SURGICAL
. and Other Interventional Techniques
Risk factors affecting operative approach, conversion, and morbidity for adrenalectomy: a single-institution series of 402 patients
James G. Bittner IV . Victoria M. Gershuni . Brent D. Matthews . Jeffrey F. Moley . L. Michael Brunt
Received: 17 July 2012/ Accepted: 31 December 2012 @ Springer Science+Business Media New York 2013
Abstract
Background Risk factors for selecting patients for open adrenalectomy (OA) and for conversion are limited in most series. This study aimed to investigate variables that are important in selecting patients for OA, predict risk of conversion from laparoscopic adrenalectomy (LA), and impact 30-day outcomes of OA and LA.
Methods A retrospective cohort study of prospectively collected data was conducted. Patients (≥16 years old) who underwent adrenalectomy in the Division of General Sur- gery at Barnes-Jewish Hospital (1993-2010) were grouped by operative approach (LA vs. OA) and compared using nonparametric tests and regression analyses (x < 0.05).
Results In total, 402 patients underwent 422 adrenalec- tomies. Compared to LA patients, those in the OA group were older (p = 0.02), had higher ASA scores (p = 0.04), larger tumor size (p < 0.01), and fewer functioning lesions (p < 0.01). OA patients more often required concurrent procedures (p < 0.01), had a longer operative time (p = 0.04), more intraoperative complications (p = 0.02), higher estimated blood loss (EBL), and larger transfusion requirement. Preoperative factors that predicted selection for OA were higher patient age (p = 0.01), higher ASA score (p = 0.03), larger tumor size (p < 0.01), nonfunctioning
Presented at the SAGES 2012 Annual Meeting, San Diego, CA, March 7-10, 2012.
J. G. Bittner IV () . V. M. Gershuni .
B. D. Matthews . J. F. Moley . L. M. Brunt Section of Minimally Invasive Surgery, Department of Surgery, Washington University School of Medicine, 660 S. Euclid Avenue, Campus Box 8109, St. Louis, MO 63111, USA e-mail: jbittner4@gmail.com
L. M. Brunt
e-mail: bruntm@wustl.edu
lesion (p < 0.01), diagnosis of adrenocortical carcinoma (p < 0.01), and the need for concomitant procedures (p < 0.01). Conversion to open or hand-assisted approach occurred in 6.2 % of LA patients. Preoperative risks for conversion included large tumor size (>8 cm) and need for concomitant procedures (p < 0.01). Multivariate analysis revealed that large indeterminate adrenal mass, adrenocor- tical carcinoma, tumor size (>6 cm), an open operation, conversion, concomitant procedures, operative time >180 min, and EBL >200 mL were predictors of 30-day morbidity.
Conclusions Adrenal tumor size and need for concurrent procedures significantly impact the selection of patients for OA, the likelihood of conversion, and perioperative mor- bidity. These metrics should be considered when assessing operative approach and risks for adrenalectomy.
Keywords Adrenalectomy . Open adrenalectomy . Laparoscopic adrenalectomy · Conversion . Risk factors
Successful laparoscopic adrenalectomy (LA) was first reported in 1992 [1]. As the number of incidental adrenal lesions and clinical experience with laparoscopy have grown, the frequency of and indications for LA have chan- ged. Over time more patients with increasingly larger benign and indeterminate adrenal lesions were selected for LA. A number of case series, cohort studies, and database reviews have demonstrated less pain, shorter hospital stay, and fewer postoperative complications following LA [2-4]. However, the majority of adrenalectomy procedures during the early and mid-1990s were still performed using an open approach.
From 1998 through 2006, the number of adrenal resec- tions in the United States increased significantly from 3,241 to 5,019 cases according to a recent review of the
Nationwide Inpatient Sample [5]. Over the same period, a majority (83 %) of adrenalectomies were performed using an open approach, with most cases completed at non- teaching, low-volume hospitals. In this nationwide sample, open adrenalectomy (OA) was associated with a longer hospital stay and higher in-hospital morbidity and mortality compared to LA [5]. Additionally, a review of the Veterans Affairs National Surgical Quality Improvement Program database revealed that patients who underwent OA (n = 311) compared to LA (n = 358) were more likely to be older, have a higher American Society of Anesthesiol- ogists (ASA) classification, and harbor malignancy. Even after adjusting for confounders, OA was associated with increased operative time, transfusion requirement, hospital length of stay, and 30-day morbidity [6].
During the last decade LA replaced OA as the preferred method for removal of most adrenal tumors. This paradigm shift has occurred as a result of multiple factors, including greater surgeon experience with advanced laparoscopy, improved technology, and better short-term patient out- comes. A growing body of literature suggests that experi- enced surgeons are performing LA for a wider range of indications than ever before [3, 4, 7]. However, few studies clearly characterized the preoperative risk factors that may influence the choice of operative approach and, thereby, affect outcomes after adrenalectomy. To that end, we sought to investigate predictive factors for selecting patients for OA, the rate and risk for conversion to an open approach, and resultant 30-day morbidity. The study was conducted at a large, urban university referral center whose surgeons have substantial experience with adrenalectomy.
Materials and methods
A retrospective cohort study using a prospectively col- lected database was conducted under Institutional Review Board approval. Patients were included if they were at least 16 years old and underwent adrenalectomy for any indi- cation by surgeons in the Sections of Minimally Invasive Surgery, Endocrine and Oncologic Surgery, and Hepatob- iliary-Pancreatic and Gastrointestinal Surgery at Wash- ington University School of Medicine/Barnes-Jewish Hospital from January 1, 1993 through December 31, 2010. Also included were patients who underwent adrenalectomy as the primary procedure and had concomitant procedures at the time of adrenalectomy.
