- Email: [email protected]
Patient selection and outcomes of laparoscopic transabdominal versus posterior retroperitoneal adrenalectomy among surgeons in the Collaborative Endocrine Surgery Quality Improvement Program (CESQIP)
Surgery xxx (2019) 1e7
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Patient selection and outcomes of laparoscopic transabdominal versus posterior retroperitoneal adrenalectomy among surgeons in the Collaborative Endocrine Surgery Quality Improvement Program (CESQIP) Annette Pascual Marrero, MD, MPHa,*, Hadiza S. Kazaure, MDa, Samantha M. Thomas, MSb,c, Michael T. Stang, MDa, Randall P. Scheri, MDa a
Department of Surgery, Duke University Medical Center, Durham, NC Department of Biostatistics and Bioinformatics, Duke University, Durham, NC c Duke Cancer Institute, Duke University, Durham, NC b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 12 March 2019 Available online xxx
Background: Laparoscopic adrenalectomy can be performed using a transabdominal or posterior retroperitoneal approach. Choosing the optimal approach can be challenging. Methods: Using data from the Collaborative Endocrine Surgery Quality Improvement Program (2014 e2018), baseline patient characteristics and outcomes were compared with bivariate methods; univariate and multivariate analyses were used to estimate the association between operative approach and complication risk. Results: Among 833 patients, 35.3% underwent posterior retroperitoneal. Median age was 54 years. Patients undergoing posterior retroperitoneal had lesser rates of body mass index >40 (9.2% vs 17.4%, P ¼ .001), smaller nodules (median 2.4 vs 3.2 cm, P < .001), and more commonly right-sided nodules (46.6% vs 36.9%, P ¼ .02). Posterior retroperitoneal was associated with a lesser rate of conversion to an open procedure (0.7% vs 4.1%, P ¼ .004), less complications (3.1% vs 8.7%, P ¼ .002), and shorter hospital stay (48 h: 92.2% vs 76.6%, P < .001), but a greater rate of capsular disruption (12.6% vs 7.6%, P ¼ .02). For posterior retroperitoneal cases with capsular disruption, median nodule size was 2.2 cm, and 16.2% were metastatic tumors. After multivariate adjustment, posterior retroperitoneal was 2.2 times as likely to result in capsular disruption as transabdominal (95% confidence interval, 1.04e4.79, P ¼ .04). Conclusion: This study revealed a greater rate for capsular disruption during posterior retroperitoneal even for small tumors. Our findings from the Collaborative Endocrine Surgery Quality Improvement Program (2014e2018) suggests that posterior retroperitoneal should be used selectively, especially when a malignancy is suspected. © 2019 Elsevier Inc. All rights reserved.
Introduction Laparoscopic adrenalectomy has emerged as a minimally invasive standard of care for most adrenal pathologies with improved operative outcomes compared to open adrenalectomy.1e4 As surgeons become more experienced with minimally invasive techniques, a variety of operative approaches for laparoscopic
This study was accepted for podium presentation at the 40th annual meeting of the American Association of Endocrine Surgeons, Los Angeles, CA. * Reprint requests: Annette Pascual Marrero, MD, MPH, Department of Surgery, Duke University Medical Center, Box 2945, Durham, NC 27710. E-mail address: [email protected] (A.P. Marrero). https://doi.org/10.1016/j.surg.2019.03.034 0039-6060/© 2019 Elsevier Inc. All rights reserved.
adrenalectomy has emerged. These include transabdominal or posterior retroperitoneal approaches using either classic laparoscopic or robot-assisted techniques. Contrary to the posterior retroperitoneal approach to adrenalectomy (PRA), in which the operative space is small and requires surgeon familiarity with unconventional visualization of anatomic structures, the adrenal gland can be accessed transabdominally (TA) in a wider working space and with conventional anatomic landmarks. Several studies indicate that PRA is beneficial in patients with previous abdominal operations and those with bilateral adrenal pathologies.4e8 Existing literature provides conflicting results when postoperative outcomes of PRA are compared to TA. Rok Lee et al have
2
A.P. Marrero et al. / Surgery xxx (2019) 1e7
reported that PRA is associated with earlier recovery of bowel function, possibly leading to a shorter hospital stay and earlier recovery.7 In their study, Walz et al found that PRA resulted in a lesser need for postoperative analgesia and greater rates of early mobilization.9 Berber et al showed similar mortality and conversion rates and a comparable duration of hospital stay. Among patients who underwent PRA, a lesser duration of surgery and less blood loss were observed; overall, the authors concluded that TA and PRA have similar outcomes. They recommended PRA as the most favorable approach for patients with small tumors, lesser body mass index (BMI), bilateral adrenal pathologies, and those with a history of abdominal operations.10 At a multi-institutional level, the optimal laparoscopic strategy based on adrenal tumor size, diagnosis, and laterality, is unclear. The Collaborative Endocrine Surgery Quality Improvement Program (CESQIP) is an interinstitutional collaborative effort among endocrine surgeons to gather patient data for tracking, collection, and analysis. Surgeons receive feedback and access clinical data to improve quality of service and outcomes and decrease costs. The dataset provides the power and surgical and clinical details in addition to a cross-section of high-volume surgeons necessary to better characterize factors associated with choice of operative approach of adrenalectomy and postoperative outcomes. The aims of this study were to evaluate the baseline and clinical characteristics and outcomes of laparoscopic adrenalectomies among CESQIP surgeons based on operative approach (TA versus PRA) and to compare risks associated with each technique. Approximately 37 participating sites reported adrenalectomies outcomes in CESQIP. Methods All patients who underwent laparoscopic TA or PRA between January 1, 2014, and April 26, 2018, captured in the CESQIP adrenalectomy database were included in this study. Patients <18 years old and those who underwent remedial adrenalectomy were excluded. Baseline characteristics included age, race, sex, BMI, hypertension, and diabetes mellitus. The CESQIP database classifies BMI ( or >40 kg/m2) as a binary categorical variable. Preoperative characteristics included familial versus sporadic adrenal pathology, nodule size, tumor type, Cushing’s classification, and metastatic disease type. Familial or genetic syndromes included multiple endocrine neoplasia (MEN types 1 and 2), neurofibromatosis (type 1), succinate dehydrogenase mutation, and Von Hippel Lindau syndrome. Operative variables were surgeon and facility volume, duration of the operation, and tumor laterality. Outcomes were occurrence of capsular disruption, conversion to an open procedure, incidence of any complication (visceral injury, hemorrhage requiring transfusion, postoperative bleeding requiring transfusion, unplanned return to operating room, surgical site infection, pulmonary embolism, deep vein thrombosis, sepsis, myocardial infarction, cardiac arrest, urinary tract infection, acute renal failure, pneumonia, ventilator >48 h, or hypertensive crisis), presence of adrenal insufficiency, and duration of hospital stay. In CESQIP, capsular disruption is captured from intraoperative data and was defined as evidence of gross macroscopic intraoperative disruption of the gland’s capsule. Because of its rarity, complications were analyzed as a categorical variable based on the occurrence of 1 of any of the complications listed above. Categorical variables were reported as N (%) and continuous variables as median (interquartile range, IQR). Differences between study groups were tested using the c2 test or Fisher exact test or t test, as appropriate. For nodule size, patients indicated as not having a measurable nodule by preoperative imaging were defined as size 0 cm. Univariate logistic regression was utilized to estimate
