Check for updates
Perioperative morbidity and mortality in dogs with invasive adrenal neoplasms treated by adrenalectomy and cavotomy
Philipp D. Mayhew BVM&S, DACVS1 | Sarah E. Boston DVM, DACVS2 | Allison L. Zwingenberger DVM, MAS, DACVR, DECVDI1 | Michelle A. Giuffrida VMD, DACVS1 ® | Jeffrey J. Runge VMD, DACVS3 | David E. Holt BVSc, DACVS3 | Joseph S. Raleigh DVM1 | Ameet Singh DVM, DVSc, DACVS4 | William T. N. Culp VMD, DACVS1 ® | J. Brad Case DVM, DACVS5 | Michele A. Steffey DVM, DACVS1 | Ingrid M. Balsa DVM, DACVS-SA1
1School of Veterinary Medicine, University of California-Davis, Davis, California
2VCA Canada, 404 Veterinary Emergency and Referral Hospital, Newmarket, Ontario, Canada
3Matthew J. Ryan Veterinary Hospital at the University of Pennsylvania, Philadelphia, Pennsylvania
4Departments of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, Ontario, Canada
5School of Veterinary Medicine, University of Florida, Gainesville, Florida
Correspondence
Philipp D. Mayhew, Department of Surgical and Radiological Sciences, University of California-Davis, One Shields Road, Davis, CA 95616. Email: philmayhew@gmail.com
Abstract
Objective: To report the morbidity and mortality associated with adrenalectomy with cavotomy for resection of invasive adrenal neoplasms in dogs and evaluate risk factors for perioperative outcomes.
Study design: Retrospective study.
Animals: Forty-five client-owned dogs.
Methods: Dogs that underwent open adrenalectomy with cavotomy for resection of adrenal masses with tumor thrombus extending into the vena cava were included. Clinicopathologic data were harvested from medical records. Selected clinical, imaging, and operative variables were statistically evaluated as risk factors for packed red blood cell transfusion, nephrectomy, perioperative death, and overall survival.
Results: Thirty-six of 45 masses were pheochromocytomas, 7 were adrenocortical carcinomas, and 2 were unknown type. Caval thrombus terminated prehepatically in 21 of 45 dogs and extended beyond the porta hepatis but terminated prediaphragmatically (intrahepatic prediaphragmatic location) in 15 dogs and thrombi extended postdiaphragmatically in 5 dogs. Thirty-four (76%) dogs were discharged from the hospital, and 11 (24%) dogs died or were euthanized prior to discharge. Median overall survival time for all 45 dogs was 547 days (95%CI 146-710). Bodyweight, tumor type, and size and extent of caval thrombus did not affect survival to discharge, but postdiaphragmatic (rather than prediaphragmatic) thrombus termination was associated with a greater risk of death.
Conclusion: Long-term survival was common in dogs that survived the periopera- tive period. Postdiaphragmatic thrombus extension affected the prognosis for over- all survival.
Clinical significance: Findings of this study help to stratify operative risk in dogs with adrenal neoplasia and caval invasion.
1 INTRODUCTION |
Approximately 20%-48% of adrenal neoplasms exhibit vas- cular invasion into the vena cava, phrenicoabdominal veins, or renal vasculature, with pheochromocytomas more likely than adrenocortical tumors to invade.1-9 Outcome data in the veterinary literature for dogs with invasive adrenal tumors are limited to small cohorts of dogs.2,3,5,7-9 Debate exists regarding whether vena caval invasion increases the risk of morbidity and mortality. In 1 study that included 10 dogs requiring cavotomy for thrombus removal, there was no dif- ference in mortality between invasive and noninvasive adre- nal masses; 69% and 81% of dogs with and without tumor thrombi in the caudal vena cava, respectively, survived the perioperative period.3 However, in a more recent study that included 14 dogs with caval thrombi, those with invasion had a significantly increased risk of perioperative mortality.9 In that study only 4 of 14 (28%) dogs requiring a cavotomy for thrombectomy survived to 14 days postoperatively, and mortality was 100% (7/7) in dogs that had caval thrombi extending cranial to the liver.9 Care should be taken in com- paring these 2 canine cohorts because they contained differ- ent populations of dogs with differing degrees of tumor thrombus extension. Data on larger number of dogs under- going this procedure are required to provide accurate prog- nostic information to allow owners to make informed decisions regarding care of their dogs.
The objective of this study was to report morbidity, mor- tality, and complications in a cohort of dogs with invasive adrenal neoplasms that underwent adrenalectomy with cav- otomy. A secondary objective was to evaluate risk factors for perioperative and long-term outcomes in dogs undergo- ing the procedure.
2 MATERIALS AND METHODS |
2.1 | Case selection
Dogs that underwent open adrenalectomy with a venotomy for resection of tumor thrombus within the vena cava at 4 dif- ferent academic institutions between March 2, 2004 and April 11, 2017 were included in the study.