Patient cohorts were grouped by operative approach (laparoscopic vs. open) and perioperative variables were analyzed using an intention-to-treat strategy. Preoperative variables included patient age, gender, body mass index (BMI), ASA score, history of prior abdominal operations, side of the lesion, radiographic tumor size (maximum
diameter), tumor functioning status as determined by bio- chemical evaluation, indication for operation as docu- mented in the medical record, and frequency and type of concomitant procedures, which were defined as resection of one or more hollow viscous or solid organ(s) in addition to adrenalectomy. Of note, cholecystectomy was analyzed as an independent concomitant procedure and was not considered a multiorgan resection.
Intraoperative variables were total operative time, esti- mated blood loss (EBL), transfusion requirement, use of ultrasound, drain placement, complications, and indication for conversion when applicable. Operative time was defined as the period from skin incision to application of wound dressings. EBL constituted all blood lost during the entire operative period, including adrenalectomy and con- comitant procedures when applicable. Review of operative dictations and anesthesia records for patients who under- went concomitant procedures did not produce sufficient clarity to report EBL specific to adrenalectomy. Conver- sion was defined as the transition from a laparoscopic to an open approach or from a laparoscopic to a hand-assisted technique. Intraoperative complications were categorized by the Accordion Severity Grading System [8]. Immediate postoperative and 30-day outcome variables included specimen size and histology by gross and surgical pathol- ogy examination, respectively, the need for blood transfu- sion, hospital length of stay (LOS), complication frequency and severity classified by the Accordion Severity Grading System, and overall morbidity and mortality.
With regard to surgical technique, laparoscopic adre- nalectomy was performed preferentially using the lateral transabdominal approach as previously described [1]. A small number of cases in the last 2 years of the study were performed via a retroperitoneal endoscopic approach according to the method described by Walz et al. [9]. The decisions regarding need for and types of concomitant procedures were based on patient factors, attendant pathology, intraoperative findings, and surgeon judgment.
Statistical analyses were performed using IBM SPSS ver. 20.0 for Windows (IBM Corporation, Inc., Armonk, NY). Since not all of the data were normally distributed, demo- graphic, perioperative, and outcome variables were com- pared between OA and LA groups using nonparametric tests. Patients who underwent OA were compared with those who underwent LA using x2 and Fisher’s exact tests for cate- gorical variables and the Kruskal-Wallis or Mann-Whitney U test for continuous variables as appropriate. Age was categorized by 10-year periods, radiographic tumor size by 1-cm intervals, operative time by 60-min periods, and EBL by 50- and 100-mL volumes. Next, binary logistic regression was used for multivariate analysis of select categorized variables as a way to determine independent predictors of initial selection for OA and conversion. For these two
logistic regression models, preoperative categorized vari- ables included only age, gender, BMI, ASA score, lesion side, history of MEN-2, history of prior abdominal operation, diagnosis of a nonfunctioning adrenal lesion, the indication for operation, the radiographic tumor size, and the decision to perform concomitant procedures at the time of adrenal- ectomy. Alternatively, all pre-, intra-, and postoperative variables were used for determining independent predictors of 30-day morbidity. The Hosmer-Lemeshow test was used to confirm the goodness of fit for each model. Results were reported with odds ratio (OR) plus 95 % confidence interval (CI) for logistic regression analyses. Unless otherwise indi- cated, data were expressed as mean and standard deviation. Equal variances were not assumed, all tests were two-tailed, and significance was set at the x = 0.05 level.
Results
In total, 402 patients underwent 422 adrenalectomies. LA patients (n = 356) underwent adrenalectomy by the trans- abdominal (n = 343) or the retroperitoneal (13) approach.
The other cohort comprised patients initially selected for OA (n = 46). Overall and cohort-specific patient demographics are given in Table 1.The mean age of patients who under- went adrenalectomy was 50 years, 61 % were female, 52 % were left-sided tumors, and 63 % were functioning lesions. The LA and OA cohorts were similar with respect to BMI and proportions that were female and had left-sided lesions, a history of multiple endocrine neoplasia (MEN-2), and prior abdominal surgery. However, LA and OA cohorts differed significantly across multiple other preoperative variables. Compared to the LA group, patients in the OA group were approximately 6 years older on average, had a slightly higher mean ASA score, had a much larger radiographic tumor size, and were much less likely to have a functioning adrenal lesion. Residents and/or fellows were involved in all cases and therefore were not a factor in the selection of patients for operative approach or outcomes.
