ORIGINAL ARTICLE
Laparoscopic versus open surgery in stage I-III adrenocortical carcinoma - a retrospective comparison of 32 patients
ALEXANDER FOSSÅ1, BÅRD I. RØSOK2, AIRAZAT M. KAZARYAN2,5, HARALD JR. HOLTE1, BJØRN BRENNHOVD3, OLA WESTERHEIM4, IRINA P. MARANGOS2,5 & BJØRN EDWIN2,5
1 Department of Oncology, Oslo University Hospital, Oslo, Norway, 2Department of Gastrointestinal Surgery, Oslo University Hospital, Oslo, 3Department of Urology, Oslo University Hospital, Oslo, Norway, 4Department of Breast and Endocrine Surgery, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Oslo, Norway and 5 Interventional Unit Rikshospitalet, Division of Diagnostics and Intervention, Oslo University Hospital, Oslo, Norway
Abstract
Laparoscopic surgery (LS) for resectable adrenocortical carcinoma (ACC) has been questioned due to uncertainty with regard to long-term oncological outcome. We analyzed the experience with LS compared to open surgery (OS) at Oslo University Hospital (OUH). Material and methods. Between 1998 and 2011 32 patients were identified with ACC stage I-III operated either by LS (17 patients) or OS (15 patients). Patients’ records were reviewed retrospectively with regard to pre- and intraoperative findings, short-term surgical outcome, relapse and survival. The patients in the LS group had significantly smaller tumors and higher body mass index, otherwise the groups did not differ significantly. Thirty-one patients had been operated at surgical departments of the OUH, and all had been followed at OUH. Results. Short-term outcome favored LS by significantly shorter operation time, lower blood loss and need for transfusions, fewer postoperative compli- cations and shorter hospitalization. The completeness of resection was similar in both groups with R0 resection accom- plished in 12 patients in the LS group and 12 in the OS group. Twelve and 15 patients have relapsed in the LS and OS groups, respectively, with a similar pattern of relapse (local, peritoneal or distant). Median progression-free survival (15.2 months for LS vs. 8.1 months for OS) and median overall survival (103.6 months for LS vs. 36.5 months for OS) were not significantly different. Discussion. LS seems to offer short-term advantages and similar long-term outcome compared to OS in patients with resectable ACC stage I-III.
Adrenal cortical carcinoma (ACC) is a rare neoplasm with poor prognosis [1,2]. Many patients are diag- nosed with large primary tumors that may invade adjacent structures or show vascular infiltration with venous tumor thrombi, even when overt metastases are not present. Symptoms of hypercortisolism or androgen secretion or radiological work-up for other reasons may lead to earlier diagnosis in some cases.
At the same time complete operative resection of localized disease is the only treatment with potential to cure patients with ACC and the prognosis corre- lates with stage at diagnosis [1,3,4]. Normally, com- plete resection can be attempted in tumors without invasion of surrounding structures, adjacent organs, or spread to lymph nodes or distant organs (stage I
and II). In a recent analysis of register data from Germany five-year survival in stage I patients was 84% and 63% for stage II patients [1]. When the tumor invades adjacent tissues or organs, presents with lymph node metastases or endovascular tumor thrombi (stage III disease) complete resection may be still be attempted but survival in these patients is much lower [1,3].
Over the last decades, adrenal laparoscopic sur- gery (LS) has become an alternative to open surgery (OS), first in benign adrenal tumors, and later for resec- tion of suspected or verified malignant tumors [5]. The short-term benefit of laparoscopic approaches has been demonstrated in benign disorders, but the long-term outcome in ACC has been questioned. In
Acta Oncol Downloaded from informahealthcare.com by University of Nebraska on 01/03/15 For personal use only.
ACC cure is critically dependent on the removal of all tumor with an adequate margin and patients with microscopically free margins at operation (R0) do better than patients with micro- or macroscopical residual disease (R1 or R2 resection) [1,3]. Early reports with small numbers of patients operated laparoscopically have been inconclusive as summa- rized recently [6]. A report from USA with a higher number of patients suggested that LS may result in lower rates of complete resection, more frequent intra-operative tumor spill and shorter times to recurrence compared to patients selected for OS [7]. Two studies from Germany and Italy led to opposite conclusions stating that patients after LS or OS have equal outcomes in terms of recurrence-free and over- all survival [8,9].The conflicting results in different series comparing LS and OS for ACC may, among other reasons, be due to the limited number of patients in most studies, their retrospective nature, referral bias and different levels of expertise with surgical approaches for this rare disease [6].
