neurology of the pituitary gland - jnnp’s ambition is to...

REVIEW: NEUROLOGY AND MEDICINE

Neurology of the pituitary gland

J R Anderson, N Antoun, N Burnet, K Chatterjee, O Edwards, J D Pickard, N Sarkies

This review will focus on those aspects ofpituitary disease immediately relevant to neu-rologists and neurosurgeons when assessingand counselling patients. It is essential to adopta multidisciplinary approach to the diagnosisand management of pituitary disease asemphasised by the recently published guide-lines from the Royal College of Physicians ofLondon.1–4

Range of pathology presenting in thesellar regionThe commonest lesions presenting in thisregion are pituitary tumours (incidence of15–20/million/year), including adenomas andcraniopharyngiomas, aneurysms, and meningi-omas, but many other diseases need to be con-sidered (table 1).

Neurological presentations of pituitarydisease”Pituitary incidentalomas” may be disclosedwhen investigating unrelated disease (fig 1).Although figures from 5% to 27% have beenquoted for the incidence of subclinical adeno-mas at postmortem, far fewer are of significantsize—that is, over 5 mm in diameter withdeviation of the stalk and unilateral enlarge-ment of the gland. Careful endocrine andvisual assessments are required and, where noabnormalities are found, most can be managedconservatively with follow up MRI.5–8

Pituitary disease often presents insidiouslyand in retrospect might have been detectedearlier. The symptoms of hormonal hyperse-cretion in endocrinologically active tumourswill obviously present before evidence ofsuprasellar or parasellar extension. Althoughsomatic changes usually bring the growth hor-mone secreting adenoma to medical attentionfirst, the neurologist may encounter nerveentrapment (particularly the carpal tunnel syn-drome), proximal myopathy (weakness dispro-portionate to the increased body mass),peripheral neuropathy (muscle atrophy, distalsensory loss, and neuropathic joints) and thepsychological and emotional sequelae of thedisease.9 It has been debated whether the emo-tional sequelae reflect a specific interactionbetween growth hormone and the limbicsystem.10 The headaches are often bitemporal,periorbital, or referred to the vertex, and do notseem to correlate with the size of the mass.11

Fronto-occipital headaches have also beendescribed that may be exacerbated by cough-ing. More remotely, acromegaly may presentwith the neurological complications of diabetesmellitus and hypertension. Some patients maypresent with symptoms drawn from more thanone endocrinopathy when there is dual secre-tion by the pituitary adenoma—for example, ofgrowth hormone and prolactin. Finally, not allgrowth hormone hypersecretion is the productof a pituitary adenoma—a rapidly progressive

DiVerential diagnosis of neoplasms and “tumour-like” lesions of the sellar region2

Tumours of adenohypophyseal origin Cysts, hamartomas, and malformationsPituitary adenoma Rathke’s cleft cystPituitary carcinoma Arachnoid cyst

Tumours of neurohypophyseal origin Epidermoid cystGranular cell tumour Dermoid cystAstrocytoma of posterior lobe and/or stalk (rare) Gangliocytoma

Tumours of non-pituitary origin Empty sella syndromeCraniopharyngioma Metastatic tumoursGerm cell tumours CarcinomaGlioma (hypothalamic, optic nerve/chiasm, infundibulum) PlasmacytomaMeningioma LymphomaHaemangiopericytoma LeukaemiaChordoma Inflammatory conditionsHaemangioblastoma Infection/abscessLipoma MucocoeleGiant cell tumour of bone Lymphocytic hypophysitisChondroma SarcoidosisFibrous dysplasia Langerhans’ cell histiocytosisSarcoma (chrondrosarcoma, osteosarcoma, fibrosarcoma) Giant cell granulomaPostirradiation sarcomas Vascular lesionsParaganglioma Internal carotid artery aneurysmsSchwannoma Cavernous angiomaGlomangiomaEsthaesioneuroblastomaPrimary lymphomaMelanoma

J Neurol Neurosurg Psychiatry 1999;66:703–721 703

Department ofNeuropathology,Box 235J R Anderson

Department ofNeuroradiology,Box 219N Antoun

Department ofNeuro-oncology,Box 193N Burnet

Department ofDiabetes andEndocrinology Box 157K Chatterjee

Department ofDiabetes andEndocrinology, Box 49O Edwards

Department ofNeurosurgery, Box 167J D Pickard

Department ofOphthalmology, Box41, Addenbrooke’sHospital, Cambridge,UKN Sarkies

Correspondence to:Professor JD Pickard,Department ofNeurosurgery, Box 167,Addenbrooke’s Hospital,Cambridge CB2 2QQ, UK.Telephone 0044 1223336946.

Received 2 Novembr 1998and in revised form27 January 1999Accepted 4 February 1999

on 3 May 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.66.6.703 on 1 June 1999. D

ownloaded from

http://jnnp.bmj.com/

history over a few months suggests an ectopictumour.

The clinical manifestations of Cushing’ssyndrome are well known. The patient usuallypresents at the microadenoma stage withremarkably little if any abnormality on MRI.Symptoms can fluctuate, particularly in teen-agers. Weakness from proximal myopathy, psy-chiatric disturbances including anxiety andpanic attacks, the neurological consequences ofobesity and hypertension, and benign intracra-nial hypertension12 may bring the patient to aneurologist.

A neurologist will usually see prolactinomasand endocrinologically inactive tumours whenthey have already reached a size suYcient toimpinge on the normal pituitary and extrasellarstructures (visual failure, ophthalmoplegia,trigeminal sensory loss, and hydrocephalus).Hypogonadal symptoms, weakness, fatigue, or

headache may be the presenting feature.Hypogonadism may be compounded by mildhyperprolactinaemia resulting from the eVectof raised intrasellar pressure on the normalpituitary gland (see later). The neurologicalmanifestations of hypothyroidism include car-pal tunnel syndrome, slowing of cerebration,and, less commonly, myotonia and myopathy.