Among patients selected for LA, the most common indication was a functioning adrenal lesion, specifically pheochromocytoma (Table 1). The vast majority of patients with pheochromocytoma (96.2 %), cortisol-producing adenoma (97.7 %), and aldosterone-producing adenoma
| Overall (n = 402) | LA (n = 356) | OA (n = 46) | p value | |
|---|---|---|---|---|
| Age (years) | 50.0 ± 14.5 | 49.4 ± 14.4 | 55.0 ± 14.4 | 0.02 |
| Female (%) | 245 (60.9) | 221 (62.1) | 24 (52.2) | 0.20 |
| BMI (kg/m2) | 30.2 ± 7.5 | 30.3 ± 7.5 | 29.0 ± 7.4 | 0.21 |
| ASA | 2.5 ± 0.6 | 2.5 ± 0.6 | 2.7 ± 0.6 | 0.04 |
| MEN-2 syndrome (%) | 34 (8.4) | 32 (9) | 2 (4.3) | 0.29 |
| Prior abdominal operation (%) | 184 (45.8) | 164 (46.1) | 20 (43.5) | 0.74 |
| Lesion side (%) | 1.00 | |||
| Left | 209 (52) | 185 (52) | 24 (52) | |
| Right | 173(43) | 154 (43) | 19 (41) | |
| Bilateral | 20 (5) | 17 (5) | 3 (7) | |
| Radiographic tumor size (cm) | 3.8 ± 3.4 | 3.2 ± 2.1 | 8.5 ± 5.1 | <0.01 |
| Functioning lesion (%) | 252 (62.7) | 240 (67.4) | 12 (26.1) | <0.01 |
| Indication for operation (%) | ||||
| Pheochromocytoma | 106 (26.4) | 102 (28.7) | 4 (8.7) | <0.01 |
| Aldosterone-producing adenoma | 88 (21.9) | 88 (24.7) | 0 (0) | <0.01 |
| Large indeterminate mass | 56 (13.9) | 36 (10.1) | 20 (43.5) | <0.01 |
| Nonfunctioning adenoma | 44 (10.9) | 43 (12.1) | 1 (2.2) | 0.04 |
| Cortisol-producing adenoma | 43 (10.7) | 42 (11.8) | 1 (2.2) | 0.07 |
| Metastasis | 26 (6.5) | 20 (5.6) | 6 (13) | <0.01 |
| Myelolipoma | 16 (4) | 13 (3.7) | 3 (6.5) | 0.41 |
| Cushing's disease | 9 (2.2) | 9 (2.5) | 0 (0) | 0.40 |
| Adrenocortical carcinoma | 8 (2) | 1 (0.3) | 7 (15.2) | <0.01 |
| Primary adrenal hyperplasia | 6 (1.5) | 2 (0.6) | 4 (8.7) | <0.01 |
Comparison between laparoscopic and open adrenalectomy groups with significance set at p < 0.05
LA laparoscopic adrenalectomy, OA open adrenalectomy, BMI body mass index, ASA American Society of Anesthesiologists classification, MEN-2 multiple endocrine neoplasia II
(100 %) were approached laparoscopically. In contrast, the most common indications for OA were large indeterminate adrenal mass and adrenocortical carcinoma (Table 1). Overall, most patients with a preoperative diagnosis of large indeterminate adrenal mass (64.9 %) and adrenocortical carcinoma (87.5 %) were selected for OA.
A comparison of intraoperative outcomes for LA and OA cohorts appears in Table 2. Patients initially selected for OA had a longer mean operative time than those who had LA (mean difference of 38 min); however, a larger proportion of OA patients (63 %) also underwent concur- rent procedures at the time of adrenalectomy. Of those 29 OA patients who underwent concurrent procedures, 69 % required multiorgan resection (Table 3). In the LA group, operative times for left (154 ± 62 min) and right (148 ± 51 min) adrenal lesions were not significantly different. OA was associated with more intraoperative complications, greater estimated blood loss (>100 mL, 18.3 vs. 72 %, p < 0.01), and a larger number of blood transfusions. For example, only 2 % of LA patients had an EBL >500 mL compared to 23.9 % of OA patients.
Preoperative variables associated with selection for OA included larger radiographic tumor size (p <0.01), the presence of a nonfunctioning lesion (OR 5.9, CI 2.9-11.7, p < 0.01), the diagnosis of adrenocortical carcinoma (OR 117, CI 6-2,124, p < 0.01), and the decision to perform concomitant procedures such as multiorgan resection (OR 7.2, CI 3.8-13.9, p < 0.01). Additionally, patients selected for OA were older (p = 0.01) and had a higher ASA score (p = 0.03).
Conversion to open or hand-assisted adrenalectomy occurred in 22 of 356 (6.2 %) LA patients, who remained in the LA cohort for outcomes analysis. Specifically, of the
356 patients, conversion from LA to OA occurred in 15 (4.2 %) and from LA to a hand-assisted approach in 7 (2 %) patients (Table 4). Most conversions were related to large tumor size and/or difficulty with dissection, including four cases in which the tumor extended behind and/or abutted the inferior vena cava. The rates of conversion remained relatively consistent during the study (Fig. 1).
