OCR Text |
Show Journal of Neuro- Ophthalmology 16( 4): 252- 257', 1996. © 1996 Lippincott- Raven Publishers, Philadelphia Chiasmal Herniation as a Complication of Bromocriptine Therapy Pamela Taxel, M. D., David M. Waitzman, M. D., J. Frederick Harrington, Jr., M. D., Robert H. Fagan, B. S., Naomi F. Rothfield, M. D., Harry H. Chen, M. D., and Carl D. Malchoff, M. D., Ph. D. Introduction: Medical treatment of macroprolactinomas with dopamine agonists decreases tumor mass and improves visual defects. We report an unusual complication of a macroprolactinoma responding to bromocriptine: a visual field defect caused by downward herniation of the optic chiasm. Materials and Methods: A 64- year- old woman was found to have a 4.5 cm macroprolactinoma with superior displacement of the optic chiasm, bitemporal hemiano-pia, and serum prolactin concentration ( P) of 17,060 u. g/ L. Bromocriptine was initiated at 2.5 mg/ day and increased to 7.5 mg/ day over 2 months. Results: After 2 months, visual fields improved significantly and tumor height decreased to 3 cm with resolution of the optic chiasm displacement. P decreased to 1,180 jjug/ L. After 5 months of therapy, visual fields were normal, and P was 8 | xg/ L. After 8 months of therapy, new bilateral visual defects were observed. Magnetic resonance imaging ( MRI) revealed further decrease of the tumor height to 1.5 cm, and inferior and leftward traction of the optic chiasm as the probable mechanism for the new visual field deficit. P was < 1 | xg/ L. Bromocriptine was decreased to 5 mg/ day to allow reduced traction on the optic chiasm and its blood supply. Over the next 4 months, visual field abnormalities resolved. Conclusions: We report the development of a visual field abnormally that is explained by chiasmal herniation caused by a shrinking macroprolactinoma. This complication resolved with a decrease in the bromocriptine dose. We suggest that patients undergoing bromocriptine therapy for macroprolactinomas be followed for this potential complication. Key Words: Macroprolactinoma- Bromocriptine- Visual field defect- Optic chiasm- Prolactin. Manuscript received December 13, 1995; accepted March 22, 1996. From the Departments of Medicine ( P. T., N. F. R., C. D. M.), Surgery ( J. F. H., R. H. F.), Radiology ( H. H. C.), and Neurology ( D. M. W.), University of Connecticut Health Center, Farming-ton, and Department of Neurology ( D. M. W.), Veterans Administration Connecticut, Newington Campus, Newington, CT, U. S. A. Address correspondence to Dr. Carl D. Malchoff, University of Connecticut Health Center, Farmington, CT 06030- 1110, U. S. A. The medical treatment of pituitary macroprolactinomas with ergot alkaloids is a well- accepted form of therapy. The use of bromocriptine therapy results in tumor shrinkage in 90% of cases ( 1,2). Tumor size reductions of > 50% occur in 40- 85% of cases ( 3). Serum prolactin concentrations ( P) fall to < 10% of pretreatment values in 90% of patients and normalize in 66% ( 1). Bromocriptine therapy leads to visual field improvement in 80- 90% of patients with significant visual field abnormalities ( 4). These visual improvements often precede radiologic evidence of tumor shrinkage ( 4). In general, bromocriptine is without significant complications. We report an unusual complication associated with bromocriptine therapy and reduction in tumor size. In a patient with a large macroprolactinoma being treated with bromocriptine, we observed the expected resolution of significant visual field defects followed by the unexpected development of n ew visual field defects. The development of new visual field defects in a patient responding to bromocriptine with tumor shrinkage is a previously u n r e p o r t e d complication. The potential mechanism and successful treatment are described. CASE REPORT A 64- year- old white woman was referred for a macroprolactinoma. She developed amenorrhea at age 27 years after the birth of her last child. Galactorrhea was not present, and the etiology of the amenorrhea remained undetermined. Hypothyroidism was diagnosed soon thereafter, and thyroxine replacement was begun at 0.075 mg/ day. Family history was negative for kidney stones, hypercalcemia, or excessive peptic ulcer disease. Headaches and decreased visual acuity with loss 252 BROMOCRIPTINE COMPLICATION 253 of the temporal half of the right visual field developed gradually after age 61 years. Generalized pain and shoulder stiffness, along with an elevated sedimentation rate, led to a diagnosis of polymyalgia rheumatica within the