The neurological consequences of hyperthy-roidism including exophthalmos may herald athyrotrophinoma, the great majority of whichare macroadenomas and hence may be accom-panied by headache, visual field defect, andgalactorrhoea. However, other conditions maybe associated with hyperthyroidism plus adetectable concentration of serum thyro-trophin and specialist expertise is required todiVerentiate them.13

Children with pituitary and third ventricularlesions may present with precocious puberty,growth retardation, diabetes insipidus, denialof visual failure, and the symptoms and signs ofeither high or low pressure hydrocephalus.

Some diagnostic pitfallsAlthough most patients with a pituitary prob-lem fall within well defined categories, aminority present with manifestations of one ofthe myriad of small print diagnoses. Ectopictumours may cause not only Cushing’s syn-drome but also acromegaly. Primary hypo-thyroidism may be accompanied by pituitarygland swelling3 14 as the result of overstimula-tion from the hypothalamus leading to visualsymptoms, hypopituitarism, headache, and lesscommonly, diabetes insipidus, syndrome of

Figure 1 (A) Normalappearance of pituitarygland. Sagittal TI weightedimage. Note the bright signalfrom the posterior lobe of thepituitary. More posteriorlythe dorsum sella (arrowhead) is shown as linear highsignal (marrow) surroundedby low signal (dense cortex).(B) Microadenoma on theleft side. This has a lowersignal than normal pituitary.The normal pituitary andcavernous sinus showenhancement aftergadolinium injection.

Figure 2 (A) Pituitary haemorrhage which presentedwith apoplexy. Coronal TI weighted image. High signalblood within an enlarged sella. Infundibulum deviated tothe left. (B) Infarction of macroadenoma. Postgadoliniumcoronal TI weighted image. The optic chiasm is stretchedand displaced superiorly (arrow head). Enhancement inresidual viable adenoma superiorly and to the left. The verybright signal represents small areas of haemorrhage

704 Anderson, Antoun, Burnet, et al



j.com/

J Neurol N


ownloaded from


innappropriate antidiuretic hormone secretion(SIADH), and precocious puberty. To add tothe confusion, one third of such female patientsmay have menstrual irregularities as a reflec-tion of modest hyperprolactinaemia leading toa misdiagnosis of prolactinoma. The pituitaryswelling resolves with thyroxine and surgery isunnecessary.

The stalk compression syndrome refers tothe symptoms and signs of hyperprolactinae-mia in the presence of moderately raised prol-actin concentrations and a sellar or suprasellarmass but distortion of the pituitary stalk alonedoes not correlate with prolactinconcentrations.2 3 15 It is not clear why it is theprolactin inhibitory system alone of thehypothalamo-anterior pituitary control sys-tems that is predominantly aVected. Directmeasurements of intrasellar pressure madeduring trans-sphenoidal surgery have shownthat the intrasellar pressure may be raisedabove the presumed perfusion pressure of theanterior pituitary gland.16 The anterior pitui-tary is only supplied by thin walled portal ves-sels (derived from the superior and inferiorhypophyseal arteries) that can easily be ob-structed by CSF/intrasellar pressures aboveabout 20 mm Hg. Dynamic CT after contrastinjection discloses rapid enhancement of theposterior pituitary (direct arterial supply fromthe inferior hypophyseal arteries) followed byprogressive enhancement of the anterior pitui-tary as the contrast supplied through the portalcirculation extravasates through the defectiveblood-brain barrier in the anterior pituitarygland.17 There is no specific concentration ofprolactin that distinguishes stalk compressionfrom a true prolactinoma but values of2500–5000 mU/l are often quoted despitemany exceptions in the literature. It is moreimportant to take the concentration in the con-text of the imaging.18 If the macroadenoma is aprolactinoma, the prolactin concentrationshould be much greater than 2500 mU/l. Ifthere is no significant intrasellar or suprasellarmass and the concentration is less than 2500,this cannot be the result of stalk compressionor raised intrasellar pressure. In addition thereare many other causes of hyperprolactinaemiato be considered. Hopefully, the days are gonewhen patients with symptomatic hyperprolacti-naemia were treated with dopamine agonistswithout any radiological imaging—bromocriptine will suppress the galactorrhoeaof stalk compression hyperprolactinaemiawhereas the craniopharyngioma gets bigger!Immunocytochemistry on the excised tumouris the final arbiter.

Lymphocytic hypophysitis often presents atthe end of pregnancy with visual failure and ithas also been described in men. MRI shows anenlarged homogeneously enhancing pituitarygland but a specific diagnosis cannot be madeby imaging alone. Management requires care-ful coordination between obstetrician, endo-crinologist, ophthalmologist, andneurosurgeon—early delivery may lead to earlyshrinkage and avoid trans-sphenoidal decom-pression provided vision has not deterioratedtoo far. Lymphocytic hypophysitis represents

an autoimmune attack on the adenohypophysisand manifests histologically as diVuse lym-phocytic infiltration eVacing and destroying thenormal glandular elements with progression tofibrotic scarring.19 The infiltrate is polyclonal,predominantly T cell, but often includeslymphoid germinal centres, plasma cells, andfoamy macrophages. A similar lymphocyticinflammatory process may be confined to theneurohypophyseal system, sparing the anteriorpituitary. By contrast, granulomatous hypo-physitis is characterised by non-caseatingepithelioid granulomata and multinucleateLanghans-type giant cells. When restricted tothe anterior pituitary this pathology probablyoverlaps with lymphocytic hypophysitis20 but inthe posterior pituitary it is more often a mani-festation of neurosarcoidosis and invariablyassociated with involvement of the hypothala-mus and basal meninges. In any inflammatoryprocess, unusual infections, particularly myco-bacterial or fungal, must be excluded but inwestern countries these are unlikely except inthe context of immunosuppression.