2.2 | Diagnostic evaluation
Clinicopathological data were recorded for each dog. Data recorded included history, clinical signs, physical examination,
and the results of laboratory testing prior to surgery. Diagnos- tic imaging findings including the results of plain thoracic radi- ography and abdominal ultrasonography were recorded from the medical record. Results from all computed tomography angiography (CTA) performed prior to surgery were obtained and were read by 1 board-certified radiologist (ALZ). Data from diagnostic imaging studies that were recorded included side affected, mass mineralization, maximal tumor diameter, tumor volume excluding thrombus, volume of thrombus, max- imum tumor length (caval component), maximal thrombus width (caval component), and termination point of the caval thrombus. The termination of the caval thrombus was classified from results of CTA or from ultrasound reports or images that were available for review as prehepatic (with ter- mination of thrombus caudal to porta hepatis), intrahepatic prediaphragmatic (with termination of thrombus cranial to porta hepatis but caudal to diaphragm), or postdiaphragmatic (with termination of thrombus cranial to diaphragm). In cases in which CTA was performed, 3 ratios were calculated: ratio of maximum thrombus width to normal caval width (measured at a location just cranial to the termination of the tumor throm- bus), maximal thrombus length to the length of the L3 verte- bra, and maximum thrombus width to the length of the L3 vertebra.
2.3 | Anesthesia and surgery
All dogs underwent a ventral midline celiotomy, with anes- thesia protocol determined by the attending anesthesiologist. In addition, some dogs underwent either a right or a left paracostal incision or a caudal sternotomy to improve visual- ization of the surgical site. Dogs underwent a variety of dif- ferent adrenalectomy techniques that broadly followed previously published techniques.1º The primary tumor was always resected from its attachments to surrounding tissues first until only the stalk of the thrombus in the phrenicoabdominal vein remained attached. Briefly, Rummel tourniquet’s made of umbilical tape, silastic vas- cular loops, or suture material were placed around the vena cava cranial and caudal to the extent of the tumor throm- bus when indicated. In addition, tourniquets were placed around 1 or both renal veins. In cases with tumor thrombus that extended beyond the porta hepatis, Rummel tourni- quets were placed cranial to the diaphragm, on the thoracic caudal vena cava, and/or on the thoracic descending aorta. After inflow occlusion to the vascular segment containing the thrombus had been achieved by tightening the relevant
tourniquets, an incision large enough to extract the throm- bus was made into the vena cava and/or renal veins. In some cases, after thrombectomy was complete, a Satinsky vascular clamp was placed across the defect in the vessel wall, and tourniquets were loosened to allow partial flow in the vena cava to be reestablished. However, in other cases, inflow occlusion was maintained until the defect in the vena cava/renal vein had been closed. The defect in the vessel wall was closed primarily by suturing. Other indi- cated surgical procedures were performed as required. Sur- gical time was recorded in all cases from the time of skin incision to completion of closure.
2.4 | Outcome assessment
Number and type of blood product transfusions were recorded. Intraoperative and postoperative complications and long-term outcome were recorded from the medical record or by telephone calls with the owners. Owners were asked to confirm whether their dog was alive or dead and whether any recurrence of adrenal-related disease had occurred since the time of surgery. If a dog had died, the owner was asked whether cause of death was related to adre- nal disease.
2.5 | Statistical analysis
Descriptive statistics were calculated. Distribution of con- tinuous variables was assessed with the Shapiro-Wilk test. Categorical and continuous tumor variables were compared according to tumor type with Fisher’s exact or x2 tests, and Wilcoxon rank sum tests, respectively. Simple logistic regression was used to examine possible risk factors for receiving a packed red blood cell transfusion, having a nephrectomy, and surviving to hospital discharge. Kaplan- Meier product limit method was used to estimate overall survival among all dogs and in the subgroup of dogs that survived to hospital discharge; dogs alive or lost to follow- up were censored at their last known dates alive. Multivari- able Cox regression was used to model prognostic factors associated with overall survival. Univariable analyses were performed, and covariates were retained in the multivari- able model by using a backward selection approach if their likelihood ratio test or Wald test P values were ≤.05 or if covariates were identified as confounders (defined as >15% change in hazard ratio). Proportional hazards assumptions were evaluated with Shoenfeld residuals by using the score test, with P > .05 considered acceptable. All tests were 2-sided, and P < . 05 was considered statisti- cally significant.
3 RESULTS |
3.1 | General characteristics
Medical records review yielded 45 dogs that met the criteria for inclusion in the study (Table 1). In 11 dogs, the adrenal masses were found incidentally as a result of either a well- ness examination or the diagnostic workup for an unrelated condition. In 3 dogs, the adrenal mass was diagnosed during further workup of elevations in hepatocellular enzyme levels according to serum biochemical evaluation. In dogs with clinical signs, the following were noted on presentation: polydipsia with polyuria (n = 17), anorexia or hyporexia (10), lethargy (10), vomiting (8), diarrhea (8), weakness or collapse episodes (6), increased respiratory rate or excessive panting (4), increased appetite (2), and cough (1).
At physical examination, 11 dogs had clinical abnormali- ties. Physical examination abnormalities detected the follow- ing: heart murmur (n = 9), cutaneous or subcutaneous masses (8), tense abdomen or abdominal pain (6), alopecia or thinning haircoat (5), palpable abdominal mass (4), dis- tended abdomen (3), organomegaly (2), and hindlimb neuro- logical deficits (2).
