Repeat transsphenoidal surgery for Cushing's disease

Transcription

1 J Neurosurg 71: , 1989 Repeat transsphenoidal surgery for Cushing's disease ROBERT B. FRIEDMAN, M.D., EDWARD H. OLDFIELD~ M.D., LYNNETTE K. NIEMAN, M.D., GEORGE P. CHROUSOS, M.D., JOHN L. DOPPMAN, M.D., GORDON B. CUTLER, JR., M.D., AND D. LYNN LORIAUX, M.D., PH.D. Surgical Neurology Branch, National Institute of Neurological Disorders and Stroke, Developmental Endocrinology Branch, National Institute of Child Health and Human Development, and Department of Diagnostic Radiology, The Clinical Center, National Institutes of Health, Bethesda, Maryland ~" Transsphenoidal resection of adrenocorticotrophic hormone (ACTH)-producing pituitary adenomas has the potential of curing most patients with Cushing's disease. However, transsphenoidal exploration of the pituitary is not always curative, and patients who have remission of hypercortisolism following surgery occasionally develop a recurrence. Whether repeat pituitary surgery should be performed for or Cushing's disease has not been evaluated previously. To determine the efficacy of transsphenoidal surgery in or Cushing's disease, we performed transsphenoidal surgery in 31 patients (22 women and nine men) who had previously undergone a transsphenoidal operation and two female patients who had had previous pituitary irradiation only. In 24 (73%) of the 33 patients, remission of hypercortisolism was achieved by surgery. Although preoperative computerized tomography (CT) scanning identified an adenoma in only three of the 33 patients, in 20 patients a discrete adenoma was identified at pituitary exploration. The incidence of remission of hypercortisolism was greatest if an adenoma was identified at surgery and the patient received selective adenomectomy (19, or 95% of 20 patients), if there was evidence at surgery or by preoperative CT scanning that the previous surgical exposure of the pituitary was incomplete (seven, or 78% of nine patients), if an adenoma was seen on preoperative CT scanning (three of three patients), or if the patient had had prior pituitary irradiation without surgery (two of two patients). In contrast, only five (42%) of 12 patients who received subtotal or total hypophysectomy had remission of hypercortisolism. Surgically induced hypopituitarism occurred in six (50%) of these 12 patients, but in only one (5%) of the 20 patients who underwent selective adenomectomy. Three (13%) of the 24 patients who were in remission from hypercortisolism following repeat surgery developed hypercortisolism 10 to 47 months postoperatively. Repeat transsphenoidal exploration of the pituitary and treatment limited to selective adenomectomy should be considered in patients with hypercortisolism despite previous pituitary treatment. If an adenoma is identified during surgery, the chance of remission of Cushing's disease is high and the risk of hypopituitarism is low; however, if no adenoma can be found and partial or complete hypophysectomy is performed, remission of hypercortisolism is less likely and the risk of hypopituitarism is about 50%. KEY WORDS 9 Cushing's disease 9 Cushing's syndrome 9 transsphenoidal surgery 9 hypercortisolism ~ pituitary adenoma 9 microadenomectomy 9 Nelson's syndrome T RANSSPHENOIDAL resection of adrenocorticotrophic hormone (ACTH)-producing pituitary adenomas has the potential of curing most patients with Cushing's disease, However, cure rates with transsphenoidal pituitary exploration generally range from 20% to 80%, depending upon the institution and experience of the surgeon Therefore, many patients who undergo pituitary exploration are not cured of hypercortisolism or have hypercortisolism after an interval of apparent cure. These patients require further evaluation and treatment. Potential reasons for or Cushing's syndrome after previous pituitary surgery include: misdiagnosis in patients with ectopic ACTH syndrome, ectopic corticotropin-releasing hormone secretion, or pseudo-cushing's syndrome associated with depression or alcoholism; failure to locate small adenomas at operation; and incomplete removal of an adenoma. The latter two sources of surgical failure should potentially be correctable with repeat pituitary surgery. Although radiation, medical, and surgical therapies can be used for patients with Cushing's disease despite 520 J. Neurosurg. / Volume 71/October, 1989