The symptoms and signs of extrasellarextension depend on its direction:x Downwards—nasal obstruction or epistaxis;

CSF leakx Upwards—visual failure (see later); hypotha-

lamic disturbance, third ventricular obstruc-tion with acute hydrocephalus (symptomsand signs of raised intracranial pressure), orchronic hydrocephalus (gait apraxia, slowingof mentation, and incontinence)

x Laterally—cavernous sinus syndrome; com-plex partial seizures.

Pituitary apoplexy (fig 2)As pituitary adenomas enlarge, they acquire adirect arterial blood supply in addition to theportal blood supply but their blood supplyremains tenuous.3 21–23 There is a range of pres-entation of haemorrhagic infarction from sub-clinical infarction, noted at operation or onpreoperative imaging (the prevalence of haem-orrhage within adenomas is 9%-43% on MRIof which less than 30% are symptomatic), tofull blown apoplexy that mimics subarachnoidhaemorrhage. Indeed, the diagnosis may ini-tially be missed. Many of these patients maynot be aware of a pre-existing adenoma. Inaddition to abrupt headache with or withoutloss of consciousness, the clinical featuresinclude visual failure,24 cavernous sinus syn-drome, focal neurological signs, and acutepanhypopituitarism, sometimes with low bloodpressure. Pituitary apoplexy can occur with arelatively small adenoma and is often spontane-ous but may be precipitated by endocrine stresstests, bromocriptine,25 26 and anticoagulants.The unenhanced emergency CT usually showsan enlarged pituitary fossa with some blood orspeckled density within the mass. Magneticresonance imaging will then confirm the diag-nosis and angiography is rarely needed toexclude a vascular abnormality. Haemorrhageappears as high signal on both T1 and T2weighted imaging, and can be confused withcraniopharyngioma and Rathke’s cleft cyst,especially as the signal changes can persist for

Neurology of the pituitary gland 705



j.com/

J Neurol N


ownloaded from


3–12 months, to be replaced by an area ofhypointensity in the chronic phase. Manage-ment includes resuscitation, steroid replace-ment, and early transphenoidal decompressionwith craniotomy reserved for the very smallminority of patients with a dumb-belltumour.27

Empty sella syndrome (fig 3)Empty or partial empty sella are terms usedwhen the pituitary gland does not fully occupythe sella, and this is found in 6% of routinepostmortems and MRI examinations when amidline sagittal image is obtained. It is morecommon in multiparous women, presumablydue to repeated enlargement and involution ofthe gland with each pregnancy. Empty sella is aradiological diagnosis based on CT or MRIperformed for headache, CSF leak, or latevisual deterioration after surgery or radiationtherapy for a pituitary lesion. The radiologicaldiagnosis of an empty sella should be madewith care—a cystic craniopharyngioma orRathke’s cleft cyst may sometimes mimic thenormal anatomical contours both within andabove the sella. This distinction is particularlyimportant in those 18% of children withneuroendocrinological abnormalities whocould have an empty sella, cyst, orcraniopharyngioma.28 29 This should be less ofa problem with MRI than it was with CT. Pri-mary empty sella conventionally is not associ-ated with significant endocrine or visual abnor-

malities. Sensitive dynamic endocrine testingmay disclose abnormalities but hormone re-placement is seldom indicated.30 An emptysella is commonly seen in patients with benignintracranial hypertension31 and may be thesequel to pituitary apoplexy. About 50% ofadult patients with a primary empty sella haveantipituitary antibodies that indicate previousautoimmune hypophysitis.32 Postpartumnecrosis of the pituitary gland may be followedby an empty fossa. Spontaneous CSF rhinor-rhoea may be associated with an empty sellawith or without raised intracranial pressure.33

When there has been a previous cause for theempty sella, then there may be an associatedendocrine abnormality (either hypopituitarismor hormone hypersecretion) if there is anaccompanying pituitary adenoma, particularlywhen shrunk either by bromocriptine or soma-tostatin. Prior intrasellar surgery and radio-therapy may lead to an empty sella.

The optic chiasm may herniate down intothe empty fossa but there is considerabledebate as to whether such herniation itself canbe responsible for visual field defects.34 Radia-tion necrosis and arachnoiditis are more likelyto be the cause of visual failure in such circum-stances. Although some good results from chi-asmapexy have been reported,35 both for therelief of intractable headache and for visualfailure, there have certainly been disasters,which inevitably go underreported.

Diabetes insipidus and SIADHNeither of these fluid balance disturbances arecommon presentations of anterior pituitarypathology—they usually indicate hypothalamicor pituitary stalk pathology. The posteriorpituitary gland and stalk are supplied by anarterial circulation and the rise in intrasellarpressure produced by a pituitary adenoma isinsuYcient to compromise such an arterial cir-culation except possibly after pituitary apo-plexy. Vasopressin and oxytocin are producedin hypothalamic nuclei and are transported bytheir axons to be stored as neurosecretorygranules in the posterior pituitary whichappear as a high signal intensity on MRI. Thenormal high signal (see fig 1A) is absent inalmost all patients with central and nephro-genic diabetes insipidus.36 This, however,should be interpreted with caution as thebright signal is also absent in 15%-20% of thenormal population, albeit usually in the elderlyage group who, although asymptomatic, have ahigh circulating plasma osmolality.37 The pres-ence of the bright signal indicates normal vaso-pressin stores, implying an intact neurohypo-physeal system, which eliminates centraldiabetes insipidus as a diagnostic possibility.Ectopic location of the posterior lobe is veryrare as an isolated finding with normal pituitaryfunction. Any process, however, that disturbstransport of hormones from the hypothalamusto the neurohypophysis as in transection of thestalk, after head injury or surgery, or destruc-tion of the posterior lobe, can result in theaccumulation of the high signal material proxi-mal to the site of obstruction.38 In idiopathic