3.2 | Diagnostic evaluation
Thoracic radiographs or thoracic CT was performed in 30 of 45 dogs. No dogs exhibited evidence of pulmonary meta- static disease. Reported findings on thoracic radiographs or CT included diffuse bronchointerstitial pattern (n = 2), inci- dental diaphragmatic hernia (1), cardiomegaly (1), and left atrial enlargement (1). Abdominal ultrasound was performed in all dogs, and an adrenal mass was detected in each dog. At the time of abdominal ultrasound, caval invasion was diagnosed in 38 dogs and was not diagnosed in 3 dogs. In 4 dogs, the ultrasound report was not available for review or did not state whether caval involvement was present. Right-
| Variables | Data |
|---|---|
| Age, y | 10.4 ± 2.1 years |
| Body weight, kg | 18.5± 11.8 |
| Body condition score, 1-9 | 6.1 ± 1.2 |
| Sex | Female: neutered (21), intact (2); male: castrated (21), intact (1) |
| Breed | Mixed-breed dog (14), Cavalier King Charles spaniel (4), dachshund (3), miniature dachshund (2), Labrador (2), miniature schnauzer (2), toy poodle (2), 1 each of 16 other breeds |
sided masses were present in 26 cases, left-sided masses were present in 16 dogs, and 3 dogs had bilateral adrenal tumors. Computed tomography angiography of the abdomen was performed in 38 dogs. Tumor and thrombus measure- ments were not available for all dogs but were assessed from CTA results or from ultrasound reports or images that were available for review. Tumor mineralization was observed in 9 of 37 (24.3%) dogs. The following measurements were recorded from CTA results: median maximal tumor diameter was 2.58 cm (interquartile range [IQR] 1.96-3.93, range 0.80-12.47); median tumor volume was 11.71 cm3 (IQR 4.97-24.79, range 2.45-596.79); median thrombus volume was 3.64 cm3 (IQR 0.94-11.30, range 0.12-137.58), median thrombus length was 2.57 cm (IQR 1.55-4.12, range 0.40-23.76), and median thrombus width was 1.15 cm (IQR 0.76-1.87, range 0.08-4.29). Termination of caval thrombi
was assessed from CTA or ultrasound images: thrombi ter- minated prehepatically in 21 of 41 (51.2%) dogs, terminated in an intrahepatic prediaphragmatic location in 15 of 41 (36.6% dogs), and terminated postdiaphragmatically in 5 of 41 (12.2%) dogs. Thrombus termination could not be determined in 4 dogs; thus, the thrombus terminated caudal to the diaphragm in 36 of 41 (87.8%) dogs (Table 2). In 35 dogs for which CTA measurements were available, median maximal thrombus width to normal vena cava width ratio was 1.04 (IQR 0.80-1.38, range 0.41-4.81); median maximal thrombus length to L3 vertebral length ratio was 1.34 (IQR 0.76-2.31, range, range 0.39-9.29); and median maximal thrombus width to L3 vertebral length was 0.65 (IQR 0.48-0.82, range 0.30-1.94). Measurements are pres- ented in Table 3 according to tumor type for 35 dogs in which both tumor type and CTA measurements were
| Variables | Prediaphragmatic | Intrahepatic prediaphragmatic | Postdiaphragmatic |
|---|---|---|---|
| Dogs, n (%) | 21/41 (51) | 15/41 (37) | 5/41 (12) |
| Tumor side treated | |||
| Left | 1/21 (52) | 5/15 (33) | 4/5 (80%) |
| Right | 10/21 (48) | 10/15 (67) | 1/5 (20) |
| Tumor type | |||
| Pheochromocytoma | 16/21 (76) | 11/15 (73) | 5/5 (100) |
| Cortical carcinoma | 4/21 (19) | 3/15 (20) | 0/5 (0) |
| Unknown type | 1/21 (5) | 1/15 (7) | 0/5 (0) |
| Variables | Pheochromocytoma n = 29 | Cortical carcinoma n = 6 | P value |
|---|---|---|---|
| Tumor mineralization, n (%) | 5 (17.2) | 2 (33.3) | .576 |
| Right adrenal tumor, n (%) | 14 (48.3) | 3 (50.0) | >.99 |
| Level of thrombus termination, n (%) | .708 | ||
| Prehepatic | 13 (44.8) | 3 (50.0) | |
| Intrahepatic prediaphragmatic | 11 (37.9) | 3 (50.0) | |
| Postdiaphragmatic | 5 (17.2) | 0 (0.0) | |
| Max tumor diameter, cmª | 2.55 (2.00-3.67) | 2.35 (1.36-3.83) | .512 |
| Tumor volume, ccª | 11.71 (5.12-18.45) | 12.49 (3.73-24.79) | .896 |
| Thrombus volume, ccª | 5.61 (1.52-12.57) | 1.04 (0.72-1.53) | .054 |
| Max thrombus length, cmª | 3.27 (2.01-4.67) | 1.83 (1.02-2.57) | .090 |
| Max thrombus width, cmª | 1.50 (0.97-2.02) | 0.72 (0.66-0.82) | .011 |
| Thrombus width to vena cava width ratioª | 1.22 (0.84-1.61) | 0.83 (0.67-0.90) | .020 |
| Thrombus length to L3 length ratioª | 1.35 (0.88-2.48) | 0.63 (0.59-1.17) | .092 |
| Thrombus width to L3 length ratioa | 0.67 (0.49-1.07) | 0.34 (0.29-0.52) | .024 |
ªMedian (interquartile range).