same year. She was treated with prednisone and felt better on a maintenance dose of 5 mg/ day. Due to persistent visual complaints computed tomography ( CT) and magnetic resonance imaging ( MRI) scans were ordered at age 64 years and revealed a homogeneous mass with some calcification and a height of 4.5 cm. It extended from the suprasellar cistern to the pre-pontine cistern, with destruction of the clivus and sphenoid wing, and extension to the nasopharynx. Subsequent biopsy through the nose disclosed a prolactinoma; initial serum prolactin concentration was 17,060 ( jug/ L. Visual fields revealed severe bitemporal hemianopia. Visual acuity was 20/ 30 OD and 20/ 25 OS. Ophthalmoscopic examination showed moderate pallor of the optic nerves bilaterally, right greater than left. Thyroid stimulating hormone ( TSH) was 0.2 mU/ L ( reference range 0.7- 5.0), T4 was 116 nmoles/ L ( reference range 59- 152), T3 resin uptake ratio was 0.94 ( reference range 0.89- 1.17), follicle stimulating hormone ( FSH) was 9 U/ L ( reference range for postmenopausal women 35- 151), and luteinizing hormone ( LH) was 0.8 U/ L ( reference range for postmenopausal women 10.8- 61). Serum concentrations of growth hormone, IGF- 1, ionized calcium, and phosphorus were normal. MATERIALS AND METHODS Hormone assays TSH was measured by immunoradiometric assay ( Nichols Institute, San Juan Capistrano, CA, U. S. A.); radioimmunoassays ( RIA) were used for measurement of serum concentrations of prolactin ( kit from Diagnostic Products Corp., Los Angeles, CA, U. S. A.), growth hormone ( kit from Sanofi, Arlington Heights, MN, U. S. A.), FSH and LH ( kit from CIBA- Corning, Norwood, MA, U. S. A.) and total T4 ( kit from CIBA- Corning Diagnostic, Med-field, MA, U. S. A.). IGF- 1 was measured by extraction RIA ( Nichols Institute). Ionized calcium was measured by ion analyzer ( CIBA- Corning). Phosphorus was measured by spectrophotometry. Visual fields Visual fields were measured using a 30- 2 threshold test on an automated Humphrey perimeter. Visual field results are reported as Gray scale, and total and pattern deviations. Each point in the total deviation represents the average over 6° of the visual field and has been compared to a group of age- and acuity- matched controls. The severity of the defect at each point is expressed as a statistical deviation from the mean of the control group and, thus, has an attached confidence interval. Two statistical measures were used to evaluate changes in the patient's visual fields over time; mean deviation ( MD) and change in probability ( Statpac 2 Glaucoma Change Probability program). The MD compared the patient's visual field to a group of age- matched controls; change in probability compared a single visual field to a baseline established by the initial two visual fields performed by the patient. The standard deviation for the MD was - 3 . 2 db ( p = 0.05) and was - 3.0 db for the change in probability. This suggested that the patient's reliability was excellent and not different from a group of age- matched controls. RESULTS The initial MRI results and visual field abnormalities are summarized in Fig. 1. Bromocriptine was initiated at 2.5 mg/ day for 4 days and then increased to 5 mg/ day for 2 months. Within 2 days of initiation of therapy, the patient reported subjective improvement in her vision and decreased severity of her headaches. Within 1 month, P decreased to 3,008 u. g/ L. Within 2 months, the tumor height decreased to 3 cm ( Fig. 2), and visual field examination improved dramatically. After 2 months of bromocriptine therapy at 5 mg/ day, bromocriptine was increased to 7.5 mg/ day. Over the next 3 months ( 5 months since initiation of therapy), P decreased to 8 | xg/ L, and visual field defects gradually resolved ( Fig. 2); 3 months later ( 8 months of bromocriptine therapy), a worsening of visual fields occurred in both eyes ( Fig. 3). In the left eye, this deterioration reached statistical significance. The MD ( measure of the patient versus age- matched controls) declined from a previous value of - 0.62 db to - 3.04 db, which was not statistically different from the pretreatment level of - 5.16 db using the change in probability measure ( compares fields for a single patient). The right eye MD was - 2.05 db, which was much better than the pretreatment value of - 14.08 db, but was still worse than the patient's field 2 months earlier ( MD - 0.54 db). P was undetectable. Ophthalmoscopic examination revealed mild temporal pallor OD and bowtie atrophy OS ( pallor of the temporal and nasal portions of the optic nerve suggestive of chiasmal compression). MRI revealed further involution / Neuro- Ophthalmol, Vol. 16, No. 4, 1996 254 P. TAXEL ET AL. OS • • « a >.;<•• M • • • » ' ( 2 « • * • • « • • • • « GRRYTONE SYMBOLS REV 5.3 .3 t •=• .1 41 50 2 . 