the effect of nodule size, annual surgeon, and facility volume, primary diagnosis, tumor location, BMI, and operative approach on incidence of any (1) complication and occurrence of capsular disruption. Annual surgeon volume was examined both as a continuous variable and after classification into 6 versus >6 adrenalectomies per year, based on existing literature for adrenal surgical volume.11 Because no differences were seen in modeling results for surgeon volume, only continuous values are reported. Multivariate logistic regression was used to estimate the effect of the approach to adrenalectomy on incidence of any complication and occurrence of capsular disruption after adjustment for clinically significant variables. Odds ratios (OR) and 95% confidence intervals (CI) are reported. All logistic regression models were built using the generalized estimating equations framework, and an exchangeable correlation structure was used to account for the correlation of patients treated by the same surgeon. A subgroup analysis for patients with and without capsular disruption was also conducted. No adjustments were made for multiple comparisons. All statistical analyses were conducting using SAS version 9.4 (SAS Institute, Cary, NC). Results A total of 1,145 patients were included in the CESQIP adrenalectomy database during this 4-year period. Of those, 833 (72.8%) patients underwent either laparoscopic TA or PRA and met all other inclusion criteria. From those, 539 (64.7%) underwent laparoscopic TA and 294 (35.3%) underwent PRA. Patients who underwent PRA were more likely to be older, non-black, and have a BMI 40 (all P < .05, Table I). They also were more likely to have smaller nodules (median 2.4 vs 3.2 cm, P < .001). The most prevalent diagnoses for patients who underwent PRA were: hyperaldosteronism (29.9%), pheochromocytoma (22.1%), and nonfunctioning nodule (19.0%). On the other hand, those patients who underwent TA had most commonly the following diagnoses: Cushing’s syndrome or disease (39.3%), pheochromocytoma (27.1%), and hyperaldosteronism (23.6%) among patients who underwent TA (Table II). The majority of primary diagnoses were sporadic (95.2%), whereas few were associated with a genetic mutation or syndrome (4.8%), including MEN1 (0.1%), MEN2 (1.6%), neurofibromatosis-1 (1%), succinate dehydrogenase (0.4%), von Hippel Lindau syndrome (0.5%), and other (0.8%). When compared to TA, patients who underwent PRA were more likely to have right-sided nodules (46.6% vs 36.9%, P ¼ .02). No significant difference was noted in the duration of the procedure by operative approach (P ¼ .31). The median nodule size for all patients was 2.8 cm (IQR 1.5e4.2, Table II). For patients with Cushing’s syndrome, 50.9% presented with clinical manifestations and 31.0% with subclinical diagnosis. Classification is unknown for the remainder of patients. A total of 5.2% (N ¼ 43) of all adrenalectomies were performed for adrenal metastasis; of these, 16 (37.2%) underwent PRA. Three patients with a preoperative diagnosis of adrenocortical carcinoma were identified, one of whom underwent PRA. A total of 34 facilities that performed at least one TA or PRA were included in our study; 21 (61.8%) of them performed only TAs, 4 (11.8%) performed only PRAs, and 9 (26.5%) performed both TAs and PRAs. Of the 60 CESQIP surgeons, in this study, the median number of PRAs per surgeon was 8.8 (IQR 7.0e14.6) and TAs was 7.5 (IQR 5.0e10.4) per year (P < .001). No difference was identified in facility volume for PRA versus TA (median 9.5 vs 11.4, P ¼ .06, Table III). Figure 1 compares the outcomes of both approaches. In total, 6.7% of patients experienced a complication, notably visceral injury (1.2%), hemorrhage requiring intraoperative transfusion (1.2%), adrenal insufficiency (0.8%), and hypertensive crisis (0.8%). There
A.P. Marrero et al. / Surgery xxx (2019) 1e7
3
Table I Baseline and clinical characteristics of patients who underwent laparoscopic TA or PRA in CESQIP (2014e2018)
Age (y), median (IQR) Race Non-black Black Asian Hispanic Unknown Sex Female BMI >40 Hypertension Diabetes
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
54 (43e63)
53 (41e63)
56 (47e65)
.002
562 (67.5%) 87 (10.4%) 18 (2.2%) 116 (13.9%) 46 (5.5%)
328 (60.9%) 61 (11.3%) 15 (2.8%) 95 (17.6%) 38 (7.1%)
234 (79.6%) 26 (8.8%) 3 (1.0%) 21 (7.1%) 8 (2.7%)
526 (63.1%)
353 (65.5%)
173 (58.8%)
121 (14.5%) 607 (72.9%) 213 (25.6%)
94 (17.4%) 387 (71.8%) 131 (24.3%)
27 (9.2%) 220 (74.8%) 82 (27.9%)
<.001
.06 .001 .31 .25
Data presented as N (%) unless otherwise specified. Percentages may not add up to 100 due to rounding or missing values.
was no statistical difference in complication rates; however, PRA was associated with lesser incidence rate of any (1) complication (3.1% vs 8.7%, P ¼ .002). After multivariate adjustment, the operative approach no longer was associated with complication risk (OR ¼ 0.50; 95% CI, 0.23e1.09, P ¼ .08, Fig 2). Subanalysis of cases that resulted in capsular disruption among PRA versus TA was performed. The median age of these patients was 55 years, similar to the full cohort. In this group, however, nonblack (81.8% vs 48.8%, P ¼ .005) and male patients (56.8% vs 29.3%, P ¼ .01) more commonly underwent PRA. Most patients who underwent PRA with capsular disruption had a BMI 40 (94.6% vs 78%, P ¼ .05). Among this group, the median nodule size was 2.2 cm on PRA vs 4.2 cm on TA (P < .001). For both groups, the most prevalent diagnoses were hyperfunctioning tumors (62.2% and 65.9% for PRA and TA respectively, P ¼ .39). Among PRA, hyperaldosteronism (32.4%), nonfunctioning nodule (16.2%), and metastatic disease (16.2%) were the most common tumor types compared to TA, where the most common types were hyperaldosteronism (24.4%), pheochromocytoma (24.4%), and Cushing’s syndrome (14.6%). No differences were seen in duration of operation (<2 hours, 51.4% vs 68.3% P ¼ .17), tumor laterality (left: 56.8% vs 43.9%, right: 37.8% vs 51.2%, P ¼ .43), annual surgeon volume (median 7.2 vs 10.3 cases per year, P ¼ .67), or facility volume (median 9.5 vs 9.3 cases per year, P ¼ .80) when comparing PRA versus TA on this cohort. Left sided tumors (50%) were more common among patients with capsular disruption. Only 3 patients who had capsular disruption and underwent TA had conversion to open procedure. In univariate and the multivariate analyses, PRA was associated with a greater risk of developing capsular disruption. After adjusting for covariates, patients who underwent PRA had 2 times the odds of developing capsular disruption compared to those who underwent TA (OR ¼ 2.23; 95% CI 1.04e4.79, P ¼ .04, Fig 3). All patients who underwent PRA (N ¼ 289) were classified into those with (12.8%) and without (87.2%) capsular disruption. Both groups had similar clinical characteristics without any difference on age (P ¼ .42), BMI (P ¼ .55), tumor size (P ¼ .32), laterality (P ¼ .51), or diagnosis (P ¼ .18). Capsular disruption was more prevalent among men (56.8% vs 38.9%, P ¼ .04). A total of 16.2% of patients who underwent PRA with capsular disruption had metastatic disease compared to 3.9% without capsular disruption (P ¼ .18). In addition, no difference was found in the rates of capsular disruption within institutions (P ¼ .39). In the facilities where PRA was only performed, 12.8% had capsular disruption (N ¼ 109) compared to facilities performing TA only where 8.4% had capsular disruption (N ¼ 285). In the institutions in which both approaches were performed, the rate of capsular disruption was 9.1% (N ¼ 439).