known; thrombi associated with cortical carcinomas were smaller than thrombi associated with pheochromocytomas for all measurements, although only width differences were statistically significant (P = . 011).
3.3 | Preoperative management
Thirty-one dogs were pretreated for up to 2 weeks with phenoxybenzamine at a median dose of 0.6 mg/kg (range, 0.21-1.4) by mouth every 12 hours. Three dogs were pretreated with trilostane at a median dose of 10 mg/kg (range, 4.7-25) by mouth every 12 hours. One was treated with amlodipine at a dose of 0.4 mg/kg by mouth every 24 hours. Preoperative blood pressure was measured and recorded in 25 dogs. Eleven of 25 dogs were hypertensive (>180 mm Hg systolic blood pressure), with a median sys- tolic blood pressure of 200 mm Hg (range, 185-220).
3.4 | Anesthesia and surgery
Forty dogs underwent a standard ventral midline celiotomy. Three dogs underwent ventral midline celiotomy with an additional right-sided (n = 2) or left-sided (1) paracostal incision, and 2 dogs underwent the combination of a celiotomy and caudal sternotomy. Thirty-seven dogs had complete anesthesia records available for review. Passive intraoperative hypothermia to decrease metabolic tissue demand was employed in 11 dogs, and median lowest core temperature recorded was 90.6 °F (range, 82 °F-93.5 °F). Heparin was administered to only 1 dog intraoperatively. Fifteen of 37 dogs had at least 1 episode of hypertension during surgery, with median systolic blood pressure of 212 mm Hg (range, 185-267). Twenty-seven of 37 dogs had at least 1 episode of intraoperative hypotension, with a low- est recorded systolic blood pressure of 40 mm Hg (range, 30-78). The affected adrenal gland was resected in all cases, and all dogs underwent a cavotomy for removal of the tumor thrombus. In 2 dogs with bilateral adrenal masses, only the invasive right-sided tumors were resected; in 1 additional dog with bilateral adrenal masses, bilateral adrenalectomy was performed, but only the data for the invasive left-sided tumor were included in all analyses. In all dogs, hemostasis was performed by using 1 or more of a combination of monopolar and bipolar electrosurgery, vessel-sealing, and ligation of appropriate blood vessels with suture ligation or hemoclips. Specifically, vessel-sealing devices were used in 31 dogs and were not used in 13 dogs. Rummel tourniquets were placed to minimize hemorrhage during cavotomy in 42 of 44 dogs; in 1 additional dog, a Satinsky clamp was placed on the vena cava around the entirety of the small caval thrombus that was present with no Rummel tourni- quets placed, and, in an additional case, it is unknown
whether Rummel tourniquets were not used or whether their use was simply omitted from the surgical report. In 38 dogs, Rummel tourniquets were positioned in 1 or more of the fol- lowing locations: on the vena cava cranial and caudal to the tumor thrombus and entry of the thrombus through the phrenicoabdominal vein or on the right and/or left renal vein. In 2 dogs with tumor thrombus that extended beyond the porta hepatis but terminated caudal to the diaphragm, a Rummel tourniquet was additionally placed cranial to the diaphragm. In 2 additional dogs with very extensive thrombi that extended cranial to the diaphragm, Rummel tourniquets were placed on the thoracic descending aorta, thoracic cau- dal vena cava, and abdominal vena cava cranial and caudal to the mass after caudal sternotomy. Satinsky clamps were placed on the vena cava after cavotomy and thrombus removal to partially (33/43, 76.7%) or completely (10/43, 23.3%) occlude flow in the vena cava during cavotomy clo- sure in 43 dogs with complete surgical reports. Among 39 dogs for which occlusion data and location of thrombus termination were both available, complete occlusion was per- formed in 4 of 21(19.1%) dogs with prehepatic termination, 4 of 13 (69.2%) dogs with posthepatic prediaphragmatic ter- mination, and 1 of 5 (20.0%) dogs with postdiaphragmatic ter- mination, which was not different (P = 0.863). Cavotomy closure pattern was reported for 44 dogs; in 42 dogs, a simple continuous single layer closure pattern was used, and, in 2 dogs, a simple interrupted closure pattern was used. Cav- otomy closure was performed with polypropylene suture in 32 dogs, polydioxanone in 5 dogs, polyglactin 910 in 2 dogs, and surgical silk in 1 dog in dogs in which these were noted. Suture sizes reported for cavotomy closure were 5-0 in 31 dogs, 4-0 in 8 dogs, and 6-0 in 1 dog. Caval occlusion time (either partial or complete) was reported for 17 dogs and was a median of 8 minutes (range, 2.5-42). In 1 dog that had been occluded for 15 minutes, the surgical report did not state whether occlusion was partial or complete. In the remaining 16 dogs for which occlusion time was reported, occlusion time was shorter (P = . 