5 1 36 40 8 • t •;• 3 .2 31 35 2?, 10 26 30 7y 21 25 • x .<•; A*. .•>:. 251 * * 10106 t •=• 20 • 794 • t c 316 11 15 m v w * t - . * . _ OCT 1 O 1000 6 10 MM HUH 7343 3162 1 5 • > 10000 < 0 SYM HSB DB FIG. 1. Pretreatment visual fields and MRI. A: The initial field abnormalities and MRI scan are shown. Visual fields are shown by Gray scale and by total deviations, both of which compare patients' visual fields to those of an age- matched control group. The severe bitemporal hemianopia is incongruous, right more complete than left, suggesting more involvement on the right than the left. The prolactinoma has a height of 4.5 cm. P was 17,060 ( jLg/ L. B: Graytone symbols for the Humphrey automated perimetry. These apply to Figs. 1- 4. of the tumor with herniation and angulation of the optic chiasm ( Fig. 3). To allow for decreased traction on the optic nerve and/ or its blood supply, bromocriptine was decreased to 5 mg/ day. The field defects in both eyes recovered gradually with lowering of the bromocriptine dose. One month after the decrease of the bromocriptine dose, the MD was - 2.04 db OS and - 1.24 db OD. By the end of 2 months, the left visual field continued to improve and, at that time, was not statistically different from the field obtained after 5 months of bromocriptine therapy ( MD of - 0.91 versus - 0.62 db). The visual improvement was maintained, and 4 months after decreasing the bromocriptine dose from 7.5 to 5.0 mg/ day, the visual fields were almost normal OS with an MD of - 0.13 db, and showed only slightly decreased temporal sensitivity OD ( MD of - 0.75) compared to those of age-matched controls. Tumor height remained unchanged and P remained undetectable ( Fig. 4). The changes in MD over the course of therapy with bromocriptine are shown in Fig. 5. These results show that the patient's vision improved initially on bromocriptine, worsened after 6 months of bromocriptine therapy, and then gradually improved after the dose of bromocriptine was decreased. DISCUSSION We report a visual field defect caused by herniation of the optic chiasm induced by a shrinking macroprolactinoma. The macroprolactinoma initially responded to bromocriptine therapy with the expected significant decrease in size, and the initial visual field defects resolved completely. However, following resolution of the visual field defects, new visual field defects developed unexpectedly during therapy with bromocriptine. The most likely explanation for this new visual field defect was herniation of the optic chiasm in an inferior and left direction. This complication has not been reported previously in a prolactinoma responding to bromocriptine. Several observations and statistical analysis demonstrated that the new visual field abnormality was not an artifact of testing. Sequential testing revealed new bitemporal visual field deficits 6 months after initiating bromocriptine. When the bromocriptine dose was decreased, / Neuro- Ophthalmol, Vol. 16, No. i, 1996 BROMOCRIPTINE COMPLICATION 255 OS OD • « • • - • • • • m .. • « FIG. 2. Visual field examination after 5 months of bromocriptine therapy, and MRI after 2 months of bromocriptine therapy. After 2 months of bromocriptine therapy, the prolactinoma height had decreased to 3 cm, and the optic chiasm was no longer bowed upwards. P was 3,008 | xg/ L. Visual fields were improved markedly at this time ( not shown) and, as shown, had returned almost completely to normal after 5 months of bromocriptine therapy. The patient still had a small parafoveal bitemporal hemianopia as indicated in the total deviation. P was 8 ( xg/ L. the new deficits resolved over the next 4 months. The occurrence of this new visual abnormality could not be explained by noncompliance with therapy, since P was undetectable at the time of this complication. Decreased tumor responsiveness to bromocriptine was also unlikely, as MRI demonstrated that the prolactinoma was continuing to decrease in size. Loss of vision in association with an empty sella is usually, but not always, due to herniation of the optic chiasm ( 5- 9). Herniation of the optic chiasm has previously been reported following surgical excision, radiation therapy, or a combination of the two in patients with macroadenomas ( 5,9). Additional causes of the chiasmal