2 Repeat transsphenoidal surgery for Cushing's disease TABLE 1 Previous treatment in 33 patients with or Cushing's disease* Case Age Procedure & Interval Before Recurrent vs. No. (yrs), Sex Tumor Location Present Surgery Pathology Additional Treatment Persistent H3,23ercortisolism transsphenoidal surgery 1 33, F selective adenomectomy, midline 11 mos adenoma 2 31, M pituitary exploration, no 29 mos normal pituitary 3 43, F selective adenomectomy, rt paramedian 38 mos adenoma 4 20, M resection central ~ anterior lobe, no 72 mos hyperplasia 5 38, M selective adenomectomy, rt 77 mos adenoma 6 30, F selective adenomectomy, no 64 mos normal pituitary 7 58, F pituitary exploration, NA 11 mos not available 8 52, F selective adenomectomy, It 72 mos normal pituitary 9 19, F pituitary exploration, no 18 mos normal pituitary 10 42, F selective adenomectomy, It 47 mos adenoma 11 36, F selective adenomectomy, midline 47 mos adenoma 12 24, M selective adenomectomy, midline 50 mos adenoma 13 62, M selective adenomectomy, It 10 yrs adenoma 14 43, F selective adenomectomy, adjacent to It CS 18 mos adenoma o,p'ddd, 24 mos 15 25, F selective adenomectomy, It 42 mos adenoma 16 34, F selective adenomectomy, rt 8 mos adenoma 17 48, M resection central ~ anterior lobe, no 8 yrs normal pituitary 18 51, F resection central ~ anterior lobe, no 38 mos normal pituitary 19 22, F selective adenomectomy, adjacent to rt CS 24 mos adenoma 20 21, M selective adenomectomy, posterior midline 12 mos adenoma 21 33, F selective adenomectomy, no 64 mos normal pituitary RT (46 Gy), 64 mos; o,p'ddd & aminoglutethamide, 64 mos 22 33, F three previous pituitary explorations, no 60, 36, 38 not available mos 23 33, M selective adenomectomy, no 60 mos normal pituitary 24t 32, F selective adenomectomy, It 10 mos adenoma 25 32, F total hypophysectomy, no 24 mos normal pituitary 26 46, F selective adenomectomy, It 72 mos adenoma 27 37, F resection central 89 anterior lobe, no 17 mos nomaal pituitary 28 39, F selective adenomectomy, midtine 36 mos not available 29 26, M selective adenomectomy, It 24 mos not available 30 39, F resection central 89 anterior lobe, no 39 mos adenoma 31 63, F selective adenomectomy; total hypophy- 30 mos, not available sectomy, no 18 mos radiation therapy only 32t 32, F NA 33 49, F NA RT(56Gy), 52mos bilat adrenalectomy, 66 mos RT (48 Gy), 30 mos; o.p'ddd, 30 mos RT (56 GY), 42 mos RT (43 Gy), 60 mos * RT = radiation therapy; t Gy = 100 rads; NA = not applicable; CS = cavernous sinus; o,p'ddd = 2,4'-dichlorodiphenyldichloroethane. t Cases 24 and 32 were the same patient: her disease recurred after surgery" following irradiation. previous pituitary surgery, these treatments are not always successful, may not eradicate the pituitary tumor, and may be associated with adverse side effects. ~4 Although repeat transsphenoidal surgery remains a therapeutic option for patients with or Cushing's disease, there are limited data as to the efficacy of surgery in this setting. Only four patients who received transsphenoidal surgery for or Cushing's disease after previous treatment by irradiation or surgery have been reported previously. ~ To determine the role of surgery after previous pituitary treatment, the results of pituitary' exploration were evaluated in 33 patients with or Cush- ing's disease after previous pituitary irradiation. Clinical Material and Methods pituitary surgery and/or Thirty-one patients (Cases 1 to 31:22 women and nine men, aged 19 to 63 years) with or Cushing's disease following previous pituitary surgery (with or without irradiation) received repeat transsphenoidal exploration 8 months to 10 years after their initial pituitary operation (Tables 1 and 2). Three patients had recurrence of Cushing's disease after their initial operation at the National Institutes of Health, The remaining 28 patients underwent initial surgery J. Neurosurg. / Volume 71 /October,

3 R. B. Friedman, et al. TABLE 2 Repeat transsphenoidal exploration of pituitary and outcome* Case CT Procedure & Findings; Pathology Postop Follow- Up Comments No. Scan Tumor Location Pituitary Function Period Selective Adenomectomy Only 1 normal selective adenomectomy, 5 mm; rt adenoma normal 6 mos 2 partial empty selective adenomectomy, 4.5 mm; rt adenoma normal (transient diabe- 1 mo t sella tes insipidus) 3 normal selective adenomectomy, 7 3 ram; adenoma normal 6 mos CD at 10 paramedian posteriorly mos 4 normal selective adenomectomy, 3 ram; rt adenoma panhypopituitarism; 20 mos transient diabetes insipidus 5 partial empty selective adenomectomy, 7 ram; rt adenoma normal 10 mos t sella 6 normal selective adenomectomy, 3 ram; It adenoma normal 48 mos CD at 47 mos 7 adenoma selective adenomectomy, 15 mm; It adenoma