To add to the experience with LS we have retro- spectively analyzed all patients with ACC stage I-III operated with a curative intent at the surgical depart- ments of Oslo University Hospital (OUH) from 1998 to 2011. Patients subjected to LS were compared to patients from the same period with OS. We report on preoperative findings, intra- and postoperative incidents and complications, recurrence pattern and survival.
Material and methods
Patients
From OUH registries we identified patients with the diagnosis of ACC seen from January 1998 to December 2011. Patients with stage I-III suitable for radical surgery and adequate information on surgical treatment and follow-up were included in the analy- sis. Patients had given informed consent to surgery at the time of diagnosis according to Norwegian regulations. The study was approved by the Hospi- tal’s privacy officer and regional ethical committee.
Diagnostic work-up and follow-up
Medical histories, clinical and radiological findings were reviewed from hospital charts including preop- erative endocrinological tests. Staging was performed retrospectively based on preoperative images accord- ing to the European Network for the Study of Adre- nal Tumors (ENSAT): stage I, tumor of diameter ≤5 cm; stage II, tumor diameter > 5 cm; stage III, tumor infiltration of neighboring structures, venous tumor thrombus in vena cava or vena renalis, or positive lymph nodes; and stage IV, distant metastases [10,11]. American Society of Anaesthesiology (ASA)
scores, operation and histology reports were reviewed [12]. Intra-operative unintended incidents were classified by the Satava approach for surgical errors as adapted previously for adrenal surgery and post- operative complications in accordance with the Accordion classification (Clavien-Dindo-Strasberg classification) [13-15]. Postoperative follow-up con- sisted of visits every 3-6 months with CT, ultrasound and/or plain x-rays at yearly intervals in most patients. Appropriate imaging was performed when relapse was suspected clinically. The date of clinical or radio- logical relapse was recorded, whichever occurred first, and classified as local when involving the oper- ative site or regional lymph nodes (single when involving one site or multiple when involving more sites), peritoneal when there was evidence of abdom- inal carcinomatosis or distant [9]. The occurrence of port site lesions, information on adjuvant therapy, treatment for relapse, time and cause of death were recorded.
Operation methods and results
Laparoscopic adrenalectomy was performed through a lateral transabdominal approach, as described previously [5]. In the OS group the adrenal tumor was approached through an anterior transabdomi- nal incision in 10 patients and a thoracoabdominal incision in five patients. The intention was the rad- ical removal of the affected adrenal together with tumor and surrounding fat without rupture of the tumor capsule. Concomitant en bloc resection of adjacent organs and/or veins was performed when direct invasion was suspected or tumor thrombi detected. Operation results were recorded at the time of surgery (complete macroscopic resection for tumor with or without rupture of the capsule) and by routine pathological examination for macro- or microscopic involvement of the margins of the surgical specimen.
Statistical analyses
Statistical analyses were performed using SPSS 18 (IBM Corporation, Armonk, New York). Proportions were calculated for categorical variables and medians and ranges for continuous data. Differences between groups were analyzed by x2 or Mann-Whitney U-tests. Progression-free survival (PFS) was recorded from date of operation until relapse or death of any cause, overall survival was calculated from date of operation until of death of any cause. Survival curves produced by the Kaplan-Meier method were com- pared by the log-rank test. Statistical tests were two sided and p-values below 0.05 were considered significant.
Results
Forty-seven patients with ACC were identified from January 1998 to December 2011. One patient had been operated outside of Norway, and 14 patients had primary stage IV disease. The 32 remaining patients had stage I-III and had been operated with the aim of a curative resection, all but one patient at surgical departments within OUH. Preoperative findings of the 32 patients are summarized in Table I. A preoperative computed tomography (CT) or magnetic resonance imaging (MRI) of the abdo- men was performed in all, chest imaging with plain radiograms or CT in 11 and 21 patients, respectively. Preoperative needle biopsies were done in six patients, suggesting possible ACC in two and probable ACC in four. Seventeen patients were operated by LS, 13 of these with preoperative stage I-II. The OS group consisted of 15 patients, of whom seven presented with stage I-II. The body mass index (BMI) was
significantly higher and tumor size significantly lower in the LS group. There were trends towards better performance status, lower stages and more frequent hormone secretion in the LS group. For other factors the two groups appeared balanced.