syndrome include antibiotic therapy for tuberculous meningitis ( 6), transsphenoidal hypophysectomy for breast cancer ( 7), transsphenoidal emptying of a Rathke's OS OD * « a « a. ••• « . • v. a FIG. 3. Visual field examination and MRI after 8 months of bromocriptine therapy. After 8 months of bromocriptine therapy, the height of the prolactinoma had decreased to 1.5 cm, and inferior traction and angulation of the optic chiasm had developed. New visual parafoveal and peripheral field defects had developed involving opposite surfaces of the chiasm, inferior on left ( superior defect) and superior on right ( inferior defect). Bromocriptine dose was 7.5 mg/ day and P was not detectable. / Neum- Ophthalmol, Vol. 16, No. 4, 1996 256 P. TAXEL ET Ah. OS OD : • ' ' • . • • ; • . • "• . . •• : • • •• ' • . : . : . - . • : ' • • • ' • • • ' • • ' • • . - • • . • - . ' :• . • • . • • • . : • . • • . - . - • . :< W- % . . • • K FIG. 4. Visual fields and MRI 4 months after reducing the bromocriptine dose from 7.5 to 5 mg/ day. The visual field defects caused by optic nerve traction had returned to previous baseline. The tumor height remained at 1.5 cm, and P remained undetectable. cleft cyst ( 5), sarcoidosis ( 9), and postpartum pituitary necrosis ( 9). Visual field defects included bitemporal hemianopia, unilateral temporal field cut, and virtual blindness in one eye with hemianopia in the other ( 5). Most patients are managed surgically. In a series of 11 patients, this resulted in improved vision, although there were persistent visual field abnormalities in eight patients ( 5). Surgical treatment involved elevation of the diaphragm by packing of the sella in three patients, lysis of adhesions or incision of the diaphragm in Bromocriptine I - 1 - MD | 5 mg/ d 7.5 mg/ d 5 mg/ d | months 0 2 4 10 f\ f • Right Eye • Left Eye - 5 - 14 FIG. 5. Effect of bromocriptine dose on visual fields. The mean deviation ( MD), in db, compared the patient's visual fields to those of a group of age- matched controls. The MD of each eye is shown over time ( months) in response to the indicated daily doses of bromocriptine. three patients, and placement of a hole in the lamina terminalis in one patient ( 5). Surgical treatment in the patient we describe was considered; however, we elected a conservative approach. The rationale was to reduce traction on the optic nerve and/ or its blood supply by reducing the bromocriptine dose to allow slight tumor re- expansion. We chose not to continue the current bromocriptine dose or to increase it, since further shrinkage of the tumor could cause worsening visual field defects. Within 4 months of lowering the bromocriptine dose ( 7.5 to 5.0 mg/ day), perimetry returned to normal. P remained undetectable and tumor height remained at 1.5 cm. These improvements persisted. It is of interest that the visual fields improved without detectable re- expansion of the tumor on MRI scan. This experience is similar to the course of visual field improvement in shrinking prolactinomas and other pituitary tumors; visual defects often improve without or prior to radiologic changes ( 10). This suggests that the effect of the traction on the optic chiasm and/ or its blood supply can be relieved by small changes in tumor size that are undetectable by MRI. Ischemia to the optic chiasm has been demonstrated to cause visual field defects in primary empty sella syndrome ( 8). Moreover, ischemic chiasmal syndrome with or without herniation can also result from mechanical compression on the optic chiasm, optic nerve junction, or both ( 9). It is likely that relief of ischemia occurs with a small change in tumor size. Therefore, for the first time, we were able to treat a chiasmal / Neuro- Ophthalmol, Vol. 16, No. 4, 1996 BROMOCRIPTINE COMPLICATION 257 syndrome medically by titrating the bromocriptine dose. Although the mechanism of optic nerve herniation is not clear, there are two possibilities: to account for this prolapse and traction. The optic chiasm may have prolapsed into the space vacated by the shrinking prolactinoma or it may have been pulled down by its attachment to the shrinking prolactinoma. Herniation from a shrinking mass is usually perceived to be due to traction ( 5- 7,11). For traction to occur, there must be some mechanism to explain attachment of the chiasm to the tumor. This macroprolactinoma was, likely, of longstanding duration. It probably was present at age 27, when the patient developed amenorrhea. Over the years prior to diagnosis, the patient may have developed adhesions or scarring that