Predictive factors for conversion appear in Table 5. Of note, 9 of 15 (60 %) patients converted from LA to OA had pheochromocytoma. For patients initially selected for LA, the identifiable preoperative risks for conversion included large radiographic tumor size (>8 cm) and the decision to perform concomitant procedure(s) at the time of adre- nalectomy. Intraoperative factors associated with conver- sion were operative time greater than 240 min (referent
| LA (n = 68) | OA (n = 29) | p value | |
|---|---|---|---|
| Multiorgan resection (%) | 8.8 | 69 | <0.01 |
| Liver biopsy (%) | 20.6 | 17.2 | 0.79 |
| Adhesiolysis (%) | 19.1 | 6.9 | 0.22 |
| Cholecystectomy (%) | 17.6 | 17.2 | 1.00 |
| Other (%) | 19.1 | 13.8 | 0.51 |
| Umbilical herniorrhaphy (%) | 16.2 | 3.4 | 0.10 |
| Oophorectomy (%) | 2.9 | 0 | 0.58 |
Comparison between laparoscopic and open adrenalectomy groups with significance at p < 0.05 Columns exceed 100 % as patients may have undergone multiple procedures
| Overall (n = 402) | LA (n = 356) | OA (n = 46) | p value | |
|---|---|---|---|---|
| Operative time (min) | 161 ± 69 | 159 ± 68 | 197 ± 77 | 0.04 |
| Concomitant procedure (%) | 97 (24) | 68 (19) | 29 (63) | <0.01 |
| Intraoperative ultrasound (%) | 75 (18.7) | 69 (19.4) | 6 (13) | 0.42 |
| Estimated blood loss (%) | <0.01 | |||
| ≤50 mL | 253 (62.9) | 246 (69.1) | 7 (15.2) | |
| 51-100 mL | 51 (12.7) | 45 (12.6) | 6 (13) | |
| 101-200 mL | 39 (9.7) | 31 (8.7) | 8 (17.4) | |
| 201-300 mL | 22 (5.5) | 15 (4.2) | 7 (15.2) | |
| 301-400 mL | 7 (1.7) | 6 (1.7) | 1 (2.2) | |
| 401-500 mL | 12 (3) | 6 (1.7) | 6 (13) | |
| >500 mL | 18 (4.5) | 7 (2) | 11 (23.9) | |
| Blood transfusion (%) | 12 (3) | 7 (2) | 5 (10.9) | <0.01 |
| Pathologic tumor size (cm) | 3.5 ± 3.1 | 2.9 ± 2.2 | 7.6 ± 4.9 | <0.01 |
| Drain placement (%) | 64 (15.9) | 42 (11.8) | 22 (47.8) | <0.01 |
| Operative complication (%) | 17 (4.2) | 12 (3.4) | 5 (10.9) | 0.02 |
| Conversion (%) | – | 22 (6.2) | – | – |
Comparison between laparoscopic and open adrenalectomy groups with significance at p < 0.05
LA laparoscopic adrenalectomy, OA open adrenalectomy
| Patient | Years | Age (years) | Gender | ASA | BMI | PAO | Side | Size (cm) | Pathology | Reasons for conversion | Type |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 1996 | 41 | Male | 3 | 40.8 | No | Right | 2.5 | Pheochromocytoma | Liver injury, bleeding | OA |
| 2 | 1996 | 24 | Female | 3 | 29.8 | No | Left | 3 | Metastasis | Bleeding | OA |
| 3 | 2000 | 31 | Male | 3 | 26.7 | Yes | Right | 4.2 | Pheochromocytoma | Liver injury, bleeding | OA |
| 4 | 2000 | 35 | Female | 2 | 22.1 | No | Right | 4 | Metastasis | IVC abutment, bleeding | OA |
| 5 | 2000 | 50 | Male | 3 | 39.8 | Yes | Left | 1 | Accessory spleen | Poor visualization, splenomegaly | OA |
| 6 | 2003 | 75 | Female | 2 | 31.1 | No | Right | 5 | Pheochromocytoma | Bleeding | OA |
| 7 | 2003 | 52 | Male | 2 | 30.5 | No | Right | 9 | Pheochromocytoma | IVC abutment | OA |
| 8 | 2003 | 85 | Female | 3 | 22.2 | Yes | Left | 11 | Pheochromocytoma | Poor visualization, tumor size | OA |
| 9 | 2004 | 39 | Female | 2 | 26.2 | No | Left | 6 | Pheochromocytoma | Spleen injury, bleeding | OA |
| 10 | 2004 | 72 | Female | 2 | 49.6 | No | Right | 4 | Myxoid liposarcoma | Bleeding | OA |
| 11 | 2004 | 23 | Male | 2 | 29.8 | No | Right | 7 | Pheochromocytoma | Poor visualization, bleeding | OA |
| 12 | 2004 | 60 | Male | 2 | 25.3 | Yes | Left | – | Cushing's disease | Poor visualization, bleeding | HA |
| 13 | 2005 | 44 | Male | 3 | 36.6 | No | Left | 6 | Metastasis | Difficult dissection, bleeding | HA |
| 14 | 2005 | 50 | Female | 2 | 29.9 | No | Right | 4 | Nonfunctioning adenoma | Poor visualization | HA |
| 15 | 2005 | 64 | Male | 3 | 35.3 | No | Right | 4.8 | Nonfunctioning adenoma | Poor visualization, tumor size | HA |
| 16 | 2006 | 27 | Female | 4 | 42.5 | No | Left | 7 | Pheochromocytoma | Difficult dissection | HA |
| 17 | 2006 | 47 | Male | 2 | 35.8 | No | Left | 15 | Myelolipoma | Tumor size | HA |
| 18 | 2006 | 63 | Male | 2 | 25.1 | No | Left | 7.1 | Metastasis | Difficult dissection | HA |
| 19 | 2008 | 59 | Male | 3 | 25.1 | No | Right | 4.6 | Pheochromocytoma | Difficult dissection, shock | OA |
| 20 | 2008 | 42 | Male | 3 | 30.6 | No | Bilat | 3.6 | Macronodular hyperplasia | Difficult dissection | OA |
| 21 | 2009 | 65 | Male | 3 | 22.3 | Yes | Right | 8.5 | Pheochromocytoma | IVC/duodenum abutment, difficult dissection | OA |
| 22 | 2010 | 34 | Female | 3 | 21.5 | Yes | Right | 6 | Pheochromocytoma | Cardiac arrest, rule out intra- abdominal hemorrhage | OA |
ASA American Society of Anesthesiologists classification, BMI body mass index (kg/m2), PAO previous abdominal operation, OA open adrenalectomy, HA hand-assisted adrenalectomy
10
9
8
Conversions (%)
7
n=4
6
5
n=7
HA
4
n=3
OA
3
n=5
n=6
2
n=15
1
n=4
0
1993-2000 (n=97)
2001-2005 (n=115)
2006-2011 (n=144)
1993-2011 (n=356)
121-180 min, OR 9.9, CI 1.8-53.9, p < 0.01), EBL > 100 mL (referent ≤ 50 mL, OR 26.3, CI 2.6-260, p <0.01), requirement for intraoperative blood transfusion (OR 24.5, CI 5.1-117, p < 0.01), and any intraoperative complication (OR 9.1, CI 2.5-32.9, p <0.01). There was a trend toward greater odds of conversion among LA patients with a pre- operative diagnosis of adrenal metastasis, but this metric failed to reach statistical significance.