Short-term outcome for the two groups is shown in Table II. Two patients in the LS group were con- verted to OS, but are recorded in the LS group according to the initial intent of surgery. Operation time, blood loss, need of blood transfusion and the extent of resection all favored the LS group. Intraop- erative unfavorable incidents were less frequent and less severe in the LS group, but this was to a large extent explained by the more pronounced blood loss in OS patients. For two patients with OS and five with LS violation of the tumor capsule and/or intra- operative tumor spill were recorded by the surgeon. Accidental damage to a neighboring organ (liver, pancreas, spleen or vena cava) was recorded for four
| Laparoscopic surgery n=17 | Open surgery n=15 | Significance | |
|---|---|---|---|
| Age/years median (range) | 45 (29-75) | 52 (38-71) | p =0.31 |
| Sex, male | 4 | 5 | p =0.70 |
| female | 13 | 10 | |
| BMI at diagnosis/kg/m2 median (range) | 27.2 (17.0-54.6) | 22.8 (18.9-40.2) | p =0.04 |
| ECOG status | |||
| 0 | 13 | 6 | p =0.09 |
| 1 | 4 | 7 | |
| 2 | 0 | 2 | |
| Other disease | |||
| No | 15 | 14 | |
| Adrenal adenoma | 1 | 0 | |
| Other cancer | 1 | 1 | |
| Localization | |||
| Right adrenal gland | 10 | 7 | p =0.59 |
| Left adrenal gland | 7 | 7 | |
| Bilateral | 0 | 1 | |
| Main symptoms at diagnosis | |||
| Accidental finding | 4 | 1 | p=0.38 |
| Locoregional symptoms | 5 | 9 | |
| Cushing syndrome | 6 | 3 | |
| Hyperandrogenism | 1 | 2 | |
| 'Hyperestrogenism' | 1 | 0 | |
| Hormone secretion | |||
| Glucocorticoids and/or androgens | 12 | 6 | p =0.07 |
| Estrogen | 1 | 0 | |
| No hormone secretion | 4 | 9 | |
| Size of primary tumor/cm median (range) | 8.0 (4.2-16.0) | 13.0 (6.0-24.0) | p=0.002 |
| Preoperative stage | |||
| I | 1 | 0 | p=0.06 |
| II | 12 | 7 | |
| III | 4 | 8 | |
| ASA | |||
| 1 | 2 | 3 | p =0.85 |
| 2 | 9 | 6 | |
| 3 | 5 | 5 | |
| 4 | 1 | 1 |
ASA, American Society of Anesthesiology; BMI, body mass index.
Acta Oncol Downloaded from informahealthcare.com by University of Nebraska on 01/03/15
For personal use only.
| Laparoscopic surgery n=17 | Open surgery n =15 | Significance | |
|---|---|---|---|
| Operation time/minutes | 150 (56-360) | 230 (163-344) | p=0.005 |
| Median (range) | |||
| Blood loss/ml | <400 (50-1830) | 1700 (400-10 750) | p<0.001 |
| Median (range) | |||
| Perioperative blood transfusions/units median (range) | 0 (0-11) | 3 (0-32) | p =0.04 |
| Perioperative plasma transfusions/ml median (range) | 0 (0-1600) | 400 (0-3200) | p=0.08 |
| Conversion to open surgery | 2 | ||
| Extent of surgery | |||
| Tumor only | 13 | 2 | p =0.001 |
| With en bloc resection of other organs | 3 | 9 | |
| With vascular surgery | 1 | 4 | |
| Intraoperative incidents/grade | |||
| 1 | 4 | 2 | p=0.001 |
| 2 | 10 | 1 | |
| 3 | 3 | 12 | |
| Postoperative complications/ grade | |||
| 0 | 9 | 4 | p=0.02 |
| 1 | 5 | 2 | |
| 2 | 1 | 6 | |
| 3 | 0 | 3 | |
| 4 | 2 | 0 | |
| Result of operation | |||
| R0 | 12 | 12 | p=1.0 |
| R1 | 4 | 3 | |
| R2 | 1 | ||
| Postoperative stay/days median (range) | 6 (2-28) | 13 (7-28) | p<0.001 |
patients in the OS group and two patients in the LS group. Similarly, postoperative complications were more frequent and more pronounced in OS patients. Of nine patients in the OS group with grade ≥2 complications there were four with infections, two in need of draining for pleural fluid, two requiring total parenteral nutrition and one case each of pulmonary embolism, renal insufficiency and bleeding. In the LS group there were one case each of syncope, bleed- ing requiring reoperation and pneumonia requiring ventilation support. The completeness of resection as judged by surgical and histopathological reports together where comparable. The median postopera- tive stay in hospital was significantly shorter for the LS group (6 vs. 13 days, p<0.001).