Discussion As surgeons become proficient with advanced minimally invasive techniques, the criteria for choosing the optimal approach for adrenalectomy continues to evolve. In this study, the most common approach for laparoscopic adrenalectomy among CESQIP surgeons was TA. This approach was favored for tumors that were larger, located on the left side, and in patients who had a BMI >40; diagnosis did not affect operative approach. Facilities where a TA approach predominated performed slightly more cases annually compared to those that performed mostly a PRA approach. Overall, annual surgeon volume was not associated with the risk of complications. This observation could be explained by CESQIP surgeons representing high volume surgeons whose median number of adrenalectomies is likely above the high-volume threshold advocated in existing literature11; however, surgeon volume has been associated previously with fewer postoperative complications after adrenalectomies. Anderson et al documented that surgeons performing 6 cases yearly have better outcomes compared to low volume surgeons.11 Lindeman et al also showed that high volume surgeons (4 procedures per year) have lesser complication and mortality rates compared to low volume surgeons (P < .001 and P ¼ .04).12 Several studies that have compared PRA versus TA showed no significant outcomes differences, including complications, operative time, postoperative pain control, blood loss, and duration of hospital stay.10,13,14 Mohammadi-Fallah et al showed that PRA had a lesser duration of postoperative ileus and recovery period, while Rubinstein et al showed no difference in oral intake tolerance.14,15 ski et al concluded that PRA was superior to TA in earlier Barczyn recovery, lesser operative time, less postoperative analgesia requirements and blood loss. In this study, PRA was associated with a lesser rate of conversion to an open procedure and a lesser hospital stay when compared to TA. There were no differences in operative time and overall complication risk; however, there is likely a strong selection bias for a given approach evidenced by the difference in patient clinical characteristics between both procedures.16 Of concern, there was a greater risk of capsular disruption with PRA (12.6% vs 7.6%, P ¼ .02). Based on these findings, our study neither supports the equivalence of TA to PRA nor the superiority of one operative approach over the other. Larger studies would be needed to provide conclusive results. Our study showed that PRA was performed most commonly on patients with BMI 40. Zonea et al found that PRA is a safe alternative procedure for obese patients (BMI >30 kg/m2); however, operative time, hospital stay, morbidity, and conversion rates were greater.17 In a study by Hu et al, the impact of obesity on PRA
4
A.P. Marrero et al. / Surgery xxx (2019) 1e7 Table II Clinical characteristics of patients who underwent laparoscopic TA or PRA in CESQIP (2014e2018)
Nodule size (cm), median (IQR) Diagnosis Hyperfunctioning tumor Hyperaldosteronism Pheochromocytoma Cushing's Adrenocortical carcinoma Nonfunctioning tumor Nonfunctioning nodule Metastatic disease Tumor laterality Left Right Bilateral
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
2.8 (1.5e4.2)
3.2 (1.6e4.5)
2.4 (1.4e3.6)
<.001 .65
600 (72.0%) 215 (25.8%) 211 (55.1%) 171 (20.5%) 3 (0.4%) 193 (23.2%) 150 (18.0%) 43 (5.2%)
387 (71.8%) 127 (23.6%) 146 (27.1%) 112 (39.3%) 2 (0.4%) 121 (22.4%) 94 (17.4%) 27 (5.0%)
213 (72.4%) 88 (29.9%) 65 (22.1%) 59 (20.1%) 1 (0.3%) 72 (24.5%) 56 (19.0%) 16 (5.4%)
442 (53.1%) 336 (40.3%) 50 (6%)
303 (56.2%) 199 (36.9%) 32 (5.9%)
139 (47.3%) 137 (46.6%) 18 (6.1%)
.02
Data presented as N (%) unless otherwise specified. Cushing’s (disease or syndrome). Percentages may not add up to 100 due to rounding or missing values.
Table III Operative characteristics of laparoscopic TA versus PRA in CESQIP (2014e2018)
Operative time, >2 h Annual surgeon volume Annual facility volume Annual facility PRA volume
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
405 (48.6%) 8 (5.5e10.4) 11.4 (8e14.8) 4.6 (0e8)
269 (49.9%) 7.5 (5e10.4) 11.4 (7.6e22.5) 0 (0e5)
136 (46.3%) 8.75 (7e14.6) 9.5 (8e14.8) 8 (7.6e9.5)
.31 <.001 .06 <.001
All operative and facility volumes are provided as medians; interquartile ranges are in parenthesis. Data presented as N (%) unless otherwise specified. Percentages may not add up to 100 due to rounding or missing values.
Fig 1. TA versus PRA outcomes among CESQIP surgeons (2014e2018).
outcomes was examined. These authors found that PRA is feasible for obese patients with comparable perioperative outcomes, but they also found that greater operative time was an independent risk factor for developing postoperative complications after PRA.18 Although CESQIP surgeons selected PRA less frequently for patients with a BMI >40, a patient’s BMI was not a statistically significant independent risk factor for complication in this study. Approximately 1 in 10 adrenalectomies in this study were complicated by capsular disruption, with two-thirds occurring in functional tumors. This rate of capsular disruption is a concerning with a potential risk of tumor recurrence capsule disruption was significantly more likely with a PRA approach. This observation could be explained by anatomic factors with left adrenal glands
typically resting more anteriorly in respect to the superior pole of the kidney, which makes the dissection of left adrenal vein more of a challenge in a retroperitoneal approach and potentially resulting in more excessive tumor manipulation. Additionally, the difference in sex distribution of retroperitoneal fat could affect retroperitoneal dissection. Indeed, we found that male patients undergoing PRA had greater rates of capsular disruption. A previous study has detailed how differences in patient sex affects retroperitoneal fat thickness and is associated with adverse PRA outcomes.19 Capsular disruption is particularly worrisome in operations for known malignancies because tumor spillage jeopardizes oncologic results. Malignant tumors can exhibit a desmoplastic reaction,
A.P. Marrero et al. / Surgery xxx (2019) 1e7
5
Fig 2. Multivariate model for risk of developing any complication, CESQIP (2014e2018). Model adjusted for annual facility volume, nodule size, familial, or syndromes versus sporadic diagnosis, BMI >40 vs 40, operative approach PRA versus TA. Dotted line on the x-axis denotes 1.
Fig 3. Multivariate model for risk of capsular disruption CESQIP (2014e2018).
causing inflammation and adherence of tumors to adjacent tissues with obliteration of surgical planes7; this tissue reaction could further complicate an already space-limited PRA approach. Resecting known malignancies via laparoscopic approach remains a matter of debate. Zografos et al concluded that solitary metastatic adrenal lesions could be safely remove laparoscopically; however, it is controversial
removing primary malignancies due to risk of recurrence, seeding, and/or trocar implants.20 Marangos et al supported the laparoscopic TA approach for adrenal metastasis for the minimally invasive benefits postoperatively, but advocated that it is contraindicated if the tumor invades adjacent vascular structures.21 Simutis et al discussed that PRA has safe and favorable outcomes for adrenal metastasis <6
6
A.P. Marrero et al. / Surgery xxx (2019) 1e7
cm. Results were compared with 13 published studies utilizing a TA approach and found less morbidity and blood loss.22 To our knowledge, our study is the first to perform detailed analysis of capsular disruption as a key operative parameter in laparoscopic TA and PRA. Contrary to the few studies that have nominally reported on capsular disruption, our study of CESQIP surgeons shows that capsular disruption is statistically more likely during PRA, and that this unfavorable event during PRA can occur in tumors as small as 1.2 cm in size. PRA may not be the optimal choice for malignant tumors or those with suspicious features on preoperative imaging. Overall, our results call into question the generalizability of PRA. Limitations of our study include the nature of the data being collected retrospectively and the incompleteness of data entry from participating facilities. As noted, not all facilities perform both procedures, which limits the ability to have an intrainstitutional control for variability due to procedure approach alone. Annual facility volume was included in multivariate analysis to control for operative threshold volume and the effect on the incidence or risk of postoperative complications. The CESQIP database does not have extensive data on patient comorbidities and cannot be accurately risk adjusted. There are no data collected on history of previous abdominal surgery, operative complexity, post analgesia requirements, blood loss, duration of postoperative ileus, and long-term outcomes that could enhance analyses. Conversion to an open procedure is an important endpoint, but due to a low number of events, we were unable to examine this outcome beyond univariate comparisons. Also, our data are limited to suggest that there is a learning curve among CESQIP surgeons based on the operative approach due to rarity of adrenal tumors and the short timeframe of the study (4 years); additional studies with greater study period are needed. This is the largest, multicenter study of high-volume surgeons who routinely performed adrenalectomies in which the adrenalectomy has been stratified by operative approach. Our results indicate that although facilities tend to favor one operative approach over the other, possibly based on their volume and surgeon experience, either operative approach is safe; however, PRA had a greater rate for capsular disruption even for small tumors. Therefore, the decision to choose PRA, particularly for malignant tumors, should be weighed carefully. PRA in experienced hands has low rates of complications, but surgeon experience and patient factors should be considered in deciding whether TA or PRA would be the most appropriate technique, particularly with respect to preventing capsular disruption. Funding/Support The authors have indicated that they have no funding regarding the content of this article. Conflict of interest/Disclosure The authors have nothing to disclose. Our conclusions were based on data extracted from CESQIP. Our discussion and opinions
do not necessarily represent the position of surgeons and facilities participating in CESQIP.