Postoperative 30-day outcomes following adrenalec- tomy are detailed in Table 6. On average, OA patients took more time to tolerate solid food, stayed in the hospital longer, and experienced more postoperative complications compared to LA patients. As these outcomes are influenced not only by the operative approach but also by the inherent differences between LA and OA patients, their comparison should be interpreted with caution. Over the study period, operative time and hospital LOS decreased for LA, despite
| Odds ratio | 95 % CI | p value | |
|---|---|---|---|
| Female | 2.1 | 0.3-1.4 | 0.12 |
| Lesion side | |||
| Left | Referent | ||
| Right | 1.5 | 0.1-4.0 | 0.50 |
| Bilateral | 1.2 | 0.1-10.2 | 1.00 |
| MEN-2 syndrome | 1.5 | 0.5-5.9 | 0.66 |
| Prior abdominal operation | 0.4 | 0.1-4.1 | 0.12 |
| Nonfunctioning lesion | 9.6 | 0.6-168 | 0.12 |
| Indication for operation | |||
| Pheochromocytoma | 2.2 | 0.5-10.6 | 0.35 |
| Large indeterminate mass | 2.5 | 0.4-14.4 | 0.41 |
| Nonfunctioning adenoma | Referent | ||
| Metastasis | 5.1 | 0.9-30.8 | 0.07 |
| Myelolipoma | 1.7 | 0.1-20.5 | 1.00 |
| Cushing's disease | 2.5 | 0.2-31.8 | 0.44 |
| Radiographic tumor size (cm) | |||
| ≤2 | Referent | ||
| 2.1-3 | 0.8 | 0.1-9.5 | 1.00 |
| 3.1-4 | 1.3 | 0.1-14.9 | 1.00 |
| 4.1-5 | 3.6 | 0.6-22.3 | 0.32 |
| 5.1-6 | 3.9 | 0.3-46.8 | 0.32 |
| 6.1-7 | 6.4 | 0.5-78.9 | 0.22 |
| 7.1-8 | 10.3 | 0.8-133 | 0.15 |
| >8 | 14.7 | 1.8-120 | 0.03 |
| Concomitant procedure | 5.6 | 1.1-29 | 0.04 |
| Overall (n = 402) | LA (n = 356) | OA (n = 46) | p value | |
|---|---|---|---|---|
| Solid food intake (days) | 2.3 ±.8 | 1.7 ± 1.3 | 6.9 ± 10.2 | <0.01 |
| Hospital length of stay (days) | 3.2 ± 5 | 2.5 ± 2.2 | 9.1 ± 12.1 | <0.01 |
| Readmission (%) | 24 (5.9) | 18 (5.1) | 6 (12.9) | 0.04 |
| Complication (%) | 63 (15.6) | 40 (11) | 23 (50) | <0.01 |
| Grade I | 34 (8.5) | 20 (5.6) | 14 (30.4) | |
| Grade II | 15 (3.7) | 13 (3.7) | 2 (4.3) | |
| Grade III | 4 (1) | 2 (0.6) | 2 (4.3) | |
| Grade IV | 9 (2.2) | 5 (1.4) | 4 (8.6) | |
| Grade V | 1 (0.2) | 0 (0) | 1 (2.2) |
Complications categorized by the Accordion Severity Grading Sys- tem [8]
Comparison between laparoscopic and open adrenalectomy groups with significance set at p < 0.05
LA laparoscopic adrenalectomy, OA open adrenalectomy
the fact that LA was performed for increasingly larger tumor sizes (all p < 0.01). None of the measured outcomes for OA changed during the study period.
Although the rates of hospital readmission within 30 days were similar between groups, the frequency and severity of complications differed significantly. Fifty per- cent of OA patients suffered some postoperative compli- cation. While most complications in the OA group were low grade and wound-related, a significantly greater proportion suffered severe complications compared to the LA group. The overall 30-day mortality rate was 0.2 %, which was due to the death of one patient in the OA cohort. This patient had severe Cushing’s syndrome and multiple major comorbid- ities. Also, she was thought to have an adrenocorticotropic hormone (ACTH)-producing tumor in the tail of the pan- creas. Open exploration revealed the pancreas mass to be an abscess. Bilateral adrenalectomy was attempted as defini- tive treatment of severe Cushing’s syndrome from an ectopic source of ACTH. After uncomplicated left adre- nalectomy, dissection of the right adrenal gland resulted in bleeding from the liver and right adrenal vein. During the operation, the patient suffered ventricular fibrillation and shock that did not respond to life-saving efforts.