Two patients with stage III disease received adjuvant mitotane (both after OS). No adjuvant radiotherapy was given.
Five patients have not relapsed so far and all were operated laparoscopically. The pattern of relapse was similar in both groups (Table III, p = 0.33 for Fischer’s exact test). Two patients in the LS group have experienced port site relapses. In one of these other local and peritoneal metastases were present simultaneously, and in the other port site relapse became apparent after local relapse with multiple lesions. Of the 27 patients with relapse, 11 have been offered resection of one or more lesions, nine patients
have had radiotherapy to one or more metastatic sites and 24 patients systemic treatment (mostly mitotane, streptotozocine, etoposide/doxorubicin/ciplatinum, alone or in combination/sequential order [16]).
In November 2011 median follow-up of the 12 patients alive is 29.1 months (range 11-104 months). Median PFS was 15.2 months in the LS group [95% confidence interval (CI) 2.6-27.9 months] and 8.1 months (1.8-14.4 months) in the OS group (p= 0.057, Figure 1). A subanalysis in stage I-II patients showed a similar trend towards an advantage for the LS group with median PFS of 20.1 months (4.0-36.0 months) and 8.1 months (5.1-11.2 months) in the two groups, respectively (p=0.096). Overall survival for the groups was similar. For patients with stage I-III disease median overall survival was 103.6 months (0-207.9 months) in the LS group and 36.5 months (17.4-55.6 months) in patients with OS (p=0.22). The corresponding figures in stage I-II patients were 103.6 months (0-226.9 months) and 42.4 months (7.1-77.8 months), respectively (p=0.40).
To analyze for a potential effect of treatment period, 15 patients operated between 1998 and 2005 (nine by LS) were compared to 17 patients (eight by LS) operated 2006-2010. The portion of patients operated by LS was similar in both periods. Compar- ing stage I and II patients operated by LS, eight patients were treated between 1998 and 2004 (with
| Type of relapse | Laparoscopic surgery n=12 | Open surgery n=15 |
|---|---|---|
| Local simple | 0 | 1 |
| Local multiple | 1 | 0 |
| Peritoneal | 2 | 0 |
| Distant | 6 | 9 |
| Local and distant | 3 | 5 |
R0 resection in six) and fi ve patients between 2005 and 2010. In the former group, seven have relapsed (five with distant metastases as first sign of relapse) and in the latter two have relapsed (one with distant metastases). With few patients in each group and sig- nificantly shorter follow-up of patients treated in the last period there is no significant difference in PFS or OS (data not shown).
Discussion
The major finding of the present study is the improved short-term and similar long-term oncological out- come in patients with ACC stage I-III operated by a laparoscopic approach compared to patients subjected to an open procedure.
The patients are retrospectively identified from the period 1998 to 2011 based on the diagnosis of ACC in hospital registries. OUH has been the tertiary reference center for the south-east region of Norway during this period, but we cannot exclude that patients have been operated elsewhere and not been referred to OUH. Search criteria only included a diagnosis of ACC and some cases, in particular, patients with smaller and less aggressive tumors mistaken for adenomas or deemed unclassifiable at the time of operation, may have been missed in our series, as has been suggested for other cohorts [1].