Figure 3 (A) Empty sella with normal residual pituitarytissue in the floor of the sella. (B) Enlarged partly emptysella after treatment of a macroadenoma. Note the prolapseinto the sella of both the chiasm and more anteriorally, theposterior part of the gyrus rectus of the frontal lobe.




j.com/

J Neurol N


ownloaded from


growth hormone deficiency, there is a highincidence (43%) of ectopic posteriorpituitary.39

Diabetes insipidus is more common withmetastases (33%) compared with adenomas(1%) (see fig 7F).40 Neoplasms of the posteriorpituitary are rare. Low grade gliomas thatclosely resemble astrocytomas may occur in theposterior lobe or stalk (fig 4). Benign, granulartumours, also called choristomas, contain largePAS positive granules which are solely due toabundant large lysosomes.

Visual failureANATOMY OF THE CHIASM

The chiasm (4 mm thick, 12 mm wide, and 8mm long) lies about 1 cm above the pituitaryfossa, inclining as much as 45° from thehorizontal.41 In humans, about 53% of fibresdecussate and there are about 2 million axonsfor both nerves.42 Because of the normal varia-tion in the length of the optic nerves, thechiasm may be prefixed (overlying the tubercu-lum sella anteriorly) or postfixed (overlying thetuberculum sella posteriorly).43 Small tumoursin this region or ophthalmic artery aneurysmsmay cause unilateral blindness by pressure onone intracranial optic nerve. A space of about 1cm between the dorsum sella and the chiasmwill accommodate a reasonably large tumourarising above the dorsum sella before itcompresses the chiasm. Fibres from the maculacomprise most of the fibres of the optic nerve,chiasm and optic tract. As a consequence,pituitary tumours compressing these structureswill have field defects that aVect the central 30°and so may be detected using the tangentscreen and automated perimeters which testout to 24° or 30°.

The inferior nasal fibres which cross withinthe chiasm sweep forward into the oppositeoptic nerve and thence into the opposite optictract, forming Wilbrand’s knee.44 Recent ana-tomical studies of the chiasm have emphasisedthat the exaggerated bulge of crossing fibres,which seems to involve the contralateral opticnerve, is partly an artefact caused byenucleation.45

VISUAL SYMPTOMS OF PITUITARY TUMOUR

Some patients may be unaware of their fielddefects. The most frequent complaints ofpatients with chiasmal compression from pitui-tary tumours are progressive loss of centralacuity and dimming of the visual field,especially in its temporal portion. However, thelack of a binocular temporal field may alsocause diYculties with depth perception fornear vision and diplopia. Convergence for nearvision results in crossing of the two blind hemi-fields so that an object posterior to fixationapparently disappears—so called postfixationblindness. Patients often complain of diYcultywith fine tasks such as threading a needle andcutting fingernails.46 Diplopia without anocular motor paresis results from “verticalslip”. Patients with a bitemporal hemianopiado not have a link between the remaininghemifields so that vertical or horizontal separa-tion occurs between the two intact nasal hemi-fields and visual sensory diYculties result.Diplopia may also result from involvement ofthe third, fourth, and sixth cranial nerves in thecavernous sinus causing a disturbance of ocularmotility.

FIELD DEFECTS IN CHIASMAL LESIONS

The classic pattern of a chiasmal visual fielddefect is a bitemporal depression and greater

Figure 4 (A, B) Chiasmal/hypothalamic glioma. Sagittal and coronal T1 weighted image (postgadolinium). (C, D)Astrocytoma (suprasellar/sellar); coronal TI weighted image before and after contrast medium injection




j.com/

J Neurol N


ownloaded from


near fixation, but the eVects may be peripheral,central, or a combination. The defects may beabsolute or relative.47 Patients with chiasmaldisease may have normal or reduced acuity andcolour vision. The initial field defect usuallyoccurs in the superior temporal quadrants andthen to the inferior nasal and superior nasalquadrants.

The types of field defect with chiasmallesions may be classified as:

(1) anterior angle of the chiasmWhen only a small lesion involves the crossingfibres of the ipsilateral eye, the field defect istemporal with a midline hemianopic character;if only the macular crossed fibres are aVectedthe field defect is paracentral and temporal.When both the crossed and ventral fibres fromthe contralateral eye are aVected, there is adefect in the temporal field of that eye. If thereis extensive involvement of one optic nerve, anextensive field defect or even blindness of thateye may result; when such a lesion extends toinvolve the chiasm, the earliest indication is atemporal defect in the contralateral eye.

(2) Lesions involving the body of the chiasmThese characteristically produce a bitemporalhemianopia that may be peripheral, central, ora combination of both with or without splittingof the macula. Usually the visual acuity is nor-mal. In the field of the right eye, the defectprogresses in a clockwise direction and, in theleft eye, a counterclockwise direction. Acomplete bitemporal hemianopia is rarely seenexcept after trauma to the chiasm. Mostcompressive lesions cause relative bitemporalhemianopias.

(3) Lesions at the posterior angle of the chiasmThese lesions characteristically produce bitem-poral homonymous scotomas often associatedwith peripheral bitemporal defects as well.Lesions aVecting this area often also compressthe ipsilateral optic tract.

(4) The lateral aspect of the chiasmCompression of the lateral aspect of the chiasmaVects both the uncrossed temporal fibres andthe crossed nasal fibres causing a contralateralhomonymous hemianopia.

