital stay (days), median (range) | 9.5 (4–213) |
| 90-day Clavien-Dindo classification of complications | 21 (39%) |
| Grade 1–2 (minor) | 15 (28%) |
| Grade 3–5 (major) | 6 (11%) |
| 90-day mortality | 2 (4%) |
Data are expressed as counts (percentage) unless otherwise stated.
The postoperative complications were observed in 21 patients (39%), with the majority of them being minor (28%). Major complications were observed only in 6 patients (11%), including hemorrhage (2 patients), biliary fistula (2), pulmonary embolism (1), necrotizing enterocolitis (1), and massive pleural effusion (1). No mortality was observed in the EN group. Two patients (4%) who underwent AR died during the 90-day postoperative period. The first patient underwent right hepatectomy for a tumor of 27-cm diameter occupying the entire right lobe under TVE for 58 min with 3-L blood loss, requiring 8 units of RBC transfusion. She had massive pulmonary embolism with intracardiac emboli treated with thrombolysis, pulmonary thromboembolectomy, and anticoagulation. Subsequently, she developed intra-abdominal bleeding that necessitated distal pancreatectomy with splenectomy and packing. She died on postoperative day (POD) 5. The other patient had a 13-cm lesion in the left lobe that extended to segments 1 and 8. She underwent left hepatectomy extended to segments 1 and 8 under TVE without significant intraoperative events. Postoperatively, she developed necrotizing enterocolitis and underwent right hemicolectomy with distal small bowel resection and ileostomy. Subsequently, she developed ischemic sigmoid colon perforation requiring relaparotomy and colon resection. She died on POD 213 from sepsis.
Two groups were similar in all aspects except in the number of Couinaud liver segments resected, as shown in Table 2. In the AR group, the number of segments resected was significantly more than that in the EN group. The incidence of complications and mortality during the 90-day postoperative period were not significantly different between the two groups. Independent risk factors for 90-day major complications after HH resection by both techniques were the use of TVE of the liver and operation time >4.5 h (Tables 3). Although operative time >4.5 h, use of TVE, and major hepatectomy were identified as risk factors in univariate analysis, no independent risk factor could be identified for 90-day mortality in multivariate analysis.
Table 2.
Comparison of the Two Groups: Enucleation Versus Anatomical Resection.
| Variables | Enucleation (N = 22) | Anatomical resection (N = 32) | P value |
|---|---|---|---|
| Age (years), median (range) | 51 (31–70) | 47 (33–63) | |
| Gender, female | 18 | 28 | |
| Size (cm), median (range) | 14 (7.3–30) | 14 (5–32) | |
| Number of hemangiomas, median (range) | 1.5 (1–5) | 1 (1–10) | |
| Duration of inflow control (min), median (range) | 40 (13–120) | 39 (12–130) | |
Type of inflow control
|
|||
| 1 | 6 | ||
| 4 | 2 | ||
| 12 | 17 | ||
| 5 | 7 | ||
| Couinaud segments resected, median (range) | 2 (1–5) | 2.5 (1–5) | 0.01 |
| High-risk hepatectomy | 4 | 4 | |
| Operative time (min), median (range) | 225 (135–600) | 203 (90–473) | |
| Blood loss (L), median (range) | 0.4 (0.1–11.3) | 0.4 (0.1–7.1) | |
| Hospital stay (days), median (range) | 9 (4–29) | 10 (5–213) | |
90-day morbidity
|
|||
| 6 | 9 | ||
| 3 | 3 | ||
| 90-day mortality | 0 | 2 |
Data are expressed as counts unless otherwise stated; counts were compared using Fischer's exact test; means were compared using the Mann-Whitney U test.
Table 3.