A
1.0
B 1.0
0.8
0.8
Probability
0.6
Probability
0.6
0.4
0.4
0.2
0.2
0.0
0.0
0
12
24
36
48
60
72
84
96
108
120
0
12
24
36
48
60
72
84
96
108
120
Time/months
Time/months
C
1.0
D 1.0
0.8
0.8
Probability
0.6
Probability
0.6
0.4
0.4
0.2
0.2
0.0
0.0
0
12
24
36
48
60
72
84
96
108
120
0
12
24
36
48
60
72
84
96
108
120
Time/months
Time/months
Acta Oncol Downloaded from informahealthcare.com by University of Nebraska on 01/03/15
For personal use only.
This is evident in one patient in our series who was initially diagnosed with an adenoma and later reclas- sified by the pathologist to ACC stage I when distant metastases became apparent. Important for the comparison of operation techniques, during the period 1998-2011 no uniform criteria for choice of operation technique have existed. In this retrospec- tive analysis, patients in the LS group had smaller tumors and a trend towards earlier stages, both being important for outcome of the operation and long- term prognosis [1,3]. Furthermore, the surgical departments of OUH have until 2008 been respon- sible for different geographical areas, and despite close contact between surgeons and exchange of patients, local preference and expertise of the indi- vidual surgeons will have influenced the choice of operation technique.
Ideally, the question of operation technique in curative surgery for ACC should be addressed by a prospective randomized trial, but due to the rarity of the disease it is at present unlikely that such a study will ever be performed. Retrospective controlled studies are currently the best available research despite significant limitations in all studies of this type. Our study adds to the body of evidence from similar retrospective and non-randomized compari- sons that aim to evaluate long-term oncological out- come after LS for resectable ACC. Two recent larger studies concluded that LS is equivalent to OS in these patients [8,9]. Porpiglia et al. compared patients with stage I-II that had completed radical resection by adrenalectomy alone, i.e. patients with more extensive resections, patients with non-radical sur- gery and patients with conversion to open surgery were excluded [9]. When comparing these selected groups, the outcome was equivalent in 18 patients after LS and 25 patients with OS. In the largest study to date, Brix et al. included patients with stage I-III with a tumor diameter < 10 cm identified in the Ger- man ACC register [8]. Due to the sample size, more elaborate comparisons were possible, and neither matched control nor multivariate analysis could identify an influence of the chosen operation tech- nique on long-term outcome. The latter study is noteworthy in that it aims to control for the most noticeable inherent bias present in our study and other studies: in the absence of randomization there will be a trend for smaller tumors and earlier stages among patients chosen for LS. In the absence of ran- domization in the present study, we may not exclude an impact of selection bias on the observed trend towards better PFS in patients operated laparoscop- ically, nor can we exclude that the LS group would have fared even better if they were subjected to OS.
Other studies have suggested poorer outcome for patients treated with LS [7,17]. Miller et al. reported
on 88 patients with ACC who underwent surgical resection and found similar recurrence rates in 17 and 71 patients that underwent either LS or OS (63% vs. 65%) [7]. The authors reported that 50% of patients with LS had intra-operative tumor spill and/or positive margins compared to 18% in the OS group, and a shorter time to recurrence in the former patients (9.6 vs. 19.2 months). Our results, with a similar number of patients operated laparoscopically, do not show the same tendency for intra-operative tumor spill and no difference in relapse pattern or time to relapse.
The bias introduced by referral practices and sur- gical skills in different institutions may be important when interpreting the results of some studies. For instance, in the studies of Porpiglia and Gonzales only a minority of all patients have been operated at the reporting center [9,17]. Most likely, patients with assumed higher risk or clear evidence of recurrence will be referred more often and laparoscopy may be performed more frequently at referral centers. Such trends may bias the conclusion either way. In our study, all but one of the patients have been operated in surgical departments of the OUH and 19 of the patients were operated at one surgical department. With few patients only our data do not show a clear association with treatment period and thus, within a highly specialized referral institution with previous experience in adrenal LS, no effect of a learning curve could be demonstrated.
Some authors conclude that patients with stage III tumors should be considered for OS only, at least when the tumor is larger than 10 cm [8]. How- ever, even after OS, the relapse rate in these patients is high in most series, and relapses occur relatively early. We could not, in the 12 patients with stage III disease, demonstrate a difference in relapse rates, type of recurrence, progression-free survival or overall survival in those with laparoscopy compared to those with OS. The only stage III patient in our series without a relapse was operated laparoscopi- cally for a left sided 16 cm tumor with invasion of the renal vein and has been followed postoperatively for 22 months. The number of patients is very small, and this conclusion may therefore be cautioned. However, if most patients will tend to relapse within a year regardless of operation technique, peri- and postoperative complications become more important in terms of quality of life. This is especially true since relapse is frequently treated with mitotane and/or chemotherapy; drugs associated with substantial side effects and reduced quality of life.