References 1. Elfenbein D, Scarborough J, Speicher P, Scheri RP. Comparison of laparoscopic versus open adrenalectomy results from American College of Surgeons: National Surgery Quality Improvement Project. J Surg Res. 2013;184:2160e2220. 2. Assalia A, Gagner M. Laparoscopic adrenalectomy. Br J Surg. 2004;91: 1259e1274. 3. Coste T, Caiazzo R, Torres F, et al. Laparoscopic adrenalectomy by transabdominal lateral approach 20 years of experience. Surg Endo. 2017;31: 2743e2751. 4. Ariyan C, Strong VE. The current status of laparoscopic adrenalectomy. Adv Surg. 2007;41:133e157. 5. Vrielink OM, Wevers K, Kist JW, et al. Laparoscopic anterior versus endoscopic posterior approach for adrenalectomy: A shift to a new golden standard? Langenbecks Arch Surg. 2017;402:767e773. 6. Mercan S, Seven R, Ozarmagan S. Endoscopic retroperitoneal adrenalectomy. Surgery. 1995;118:1071e1076. 7. Rok Lee C, Walz M, Park S, et al. A comparative study of the transperitoneal and posterior retroperitoneal approaches for laparoscopic adrenalectomy for adrenal tumors. Ann Surg Oncol. 2012;19:2629e2634. 8. Szydelko T, Lewandowski J, Panek W, Tupikowski K, Dembowski J, Zdrojowy R. Laparoscopic adrenalectomy ten-year experience. Central European Journal of Urology. 2012;65:71e74. 9. Walz MK, Peitgen K, Hoermann R, Giebler R, Mann K, Eigler F. Posterior retroperitoneoscopy as a new minimally invasive approach for adrenalectomy: Results of 30 adrenalectomies in 27 patients. World J Surg. 1996;20:769e774. 10. Berber E, Tellioglu G, Harvey A, Mitchell J, Milas M, Siperstein A. Comparison of laparoscopic transabdominal lateral versus posterior retroperitoneal adrenalectomy. Surgery. 2009;146:621e626. 11. Anderson K, Thomas S, Mohamed A, et al. Each procedure matters: Threshold for surgeon volume to minimize complications and decrease cost associated with adrenalectomy. Surgery. 2017;163:157e164. 12. Lindeman B, Hashimoto D, Bababekov Y, et al. Fifteen years of adrenalectomies: Impact of specialty training and operative volume. Surgery. 2018;163: 150e156. 13. Constantinides VA, Touska CP, Palazzo FF. Systematic review and meta-analysis of retroperitoneoscopic versus laparoscopic adrenalectomy. Br J Surg. 2012;99: 1639e1648. 14. Rubinstein M, Gill IS, Aron M, et al. Prospective, randomized comparison of transperitoneal versus retroperitoneal laparoscopic adrenalectomy. J Urol. 2005;174:442e445. 15. Mohammadi-Fallah MR, Mehdizadeh A, Badalzadeh A, et al. Comparison of transperitoneal versus retroperitoneal laparoscopic adrenalectomy in a prospective randomized study. J Laparoendosc Adv Surg Tech A. 2013;23: 362e366. ski M, Konturek A, Nowak W. Randomized clinical trial of posterior 16. Barczyn retroperitoneoscopic adrenalectomy versus lateral transperitoneal laparoscopic adrenalectomy with a 5-year follow-up. Ann Surg. 2014;260:740e748. 17. Zonea P, Buzga M, Ihnat P. Retroperitoneoscopic adrenalectomy in obese patients: Is it suitable? Obes Surg. 2015;25:1203e1208. 18. Hu Q, Hang Z, Ho Y, Sun C, Xu K, Xia G, Ding Q. Impact of obesity on perioperative outcomes of retroperitoneal laparoscopic adrenalectomy. Urologia Internationalis. 2015;95:361e366. 19. Kazuhide M, Ito H, Kawahara T, et al. The impact of gender difference on operative time in laparoscopic partial nephrectomy for T1 renal tumor and the utility of retroperitoneal fat thickness as a predictor of operative time. Biomed Med Central Cancer. 2016;16:944. 20. Zografos GN, Vasiliadis G, Farfaras AN, Aggeli C, Digalakis M. Laparoscopic surgery for malignant adrenal tumors. JSLS. 2009;13:196e202. 21. Marangos I, Kazaryan A, Rosseland A, et al. Should we use laparoscopic adrenalectomy for metastases? Scandinavian multicenterstudy. J Surg Oncol. 2009;100:43e47. 22. Simutis G, Lengvenis G, Beisa V, Strupas K. Endoscopic retroperitoneal adrenalectomy for adrenal metastases. Int J Endocrinol. 2014;2014:806194.
A.P. Marrero et al. / Surgery xxx (2019) 1e7
7
Discussion Dr Richard Hodin (Boston, MA): Were bilateral adrenalectomies excluded or were they included? Dr Annette Marie Pascual Marrero: They were not excluded. We had 50 patients that had bilateral disease. But we did focus on tumor laterality (right or left side), but we did not exclude those patients. There was no significant difference in either group. Dr Richard Hodin (Boston, MA): The number that stood out to me, unless I misunderstood, was that close to 25% of patients had a greater than 48-hour hospital stay after a transabdominal approach. Something is different about those patients. Why would patients stay over 2 days in the hospital after laparoscopic adrenalectomy? Dr Annette Marie Pascual Marrero: With a transabdominal approach, the patients may have a lot of adhesions and require some lysis. Then there is a higher chance of developing ileus. That could be one of the reasons. Pain could be also a reason. Dr Richard Hodin (Boston, MA): So, you do not have actual data on why they stayed that long. Dr Annette Marie Pascual Marrero: I do not have specific data. The variable is just reported as 1, 2, 3, or up to 4 days. So, we analyzed it at 2 days because usually that is what is reported in the literature. Dr Eren Berber (Cleveland Clinic): We know that when we are deciding about the specific laparoscopic approach, we use some other parameters (unfortunately not available in your database), such as body habitus and the relationship of the tumor to the renal hilum. So, I think that is a major limitation. I was surprised that even for small tumors there was more capsular laceration with the posterior approach. Is this really the just reflecting the surgeon learning curve? That was really surprising for me. Dr Annette Marie Pascual Marrero: Thank you for your question and comment. We cannot determine the learning curve
because of the short study period. In the future, if we can get data on the amount of retroperitoneal fat, it will probably help us. If there is a higher rate of capsular disruption in small tumors, it could be related to the amount of retroperitoneal fat. That is something that could be significant. For those patients that had capsular disruption, we did a subanalysis and found that 60% of them had metastatic disease. For those types of tumors, you can encounter a lot of surrounding inflammation and tissue reaction which will increase the risk of rupturing the capsule. Dr Quan-Yang Duh (San Francisco, CA): I think the way that these techniques are taught may actually explain why the capsular disruption rates are different. If you look at the standard retroperitoneal technique, there is a lot of pulling or tearing of the tissues. The use of the transabdominal technique usually involves a more precise dissection. So that may actually be the explanation for differences in capsular disruption. Regarding capsular dissection, you mentioned that we have to be careful of invasive cancer, but we also have to be careful with pheochromocytoma, because the patient can get pheochromocytosis even though it is not a cancer. Were pheochromocytomas associated with capsular disruption as well? Dr Annette Marie Pascual Marrero: There were no significant differences in the type of tumor when we subanalyzed those patients who had capsular disruption. Thank you for your comment about the way we are learning the different techniques. We know that the retroperitoneal approach generally occurs in a smaller dissection space compared to the transabdominal approach, which is often in a wider space. We need to further consider that.