normal 6 mos t 8 normal selective adenomectomy, 8 ram; rt adenoma normal 13 mos 9 normal selective adenomectomy, 8 ram; rt adenoma normal 1 mo t 10 adenoma selective adenomectomy, 14 mm; rt adenoma normal 3 mos "~ 11 normal selective adenomectomy, 6 mm; rt adenoma normal 2 mos paramedian 12 normal selective adenomectomy, 4 ram; rt adenoma normal 2 mos paramedian 13 normal selective adenomectomy, 5 mm; It adenoma normal 1 mo 14 normal selective adenomectomy, 3 ram; adja- adenoma normal 5 mos invasion of CS noted cent to It CS: CS invasion at 2nd op 15 normal selective adenomectomy, 4.0 mm; It adenoma normal 2 mos invasion of CS noted at 2nd opt 16 normal selective adenomectomy, 7.0 mm; rt adenoma normal 1 mo 17 partial empty selective adenomectomy, 1.5 mm; rt tumor in sucker, normal 3 wks sella no tissue submitted 18 partial empty selective adenomectomy, 1.0 mm; rt adenoma normal 2 wks sella 19 normal selective adenomectomy, 7 mm; adja- adenoma normal 29 mos cent to rt CS: CS invasion invasion of CS noted at 1st & 2nd op * CT = computerized tomography; CS = cavernous sinus; CD = Cushing's disease; SA = selective adenomectomy. For previous treatment see Table 1. t Inadequate exposure of pituitary at previous operation. TABLE 2 (continued) elsewhere. Two additional patients (Cases 32 and 33: both women, aged 32 and 49 years) received pituitary surgery after previous pituitary irradiation only. All operations were carried out by the same surgeon (E.H.O.). Before repeat transsphenoidal exploration was considered, all patients underwent complete endocrine and radiographic investigation to confirm the diagnosis of Cushing's disease. In all patients preoperative endocrinological studies, which included the high dose of the Liddle dexamethasone suppression test 4'9 and the corticotropin-releasing factor stimulation test, 3'12 established the diagnosis of Cushing's disease. In addition, all patients had an ACTH concentration gradient in blood samples taken from at least one of the inferior petrosal sinuses when compared to peripheral blood ACTH levels. ~3 Thus, all patients had a confirmed diagnosis of Cushing's disease before repeat transsphenoidal surgery. Patients were considered to be in remission if the morning plasma cortisol level was less than 6.0 ug/dl when the patient had been off steroid replacement therapy for at least 24 hours and the 24-hour excretion of urinary free cortisol was consistently less than 90 ug/ day. 1,6,7,14,15 Surgically induced hypopituitarism was defined as the need for replacement of one or more hormones after operation which were not required previously (with the exclusion of glucocorticoids). Patients were reevaluated at 3-month intervals for the 1st year and annually thereafter. Results Twenty-four (73%) of the 33 patients, including 22 (71%) of the 31 patients with previous transsphenoidal pituitary surgery and both of the patients with previous treatment limited to pituitary irradiation, had resolution of hypercortisolism following surgery (Table 3). 522 J. Neurosurg. / Volume 71 /October, 1989

4 Repeat transsphenoidal surgery for Cushing's disease TABLE 2 (continued)* Follow- Case Postop No. CT Scan Procedure & Findings; Tumor Location Pathology Pituitary Function Up Comments Period Partial or Total Hypophysectomy 20 adenoma no adenoma seen, resection 70%, anterior pituitary; no tumor identified granulation tissue; normal pituitary hypothyroidism; hypo- 23 mos t gonadism 21 normal no adenoma seen, resection 70%, normal anterior hypothyroid since 1st op l0 mos anterior pituitary, no tumor identified pituitary 22 normal no adenoma seen, resection 95%, normal anterior hypothyroid since 1st op 39 mos anterior pituitary; no tumor identified pituitary 23 normal multiple abnormal-appearing areas, total normalanterior panhypopituitarism 13mos stalk transection at rehypophysectomy; no tumor identified pituitary peat transsphenoidal exploration 245 normal multiple abnormal-appearing areas, total normal anterior hypothyroidism 28 mos invasion of CS hypophysectomy; adjacent to It CS pituitary noted at 2nd op 25 partial empty multiple abnormal-appearing areas, SA, normal anterior normal 7 mos t sella 1.5 mm; 80% anterior lobe removed; pituitary rt 26 normal multiple abnormal-appearing areas, SA, adenoma hypothyroidism 1 mo t 4 ram: 90% anterior lobe removed; It 27 empty sella multiple abnormal-appearing areas, SA, normal anterior normal (transient 6 mos 2 mm; 80% anterior lobe removed; rt pituitary diabetes insipidus) 28 normal no adenoma seen, 80% anterior lobe adenoma hypothyroidism 