Factors predictive of 30-day morbidity after adrenalec- tomy are given in Table 7. By multivariate analysis, the perioperative predictors of morbidity after adrenalectomy included the following indications for operation: large indeterminate adrenal mass, adrenocortical carcinoma, and primary adrenal hyperplasia, as well as increased radio- graphic tumor size (>6 cm). Other variables associated with postoperative morbidity were an open operation, need for conversion, necessity for concomitant procedures, pro- longed operative time (>180 min), and higher EBL (>200 mL). Regardless of operative approach, patients who suffered an intraoperative complication were more likely to experience a postoperative complication within 30 days (LA, p < 0.01; OA, p = 0.04).
Discussion
A substantial body of literature supports the safety and efficacy of LA as well as the demonstrable benefits of LA over OA, including less postoperative pain, shorter hospital stay, and lower morbidity [2, 3, 9-11]. At present, LA is the gold standard for removal of most benign adrenal lesions and small adrenal metastases, while OA is preferred for primary adrenal malignancy and very large adrenal masses. However, long-held indications for OA have been liberalized recently given the growing experience with LA [7, 12, 13]. For example, some case series have detailed the successful use of LA for very large benign adrenal lesions [14-18]. Of note, authors who advocate LA for liberalized indications have significant experience with LA and per- form this operation in specialized centers.
| Odds ratio | 95 % CI | p value | |
|---|---|---|---|
| Female | 1.5 | 0.8-2.9 | 0.17 |
| Lesion side | |||
| Left | Referent | ||
| Right | |||
| Bilateral | |||
| MEN-2 syndrome | 1.8 | 0.4-3.6 | 0.48 |
| Prior abdominal operation | 0.9 | 0.5-1.7 | 0.44 |
| Nonfunctioning lesion | 1.4 | 0.8-2.4 | 0.32 |
| Indication for operation | |||
| Pheochromocytoma | 2.4 | 0.7-8.8 | 0.19 |
| Aldosterone-producing adenoma | 1 | 0.2-4.1 | 1.00 |
| Large indeterminate mass | 5.7 | 1.5-20.9 | <0.01 |
| Nonfunctioning adenoma | Referent | ||
| Cortisol-producing adenoma | 1.7 | 0.4-7.9 | 0.71 |
| Metastasis | 1.7 | 0.3-9.3 | 0.66 |
| Myelolipoma | 1.9 | 0.3-12.6 | 0.61 |
| Cushing's disease | 3.8 | 0.5-27.1 | 0.20 |
| Adrenocortical carcinoma | 10 | 1.5-67 | 0.03 |
| Primary unilateral hyperplasia | 13.3 | 1.8-97 | 0.02 |
| Radiographic tumor size (cm) | |||
| ≤ 2 | Referent | ||
| 2.1-3 | 0.5 | 0.2-1.7 | 0.42 |
| 3.1-4 | 1.9 | 0.7-4.9 | 0.20 |
| 4.1-5 | 1.6 | 0.6-4.1 | 0.44 |
| 5.1-6 | 0.5 | 0.1-4.3 | 0.70 |
| 6.1-7 | 4.4 | 1.3-15.2 | 0.03 |
| 7.1-8 | 4.5 | 1.1-17.6 | 0.04 |
| >8 | 3.9 | 1.4-11.1 | 0.01 |
| Initial open approach | 8.1 | 4.2-15.8 | <0.01 |
| Concomitant procedure | 7.2 | 3.8-13.9 | <0.01 |
| Operative time (min) | |||
| ≤120 | Referent | ||
| 120-180 | 2.2 | 0.6-9.1 | 0.33 |
| 181-240 | 9.0 | 2.5-33.1 | <0.01 |
| >240 | 15.3 | 3.9-59.1 | <0.01 |
| Estimated blood loss (mL) | |||
| ≤50 | Referent | ||
| 51-100 | 1.3 | 0.5-3.6 | 0.63 |
| 101-200 | 1.7 | 0.5-5.8 | 0.51 |
| 201-300 | 4.6 | 1.5-14.1 | <0.01 |
| 301-400 | 10.7 | 3.3-35.1 | <0.01 |
| >400 | 20.3 | 7.1-58.2 | <0.01 |
| Conversion | 5.6 | 2.2-14.4 | <0.01 |
| Solid food intake (>3 days) | 34.8 | 10-115 | <0.01 |
| Hospital length of stay (>3 days) | 18.7 | 8.7-40.3 | <0.01 |
| Readmission | 1.8 | 0.6-5.5 | 0.49 |
Broadening the potential indications for LA is not without risk and may result in higher conversion rates and greater morbidity. A retrospective review of a large national database found that the number of patients selec- ted for a laparoscopic approach has increased over time, but so too have major in-hospital postoperative complica- tions. The latter finding was felt to be due, in part, to changes in patient demographics over the study period, namely, an increased number of older patients with more comorbidities undergoing adrenalectomy [5]. In order to better understand the most suitable indications for LA, it is important to confirm the appropriate indications for and risks associated with OA. This is particularly true in the United States given that the majority of patients who underwent adrenalectomy had an open approach as recently as 2006 [5].