Contents lists available at ScienceDirect
Surgery journal homepage: www.elsevier.com/locate/surg
Patient selection and outcomes of laparoscopic transabdominal versus posterior retroperitoneal adrenalectomy among surgeons in the Collaborative Endocrine Surgery Quality Improvement Program (CESQIP) Annette Pascual Marrero, MD, MPHa,*, Hadiza S. Kazaure, MDa, Samantha M. Thomas, MSb,c, Michael T. Stang, MDa, Randall P. Scheri, MDa a
Department of Surgery, Duke University Medical Center, Durham, NC Department of Biostatistics and Bioinformatics, Duke University, Durham, NC c Duke Cancer Institute, Duke University, Durham, NC b
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 12 March 2019 Available online xxx
Background: Laparoscopic adrenalectomy can be performed using a transabdominal or posterior retroperitoneal approach. Choosing the optimal approach can be challenging. Methods: Using data from the Collaborative Endocrine Surgery Quality Improvement Program (2014 e2018), baseline patient characteristics and outcomes were compared with bivariate methods; univariate and multivariate analyses were used to estimate the association between operative approach and complication risk. Results: Among 833 patients, 35.3% underwent posterior retroperitoneal. Median age was 54 years. Patients undergoing posterior retroperitoneal had lesser rates of body mass index >40 (9.2% vs 17.4%, P ¼ .001), smaller nodules (median 2.4 vs 3.2 cm, P < .001), and more commonly right-sided nodules (46.6% vs 36.9%, P ¼ .02). Posterior retroperitoneal was associated with a lesser rate of conversion to an open procedure (0.7% vs 4.1%, P ¼ .004), less complications (3.1% vs 8.7%, P ¼ .002), and shorter hospital stay (48 h: 92.2% vs 76.6%, P < .001), but a greater rate of capsular disruption (12.6% vs 7.6%, P ¼ .02). For posterior retroperitoneal cases with capsular disruption, median nodule size was 2.2 cm, and 16.2% were metastatic tumors. After multivariate adjustment, posterior retroperitoneal was 2.2 times as likely to result in capsular disruption as transabdominal (95% confidence interval, 1.04e4.79, P ¼ .04). Conclusion: This study revealed a greater rate for capsular disruption during posterior retroperitoneal even for small tumors. Our findings from the Collaborative Endocrine Surgery Quality Improvement Program (2014e2018) suggests that posterior retroperitoneal should be used selectively, especially when a malignancy is suspected. © 2019 Elsevier Inc. All rights reserved.
Introduction Laparoscopic adrenalectomy has emerged as a minimally invasive standard of care for most adrenal pathologies with improved operative outcomes compared to open adrenalectomy.1e4 As surgeons become more experienced with minimally invasive techniques, a variety of operative approaches for laparoscopic
This study was accepted for podium presentation at the 40th annual meeting of the American Association of Endocrine Surgeons, Los Angeles, CA. * Reprint requests: Annette Pascual Marrero, MD, MPH, Department of Surgery, Duke University Medical Center, Box 2945, Durham, NC 27710. E-mail address: [email protected] (A.P. Marrero). https://doi.org/10.1016/j.surg.2019.03.034 0039-6060/© 2019 Elsevier Inc. All rights reserved.
adrenalectomy has emerged. These include transabdominal or posterior retroperitoneal approaches using either classic laparoscopic or robot-assisted techniques. Contrary to the posterior retroperitoneal approach to adrenalectomy (PRA), in which the operative space is small and requires surgeon familiarity with unconventional visualization of anatomic structures, the adrenal gland can be accessed transabdominally (TA) in a wider working space and with conventional anatomic landmarks. Several studies indicate that PRA is beneficial in patients with previous abdominal operations and those with bilateral adrenal pathologies.4e8 Existing literature provides conflicting results when postoperative outcomes of PRA are compared to TA. Rok Lee et al have
2
A.P. Marrero et al. / Surgery xxx (2019) 1e7
reported that PRA is associated with earlier recovery of bowel function, possibly leading to a shorter hospital stay and earlier recovery.7 In their study, Walz et al found that PRA resulted in a lesser need for postoperative analgesia and greater rates of early mobilization.9 Berber et al showed similar mortality and conversion rates and a comparable duration of hospital stay. Among patients who underwent PRA, a lesser duration of surgery and less blood loss were observed; overall, the authors concluded that TA and PRA have similar outcomes. They recommended PRA as the most favorable approach for patients with small tumors, lesser body mass index (BMI), bilateral adrenal pathologies, and those with a history of abdominal operations.10 At a multi-institutional level, the optimal laparoscopic strategy based on adrenal tumor size, diagnosis, and laterality, is unclear. The Collaborative Endocrine Surgery Quality Improvement Program (CESQIP) is an interinstitutional collaborative effort among endocrine surgeons to gather patient data for tracking, collection, and analysis. Surgeons receive feedback and access clinical data to improve quality of service and outcomes and decrease costs. The dataset provides the power and surgical and clinical details in addition to a cross-section of high-volume surgeons necessary to better characterize factors associated with choice of operative approach of adrenalectomy and postoperative outcomes. The aims of this study were to evaluate the baseline and clinical characteristics and outcomes of laparoscopic adrenalectomies among CESQIP surgeons based on operative approach (TA versus PRA) and to compare risks associated with each technique. Approximately 37 participating sites reported adrenalectomies outcomes in CESQIP. Methods All patients who underwent laparoscopic TA or PRA between January 1, 2014, and April 26, 2018, captured in the CESQIP adrenalectomy database were included in this study. Patients <18 years old and those who underwent remedial adrenalectomy were excluded. Baseline characteristics included age, race, sex, BMI, hypertension, and diabetes mellitus. The CESQIP database classifies BMI ( or >40 kg/m2) as a binary categorical variable. Preoperative characteristics included familial versus sporadic adrenal pathology, nodule size, tumor type, Cushing’s classification, and metastatic disease type. Familial or genetic syndromes included multiple endocrine neoplasia (MEN types 1 and 2), neurofibromatosis (type 1), succinate dehydrogenase mutation, and Von Hippel Lindau syndrome. Operative variables were surgeon and facility volume, duration of the operation, and tumor laterality. Outcomes were occurrence of capsular disruption, conversion to an open procedure, incidence of any complication (visceral injury, hemorrhage requiring transfusion, postoperative bleeding requiring transfusion, unplanned return to operating room, surgical site infection, pulmonary embolism, deep vein thrombosis, sepsis, myocardial infarction, cardiac arrest, urinary tract infection, acute renal failure, pneumonia, ventilator >48 h, or hypertensive crisis), presence of adrenal insufficiency, and duration of hospital stay. In CESQIP, capsular disruption is captured from intraoperative data and was defined as evidence of gross macroscopic intraoperative disruption of the gland’s capsule. Because of its rarity, complications were analyzed as a categorical variable based on the occurrence of 1 of any of the complications listed above. Categorical variables were reported as N (%) and continuous variables as median (interquartile range, IQR). Differences between study groups were tested using the c2 test or Fisher exact test or t test, as appropriate. For nodule size, patients indicated as not having a measurable nodule by preoperative imaging were defined as size 0 cm. Univariate logistic regression was utilized to estimate