PSYCHOPHYSICAL ABNORMALITIES IN CHIASMAL

LESIONS

Recent investigation with contrast sensitivitytechniques has attempted to characterisechanges in visual function resulting fromchiasmal compression before visual acuity andvisual field are aVected. Studies of temporalcontrast sensitivity have shown both uniformand low temporal frequency loss of contrastsensitivity in patients with chiasmalcompression.48 Studies assessing spatial fre-quency sensitivity in lesions aVecting thesuprasellar region have found alterations eitherto be total or limited to low to medium spatialfrequencies only.49 A recent study of psycho-physical losses in foveal sensitivity foundsignificant losses in chromatic, luminance, andtemporal sensitivities in patients with suprasel-lar lesions without field defects.50

FUNDAL SIGNS

Optic atrophy is a late sign of chiasmalcompression from pituitary tumour; usuallythe optic discs are normal. Optic atrophy isweakly correlated with the duration of visualsymptoms and strongly correlated with persist-ent postoperative decreased visual acuity.51

There is a characteristic optic atrophy found inadvanced chiasmal compression—”bow tie” or“band” atrophy. The optic disc shows atrophyof the nasal and temporal margins with relativesparing of the superior and inferior portionswhere most spared temporal fibres (serving thenasal field) enter.47 A relative aVerent pupillarydefect may be detected when there is asymmet-ric compression of the optic nerves.

THE MECHANISM OF CHIASMAL COMPRESSION

The mechanism by which pituitary adenomasand other lesions damage selectively the decus-sating nasal fibres remains uncertain. Becausethe axons for the entire superior visual field arein the inferior portion of the chiasm, compres-sion from below would be expected to producea defect in the entire upper field; yet an altitu-dinal hemianopia is extremely rare. Thus,unless the decussating fibres are more vulner-able to compression than the uncrossed fibres,simple compression does not provide a com-plete explanation for bitemporal field defects.

There is evidence from animal studies of theeVect of direct compression on the anteriorvisual system. In a macaque monkey asuprasellar meningioma compressed the opticnerves, chiasm, and tracts dorsally causing aselective loss of small diameter fibres.52 In cats,optic nerve pressure obstruction has beenshown to disrupt large diameter fibresselectively.53 The vulnerability of the crossingfibres to compression from below may berelated to the vascular supply; in humans, thecentral chiasm derives its blood supply fromthe inferior vessels.43 Alternatively, the attach-ments of the optic chiasm are such thatexpanding tumours produce tension on thecrossing fibres only.54

THE PROGNOSIS FOR VISUAL FUNCTION AFTER

TRANS-SPHENOIDAL HYPOPHYSECTOMY

After uncomplicated surgical decompression ofthe chiasm and optic nerves, visual acuity andvisual fields usually improve rapidly over a fewdays. There may also be a slower improvementoccurring over weeks and months. Visualrecovery is usually complete by 4 months.55 56

Visual outcome is favourable in most patients;in a recent report of 67 patients, the vision wasimproved in 88%, the same in 7%, and worse in4%.55 The extent of improvement depends onwhether permanent damage to the optic nerveand chiasm has occurred. The extent of preop-erative damage may by judged from a loss ofarcuate nerve fibres surrounding the optic disc(the nerve fibre layer) and the degree of atrophyand pallor of the disc itself. The duration ofvisual loss and the degree of chiasm and opticnerve compression are the chief influences onthe preoperative visual status. In cases ofadvanced visual failure and severe opticatrophy, it is wise to caution the patient that




j.com/

J Neurol N


ownloaded from


decompression may prevent further visualdeterioration but may not improve visual func-tion.

Rarely, visual function may deteriorate aftertrans-sphenoidal hypophysectomy particularlywith second operations or after radiotherapyand all patients should be warned of this possi-bility before surgery. Visual deterioration aftersurgery is usually related to an intracapsularhaematoma, possibly as the result of infarctionof a tumour remnant, or direct surgicalmanipulation of the optic nerve or chiasm andits blood supply. Visual deterioration is morelikely after craniotomy, particularly when radi-cal removal of a capsule from the chiasm hasbeen attempted.

The other important concern in follow up iswhether patients with persistent bitemporalfield defects should be allowed to drive.According to the current criteria of the UnitedKingdom Driver and Vehicle Licensing Au-thority, if there is a bitemporal field defectaVecting the central 20°, patients should beadvised that they will not be permitted to drive.

InvestigationsENDOCRINOLOGY

Patients with pituitary tumours should beassessed by a multidisciplinary team.1 Theyoften require long term supervision, and earlyreferral to the appropriate endocrinologist issensible. The only urgent endocrine testrequired in a patient with visual failure and alarge suprasellar mass is a basal prolactin con-centration. Endocrine tests should be tailoredto the individual patient and not ordereduncritically. The conventional dynamic endo-crine tests may precipitate pituitary apoplexy,as most sleep deprived pituitary surgeons cantestify.57 Inferior petrosal sinus sampling forCushing’s syndrome uses a percutaneous bilat-eral femoral approach and requires obsessionalattention to detail if useful results are to beobtained.58 In experienced hands, the proce-dure is safe but cavernous sinus thrombosis

and brainstem infarction have been seen—heparin and a gentle contrast injection tech-nique are essential.

IMAGING59 60

Magnetic resonance is the first choice forimaging the pituitary and parasellar region asno ionising radiation is involved and boneinduced artefact is avoided. Plain skull radiol-ogy is seldom indicated except for surgicalplanning in selected cases. Computed tomog-raphy is reserved for emergency use and for theclaustrophobic, the MRI incompatible, thosetoo large for the bore of the magnet, and forspecialised indications when bone details areuseful (CSF leaks and surgical planning). Eachneuroradiology department has a standardMRI protocol but this has to be refined forspecific clinical indications and close commu-nication between radiologist and clinician isessential if errors are not to occur. Particularcare is required with long term, multidiscipli-nary follow up when miscommunication can soeasily occur with missed or mislaid scans mak-ing accurate comparisons and timely manage-ment decisions diYcult.