14 mos removed; no adenoma identified 29 normal no adenoma seen, 60% anterior lobe adenoma diabetes insipidus 20 mos removed; no tumor identified 30 normal multiple abnormal-appearing areas, SA, normal anterior normal 2 mos died 2 mos postopw 3 ram; 40% anterior lobe removed; rt pituitary 31 partial empty no adenoma seen, no evidence of any normal anterior panhypopituitarism 20 mos sella pituitary tissue in sella; no tumor pituitary identified Selective Adenomectomy Only 32z~ normal selective adenomectomy, 3 mm; It adenoma 33 partial empty multiple abnormal-appearing areas, 80% normal anterior sella anterior lobe removed on rt; rt pituitary normal normal 10 mos recurrence at 10 mos 4 mos * CT = computerized tomography; CS = cavernous sinus; CD = Cushing's disease; SA = selective adenomectomy. For previous treatment see Table 1. t Inadequate exposure of pituitary at previous operation. $ Cases 24 and 32 were the same patient: disease recurred after surgery following irradiation. w Acute bacterial meningitis after repeat surgery; developed severe diffuse cerebral vasospasm and multiple cerebral infarcts; death 2 months after surgery. Although preoperative computerized tomography (CT) scanning identified an adenoma in only three of the 33 patients, in 20 patients a discrete adenoma was identified at pituitary exploration and surgery was limited to selective adenomectomy. In 19 (95%) of these 20 patients repeat surgery induced remission of hypercortisolism (Tables 1 to 5). In contrast, of 12 patients who received subtotal (10 cases) or total (two cases) hypophysectomy when multiple sites of abnormal tissue were apparent (seven cases) or when no abnormalappearing tissue was identified (five cases), only five patients were in remission after surgery. One of the nine patients who did not respond to repeat surgery and who had pituitary insufficiency (except for hypercortisolism) preoperatively had an empty sella, as identified by preoperative CT scanning. This patient had no tissue in the sella at operation. Therefore, of the 24 patients in remission after repeat transsphenoidal surgery (mean follow-up period of 11 months, range 0.5 to 39 months), 19 underwent selective adenomectomy, four subtotal hypophysectomy, and one total hypophysectomy; whereas, of the nine patients who did not have remission, one received selective tumor resection only, six underwent subtotal hypophysectomy, one received total hypophysectomy, and one had an empty sella (Tables 4 and 5). To determine if there were identifiable factors which might have indicated preoperatively which patients were most likely to benefit from surgery, we compared a number of clinical features between the patients who were in clinical and endocrinological remission and those who were not (Tables 4 and 5). Of the features that could be determined preoperatively, only two (evidence on CT scanning of a distinct adenoma or evi- J. Neurosurg. / Volume 71/October,

5 R. B. Friedman, et al. TABLE 3 Postoperative glucocorticoid data in 33 patients in this series Case No. Plasma a.m. Cortisol (ug/dl)* Urinary Free Cortisol (tzg/day)t Preop Postop Preop Postop remission ofhyperco~isolism < < < < < < < < > < < < < < < ~ 26 < < 5 33~ 23 < failure to control hyperco~isolism died * Normal 8 to 18 #g/dl. t Twenty-four-hour excretion; normal 20 to 90/~g/day. :~ Previous treatment consisted of radiation therapy only. dence of inadequate exposure of the sella at the prior surgery) were helpful in indicating those patients who would profit from additional surgery. Pituitary adenomas were diagnosed by CT scanning based on the following previously established and generally accepted criteria: if the gland was larger than normal; if a focal area of altered tissue density was clearly apparent in the sellar tissue either before or after the administration of contrast medium; if there was an asymmetrical upward convex shape to one-half of the gland; or if the pituitary stalk was deviated away from the side of a focal expansion of the gland. Deviation of the pituitary infundibulum alone was not considered diagnostic of the presence or the site of an adenoma, as most of these patients had undergone previous surgical manipulation of the intrasellar contents, often including removal of portions of the gland. In only three patients was an adenoma identified by preopera- TABLE 4 Correlation of outcome from surgery with operative findings and previous treatment* Feature Examined Remission No Remission no. of cases 24 9 selective adenomec- 19 selective adeno- 1 selective adenomectomy