046) among 11 dogs that had partial occlusion (median 5 minutes, range 2.5-15) vs among 5 dogs that had complete occlusion (median 11 minutes, range 7-42). All occlusion times were 15 minutes or shorter with the exception of 1 dog that had complete occlusion for 42 minutes (this was 1 of the dogs that underwent nephrec- tomy but survived to discharge). Adrenalectomy with tumor thrombus resection alone was performed in 21 dogs, with the remaining dogs undergoing at least 1 other procedure at the time of the adrenalectomy; these procedures included liver biopsy (n = 13) and 1 each of splenectomy, liver lobectomy, partial lung lobectomy, cecal mass biopsy, mammary mass biopsy, cutaneous mast cell tumor resection, and gastrointesti- nal tract biopsies. Nephrectomy was performed in 8 of 45 (17.8%) dogs, 7 of 36 (19.4%) with pheochromocytoma,
1 of 7 (14.3%) with cortical carcinoma, and 0 of 2 (0.0%) with unknown tumors. Age, body weight, tumor type (pheochro- mocytoma vs cortical carcinoma), tumor side (right vs left), maximum tumor diameter, tumor volume, thrombus length, thrombus width, and location of thrombus termination were not associated with receiving a nephrectomy (Supporting Information Table 1).
Among tumors that were resected, tumor type was reported for 43 dogs and included 7 (16.3%) adrenocortical carcinomas and 36 (83.7%) pheochromocytomas. Tumors arose from the right adrenal in 23 of 45 (51.1%) dogs (18 pheochromocytoma, 4 cortical carcinoma, 1 unknown) and from the left adrenal in 22 of 45 (46.7%) dogs (18 pheo- chromocytoma, 3 cortical carcinoma, 1 unknown). The dis- tribution of tumor types did not differ between left and right sides (P = . 705) or by location of thrombus termination (Table 2; P = . 893). Median surgical time for all surgical procedures was 155 minutes (range, 45-450). Median anes- thesia time for all procedures was 255 minutes (range, 110-720).
3.5 | Complications
Intraoperative complications occurred in 16 of 45 dogs. Bleeding from the cavotomy or renal venotomy site occurred in 5 dogs and was treated by application of additional sutures, hemoclips, and/or the application of oxidized bovine cellulose hemostatic agent (Surgicel; Ethicon, Cincinnati, Ohio). Vena caval tearing occurred in 3 dogs prior to or after closure of the venotomy site. Incomplete thrombus removal because of adherence to the wall of the vena cava occurred in 1 dog. Intraoperative cardiac arrest occurred in 4 dogs. One of these dogs survived the operative period and 3 did not. None of the 4 dogs that experienced cardiac arrest intra- operatively survived to discharge from the hospital.
Blood product transfusions consisting of packed red blood cells, fresh frozen plasma, or cryoprecipitate transfu- sions were administered intraoperatively or postoperatively in 28 of 44 (63.6%) of dogs for which data were available, including 19 of 35 (54.3%) for pheochromocytoma, 7 of 7 (100.0%) for cortical carcinoma, and 2 of 2 (100.0%) for unknown tumors. Among the 24 dogs that received packed red blood cell transfusions, the median number of units administered was 1.5 (range, 1-6). Among the 14 dogs that received fresh frozen plasma transfusions, the median num- ber of units administered was 2 units (range, 1-6). One dog received 1 unit and 1 dog received 2 units of cryoprecipitate. Age, body weight, tumor type (pheochromocytoma vs cortical carcinoma), tumor side (right vs left), maximum tumor diame- ter, tumor volume, thrombus length, thrombus width, location of thrombus termination, and concurrent nephrectomy were
not associated with receiving a packed red blood cell transfu- sion (Supporting Information Table 2).
Postoperative complications occurred in 18 of 45 dogs. These included pancreatitis in 5 dogs, cardiac arrest in 4 dogs (3 of these dogs also experienced cardiac arrest intra- operatively), ongoing intra-abdominal hemorrhage in 3 dogs, regurgitation in 3 dogs, acute respiratory distress syndrome in 2 dogs, aspiration pneumonia in 1 dog, and renal failure in 1 dog.
3.6 | Outcome
Eleven dogs (24.4%) died or were euthanatized prior to hospi- tal discharge, 6 of 36 (16.7%) with pheochromocytomas, 3 of 7 (42.9%) with carcinomas, and 2 dogs with unknown tumor type. Median time in hospital postoperatively was 3 days (range, 1-10). Survival to discharge did not differ between dogs with prehepatic (17/21), intrahepatic prediaphragmatic (11/15), and postdiaphragmatic (2/5) tumor extension (P = . 167) Age, body weight, tumor type (pheochromocy- toma vs cortical carcinoma), tumor side (right vs left), maxi- mum tumor diameter, tumor volume, thrombus length, thrombus width, location of thrombus termination, concurrent nephrectomy, packed red blood cell transfusion, and caval occlusion time were not associated with survival to discharge (Supporting Information Table 3).