Clinicians should be aware of the age relatedchanges in pituitary size and characteristicsfrom childhood through puberty and preg-nancy to old age. The maximum allowablepituitary height is 6 mm in children, 8 mm inmen and postmenopausal women, 10 mm inwomen of child bearing age, and 12 mm in latepregnancy and the postpartum period.61 62

Microadenomas, defined as less than 10 mmin diameter, are 400 times more common thanmacroadenomas.63 Magnetic resonance imag-ing is emerging as a superior technique to CTfor the detection of microadenomas Thedetection rate varies in the literature from 65%to more than 90%.64 This is related to the tech-nique used, the continuing improvement inequipment, the use of contrast, and the size ofthe microadenoma. Most microadenomas arevisible by virtue of their hypointensity on TIweighted sequences in relation to the normalpituitary tissue (see fig 1B). Focal enlargementof the gland and convexity of the diaphragmasella are less specific adjuvant signs. Infundibu-lar tilt and sella floor thinning or erosion areless useful criteria.65 Microadenomas as smallas 2–3 mm can now be detected in 75%-90%of cases (with a false negative rate of about13%), which is very important for surgicalplanning in patients with pituitary drivenCushing’s disease66 when combined with bilat-eral inferior petrosal sinus sampling.58 67 SerialMRI studies have shown that about 7% ofmicroprolactinomas progressively expand over2–8 years.

The immediate T1 weighted scan aftergadolinium, with the use of thin slices, is themost useful sequence to increase the sensitivityof detection of small intrasellar lesions: theenhancing normal glandular tissue allows thelow intensity microadenoma to become moreconspicuous (fig 5).68 The ability to obtain adynamic series of images in seconds rather thanminutes has overcome some of the limitationsof conventional postcontrast imaging (fig 6).

Figure 5 Acromegaly. Adenoma secreting growth hormone. The injection of gadoliniumallows the clear distinction between the normal displaced enhancing gland (small arrowhead) and the tumour.




j.com/

J Neurol N


ownloaded from


Images obtained less than 2 minutes from thestart of injection have a detection rate of 90%compared with 60% at 5 minutes.69 70 However,some adenomas become less visible, as theyalso enhance to be isointense with the gland. Afew microadenomas show more rapid arrival ofcontrast material than the normal pituitary tis-sue suggesting perhaps the development of adirect arterial blood supply.71 Precontrast TIweighted images remain important to detectboth the few adenomas (5%) which are hyper-intense to the normal gland and haemorrhagiclesions.

Computed tomography and MRI are equiv-alent in detecting the full extent of a macroad-enoma (fig 7). The margins tend to belobulated in at least two thirds of the cases(23% in meningiomas). Sellar enlargement isseen in 94%-100% of pituitary macroadeno-mas. This is only helpful as a distinguishingfeature in a suprasellar mass when it is absent,because its absence favours neoplasms otherthan a macroadenoma72 but does not excludean aneurysm! The presence of hormone hyper-secretion and an adenoma on CT does notexclude the presence of an associated vascularabnormality or intrasellar aneurysm, particu-larly in patients with acromegaly. Visual symp-toms are seen when there is chiasmaticdisplacement greater than 8 mm above a linefrom the frontal base and posterior clinoid inthe sagittal plane and 13 mm above the uppersurface of both internal carotid arteries in thecoronal plane.73 Magnetic resonance imaging isvery helpful in defining the extent of cavernoussinus involvement, documenting the anatomi-cal eVects of medical treatment (bromocrip-

tine, somatostatin), and for surgical planning.The interpretation of early postoperative scanscan be diYcult and are best left for 3 monthsunless early radiotherapy is contemplated.74

Rathke cleft cysts and craniopharyngiomasprobably have a common origin from Rathke’spouch with occasional cases of ciliated crani-opharyngiomas showing features of both.Rathke cleft cysts are mostly small andintrasellar and lined by a single layeredcuboidal or columnar ciliated epithelium thatmay include goblet cells. Squamous metaplasiamay transform the lining or the lining maydegenerate. Occasionally larger cysts projectbeneath the optic chiasm. Such cysts are oftenasymptomatic but are increasingly recognisedwith the wider use of MRI. Small lesionsshould not be mistaken for microadenomasand the larger ones with suprasellar extensionshould be distinguished from craniopharyngi-omas as the surgical approach may be differentand postoperative radiotherapy is not neces-sary. Diagnostic features on MRI include asellar epicentre, smooth contour, the absenceof both calcification and enhancement, and,when present, an anteriorly displaced pituitarystalk. The MR signal intensities are typicallyreduced on T I and increased on T2 though inmore than half of the cases there is hyperinten-sity on T I and varying signal on T2.75–77

Craniopharyngiomas are benign squamousepithelial neoplasms that are predominantlysuprasellar but sometimes show an intrasellarextension. There are two histological patternsof craniopharyngioma, adamantinomatous andpapillary.78 The first is composed of lace-likestrands of stellate epithelial cells with a basalpallisade enclosing many cystic spaces. Degen-erative changes are associated with the deposi-tion of cholesterol crystals that confer an oilyappearance to the cyst fluid. Masses ofeosinophilic, necrotic, keratinised cells thataccumulate within the epithelium often be-come heavily calcified and may dominate thehistology of a small biopsy. The less commonpapillary craniopharyngiomas usually occur inadults and are formed by mature squamousepithelial cells encasing a fibrovascular core.79 80

This type does not calcify and only rarelyinvolves the sella. In keeping with the variedhistological constituents of craniopharyngi-omas, the MR appearances are quite diverse. Aprominent cystic component and varyingsignal intensities of the mixed cystic and solidcomponents are typical features, occurring innearly 80% of cases. Areas of homogeneous TI hyperintensities and enhancement of solid aswell as the thick wall of the cystic componentsare highly suggestive of the diagnosis (see fig7D).81 Calcification is common especially inchildhood tumours and is best demonstratedby CT. The pituitary gland can often be shownseparately from the tumour especially by MRI.Suprasellar epidermoids are much rarer tu-mours but often cannot be radiologicallydiVerentiated.