vs. subtotal mectomy tomy hypophysectomy 4 subtotal hypophy- 6 subtotal hypophysectomy sectomy 1 total hypophysec- 1 total hypophysectomy tomy 1 empty Sellat tumor identified at 19 adenomas identi- l adenoma identified surgery fled 5 multiple abnormal- 5 adenomas not appearing areas identified 3 no site of abnormality seen treatment between 21 no interval treat- 6 no interval treatoperations:~ ment ment 1 mitotane therapy 1 mitotane therapy & & pituitary radia- pituitary irradiaadenoma on preopertion 3 adenomas identition 1 bilateral adrenalectomy 1 pituitary irradiation 0 adenomas identiative sella CT scan fled fied (3 cases) vs hypercortisolism 13 4 interval between oper mos mos ations:~ (mean 48 mos) (mean 29 mos) inadequate exposure 7 patients 2 patients of pituitary at previous surgery (9 cases)~ * Excludes two patients with radiation therapy as their only previous treatment. CT = computerized tomography. ~- Patient with Sella devoid of pituitary tissue plus panhypopituitarism from first operation. z~ Does not include the two patients with prior radiation therapy but no prior surgery. tive CT scanning. These three patients were in remission after selective adenomectomy. In addition, all three of these patients and six additional patients, in whom an adenoma was not apparent on CT, had evidence on CT scanning and at surgery of inadequate exposure of the pituitary at the prior operation. This was evident when a retained sphenoid septum reached an intact portion of the sella floor or when a portion of the sella floor appeared normal and covered intact and normalappearing dura. Seven of the nine patients who had evidence of inadequate exposure of the pituitary gland at prior surgery were in remission after reoperation (Tables 2 and 4). Six of these patients were in remission after selective adenomectomy. No adenoma was identified in one patient: she is in remission after removal of 70% of the anterior pituitary gland on the side of the greater ACTH level, as demonstrated by bilateral simultaneous inferior petrosal sinus sampling. The remaining two patients were not in remission after subtotal hypophysectomy, performed because multiple ab- 524 J. Neurosurg. / Volume 71/October, 1989

6 Repeat transsphenoidal surgery for Cushing's disease TABLE 5 Selective adenomectomy vs. subtotal hypophysectomy* Procedure & Outcome No. of Cases selective adenomectomy 20 in remission 19 hypopituitarism l subtotal or total hypophysectomy 12 in remission 5 hypopituitarism 6 no tissue in sella at repeat surgery 1 in remission 0 hypopituitarism 1 * Comparison of procedures at pituitary exploration after previous pituitary surgery and/or radiation therapy. normal-appearing areas of pituitary tissue were noted during surgery. In 24 of the 33 patients, CT scanning provided no indication of an adenoma or of limited exposure of the sella at previous surgery. Seventeen (71%) of these 24 patients had remission of hypercortisolism after repeat surgery: in 14 of these, focal adenomas were identified intraoperatively despite negative CT scanning. Of the eight patients with evidence of a partially (seven cases) or completely (one case) empty sella, five were in remission after surgery. In four of these patients an adenoma was identified and selectively removed. One patient had been treated with radiation therapy only and had not received previous pituitary surgery. Two of the three patients with empty sella who were not in remission had previously undergone "total hypophysectomy" and one had previously received resection of the central third of the anterior lobe when no adenoma was identified during the prior surgery. Of the seven patients who had Cushing's disease despite previous surgery involving either total or subtotal hypophysectomy, only three (patients who had previously received partial hypophysectomy) were in remission after repeat surgery. Five of these seven patients had an empty sella on CT scanning and are described above. Both patients who previously received "total hypophysectomy" failed to respond to repeat surgery. Five patients had received resection of the central portion of the pituitary gland when no tumor was found at their prior surgery. In three of these five patients an adenoma was identified and resected from one of the remaining lateral wings of the anterior lobe at repeat surgery, and remission of hypercortisolism occurred. Six patients previously had received more than one treatment for Cushing's syndrome. One of the three patients previously with radiation therapy and transsphenoidal surgery