the effect of nodule size, annual surgeon, and facility volume, primary diagnosis, tumor location, BMI, and operative approach on incidence of any (1) complication and occurrence of capsular disruption. Annual surgeon volume was examined both as a continuous variable and after classification into 6 versus >6 adrenalectomies per year, based on existing literature for adrenal surgical volume.11 Because no differences were seen in modeling results for surgeon volume, only continuous values are reported. Multivariate logistic regression was used to estimate the effect of the approach to adrenalectomy on incidence of any complication and occurrence of capsular disruption after adjustment for clinically significant variables. Odds ratios (OR) and 95% confidence intervals (CI) are reported. All logistic regression models were built using the generalized estimating equations framework, and an exchangeable correlation structure was used to account for the correlation of patients treated by the same surgeon. A subgroup analysis for patients with and without capsular disruption was also conducted. No adjustments were made for multiple comparisons. All statistical analyses were conducting using SAS version 9.4 (SAS Institute, Cary, NC). Results A total of 1,145 patients were included in the CESQIP adrenalectomy database during this 4-year period. Of those, 833 (72.8%) patients underwent either laparoscopic TA or PRA and met all other inclusion criteria. From those, 539 (64.7%) underwent laparoscopic TA and 294 (35.3%) underwent PRA. Patients who underwent PRA were more likely to be older, non-black, and have a BMI 40 (all P < .05, Table I). They also were more likely to have smaller nodules (median 2.4 vs 3.2 cm, P < .001). The most prevalent diagnoses for patients who underwent PRA were: hyperaldosteronism (29.9%), pheochromocytoma (22.1%), and nonfunctioning nodule (19.0%). On the other hand, those patients who underwent TA had most commonly the following diagnoses: Cushing’s syndrome or disease (39.3%), pheochromocytoma (27.1%), and hyperaldosteronism (23.6%) among patients who underwent TA (Table II). The majority of primary diagnoses were sporadic (95.2%), whereas few were associated with a genetic mutation or syndrome (4.8%), including MEN1 (0.1%), MEN2 (1.6%), neurofibromatosis-1 (1%), succinate dehydrogenase (0.4%), von Hippel Lindau syndrome (0.5%), and other (0.8%). When compared to TA, patients who underwent PRA were more likely to have right-sided nodules (46.6% vs 36.9%, P ¼ .02). No significant difference was noted in the duration of the procedure by operative approach (P ¼ .31). The median nodule size for all patients was 2.8 cm (IQR 1.5e4.2, Table II). For patients with Cushing’s syndrome, 50.9% presented with clinical manifestations and 31.0% with subclinical diagnosis. Classification is unknown for the remainder of patients. A total of 5.2% (N ¼ 43) of all adrenalectomies were performed for adrenal metastasis; of these, 16 (37.2%) underwent PRA. Three patients with a preoperative diagnosis of adrenocortical carcinoma were identified, one of whom underwent PRA. A total of 34 facilities that performed at least one TA or PRA were included in our study; 21 (61.8%) of them performed only TAs, 4 (11.8%) performed only PRAs, and 9 (26.5%) performed both TAs and PRAs. Of the 60 CESQIP surgeons, in this study, the median number of PRAs per surgeon was 8.8 (IQR 7.0e14.6) and TAs was 7.5 (IQR 5.0e10.4) per year (P < .001). No difference was identified in facility volume for PRA versus TA (median 9.5 vs 11.4, P ¼ .06, Table III). Figure 1 compares the outcomes of both approaches. In total, 6.7% of patients experienced a complication, notably visceral injury (1.2%), hemorrhage requiring intraoperative transfusion (1.2%), adrenal insufficiency (0.8%), and hypertensive crisis (0.8%). There
A.P. Marrero et al. / Surgery xxx (2019) 1e7
3
Table I Baseline and clinical characteristics of patients who underwent laparoscopic TA or PRA in CESQIP (2014e2018)
Age (y), median (IQR) Race Non-black Black Asian Hispanic Unknown Sex Female BMI >40 Hypertension Diabetes
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
54 (43e63)
53 (41e63)
56 (47e65)
.002
562 (67.5%) 87 (10.4%) 18 (2.2%) 116 (13.9%) 46 (5.5%)
328 (60.9%) 61 (11.3%) 15 (2.8%) 95 (17.6%) 38 (7.1%)
234 (79.6%) 26 (8.8%) 3 (1.0%) 21 (7.1%) 8 (2.7%)
526 (63.1%)
353 (65.5%)
173 (58.8%)
121 (14.5%) 607 (72.9%) 213 (25.6%)
94 (17.4%) 387 (71.8%) 131 (24.3%)
27 (9.2%) 220 (74.8%) 82 (27.9%)
<.001
.06 .001 .31 .25
Data presented as N (%) unless otherwise specified. Percentages may not add up to 100 due to rounding or missing values.
was no statistical difference in complication rates; however, PRA was associated with lesser incidence rate of any (1) complication (3.1% vs 8.7%, P ¼ .002). After multivariate adjustment, the operative approach no longer was associated with complication risk (OR ¼ 0.50; 95% CI, 0.23e1.09, P ¼ .08, Fig 2). Subanalysis of cases that resulted in capsular disruption among PRA versus TA was performed. The median age of these patients was 55 years, similar to the full cohort. In this group, however, nonblack (81.8% vs 48.8%, P ¼ .005) and male patients (56.8% vs 29.3%, P ¼ .01) more commonly underwent PRA. Most patients who underwent PRA with capsular disruption had a BMI 40 (94.6% vs 78%, P ¼ .05). Among this group, the median nodule size was 2.2 cm on PRA vs 4.2 cm on TA (P < .001). For both groups, the most prevalent diagnoses were hyperfunctioning tumors (62.2% and 65.9% for PRA and TA respectively, P ¼ .39). Among PRA, hyperaldosteronism (32.4%), nonfunctioning nodule (16.2%), and metastatic disease (16.2%) were the most common tumor types compared to TA, where the most common types were hyperaldosteronism (24.4%), pheochromocytoma (24.4%), and Cushing’s syndrome (14.6%). No differences were seen in duration of operation (<2 hours, 51.4% vs 68.3% P ¼ .17), tumor laterality (left: 56.8% vs 43.9%, right: 37.8% vs 51.2%, P ¼ .43), annual surgeon volume (median 7.2 vs 10.3 cases per year, P ¼ .67), or facility volume (median 9.5 vs 9.3 cases per year, P ¼ .80) when comparing PRA versus TA on this cohort. Left sided tumors (50%) were more common among patients with capsular disruption. Only 3 patients who had capsular disruption and underwent TA had conversion to open procedure. In univariate and the multivariate analyses, PRA was associated with a greater risk of developing capsular disruption. After adjusting for covariates, patients who underwent PRA had 2 times the odds of developing capsular disruption compared to those who underwent TA (OR ¼ 2.23; 95% CI 1.04e4.79, P ¼ .04, Fig 3). All patients who underwent PRA (N ¼ 289) were classified into those with (12.8%) and without (87.2%) capsular disruption. Both groups had similar clinical characteristics without any difference on age (P ¼ .42), BMI (P ¼ .55), tumor size (P ¼ .32), laterality (P ¼ .51), or diagnosis (P ¼ .18). Capsular disruption was more prevalent among men (56.8% vs 38.9%, P ¼ .04). A total of 16.2% of patients who underwent PRA with capsular disruption had metastatic disease compared to 3.9% without capsular disruption (P ¼ .18). In addition, no difference was found in the rates of capsular disruption within institutions (P ¼ .39). In the facilities where PRA was only performed, 12.8% had capsular disruption (N ¼ 109) compared to facilities performing TA only where 8.4% had capsular disruption (N ¼ 285). In the institutions in which both approaches were performed, the rate of capsular disruption was 9.1% (N ¼ 439).