Risks for 90-day Major Complications.
| Variables | Univariate analysis P value | Multivariate logistic regression | ||||
|---|---|---|---|---|---|---|
| Beta | Relative risk | P value | 95% confidence interval | |||
| Lower | Upper | |||||
| Age >60 years | 0.649 | |||||
| Female gender | 0.892 | |||||
| Size of the hemangioma >20 cm | 0.018 | 0.119 | ||||
| Number >1 | 0.772 | |||||
| Anatomical resection | 0.662 | |||||
| High-risk hepatectomy | 0.023 | 0.488 | ||||
| >3 Couinaud segments resected | 0.014 | 0.303 | ||||
| Pringle maneuver | 0.086 | |||||
| Total hepatic vascular exclusion | 0.004 | 0.305 | 2.262 | 0.028 | 0.026 | 0.436 |
| Duration of inflow control >60 min | 0.091 | |||||
| Duration of surgery >4.5 h | 0.008 | 0.335 | 2.307 | 0.025 | 0.031 | 0.450 |
| Blood loss >1 L | 0.018 | |||||
| Packed red blood cell transfusion of ≥4 units | 0.013 | 0.555 | ||||
All variables in Table 1 were analyzed, and those with p <0.05 in univariate analysis were entered in the multivariate analysis.
Discussion
Statementofprinciplefindings
Both techniques of resection of giant HHs yield similar postoperative results. Although two patients who underwent AR died and none died after EN, the difference was not statistically significant. Independent predictors of 90-day major complications including mortality were use of TVE and operative time >4.5 h. The choice of technique should be based on the size and location of HH, its relation to major vascular pedicles, and simplicity of liver resection while minimizing blood loss, reducing the need for TVE and avoiding prolonged complex resections. Liberal utilization of different vascular control techniques ensures safe surgery by reducing intraoperative blood loss and thereby RBC transfusion requirements.
Comparisonwiththepublishedliterature
Cavernous hemangioma, the most common liver tumor, presents between the 3rd and 5th decades of life and occurs 4–6 times more commonly in women.1, 2, 3 An association between female hormones and liver hemangiomas has been proposed as they occur more often in multiparous women, and women who present at an younger age tend to have large tumors.3 In our study, a similar incidence was observed, with 87% of the patients between 30 and 50 years of age with female predominance (85%). It is generally agreed that asymptomatic lesions are not treated surgically owing to their benign course.2, 3, 4, 5 The size of HH in our patients ranged from 5 to 32 cm, comparable with the literature, wherein lesions as large as 28 cm were reported.3,4,11,12 Surgery is a proven curative method of treatment for symptomatic or complicated HH.5, 6, 7 The principle indications for surgery that have been described are abdominal pain, increasing size, presence of complications, and difficulty in ruling out malignancy.10,13 In the present series, the majority of patients who were operated for abdominal pain associated with documented increase in size, and the next common indication was the presence of complications, consistent with the literature.
Various types of surgical treatment for HH described in the literature include wedge resection, EN, AR along the liver planes, and liver transplantation.4,9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,22, 23, 24 Partial hepatectomy/wedge resection is usually performed for small lesions on the surface of the liver or pedunculated tumors. EN is considered the procedure of choice for HH as it has been reported to have lower incidence of complications than AT—less intraoperative blood loss, lower transfusion requirement, and no need for TVE.17, 18, 19, 20 However, in the present study, we could not observe any significant difference in blood loss, RBC transfusion, and need for different types of vascular control between the two groups. Notably, no postoperative liver failure was observed in both groups. Probable explanation could be that when surgical technique was adapted to the size and location of HH along with appropriate and liberal usage of various vascular controls, minimizing intraoperative blood loss resulted in achieving comparable perioperative outcomes in each technique in expert hands. Moreover, in the present study, AR was chosen when the entire lobe or segment was replaced by tumor, and thus, the normal liver was preserved. Further transection in both AR and EN was performed in well-defined dissection planes using CUSA™ and Ligasure™. A study by Hamaloglu et al17 has highlighted the fact that tumor localization and number were the key factors that determine the choice of resection technique. AR rather than EN was performed in difficult situations such as when the tumor was larger, was deep in liver parenchyma, was in posterior segments, was replacing the entire lobe or multiple in number, and would explain poorer outcomes in AR.10,18,19, Similarly, Gedaly et al19 concluded when safe EN could not be performed owing to tumor localization, anatomic resection should be performed expeditiously. Therefore, when surgical technique is adapted to the clinical situation, similar perioperative outcomes could be expected in EN and AR with respect to blood loss, transfusion requirement, hospital stay, and postoperative morbi-mortality.