and one of the two patients who had received more than one prior surgical exploration of the pituitary were in remission after repeat surgery. The sixth patient, who had previously received bilateral adrenalectomy and transsphenoidal surgery, exhibited hypercortisolism despite the repeat operation reported here. The mean interval between transsphenoidal operations was greater in patients with a successful outcome after reoperation (48 months for patients in remission vs. 29 months for the group with no remission). However, the difference was not significant (p < 0.10, t-test for unpaired samples). An adenoma was identified at the previous surgery and again at repeat operation in 12 patients. The fact that the tissue considered adenomatous at the previous surgery was indeed tumor was confirmed by either a pathological diagnosis of adenoma or the temporary resolution of hypercortisolism in all 12 patients. The adenoma at repeat operation was at or adjacent to the site of the initial adenoma in 11 of the 12 patients. In one patient the tumor site at repeat surgery was in the opposite half of the pituitary from that reported at the initial operation. Reoperation produced hypopituitarism in several patients. Partial or complete loss of pituitary function occurred in four of the 10 patients who received subtotal hypophysectomy, in both patients who received total hypophysectomy, in the patient in whom no tissue was found in the sella at repeat operation, and in one of 20 patients who received selective adenomectomy (Tables 2 and 5). Therefore, some compromise in pituitary function occurred in six of the 12 patients who received partial or complete hypophysectomy, but in only one of 20 patients who received selective adenomectomy. Hypopituitarism after repeat surgery did not correlate with an empty sella on preoperative CT scanning (one of eight patients), prior "total hypophysectomy" (one of two patients), prior central wedge resection of the anterior lobe (one of five patients), more than one prior transsphenoidal operation (one of two patients), previous transsphenoidal surgery and radiation therapy ( of three patients), or prior therapy limited to pituitary radiation ( of two). Cerebrospinal fluid rhinorrhea requiring repeat transsphenoidal exploration occurred in one patient (3%). Another patient developed acute bacterial meningitis 7 days after unsuccessful surgery; although the meningitis was successfully treated, she developed severe diffuse cerebral vasospasm and multiple bilateral cerebral hemispheric infarcts and died 2 months after surgery. Discussion There is no ideal therapy for patients with Cushing's disease who are not cured after previous pituitary treatment. The alternatives include bilateral adrenalectomy, radiotherapy, and several medical therapies; however, is ideal since is free from morbidity or has uniform success in controlling the hypercortisolism. Successful bilateral adrenalectomy cures hypercortisolism, but leaves the patient permanently dependent upon exogenous mineralo- and glucocorticoids, with J. Neurosurg. / Volume 71 /October,

7 R. B. Friedman, et al the risk of life-threatening acute adrenal crisis as well as a 10% to 30% risk of developing Nelson's syndrome.~~ Pituitary radiotherapy alone induces remission of hypercortisolism in only about 20% of adult patients with Cushing's disease. In addition, over half" of the patients who have received similar doses of pituitary radiation for treatment of other types of pituitary adenomas, such as growth hormone-secreting tumors, have developed some degree of pituitary insufficiency within I0 years. 5 Medical therapies utilizing 2,4'-dichlorodiphenyldichloroethane (mitotane), aminoglutethimide, trilostane, or ketoconazole are useful for the acute control of hypercortisolism and have been used for long-term management in selected patients, usually in conjunction with pituitary irradiation. TM However, excess cortisol production usually resumes when these drugs are discontinued. In addition, significant side effects often occur during medical treatment including: nausea and vomiting, somnolence, and dermatitis with mitotane; goiter, somnolence, dermatitis, and bone marrow depression with animoglutethimide; abortion, miscarriage, and decreased gonadal function with trilostane; and gonadal dysfunction, gynecomastia, and fatal hepatic necrosis with ketoconazole. Other complications, which are not now apparent, may be recognized in the future with continued use of the two relatively new agents, trilostane and ketoconazole. The