Discussion As surgeons become proficient with advanced minimally invasive techniques, the criteria for choosing the optimal approach for adrenalectomy continues to evolve. In this study, the most common approach for laparoscopic adrenalectomy among CESQIP surgeons was TA. This approach was favored for tumors that were larger, located on the left side, and in patients who had a BMI >40; diagnosis did not affect operative approach. Facilities where a TA approach predominated performed slightly more cases annually compared to those that performed mostly a PRA approach. Overall, annual surgeon volume was not associated with the risk of complications. This observation could be explained by CESQIP surgeons representing high volume surgeons whose median number of adrenalectomies is likely above the high-volume threshold advocated in existing literature11; however, surgeon volume has been associated previously with fewer postoperative complications after adrenalectomies. Anderson et al documented that surgeons performing 6 cases yearly have better outcomes compared to low volume surgeons.11 Lindeman et al also showed that high volume surgeons (4 procedures per year) have lesser complication and mortality rates compared to low volume surgeons (P < .001 and P ¼ .04).12 Several studies that have compared PRA versus TA showed no significant outcomes differences, including complications, operative time, postoperative pain control, blood loss, and duration of hospital stay.10,13,14 Mohammadi-Fallah et al showed that PRA had a lesser duration of postoperative ileus and recovery period, while Rubinstein et al showed no difference in oral intake tolerance.14,15 ski et al concluded that PRA was superior to TA in earlier Barczyn recovery, lesser operative time, less postoperative analgesia requirements and blood loss. In this study, PRA was associated with a lesser rate of conversion to an open procedure and a lesser hospital stay when compared to TA. There were no differences in operative time and overall complication risk; however, there is likely a strong selection bias for a given approach evidenced by the difference in patient clinical characteristics between both procedures.16 Of concern, there was a greater risk of capsular disruption with PRA (12.6% vs 7.6%, P ¼ .02). Based on these findings, our study neither supports the equivalence of TA to PRA nor the superiority of one operative approach over the other. Larger studies would be needed to provide conclusive results. Our study showed that PRA was performed most commonly on patients with BMI 40. Zonea et al found that PRA is a safe alternative procedure for obese patients (BMI >30 kg/m2); however, operative time, hospital stay, morbidity, and conversion rates were greater.17 In a study by Hu et al, the impact of obesity on PRA
4
A.P. Marrero et al. / Surgery xxx (2019) 1e7 Table II Clinical characteristics of patients who underwent laparoscopic TA or PRA in CESQIP (2014e2018)
Nodule size (cm), median (IQR) Diagnosis Hyperfunctioning tumor Hyperaldosteronism Pheochromocytoma Cushing's Adrenocortical carcinoma Nonfunctioning tumor Nonfunctioning nodule Metastatic disease Tumor laterality Left Right Bilateral
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
2.8 (1.5e4.2)
3.2 (1.6e4.5)
2.4 (1.4e3.6)
<.001 .65
600 (72.0%) 215 (25.8%) 211 (55.1%) 171 (20.5%) 3 (0.4%) 193 (23.2%) 150 (18.0%) 43 (5.2%)
387 (71.8%) 127 (23.6%) 146 (27.1%) 112 (39.3%) 2 (0.4%) 121 (22.4%) 94 (17.4%) 27 (5.0%)
213 (72.4%) 88 (29.9%) 65 (22.1%) 59 (20.1%) 1 (0.3%) 72 (24.5%) 56 (19.0%) 16 (5.4%)
442 (53.1%) 336 (40.3%) 50 (6%)
303 (56.2%) 199 (36.9%) 32 (5.9%)
139 (47.3%) 137 (46.6%) 18 (6.1%)
.02
Data presented as N (%) unless otherwise specified. Cushing’s (disease or syndrome). Percentages may not add up to 100 due to rounding or missing values.
Table III Operative characteristics of laparoscopic TA versus PRA in CESQIP (2014e2018)
Operative time, >2 h Annual surgeon volume Annual facility volume Annual facility PRA volume
All patients (N ¼ 833)
Transabdominal (N ¼ 539)
Retroperitoneal (N ¼ 294)
P value
405 (48.6%) 8 (5.5e10.4) 11.4 (8e14.8) 4.6 (0e8)
269 (49.9%) 7.5 (5e10.4) 11.4 (7.6e22.5) 0 (0e5)
136 (46.3%) 8.75 (7e14.6) 9.5 (8e14.8) 8 (7.6e9.5)
.31 <.001 .06 <.001
All operative and facility volumes are provided as medians; interquartile ranges are in parenthesis. Data presented as N (%) unless otherwise specified. Percentages may not add up to 100 due to rounding or missing values.
Fig 1. TA versus PRA outcomes among CESQIP surgeons (2014e2018).
outcomes was examined. These authors found that PRA is feasible for obese patients with comparable perioperative outcomes, but they also found that greater operative time was an independent risk factor for developing postoperative complications after PRA.18 Although CESQIP surgeons selected PRA less frequently for patients with a BMI >40, a patient’s BMI was not a statistically significant independent risk factor for complication in this study. Approximately 1 in 10 adrenalectomies in this study were complicated by capsular disruption, with two-thirds occurring in functional tumors. This rate of capsular disruption is a concerning with a potential risk of tumor recurrence capsule disruption was significantly more likely with a PRA approach. This observation could be explained by anatomic factors with left adrenal glands
typically resting more anteriorly in respect to the superior pole of the kidney, which makes the dissection of left adrenal vein more of a challenge in a retroperitoneal approach and potentially resulting in more excessive tumor manipulation. Additionally, the difference in sex distribution of retroperitoneal fat could affect retroperitoneal dissection. Indeed, we found that male patients undergoing PRA had greater rates of capsular disruption. A previous study has detailed how differences in patient sex affects retroperitoneal fat thickness and is associated with adverse PRA outcomes.19 Capsular disruption is particularly worrisome in operations for known malignancies because tumor spillage jeopardizes oncologic results. Malignant tumors can exhibit a desmoplastic reaction,
A.P. Marrero et al. / Surgery xxx (2019) 1e7
5
Fig 2. Multivariate model for risk of developing any complication, CESQIP (2014e2018). Model adjusted for annual facility volume, nodule size, familial, or syndromes versus sporadic diagnosis, BMI >40 vs 40, operative approach PRA versus TA. Dotted line on the x-axis denotes 1.
Fig 3. Multivariate model for risk of capsular disruption CESQIP (2014e2018).
causing inflammation and adherence of tumors to adjacent tissues with obliteration of surgical planes7; this tissue reaction could further complicate an already space-limited PRA approach. Resecting known malignancies via laparoscopic approach remains a matter of debate. Zografos et al concluded that solitary metastatic adrenal lesions could be safely remove laparoscopically; however, it is controversial
removing primary malignancies due to risk of recurrence, seeding, and/or trocar implants.20 Marangos et al supported the laparoscopic TA approach for adrenal metastasis for the minimally invasive benefits postoperatively, but advocated that it is contraindicated if the tumor invades adjacent vascular structures.21 Simutis et al discussed that PRA has safe and favorable outcomes for adrenal metastasis <6
6
A.P. Marrero et al. / Surgery xxx (2019) 1e7
cm. Results were compared with 13 published studies utilizing a TA approach and found less morbidity and blood loss.22 To our knowledge, our study is the first to perform detailed analysis of capsular disruption as a key operative parameter in laparoscopic TA and PRA. Contrary to the few studies that have nominally reported on capsular disruption, our study of CESQIP surgeons shows that capsular disruption is statistically more likely during PRA, and that this unfavorable event during PRA can occur in tumors as small as 1.2 cm in size. PRA may not be the optimal choice for malignant tumors or those with suspicious features on preoperative imaging. Overall, our results call into question the generalizability of PRA. Limitations of our study include the nature of the data being collected retrospectively and the incompleteness of data entry from participating facilities. As noted, not all facilities perform both procedures, which limits the ability to have an intrainstitutional control for variability due to procedure approach alone. Annual facility volume was included in multivariate analysis to control for operative threshold volume and the effect on the incidence or risk of postoperative complications. The CESQIP database does not have extensive data on patient comorbidities and cannot be accurately risk adjusted. There are no data collected on history of previous abdominal surgery, operative complexity, post analgesia requirements, blood loss, duration of postoperative ileus, and long-term outcomes that could enhance analyses. Conversion to an open procedure is an important endpoint, but due to a low number of events, we were unable to examine this outcome beyond univariate comparisons. Also, our data are limited to suggest that there is a learning curve among CESQIP surgeons based on the operative approach due to rarity of adrenal tumors and the short timeframe of the study (4 years); additional studies with greater study period are needed. This is the largest, multicenter study of high-volume surgeons who routinely performed adrenalectomies in which the adrenalectomy has been stratified by operative approach. Our results indicate that although facilities tend to favor one operative approach over the other, possibly based on their volume and surgeon experience, either operative approach is safe; however, PRA had a greater rate for capsular disruption even for small tumors. Therefore, the decision to choose PRA, particularly for malignant tumors, should be weighed carefully. PRA in experienced hands has low rates of complications, but surgeon experience and patient factors should be considered in deciding whether TA or PRA would be the most appropriate technique, particularly with respect to preventing capsular disruption. Funding/Support The authors have indicated that they have no funding regarding the content of this article. Conflict of interest/Disclosure The authors have nothing to disclose. Our conclusions were based on data extracted from CESQIP. Our discussion and opinions
do not necessarily represent the position of surgeons and facilities participating in CESQIP.