Suprasellar and intrasellar meningiomasarise from the diaphragma, tuberculum, andthe dorsum sella as well as the dura of the adja-cent cavernous sinuses (see fig 7E). They are

Figure 6 Microadenoma. Dynamic MRI. 3D F SPGR 3 mm thick slices. Temporalresolution of 20 seconds. Contrast injected after the first image. Note early enhancement ofnormal gland (first row). Adenoma and gland became almost indistinguishable at laterimages (bottom row)




j.com/

J Neurol N


ownloaded from


isointense with the grey matter on T1 weightedimages but invariably show homogeneousdense enhancement. Obtuse dural margins anddural tail enhancement, in 68% of lesionsinvolving the sella, are helpful in the preopera-tive diagnosis although they are not specific.72 82

Adjacent hyperostosis, best seen on CT, ispresent in more than one third of cases and is ahelpful sign.

Hypothalamic/optic gliomas in the paediat-ric age group and young adults are slowlygrowing, benign pilocytic astrocytomas. Theyare related to neurofibromatosis type I in atleast 30% but invariably if there are bilateral

optic nerve gliomas. In adults, hypothalamicgliomas tend to be more aggressive. The fullextent of the lesion and invasion of the optictracts, lateral geniculate bodies, and opticradiation can be readily assessed by MRI (seefig 4).

The table lists the many other lesions thatmay present in the sellar region which unfortu-nately do not have pathognomonic imagingcharacteristics (fig 8) even though their man-agement may be diVerent, reflecting, for exam-ple, the radiosensitivity of most germinomas83

and the disseminated nature of a metastasis.Finally, the development of specific radiolig-

Figure 7 (A) Intrasellar macroadenoma. Extension into right cavernous sinus. (B) Macroadenoma. Coronal T1weighted image. The suprasellar extension is compromising the left optic nerve (arrow head). Adenoma is extending into thesphenoid sinus. (C) Macroadenoma. Optic chiasm is displaced and stretched resulting in classic bitemporal hemianopia.Note the invasion of clivus and extension to sphenoid sinus. (D) Craniopharyngioma. Sagittal T1 weighted image. Largewell defined lobular mass. The cystic component shows high signal and the fluid level is arrowed. Suprasellar calcificationwas seen on CT. (E) Meningioma. Sagittal TI weighted image postgadolinium. Enhancing lesion with anterior extensionbased on planum sphenoidale. The normal size of the pituitary fossa makes macroadenoma very unlikely. (F) Metastasis(bronchial carcinoma) Sagittal TI weighted image postgadolinium. Sellar and suprasellar enhancing mass extending intothe hypothalamic region.




j.com/

J Neurol N


ownloaded from


ands suitable for positron emission tomogra-phy (PET) imaging may contribute to thefunctional diagnosis and localisation of pitui-tary lesions and their follow up.

ManagementManagement ranges from the conservativefinding of “incidentalomas” through medical,surgical, and radiotherapeutic strategies de-pending on the tumour type, size, and invasive-ness. The objectives of treatment includerestoration of a feeling of wellbeing, relief ofpressure symptoms, treatment of hormoneexcess symptoms, restoration of or substitutionfor specific system deficits, prevention oftumour regrowth, and reversal of increasedlong term mortality. Results between series canbe compared only if the anatomical, neu-ropathological, and endocrine characteristicsare specified. The definition of a “cure”depends on the specific lesion:x Non-functioning tumours—no remnant on

sequential MR scans; lack of enlargement ofa known remnant after surgery and radio-therapy suggests that control has beenachieved but longterm follow up is stillrequired.

x Acromegaly—the criteria have become morestringent and now include a glucose sup-pressed growth hormone less than 2 ng/mland normalisation of insulin-like growth fac-tor (IGF-1).

x Cushing’s disease—reversal of hypercorti-solaemia but care has to be taken with theinterpretation of the results.

x Prolactinoma—correction of hyperprolacti-naemia; maintained shrinkage of adenomaon sequential MR.The following sections show some of the dif-

ficulties.

Grading systemsHardy’s original and robust classification wasbased on skull radiography and CT and for allpractical purposes requires only minor modifi-cations to take account of information fromMRI despite eVorts to create more elaboratesystems84–86:

I Microadenoma less than 10 mm indiameter

II Macroadenoma without suprasellar exten-sion

III Macroadenoma with suprasellar extensionIV Macroadenoma with localised destruc-

tion of the floor of the sellaV Macroadenoma with diVuse invasionVI Giant adenomas extending more than 40

mm above the planum sphenoidale or withmultidirectional extensions.

Any grading system will be imperfect becauseno imaging modality can yet distinguish dis-placement of the cavernous sinus from invasionof the dura. Such invasion can occur even witha small, laterally placed microadenoma.

Figure 8 (A) Germinoma. Sagittal T1 weighted imagepostgadolinium. Enhancing sellar and infundibular mass.Second lesion is present in the pineal region. Adolescentfemale presented with diabetes inspidus and failure to gainheight. (B) Tuberculous meningitis. Axial T1 weightedimage postgadolinium. Indolent TB meningitis presentedwith deteriorating vision. Suprasellar tuberculoma withmultiple small abscesses spreading mainly through theperivascular spaces. (C) Chondrosarcoma. Coronal T1weighted image after gadolinium. Skull base tumourextending into the right cavernous sinus displacing thenormal pituitary.




j.com/

J Neurol N


ownloaded from


Neuropathology and biological behaviourThe original classification of adenomas as aci-dophil (GH ± prolactin), basophil (adrenocor-tictrophin (ACTH)) and chromophobe (prol-actin, non-secretory, á-subunits) was based ontinctorial staining properties largely deter-mined by the nature and density of thesecretory granules. The combined approachesof electron microscopy, immunocytochemistry,and molecular biology have served to elucidateand characterise the plurihormonal nature ofmany adenomas with expansion of their classi-fication to include at least 16 diVerent types.87