The unique feature of the present study is that the surgical resection of HH was performed in specialized liver units using standardized techniques in a significant number of patients over a prolonged period. To the best of our knowledge, this is the first study to identify specifically independent predictors of major postoperative complications after HH resection. The major limitation is the inherent selection bias in the choice of surgical technique as the nonhomogenous population from 2 different centers with single and multiple, and/or right and left lobe, hemangiomas were included in the present study. It would be ideal to match for comparison single tumors based on their size and location, but it would be statistically impossible because of small numbers. These facts should be considered while interpreting the predictive values of this study.
Both the techniques, namely, EN and AR, could be used to resect HHs, with similar postoperative outcomes. Instead of being dogmatic, the choice of technique should be based on the tumor size, its location, and simplicity of resection. EN could be used for peripheral lesions, whereas AR could be performed for centrally located tumors and/or those complete replacing a liver segment/lobe. Prolonged complex hepatic resections necessitating the use of TVE should be avoided. Probably, the indication of HH resection should be proposed early especially in growing asymptomatic hemangiomas to avoid complex resection later.
CRediT authorship contribution statement
Muthukumarassamy Rajakannu: Data collection, Formal analysis, and manuscript preparation. Gerard Pascal: Data collection. Denis Castaing: Patient recruitment and manuscript preparation. Eric Vibert: Patient recruitment and manuscript preparation. Christian Ducerf: Patient recruitment and manuscript preparation. Jean-Yves Mabrut: Patient recruitment and manuscript preparation. Jacques Baulieux: Study design and final approval of manuscript. René Adam: Study design and final approval of manuscript.
Conflicts of interest
The authors have none to declare.
Financial support
No funding support to declare.
References
- 1.Ishak K.G., Rabin L. Benign tumors of the liver. Med Clin. 1975;59:995–1013. doi: 10.1016/s0025-7125(16)31998-8. [DOI] [PubMed] [Google Scholar]
- 2.Schwartz S.I. Cavernous hemangioma of the liver: a single institution report of 16 resections. Ann Surg. 1987;205:456–465. doi: 10.1097/00000658-198705000-00003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Di Biscegli A.M., Befeler A.S. Tumors and cysts of the liver. In: Feldman M., Friedman L.S., Brandt L.J., editors. Gastrointestinal and Liver Disease. Saunders Elseiver; Philadelphia: 2010. pp. 1586–1587. [Google Scholar]
- 4.Farges O., Daradkeh S., Bismuth H. Cavernous hemangiomas of the liver: are there any indications for resection? World J Surg. 1995;19:19–24. doi: 10.1007/BF00316974. [DOI] [PubMed] [Google Scholar]
- 5.Hoekstra L.T., Bieze M., Erdogan D., Roelofs J.J., Beuers U.H., van Gulik T.M. Management of giant liver hemangiomas: an update. Expet Rev Gastroenterol Hepatol. 2013;7:263–268. doi: 10.1586/egh.13.10. [DOI] [PubMed] [Google Scholar]
- 6.Toro A., Mahfouz A.E., Ardiri A. What is changing in indications and treatment of hepatic hemangiomas. A review. Ann Hepatol. 2014;13:327–339. [PubMed] [Google Scholar]
- 7.Schneldorfer T., Ware A.L., Smoot R. Management of giant hemangiomas of the liver: resection versus observation. J Am Coll Surg. 2010;211:724–730. doi: 10.1016/j.jamcollsurg.2010.08.006. [DOI] [PubMed] [Google Scholar]