References 1. Elfenbein D, Scarborough J, Speicher P, Scheri RP. Comparison of laparoscopic versus open adrenalectomy results from American College of Surgeons: National Surgery Quality Improvement Project. J Surg Res. 2013;184:2160e2220. 2. Assalia A, Gagner M. Laparoscopic adrenalectomy. Br J Surg. 2004;91: 1259e1274. 3. Coste T, Caiazzo R, Torres F, et al. Laparoscopic adrenalectomy by transabdominal lateral approach 20 years of experience. Surg Endo. 2017;31: 2743e2751. 4. Ariyan C, Strong VE. The current status of laparoscopic adrenalectomy. Adv Surg. 2007;41:133e157. 5. Vrielink OM, Wevers K, Kist JW, et al. Laparoscopic anterior versus endoscopic posterior approach for adrenalectomy: A shift to a new golden standard? Langenbecks Arch Surg. 2017;402:767e773. 6. Mercan S, Seven R, Ozarmagan S. Endoscopic retroperitoneal adrenalectomy. Surgery. 1995;118:1071e1076. 7. Rok Lee C, Walz M, Park S, et al. A comparative study of the transperitoneal and posterior retroperitoneal approaches for laparoscopic adrenalectomy for adrenal tumors. Ann Surg Oncol. 2012;19:2629e2634. 8. Szydelko T, Lewandowski J, Panek W, Tupikowski K, Dembowski J, Zdrojowy R. Laparoscopic adrenalectomy ten-year experience. Central European Journal of Urology. 2012;65:71e74. 9. Walz MK, Peitgen K, Hoermann R, Giebler R, Mann K, Eigler F. Posterior retroperitoneoscopy as a new minimally invasive approach for adrenalectomy: Results of 30 adrenalectomies in 27 patients. World J Surg. 1996;20:769e774. 10. Berber E, Tellioglu G, Harvey A, Mitchell J, Milas M, Siperstein A. Comparison of laparoscopic transabdominal lateral versus posterior retroperitoneal adrenalectomy. Surgery. 2009;146:621e626. 11. Anderson K, Thomas S, Mohamed A, et al. Each procedure matters: Threshold for surgeon volume to minimize complications and decrease cost associated with adrenalectomy. Surgery. 2017;163:157e164. 12. Lindeman B, Hashimoto D, Bababekov Y, et al. Fifteen years of adrenalectomies: Impact of specialty training and operative volume. Surgery. 2018;163: 150e156. 13. Constantinides VA, Touska CP, Palazzo FF. Systematic review and meta-analysis of retroperitoneoscopic versus laparoscopic adrenalectomy. Br J Surg. 2012;99: 1639e1648. 14. Rubinstein M, Gill IS, Aron M, et al. Prospective, randomized comparison of transperitoneal versus retroperitoneal laparoscopic adrenalectomy. J Urol. 2005;174:442e445. 15. Mohammadi-Fallah MR, Mehdizadeh A, Badalzadeh A, et al. Comparison of transperitoneal versus retroperitoneal laparoscopic adrenalectomy in a prospective randomized study. J Laparoendosc Adv Surg Tech A. 2013;23: 362e366. ski M, Konturek A, Nowak W. Randomized clinical trial of posterior 16. Barczyn retroperitoneoscopic adrenalectomy versus lateral transperitoneal laparoscopic adrenalectomy with a 5-year follow-up. Ann Surg. 2014;260:740e748. 17. Zonea P, Buzga M, Ihnat P. Retroperitoneoscopic adrenalectomy in obese patients: Is it suitable? Obes Surg. 2015;25:1203e1208. 18. Hu Q, Hang Z, Ho Y, Sun C, Xu K, Xia G, Ding Q. Impact of obesity on perioperative outcomes of retroperitoneal laparoscopic adrenalectomy. Urologia Internationalis. 2015;95:361e366. 19. Kazuhide M, Ito H, Kawahara T, et al. The impact of gender difference on operative time in laparoscopic partial nephrectomy for T1 renal tumor and the utility of retroperitoneal fat thickness as a predictor of operative time. Biomed Med Central Cancer. 2016;16:944. 20. Zografos GN, Vasiliadis G, Farfaras AN, Aggeli C, Digalakis M. Laparoscopic surgery for malignant adrenal tumors. JSLS. 2009;13:196e202. 21. Marangos I, Kazaryan A, Rosseland A, et al. Should we use laparoscopic adrenalectomy for metastases? Scandinavian multicenterstudy. J Surg Oncol. 2009;100:43e47. 22. Simutis G, Lengvenis G, Beisa V, Strupas K. Endoscopic retroperitoneal adrenalectomy for adrenal metastases. Int J Endocrinol. 2014;2014:806194.
A.P. Marrero et al. / Surgery xxx (2019) 1e7
7
Discussion Dr Richard Hodin (Boston, MA): Were bilateral adrenalectomies excluded or were they included? Dr Annette Marie Pascual Marrero: They were not excluded. We had 50 patients that had bilateral disease. But we did focus on tumor laterality (right or left side), but we did not exclude those patients. There was no significant difference in either group. Dr Richard Hodin (Boston, MA): The number that stood out to me, unless I misunderstood, was that close to 25% of patients had a greater than 48-hour hospital stay after a transabdominal approach. Something is different about those patients. Why would patients stay over 2 days in the hospital after laparoscopic adrenalectomy? Dr Annette Marie Pascual Marrero: With a transabdominal approach, the patients may have a lot of adhesions and require some lysis. Then there is a higher chance of developing ileus. That could be one of the reasons. Pain could be also a reason. Dr Richard Hodin (Boston, MA): So, you do not have actual data on why they stayed that long. Dr Annette Marie Pascual Marrero: I do not have specific data. The variable is just reported as 1, 2, 3, or up to 4 days. So, we analyzed it at 2 days because usually that is what is reported in the literature. Dr Eren Berber (Cleveland Clinic): We know that when we are deciding about the specific laparoscopic approach, we use some other parameters (unfortunately not available in your database), such as body habitus and the relationship of the tumor to the renal hilum. So, I think that is a major limitation. I was surprised that even for small tumors there was more capsular laceration with the posterior approach. Is this really the just reflecting the surgeon learning curve? That was really surprising for me. Dr Annette Marie Pascual Marrero: Thank you for your question and comment. We cannot determine the learning curve
because of the short study period. In the future, if we can get data on the amount of retroperitoneal fat, it will probably help us. If there is a higher rate of capsular disruption in small tumors, it could be related to the amount of retroperitoneal fat. That is something that could be significant. For those patients that had capsular disruption, we did a subanalysis and found that 60% of them had metastatic disease. For those types of tumors, you can encounter a lot of surrounding inflammation and tissue reaction which will increase the risk of rupturing the capsule. Dr Quan-Yang Duh (San Francisco, CA): I think the way that these techniques are taught may actually explain why the capsular disruption rates are different. If you look at the standard retroperitoneal technique, there is a lot of pulling or tearing of the tissues. The use of the transabdominal technique usually involves a more precise dissection. So that may actually be the explanation for differences in capsular disruption. Regarding capsular dissection, you mentioned that we have to be careful of invasive cancer, but we also have to be careful with pheochromocytoma, because the patient can get pheochromocytosis even though it is not a cancer. Were pheochromocytomas associated with capsular disruption as well? Dr Annette Marie Pascual Marrero: There were no significant differences in the type of tumor when we subanalyzed those patients who had capsular disruption. Thank you for your comment about the way we are learning the different techniques. We know that the retroperitoneal approach generally occurs in a smaller dissection space compared to the transabdominal approach, which is often in a wider space. We need to further consider that.