Earlier case series and retrospective studies have com- pared outcomes following LA and OA. Lee et al. [6] col- lected adrenalectomy data on 669 patients from 123 Veterans Affairs Medical Centers and 14 university hos- pitals over 3 years and then compared short-term outcomes and analyzed risks for morbidity. The authors found that patients who underwent OA were older, had a higher ASA classification, and more commonly harbored malignancy. The open approach was associated with longer operative time, the need for blood transfusion, longer hospital stay, and greater 30-day morbidity. OA resulted in more respi- ratory, cardiac, renal, and wound complications compared to LA. However, the authors were unable to differentiate between indications for operation and types of functioning tumors, detail the pathologic findings, or delineate which patients had an adrenalectomy for adrenocortical carci- noma versus concomitant adrenalectomy for metastatic disease. Therefore, they were not able to adjust for the effects of pathology and concomitant resection on clinical outcomes. Other authors confirmed that older age may be associated with adverse outcomes after adrenalectomy [19].
One aim of the present study was to investigate the impact of preoperative variables, normally excluded from national databases, on outcomes after adrenalectomy. Similar to published results, our study confirmed that patients who underwent OA were older, had a higher ASA classification, and more commonly harbored malignancy. However, data not available in national databases that were found to differentiate OA from LA patients in our series included preoperative tumor size, functioning status of the lesion, and certain indications for operation. Specifically, patients who presented with larger tumors, nonfunctioning lesions, and/or a preoperative diagnosis of adrenocortical carcinoma were at greater odds of being selected for OA.
A second study aim was to define the rate and risk of conversion among patients initially selected for LA. In our series, the conversion rate to OA was 4.2 %, consistent with previous reports [20, 21], and an additional 7 patients (2.0 %) were converted to a hand-assisted approach. For patients initially selected for LA, the identifiable preoper- ative risks for conversion included large radiographic tumor size (>8 cm) and the need for concomitant proce- dure(s). The statistically increased odds of conversion among patients with tumors over 8 cm, but not 5-6 cm as has been reported previously [21, 22], may represent sur- geon experience and willingness to attempt LA for increasingly larger lesions.
Like the results of Lee et al. [6], patients in our study who were initially selected for OA had a longer mean operative time, higher EBL, more transfusions, longer LOS, and greater morbidity than those who had LA. Also, more patients in the OA group required concomitant procedures (multiorgan resection). Besides operative approach, other variables that predicted 30-day morbidity after adrenalec- tomy included the preoperative diagnosis of large indeter- minate adrenal mass, adrenocortical carcinoma, or primary adrenal hyperplasia; radiographic tumor size (>6 cm); and the need for concomitant procedures (multiorgan resection). Thus, defining adrenal pathology simply as functioning or nonfunctioning and benign or malignant may be inadequate when assessing clinical outcomes.
These results suggest that future reporting of clinical outcomes should include specific adrenal tumor histology and size. Moreover, the need for and type of concomitant procedures performed during adrenalectomy should be included as a surrogate for extent of disease and case complexity. Databases should include information about adrenal tumor-functioning status and histology, as both nonfunctioning status and primary adrenal malignancy impact the choice of operative approach and short-term outcomes. A focus of future studies may be to investigate the impact of surgeon volume or practice type on outcomes of LA [23].
Although this study confirmed recent national trends and highlighted specific variables that predict the choice of operative approach and 30-day morbidity for adrenalec- tomy, it does have several limitations. Recent analyses of patients who underwent LA for various indications revealed that certain patient-level factors significantly contribute to 30-day morbidity. These factors included race/ethnicity, BMI, serum albumin level, presence of a bleeding disorder, use of corticosteroids, baseline hemat- ocrit, and chronic obstructive pulmonary disease [11, 24]. One shortcoming in the present study is that not all of these patient factors, which may impact short-term morbidity
after LA, were included in the analysis. Thirteen patients (3.6 %) in the LA cohort underwent a retroperitoneal approach; yet, for the purpose of statistical analysis, they were considered similar to the 343 patients who underwent a transabdominal approach. The statistical impact of this decision was not assessed. It is unlikely that patient age and ASA score are clinically relevant selection factors for choosing an operative approach to adrenalectomy. Other study limitations included the inability to compare surgeon or other adrenalectomy outcomes beyond conversion rates based on total experience with LA. However, all surgeons in this series were experienced with OA and LA and likely beyond the upper inflection point of their learning curve, certainly for the latter years of the study. Since all proce- dures were performed in a large, urban, university hospital serving as a tertiary referral center, results may not be applicable across different clinical settings. Lastly, given the study design, there was no attempt to correlate short- term objective outcome measures with subjective symptom or quality-of-life scores.
Despite these limitations, the present study elucidated predictive factors that should be considered when selecting an operative approach to adrenalectomy. Additionally, the study has further defined rates and risk factors of conver- sion and reported differences in 30-day morbidity between laparoscopic and open adrenalectomy. Future multicenter investigations are needed to explore the differences in selection of an operative approach, conversion rates, peri- operative outcomes, and their relationship to nationally tracked databases.
Conclusions
Adrenal tumor size and the need to undertake concurrent procedures significantly impact the selection of patients for OA, the likelihood of conversion, and perioperative mor- bidity. These factors should be considered when assessing operative approach and risks for adrenalectomy.
Acknowledgments The authors recognize Mary Quasebarth, RN, for her assistance with IRB protocol submission. James G. Bittner IV, MD, is supported by an MIS Clinical Fellowship educational grant from the Foundation for Surgical Fellowships.