Median overall survival time for all 45 dogs was 547 days (95%CI 146-710). Among the 34 dogs that sur- vived to hospital discharge, 20 died during the follow-up period, 6 were lost to follow-up, and 8 were still alive at the time of data collection. Median survival time for the 34 dogs that were discharged from the hospital was 690 days (95%CI 510-1162). Results of univariable and multivariable Cox regression are presented in Table 4. After adjustment for age and tumor diameter, carcinoma tumor type (hazard ratio 3.74, 95%CI 1.18-10.94; P = . 025) and postdiaphragmatic thrombus termination were associated with greater hazard of death. Dogs with postdiaphragmatic thrombus termination had greater hazard of death compared with dogs with prehepatic termination (hazard ratio 4.72, 95%CI 1.27-17.47; P = . 020) and with dogs with intrahepatic prediaphragmatic termination (hazard ratio 3.53, 95%CI 1.02-12.28; P = . 047). There was no difference in estimated survival between prehepatic termi- nation vs intrahepatic prediaphragmatic termination (hazard ratio 1.33, 95%CI 0.56-3.21; P = . 519). Kaplan-Meier (unadjusted) survival estimates according to location of tumor thrombus termination are presented in Figure 1.
Recurrence of adrenal neoplasia was highly suspected on the basis of imaging of recurrent nodules or masses in the location of the previous adrenalectomy site according to postoperative ultrasound or CTA in 5 dogs (4 in which pheochromocytoma had been originally diagnosed and 1 in
| Variables | Univariable | Multivariable | ||||
|---|---|---|---|---|---|---|
| Hazard ratio | 95%CI | P value | Hazard ratio | 95%CI | P value | |
| Age, y | 1.15 | 0.95-1.39 | .140 | 1.38 | 1.08-1.75 | .010 |
| Body weight, kg | 1.01 | 0.98-1.05 | .387 | ... | ... | ... |
| Carcinomaª | 2.60 | 1.01-6.65 | .047 | 3.59 | 1.18-10.94 | .025 |
| Right-sided tumor | 0.70 | 0.34-1.44 | .330 | ... | ... | ... |
| Maximum tumor diameter, cm | 1.14 | 1.00-1.31 | .057 | 1.27 | 0.98-1.66 | .075 |
| Packed red blood cell transfusion | 1.58 | 0.73-3.38 | .243 | ... | ... | ... |
| Nephrectomy | 1.20 | 0.49-2.96 | .688 | ... | ... | |
| Thrombus termination Prehepatic | Reference | Reference | ||||
| Intrahepatic prediaphragmatic | 1.16 | 0.51-2.64 | .716 | 1.33 | 0.56-3.21 | .519 |
| Postdiaphragmatic | 2.97 | 1.00-8.76 | .049 | 4.72 | 1.27-17.47 | .020 |
… , no data.
ªVersus pheochromocytoma.
1.00
0.75
0.50
0.25
0.00
0
500
1000
1500
2000
Days
Pre-hepatic
Intra-hepatic pre-diaphragmatic
Post-diaphragmatic
which adrenocortical carcinoma had been originally diag- nosed) in which diagnosis was given at a median of 578 days (range 112-1470). In none of these dogs was recurrence confirmed by biopsy or resection of the masses or nodules.
4 DISCUSSION |
Perioperative complications and death in dogs with invasive adrenal tumors were not found to be associated with differ- ences in patient size or age, tumor type, or dimensions or extent of tumor thrombus. In addition, tumor dimensions and thrombus extension were not found to be prognostic for short-term outcomes, although it should be noted that, in
88% of dogs, the tumor thrombus terminated caudal to the diaphragm, with only 5 dogs having thrombus extending beyond the diaphragm. A larger sample is required to state definitively whether extensive tumor thrombus extension implies a greater surgical risk. Extensive thrombus extension should therefore not constitute a factor that excludes these dogs from being recommended for surgical resection, but it should be remembered that, with only 40% survival in dogs with postdiaphragmatic thrombus extension compared with 73%-81% survival in dogs with thrombus that terminated prediaphragmatically, it may be that a larger sample would have confirmed a worse prognosis in the postdiaphragmatic thrombus cohort.
The probability that blood products will be required dur- ing resection of an invasive adrenal mass is substantial; 63% of dogs in this study received at least 1 type of blood prod- uct. Surgeons and owners should be prepared for this possi- bility, and, because of these findings, the authors suggest that these cases be treated only in centers where blood prod- ucts are readily available. Animal size as well as tumor or thrombus dimensions were not found to be significant risk factors for requiring a packed red blood cell transfusion.