results of this study provide guidelines for patient selection for repeat surgery for Cushing's disease and for intraoperative decisions that influence the outcome of surgery. Nineteen of the 20 patients in whom an adenoma was identified at surgery were in remission after repeat surgery. Only one of these patients suffered a compromise in pituitary function as a result of pituitary reexploration. Positive preoperative CT scanning suggests a likelihood of successful reoperation, as all three patients in this series with an identifiable tumor on pituitary CT scans had remission of hypercortisolism. However, CT scanning did not reveal an adenoma in 30 of the 33 patients, including 17 (85%) of the 20 patients in whom a discrete adenoma was identified at surgery. This low incidence of detection of the adenoma by CT scanning is comparable to our experience in patients with Cushing's disease without prior treatment; 35 (70%) of 50 patients had normal CT scans. 16 A CT scan may have predictive value even if the adenoma is not demonstrated. Seven of the nine patients with CT evidence of inadequate surgical exposure of the pituitary at the prior operation were in remission after repeat surgery. However, our experience indicates that, even if the preoperative CT imaging does not permit identification of an adenoma or show evidence of inadequate previous pituitary exposure, the 71% incidence of remission (in 17 of 24 patients) is sufficient to justify pituitary exploration. The outcome of repeat surgery was less rewarding when no definitive identification of an adenoma was made and partial or complete hypophysectomy was carried out. Of the lo patients who received subtotal hypophysectomy after intraoperative identification of multiple abnormal-appearing sites, or in whom no site of abnormality could be identified, only four were in remission and four developed hypopituitarism; this was also true of both patients who received total hypophysectomy, only one of whom was in remission. It has been common practice with initial pituitary surgery to perform a subtotal or total hypophysectomy when no adenoma is found. 1'6"7'17 Our results indicate that this procedure succeeds infrequently with repeat surgery (five (42%) of 12 patients). Since this approach often produces hypopituitarism, it should be selected carefully if it is to be employed at all. Our remission rate of four among l0 patients undergoing subtotal pituitary resection probably does not warrant the risk of loss of pituitary function (50%) to younger patients in their reproductive years, and can rarely be justified in older patients. The causes of tumor and surgical failures with other types of pituitary tumors (such as dural invasion and the large size of many growth hormonesecreting, prolactin-secreting, and nonsecreting tumors) are not necessarily the same as those that result in or Cushing's disease; in the latter cases, failure to find the tumor at initial surgery or fhilure to remove all of the tumor within the gland seems to be responsible for most treatment failures. Patients with ACTH-secreting tumors usually have very small tumors at initial consultation and at presentation with or disease; 12 of the 20 tumors identified in the patients reported here were 5 mm or less in greatest diameter. The results of the current study suggest that the success of repeat pituitary exploration is greater in patients with longer intervals between the first and last operations. Since many unsuccessful initial pituitary explorations do not expose a visible adenoma, the additional time may permit some of the smaller lesions to reach a size that will permit identification at a later operation. Furthermore, 11 of the 12 tumors that had been confirmed at prior operation (by pathology or by temporary resolution of hypercortisolism) were found at the same site at repeat surgery; this indicates that many patients with or Cushing's disease have adenomas that are incompletely removed at initial surgery rather than another explanation of Cushing's disease, such as a hypothalamie etiology, corticotroph hyperplasia, or multiple sites of adenoma. Repeat surgery was usually curative, whether or not radiographic imaging demonstrated an adenoma. The remission rates of 71% for reoperated patients and of 73 % in all patients who had received either prior surgery or pituitary irradiation compare favorably with other treatments for or Cushing's disease. However, three of the 20 patients who were in early remission developed hypercortisolism within 10 months of what initially seemed to be successful surgery. A longer follow-up period will be required to determine the incidence of permanent remission. Fur- 526 J. Neurosurg. / Volume 71 /October, 1989