Declaration of interest: The authors report no conflicts of interest. The authors alone are respon- sible for the content and writing of the paper.
Acta Oncol Downloaded from informahealthcare.com by University of Nebraska on 01/03/15 For personal use only.
References
[1] Fassnacht M, Allolio B. Clinical management of adrenocor- tical carcinoma. Best Pract Res Clin Endocrinol Metab 2009;23:273-89.
[2] Zini L, Porpiglia F, Fassnacht M. Contemporary management of adrenocortical carcinoma. Eur Urol 2011;60:1055-65.
[3] Soreide JA, Brabrand K, Thoresen SO. Adrenal cortical carci- noma in Norway, 1970-1984. World J Surg 1992;16:663-7.
[4] Vassilopoulou-Sellin R, Schultz PN. Adrenocortical carci- noma. Clinical outcome at the end of the 20th century. Cancer 2001;92:1113-21.
[5] Kazaryan AM, Marangos IP, Rosseland AR, Røsok BI, Villanger O, Pinjo E, et al. Laparoscopic adrenalectomy: Norwegian single-center experience of 242 procedures. J Laparoendosc Adv Surg Tech A 2009;19:181-9.
[6] Porpiglia F, Miller BS, Manfredi M, Fiori C, Doherty GM. A debate on laparoscopic versus open adrenalectomy for adrenocortical carcinoma. Horm Cancer 2011;2:372-7.
[7] Miller BS, Ammori JB, Gauger PG, Broome JT, Hammer GD, Doherty GM. Laparoscopic resection is inap- propriate in patients with known or suspected adrenocortical carcinoma. World J Surg 2010;34:1380-5.
[8] Brix D, Allolio B, Fenske W, Agha A, Dralle H, Jurowich C, et al. Laparoscopic versus open adrenalectomy for adreno- cortical carcinoma: Surgical and oncologic outcome in 152 patients. Eur Urol 2010;58:609-15.
[9] Porpiglia F, Fiori C, Daffara F, Zaggia B, Bollito E, Volante M, et al. Retrospective evaluation of the outcome of open versus laparoscopic adrenalectomy for stage I and II adrenocortical cancer. Eur Urol 2010;57:873-8.
[10] Fassnacht M, Johanssen S, Quinkler M, Bucsky P, Willenberg HS, Beuschlein F, et al. Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: Proposal for a Revised TNM Classification. Cancer 2009;115:243-50.
[11] Lughezzani G, Sun M, Perrotte P, Jeldres C, Alasker A, Isbarn H, et al. The European Network for the Study of Adrenal Tumors staging system is prognostically superior to the international union against cancer-staging system: A North American validation. Eur J Cancer 2010;46:713-9.
[12] Dripps RD, Lamont A, Eckenhoff JE. The role of anesthesia in surgical mortality. JAMA 1961;178:261-6.
[13] Kazaryan AM, Marangos IP, Rosok BI, Rosseland AR, Edwin B. Impact of body mass index on outcomes of laparoscopic adrenal surgery. Surg Innov 2011;18:258-67.
[14] Satava RM. Identification and reduction of surgical error using simulation. Minim Invasive Ther Allied Technol 2005;14:257-61.
[15] Porembka MR, Hall BL, Hirbe M, Strasberg SM. Quantita- tive weighting of postoperative complications based on the accordion severity grading system: Demonstration of poten- tial impact using the American College of Surgeons national surgical quality improvement program. J Am Coll Surg 2010;210:286-98.
[16] Fassnacht M, Terzolo M, Allolio B, Baudin E, Haak H, Berruti A, et al. Combination chemotherapy in advanced adrenocortical carcinoma. N Engl J Med 2012;366:2189-97.
[17] Gonzalez RJ, Shapiro S, Sarlis N, Vassilopoulou-Sellin R, Perrier ND, Evans DB, et al. Laparoscopic resection of adre- nal cortical carcinoma: A cautionary note. Surgery 2005; 138:1078-85.