An important complication of adrenalectomy is the requirement to perform a concurrent ipsilateral nephrectomy, which can be required in dogs with or without vascular inva- sion. Nephrectomy may be required in cases in which tumor thrombus extends into the renal vein or in cases in which inadvertent damage to the renal vessels causes profuse hem- orrhage during dissection. Nephrectomy was required in 8 of 45 (17.8%) dogs in this cohort, which is similar to other pub- lished studies.2,3,5,7-9 Tumor side, size, or type did not affect the risk of having to perform a nephrectomy. Nephrectomy
also was not a risk factor for long-term survival in contrast to at least 1 other study in which dogs undergoing concur- rent nephrectomy did experience shorter long-term survival.5 It may be that failure to detect a statistical difference in out- comes between dogs underwent a nephrectomy vs those that did not was simply a type II statistical error caused by inade- quate case numbers.
Long-term survival of dogs discharged from the hospital was commonplace in this study, just as has been previously reported.2,3,5,7-9 Most of the dogs in the study died of causes unrelated to adrenal disease, although extensive follow-up and investigation of the cause of death was not possible for most dogs. Diagnosis of metastatic disease was not con- firmed in any dog in this study at the time of adrenalectomy, although at least 1 dog had confirmed metastatic adrenocor- tical carcinoma at necropsy, and 6 others had suspected but unconfirmed local recurrence or distant metastasis. Meta- static disease was found in regional lymph nodes in 12% of dogs and at distant sites in 24% of dogs in necropsy studies of dogs with pheochromocytoma,11 suggesting a more aggressive biological behavior than is often suspected and suggesting that underdiagnosis of metastatic lesions may be common when antemortem diagnostic evaluation for gross metastasis alone is considered in these cases. In 1 previously published study with dogs with adrenal tumors, masses or nodules were diagnosed at the time of presurgical imaging in 7 dogs and were resected or biopsied in each case. In all 7 dogs, histopathological analysis of lesions were either met- astatic cortical carcinoma or pheochromocytoma of the liver, pancreas, or kidney, and these dogs had a significantly shorter long-term survival compared with those dogs in which no metastatic lesions were detected.8
In this study, dogs with postdiaphragmatic caval throm- bus extension had shorter overall survival compared with those with prehepatic or intrahepatic prediaphragmatic thrombus extension. There was no difference in overall survival between dogs with prehepatic vs intrahepatic prediaphragmatic thrombus extension. It is impossible to conclude whether this finding is related to long-term compli- cations, the severity of the disease at the time of diagnosis, or the development of metastatic disease because, in most cases, necropsy of dogs that died was not performed. It may be that dogs with very extensive caval thrombosis have a greater incidence of residual disease or distant microscopic disease that is not detected at the time of surgery. Because adrenal tumors rarely invade the vena cava in man, the clos- est human corollary to invasive canine adrenal tumors is renal cell carcinoma. These tumors are different because invasion occurs through the renal vein but variably ascends through the inferior vena cava and, as they become more extensive, will involve the hepatic vena cava, cross the dia- phragm, and eventually enter the right atrium. The literature
is also mixed regarding whether thrombus extension affects prognosis, with some researchers finding no difference in 5-year survival between persons with renal cell carcinoma with either prehepatic or posthepatic extension.12,13 How- ever, in some human studies, thrombus extension has been found to play a significant role in long-term prognosis, with 1 article reporting 5-year survival rates of 42%, 20%, and 0% for level 1 (prehepatic), 2 (intrahepatic but not crossing into right atrium) and 3 (thrombus in right atrium) thrombus extension, respectively.14 An explanation for this finding is also not readily available in the human literature and is likely multifactorial. For the dogs in this study, it is difficult to make robust arguments regarding the effects of tumor throm- bus extension on prognosis because relatively small numbers of dogs with postdiaphragmatic thrombus extension were surgically treated.
The decision to perform surgical resection in dogs with invasive adrenal masses is often challenging. In select tumors with small caval thrombi, surgical modifications such as phrenicoabdominal venotomy may obviate the requirement to perform cavotomy, which, in turn, might reduce surgical morbidity in these cases.15 However, per- haps the larger question is whether alternatives to surgical management could provide less morbid results in these dogs. The results of conservative management are difficult to quantitate because data are scarce and few investigators report the results of noninterventional management. Authors of one small study reported prolonged survival (mean, 13.1 months) in a group of dogs with mainly noninvasive non-cortisol-producing tumors.16 Only 3 dogs in this study had vascular invasion into the vena cava, and, although none of the primary tumors grew significantly during the follow-up period, caval thrombosis was progressive in all 3.16 The results of hypofractionated stereotactic volumetric- modulated arc radiotherapy were also reported recently in 9 dogs with invasive adrenal tumors.17 These dogs received between 30 and 45 Gy in 3 or 5 consecutive daily fractions and experienced a median survival time of 1030 days. These results are very encouraging and, if repeatable, may provide a new therapeutic approach to these challenging cases, especially in dogs in which surgery is deemed too risky or for owners who would prefer to avoid the risk of surgery.
There are several limitations to the study. Follow-up was not available for all dogs, and some surgery reports were not as complete as others, so some datapoints were not available to report. A control group of dogs without caval invasion was not included because most noninvasive adrenal masses are treated laparoscopically in the authors institutions. A multitude of surgeons performed procedures at 4 different academic teaching hospitals, so standardization of therapy was impossible.