Disclosures James G. Bittner IV, MD, and L. Michael Brunt, MD, has received honoraria from Ethicon Endo-Surgery, Inc. for teaching and speaking. Brent D. Matthews, MD, is a consultant for Atrium Medical Corporation and Ethicon, Inc. He has received honoraria and research/equipment support from Atrium Medical Corporation, Ethi- con Endo-Surgery, Inc., Karl Storz Endoscopy, Stryker Endoscopy, and W.L. Gore & Associates, Inc. Victoria M. Gershuni, MS, and Jeffrey F. Moley, MD, have no conflicts of interest or financial ties to disclose.
References
1. Gagner M, Lacroix A, Bolte E (1992) Laparoscopic adrenalec- tomy in Cushing’s syndrome and pheochromocytoma. N Engl J Med 327:1033
2. Brunt LM, Doherty GM, Norton JA, Soper NJ, Quasebarth MA et al (1996) Laparoscopic adrenalectomy compared to open adrenalec- tomy for benign adrenal neoplasms. J Am Coll Surg 183:1-10
3. Brunt LM, Moley JF, Doherty GM, Lairmore TC, DeBenedetti MK et al (2001) Outcomes analysis in patients undergoing lap- aroscopic adrenalectomy for hormonally active adrenal tumors. Surgery 130:629-634 (discussion 634-625)
4. Brunt LM (2002) The positive impact of laparoscopic adrenal- ectomy on complications of adrenal surgery. Surg Endosc 16:252-257
5. Murphy MM, Witkowski ER, Ng SC, McDade TP, Hill JS et al (2010) Trends in adrenalectomy: a recent national review. Surg Endosc 24:2518-2526
6. Lee J, El-Tamer M, Schifftner T, Turrentine FE, Henderson WG et al (2008) Open and laparoscopic adrenalectomy: analysis of the National Surgical Quality Improvement Program. J Am Coll Surg 206:953-959; discussion 959-961
7. Henneman D, Chang Y, Hodin RA, Berger DL (2009) Effect of laparoscopy on the indications for adrenalectomy. Arch Surg 144:255-259; discussion 259
8. Strasberg SM, Linehan DC, Hawkins WG (2009) The accordion severity grading system of surgical complications. Ann Surg 250:177-186
9. Walz MK, Alesina PF, Wenger FA, Deligiannis A, Szuczik E et al (2006) Posterior retroperitoneoscopic adrenalectomy- results of 560 procedures in 520 patients. Surgery 140:943-948 (discussion 948-950)
10. Nguyen PH, Keller JE, Novitsky YW, Heniford BT, Kercher KW (2011) Laparoscopic approach to adrenalectomy: review of per- ioperative outcomes in a single center. Am Surg 77:592-596
11. Gupta PK, Natarajan B, Pallati PK, Gupta H, Sainath J et al (2011) Outcomes after laparoscopic adrenalectomy. Surg Endosc 25:784-794
12. Brunt LM, Lairmore TC, Doherty GM, Quasebarth MA, DeBenedetti M et al (2002) Adrenalectomy for familial pheo- chromocytoma in the laparoscopic era. Ann Surg 235:713-720; discussion 720-711
13. Mazzaglia PJ, Vezeridis MP (2010) Laparoscopic adrenalectomy: balancing the operative indications with the technical advances. J Surg Oncol 101:739-744
14. Asari R, Koperek O, Niederle B (2012) Endoscopic adrenalec- tomy in large adrenal tumors. Surgery 152(1):41-49
15. Perry KA, El Youssef R, Pham TH, Sheppard BC (2010) Lapa- roscopic adrenalectomy for large unilateral pheochromocytoma: experience in a large academic medical center. Surg Endosc 24:1462-1467
16. Boylu U, Oommen M, Lee BR, Thomas R (2009) Laparoscopic adrenalectomy for large adrenal masses: pushing the envelope. J Endourol 23:971-975
17. Ippolito G, Palazzo FF, Sebag F, Thakur A, Cherenko M et al (2008) Safety of laparoscopic adrenalectomy in patients with large pheochromocytomas: a single institution review. World J Surg 32:840-844 (discussion 845-846)
18. Novitsky YW, Czerniach DR, Kercher KW, Perugini RA, Kelly JJ et al (2003) Feasibility of laparoscopic adrenalectomy for large adrenal masses. Surg Laparosc Endosc Percutan Tech 13:106-110
19. Kazaure HS, Roman SA, Sosa JA (2011) Adrenalectomy in older Americans has increased morbidity and mortality: an analysis of 6,416 patients. Ann Surg Oncol 18:2714-2721
20. Gaujoux S, Bonnet S, Leconte M, Zohar S, Bertherat J et al (2011) Risk factors for conversion and complications after uni- lateral laparoscopic adrenalectomy. Br J Surg 98:1392-1399
21. Shen ZJ, Chen SW, Wang S, Jin XD, Chen J et al (2007) Pre- dictive factors for open conversion of laparoscopic adrenalec- tomy: a 13-year review of 456 cases. J Endourol 21:1333-1337
22. Shen WT, Kebebew E, Clark OH, Duh QY (2004) Reasons for conversion from laparoscopic to open or hand-assisted adrenal- ectomy: review of 261 laparoscopic adrenalectomies from 1993 to 2003. World J Surg 28:1176-1179
23. Maccabee DL, Jones A, Domreis J, Deveney CW, Sheppard BC (2003) Transition from open to laparoscopic adrenalectomy: the need for advanced training. Surg Endosc 17:1566-1569
24. Kazaure HS, Roman SA, Sosa JA (2011) Obesity is a predictor of morbidity in 1,629 patients who underwent adrenalectomy. World J Surg 35:1287-1295