8 Repeat transsphenoidal surgery for Cushing's disease therrnore, complications included hsq~opituitarism in eight (26%) of the 31 patients with previous pituitary surgery and cerebrospinal fluid rhinorrhea requiring operative repair in one (3 %). One patient died 2 months after repeat surgery of complications from acute bacterial meningitis that developed 1 week after surgery. Finally, we emphasize that all patients in this series were shown to have Cushing's disease by complete endocrine testing (which included positive results with the Liddle dexamethasone suppression test, the corticotropin-releasing factor stimulation test, and bilateral simultaneous inferior petrosal sinus sampling) to distinguish Cushing's disease from other causes of Cushing's syndrome before consideration of repeat pituitary surgery. Thus, the potential of submitting patients with ectopic ACTH secretion, or other causes of Cushing's syndrome other than Cushing's disease, was minimal. Furthermore, although repeat transsphenoidal surgery was frequently effective, producing a 73% overall remission rate, the incidence of remission was appreciably less than that achieved (> 90%) by us in a larger group of patients with Cushing's disease who had not received previous pituitary treatment (unpublished data). Repeat pituitary exploration should be considered in patients with or Cushing's disease, despite previous pituitary treatment. Most patients who receive repeat pituitary surgery limited to selective adenomectomy when an adenoma is identified during surgery have resolution of hypercortisolism; lasting adverse effects of surgery limited to selective adenomectomy are uncommon. Partial or total hypophysectomy when no adenoma can be identified is less rewarding and should be considered only in special circumstances, if at all. References i. Boggan JE, Tyrrell JB, Wilson CB: Transsphenoidal microsurgical management of Cushing's disease. Report of 100 cases. J Neurosurg 59: , Burch W: A survey of results with transsphenoidal surgery in Cushing's disease. N Engl J Med 308: , 1983 (Letter) 3. Chrousos GP, Schulte HM, Oldfield EIt, et al: The corticotropin-releasing factor stimulation test: an aid in the evaluation of patients with Cushing's syndrome. N Engl J Med 310: , Crapo L: Cushing's syndrome: a review of diagnostic tests. Metabolism 28: , Eastman RC, Gordon P, Roth J: Conventional supervoltage irradiation is an effective treatment for acromegaly. J Clin Metab 48:931, 1979 (Abstract) 6. Hardy J: Cushing's disease: 50 years later. Can J Neuroi Sci 4: , Hardy J: Transsphenoidal microsurgery of the normal and pathological pituitary. Clin Neurosurg 16:185-2 t 7, Laws ER Jr, Fode NC, Redmond MJ: Transsphenoidal surgery following unsuccessful prior therapy. An assessment of benefits and risks in 158 patients, J Neurosurg 63: , Liddle GW: Tests and pituitary-adrenal suppressibility in the diagnosis of Cushing's syndrome. J Clin Endocrinol Metab 21: , 196 l t0. Nelson DH, Meakin JW, Dealy JB Jr, et al: ACTHproducing tumor of the pituitary gland. N Engl J Med 259: , Nelson DH, Meakin JW, Thorn GW: ACTH-producing pituitary tumors following adrenalectomy for Cushing's syndrome. Ann Intern Med 52: , Nieman LK, Chrousos GP, Oldtield EH, et al: The ovine corticotropin-reteasing hormone stimulation test and the dexamethasone suppression test in the differential diagnosis of Cushing's syndrome. Ann Intern Med 105: , Oldfield EH, Chrousos GP, Schulte HM, et al: Preoperative lateralization of ACTH-secreting microadenomas by bilateral and simultaneous inferior petrosal sinus sampling. N Engl J Med 312: , Orth DN, Liddle GW: Results of treatment in 108 patients with Cushing's syndrome. N Engl J Med 285: , Salassa RM, Laws ER Jr, Carpenter PC, et at: Transsphenoidal removal of pituitary microadenoma in Cushing's disease. Mayo Ciin Proc 53:24-28, Saris SC, Patronas N, Loriaux DL, et al: Pituitary CT scanning in Cushing's syndrome. Radiology 162: , Wilson CB, Tyrrell JB, Fitzgerald PA, et al: Neurosurgical aspects of Cushing's disease and Nelson's syndrome, in Tindall GT, Collins WF (eds): Clinical Management of Pituitary Disorders. New York: Raven Press, 1979, pp Manuscript received May 13, Accepted in final form March 8, Address reprint requests to: Edward H. Oldfield, M.D., Building 10, Room 5D-37, National Institutes of Health, Bethesda, Maryland J. Neurosurg. / Volume 71/October,