Perioperative mortality remains substantial in dogs with invasive adrenal neoplasia, and dogs with postdiaphragmatic tumor thrombus extension represent a subgroup of dogs with a worse prognosis for overall survival.
CONFLICT OF INTEREST
The authors declare no conflicts of interest related to this report.
ORCID
Michelle A. Giuffrida ® https://orcid.org/0000-0002-6862- 2653
William T. N. Culp ® https://orcid.org/0000-0001-6132- 156X
Michele A. Steffey ® https://orcid.org/0000-0003-0852-0644
REFERENCES
1. Barthez PY, Marks SL, Woo J, Feldman EC, Matteucci M. Pheo- chromocytoma in dogs: 61 cases (1984-1995). J Vet Intern Med. 1997;11:272-278.
2. Anderson CR, Birchard SJ, Powers BE, Belandria GA, Kuntz CA, Withrow SJ. Surgical treatment of adrenocortical tumors: 21 cases (1990-1996). J Am Anim Hosp Assoc. 2001;37:93.
3. Kyles AE, Feldman EC, De Cock HEV, et al. Surgical manage- ment of adrenal gland tumors with and without associated tumor thrombi in dogs: 40 cases. J Am Vet Med Assoc. 2003;223: 654-662.
4. Herrera MA, Mehl ML, Kass PH, Pascoe PJ, Feldman EC, Nelson RW. Predictive factors and the effect of phenoxybenzamine on outcome in dogs undergoing adrenalectomy for pheochromocy- toma. J Vet Intern Med. 2008;22:1333-1339.
5. Schwartz P, Kovak JR, Koprowski A, Ludwig LL, Monette S, Bergman PJ. Evaluation of prognostic factors in the surgical treat- ment of adrenal gland tumors in dogs: 41 cases (1999-2005). J Am Vet Med Assoc. 2008;232:77-84.
6. Schultz RM, Wisner ER, Johnson EG, Macleod JS. Contrast- enhanced computed tomography as a preoperative indicator of vas- cular invasion from adrenal masses in dogs. Vet Rad Ultrasound. 2009;50:625-629.
7. Lang JM, Schertel E, Kennedy S, Wilson D, Barnhart M, Danielson B. Elective and emergency surgical management of
adrenal gland tumors: 60 cases (1999-2006). J Am Anim Hosp Assoc. 2011;47:428-435.
8. Massari F, Nicoli S, Romanelli G, Buracco P, Zini E. Adrenalec- tomy in dogs with adrenal gland tumors: 52 cases (2002-2008). J Am Vet Med Assoc. 2011;239:216-221.
9. Barrera JS, Bernard F, Ehrhart EJ, Withrow SJ, Monnet E. Evalua- tion of risk factors for outcome associated with adrenal gland tumors with or without invasion of the caudal vena cava treated via adrenalectomy in dogs: 86 cases (1993-2009). J Am Vet Med Assoc. 2013;242:1715-1721.
10. Adin CA, Nelson RW. Adrenal glands. In: Johnston SA, Tobias KM, eds. Veterinary Surgery Small Animal. 2nd ed. St Louis, MO: Elsevier; 2018:2281-2291.
11. Gilson SD, Withrow SJ, Wheeler SL, Twedt DC. Pheochromocy- toma in 50 dogs. J Vet Intern Med. 1994;8:228-232.
12. Tsuji Y, Goto A, Hara I, et al. Renal cell carcinoma with extension of tumor thrombus into the vena cava: surgical strategy and prog- nosis. J Vasc Surg. 2001;33:789-796.
13. Nesbitt JC, Soltero ER, Dinney PN, et al. Surgical management of renal cell carcinoma with inferior vena cava tumor thrombus. Ann Thorac Surg. 1997;63:1592-1600.
14. Skinner DG, Pritchett RT, Lieskoxsky G, Boyd SD, Stiles QR. Vena caval involvement by renal cell carcinoma. Ann Surg. 1989; 210:387-394.
15. Mayhew PD, Culp WTN, Balsa IM, Zwingenberger AL. Phrenicoabdominal venotomy for tumor thrombectomy in dogs with adrenal neoplasia and suspected vena caval invasion. Vet Surg. 2018;47:227-235.
16. Arenas C, Perez-Alenza D, Melian C. Clinical features, outcome and prognostic factors in dogs diagnosed with non-cortisol- secreting adrenal tumours without adrenalectomy: 20 cases (1994-2009). Vet Rec. 2013;173:501.
17. Dolera M, Malfassi L, Pavesi S, et al. Volumetric-modulated arc stereotactic radiotherapy for canine adrenalocortical tumors with vascular invasion. J Small Anim Pract. 2016;57:710-717.
SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of this article.
How to cite this article: Mayhew PD, Boston SE, Zwingenberger AL, et al. Perioperative morbidity and mortality in dogs with invasive adrenal neoplasms treated by adrenalectomy and cavotomy. Veterinary Surgery. 2019;1-9. https://doi.org/10.1111/vsu.13221