- 8.Adam Y.G., Huvos A.G., Fortner J.G. Giant hemangiomas of the liver. Ann Surg. 1970;172:239–245. doi: 10.1097/00000658-197008000-00010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Baer H.U., Dennison A.R., Mouton W., Stain S.C., Zimmermann A., Blumagart L.H. Enucleation of giant hemangiomas of the liver. Technical and pathologic aspects of a neglected procedure. Ann Surg. 1992;216:673–676. doi: 10.1097/00000658-199212000-00009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yoon S.S., Charny Ck, Fong Y. Diagnosis, management, and outcomes in 115 patients with hepatic hemangioma. J Am Coll Surg. 2003;197:392–402. doi: 10.1016/S1072-7515(03)00420-4. [DOI] [PubMed] [Google Scholar]
- 11.Erdogan D., Busch O., van Selden O. Management of liver hemangiomas according to size and symptoms. J Gastroenterol Hepatol. 2007;22:1953–1958. doi: 10.1111/j.1440-1746.2006.04794.x. [DOI] [PubMed] [Google Scholar]
- 12.Duxbury M.S., Garden O.J. Giant haemangioma of the liver: observation or resection? Dig Surg. 2010;27:7–11. doi: 10.1159/000268108. [DOI] [PubMed] [Google Scholar]
- 13.Fergusson J., Tait N. Resection for hepatic hemangiomas. J Gastroenterol Hepatol. 2007;22:1699–1705. doi: 10.1111/j.1440-1746.2007.05089.x. [DOI] [PubMed] [Google Scholar]
- 14.Jiang G., Chen Z., Prasoon P. Surgical management for giant liver hemangiomas greater than 20 cm in size. Gut Liver. 2011;5:228–233. doi: 10.5009/gnl.2011.5.2.228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Alper A., Ariogui O., Emre A. Treatment of liver hemangiomas by enucleation. Arch Surg. 1998;123:660–661. doi: 10.1001/archsurg.1988.01400290146027. [DOI] [PubMed] [Google Scholar]
- 16.Giuliante F., Ardito F., Vellone M. Reappraisal of surgical indications and approach for liver hemangioma: single-center experience on 74 patient. Am J Surg. 2011;201:741–748. doi: 10.1016/j.amjsurg.2010.03.007. [DOI] [PubMed] [Google Scholar]
- 17.Hamaloglu E., Altun H., Ozdemir A. Giant liver hemangioma: therapy by enucleation or liver resection. World J Surg. 2005;29:890–893. doi: 10.1007/s00268-005-7661-z. [DOI] [PubMed] [Google Scholar]
- 18.Lerner S.M., Hiatt J.R., Salamandra J. Giant cavernous liver hemangiomas: effect of operative approach on outcome. Arch Surg. 2004;139:818–823. doi: 10.1001/archsurg.139.8.818. [DOI] [PubMed] [Google Scholar]
- 19.Gedaly R., Pomposelli J., Pomfret E.A. Cavernous hemangioma of the liver anatomic resection vs enucleation. Arch Surg. 1999;134:407–411. doi: 10.1001/archsurg.134.4.407. [DOI] [PubMed] [Google Scholar]
- 20.Singh R.K., Kapoor S., Sahni P., Chattopadhyay T.K. Giant haemangioma of the liver: is enucleation better than resection? Ann R Coll Surg Engl. 2007;89:490–493. doi: 10.1308/003588407X202038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dindo D., Demartines N., Clavien P.A. 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. doi: 10.1097/01.sla.0000133083.54934.ae. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Miura J.T., Amini A., Schmocker R. Surgical management of hepatic hemangiomas: a multi-institutional experience. HPB. 2014;16:924–928. doi: 10.1111/hpb.12291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Choi J., Lee Y.J., Hwang D.W., Chon S.H., Nagpal A., Park K.M. Surgical treatment of giant hepatic hemangiomas: technical point of view. Am Surg. 2011;77:48–54. [PubMed] [Google Scholar]
- 24.Vagefi P.A., Klein I., Gelb B. Emergent orthotopic liver transplantation for hemorrhage from a giant cavernous heptic hemangioma: case report and review. J Gastrointest Surg. 2011;15:209–214. doi: 10.1007/s11605-010-1248-1. [DOI] [PMC free article] [PubMed] [Google Scholar]