attached the chiasm to the prolactinoma. The presence of calcifications in the tumor suggests previous hemorrhage and scarring and supports this possibility. The second possibility is that bromocriptine-induced shrinkage was responsible for fibrosis and, thus, attachment of the prolactinoma to the optic chiasm. Bromocriptine- induced fibrosis associated with tumor shrinkage has been identified in resected prolactinomas that had been pretreated with bromocriptine ( 12- 15). Therefore, in theory, it is possible that attachment of the optic nerve to the tumor occurred secondary to bromocriptine-induced tumor fibrosis. However, this latter possibility seems less likely, since chiasmal herniation with bromocriptine therapy is so rare. If bromocriptine were the cause of attachment of the optic chiasm to the prolactinoma, then chiasmal herniation would be observed more commonly. Therefore, it seems most likely that the attachment of the optic nerve to the tumor occurred prior to treatment with bromocriptine. In summary, we have described the development of a new visual field defect caused by herniation of the optic chiasm in a subject with a prolactinoma that was shrinking in response to bromocriptine. This new defect responded to a decrease in the bromocriptine dose. To our knowledge, this case represents a previously unreported complication of bromocriptine- induced tumor shrinkage and the first report of medical management of a chiasmal syndrome. We conclude that patients receiving bromocriptine therapy for mac-roprolactinomas should be followed carefully for development of new visual defects, even when the tumor is shrinking and initial defects are resolving. Acknowledgement: This work was supported, in part, by the General Clinical Research Center ( NIH Grant IM01RR06192), NIH Grant R29 DK- 42840, NEI Grant R29 EY- 09481- 04, NIH Grant AR- 20621, and a grant from the Donaghue Foundation, Hartford, Connecticut, U. S. A. * tWe would like to thank Mary Lee Vance, M. D., for her comments and input, and loanne D'Aprile, Laurie Amara, and Sue Hill for the typing of this manuscript. REFERENCES 1. Molitch ME, Elton RL, Blackwell RE, et al. Bromocriptine as primary therapy for prolactin- secreting macroadenomas: results of a prospective multi- center study. / Clin Endocrinol Metab 1985; 60: 698- 705. 2. Wass JAH, Williams J, Charlesworth M, et al. Bromocriptine in management of large pituitary tumors. Br Med } 1982; 284: 1908. 3. Bevan IS, Webster J, Burke CW, Scanlon MF. Dopamine agonists and pituitary tumor shrinkage. Endocr Rev 1992; 13: 220- 40. 4. Molitch ME. The pituitary. In: Melmed S, ed. Prolactinomas. Boston: Blackwell Scientific Publications, 1995: 443- 77. 5. Fisher EG, DeGirolami U, Suojanen JN. Reversible visual deficit following debulking of a Rathke's cleft cyst: a tethered chiasm? / Neurosurg 1994; 81: 459- 62. 6. Scott RM, Sonntag VKH, Wilcox LM, Adelman LS, Rockel TH. Visual loss from optochiasmatic arachnoiditis after tuberculous meningitis. Case report. / Neurosurg 1977; 46: 524- 6. 7. Decker RE, Carras R. Transsphenoidal chiasmapexy for correction of posthypophysectomy traction syndrome of optic chiasm. Case report. / Neurosurg 1977; 46: 527- 9. 8. Cupps TR, Woolf PD. Primary empty sella syndrome with panhypopituitarism, diabetes insipidus, and visual field defects. Acta Endocrinologica 1978; 89: 445- 60. 9. Lee F. Ischemic chiasma syndrome. Am } Neuroradiol 1983; 4: 777- 80. 10. Vance ML, Ridgway EC, Thorner MO. Follicle- stimulating hormone- and alpha- subunit- secreting pituitary tumor treated with bromocriptine. / Clin Endocrinol Metab 1985; 61: 580- 4. 11. Kaufman B, Tomsak RL, Kaufman BA, et al. Herniation of the suprasellar visual system and third ventricle into empty sellae: morphologic and clinical considerations. Am ] Neuroradiol 1989; 10: 65- 76. 12. Bevan JS, Adams CBT, Burke CW, et al. Factors in the outcome of transsphenoidal surgery for prolactinoma and non- functioning pituitary tumor including preoperative bromocriptine therapy. Clin Endocrinol 1982; 26: 541- 56. 13. Esiri MM, Bevan JS, Burke CW, Adams CBT. Effect of bromocriptine treatment on fibrous tissue content of prolactin-secreting and nonfunctioning macroadenomas of the pituitary gland. / Clin Endocrinol Metab 1986; 63: 383- 8. 14. Landolt AM, Osterwalder V. Perivascular fibrosis in prolactinomas: is it increased by bromocriptine? / Clin Endocrinol Metab 1984; 58: 1179- 83. 15. Molitch ME. Pathologic hyperprolactinemia. Endocrinol Metab Clin North Am 1992; 21: 877- 901. / Neum- Ophthalmol, Vol. 16, No. 4, 1996 |