Molecular biological studies have shown thatpituitary adenomas are mostly monoclonal,initiated by genomic mutation of a singleadenohypophyseal cell and stimulated to pro-liferate by various hypothalamic, peripheralendocrine, and paracrine factors. Individualtumour cells often contain and may secretemore than one active hormone, in addition tofragments of various precursor molecules. Forexample, many growth hormone adenomasalso contain prolactin and thyrotrophin andmany secrete the non-functioning á-subunitthat is common to all glycoprotein hormones asconfirmed by studies of primary monolayercultures of human anterior pituitary cells. Suchfindings are important for advancing ourunderstanding of the role of cell specifictranscription factors such as PIT-1 in thedevelopment of pituitary adenomas.88 Indi-vidual adenomas may show either a monomor-phous or mixed cell population with variationin fine structure and immunophenotype. Cyto-plasmic granularity and positive immunostain-ing indicate hormone synthesis and storage butnot necessarily release and hence there may bepoor correlation with serum concentrations ofthe paramount hormones or their systemicendocrine eVects. Whereas macroadenomastend to be endocrinologically more inert, thereis emerging evidence of greater proliferativeactivity and other diVerences from smaller butsystemically more active adenomas.89 Sparselygranulated tumours may be less well diVerenti-ated and more aggressive than the denselygranulated adenoma as has been documentedin acromegaly. To complicate matters, the stemcell chromophobe adenoma synthesises bothgrowth hormone and prolactin but, unlike thecommon prolactinoma, does not respond tobromocriptine, and shows both oncocyticchange and more aggressive growth.90 An unu-sual growth hormone variant contains bothtypical adenomatous cells and also clusters oflarge, sometimes dysmorphic, neurons inter-preted as a gangliocytoma. The neuronscontain growth hormone releasing hormoneand support the hypothesis that hypothalamicreleasing hormones promote tumorigenesis inthe anterior pituitary.91 Cushing’s disease maybe produced by a tiny microadenoma butislands of hyperplasia may be seen—theconcept of endocrine cell hyperplasia preced-ing neoplasia is fully accepted elsewhere, forexample, in thyroid C cell hyperplasia. Thereare at least two other unusual ACTH adeno-mas that are clinically silent, usually large, andonly established by immunostaining to synthe-

sise ACTH. One is a basophilic granulatedtumour that has a marked tendency to undergospontaneous haemorrhage and is a cause ofpituitary apoplexy. The other is a chromo-phobe adenoma showing only weak ACTHimmunoreactivity but stronger reactivity forâ-endorphin, which is also a derivative of theACTH precursor pro-opiomelanocortin.92

Thyrotrophinomas are the rarest of all adeno-mas and only a few are hyperfunctioning.Immunostaining may reveal sparse thyrotrophsin growth hormone adenomas. Both thyro-trophinomas and prolactinomas may becomeheavily calcified.93

Some 20%-30% of all adenomas are endo-crinologically inactive and include the silentACTH adenomas, growth hormone tumours,gonadotrophinomas, null cell adenomas, andoncocytomas.94 Null cell tumours resemblechromophobe adenomas but show negligibleimmunoreactivity or ultrastructural evidenceof hormone production. Oncocytic cells con-tain abundant large mitochondria that impartcytoplasmic granularity and eosinophilia whichcould be mistaken for an acidophilic growthhormone adenoma.95 Such tumours maymimic prolactinomas clinically but prolactinae-mia is due to raised intrasellar pressure or stalkcompression.

Aggressive adenomas and carcinomas— mi-croscopical extrasellar extension of pituitaryadenomas is common but these tumours tendto displace brain, optic nerves/chiasm, and ves-sels rather than invade them. Adenomas thatinvade the sphenoid or cavernous sinuses canbe regarded as unduly aggressive and pituitarycarcinomas are defined by evidence of cranios-pinal, or systemic metastases, or both. As withother endocrine tumours, nuclear and cellularpleomorphism alone are not reliable markers ofinvasive potential, but frequent mitoses gener-ally constitute an adverse prognostic sign. Vari-ous cell proliferation markers, including prolif-erating cell nuclear antigen, Ki-67 or MIB 1staining for Ki67 antigen have been employedin attempts to predict the likelihood ofrecurrence. High proliferation indices tend tocorrelate with more aggressive growth butsampling error is a problem and presently, par-ticularly in tiny biopsies, the results are not suf-ficiently reliable to provide a clear prediction inan individual case.89 96 The very rare pituitarycarcinomas are most often prolactinomas orACTH secreting tumours emerging in Nel-son’s syndrome.97 Molecular biological studiesdisclose multistep mechanisms oftumorigenesis—for example, ras oncogene mu-tation and activation are uncommon in pri-mary pituitary tumours but have been identi-fied in metastases.98 Proliferation indices andp53 expression also tend to be higher in themetastases.97

Medical managementDopamine agonists are the first line of manage-ment in patients with hyperprolactinaemia—particularly those with a microprolactinoma.Cabergoline, a newer ergot based dopamineagonist, is taken twice weekly as opposed to dailywith bromocriptine and is associated with fewer




j.com/

J Neurol N


ownloaded from


side eVects. Quinagolide is a non-ergotdopamine agonist which has been shown to beeVective—particularly in bromocriptine resist-ant cases—but can be associated with neuropsy-chiatric side eVects when administered in highdosage. Medical therapy is also the mainstay ofmanagement for most macroprolactinomas, butwith careful coordination between endocrinolo-gists and pituitary surgeons. Clearly, pituitaryapoplexy into a macroprolactinoma is an indica-tion for urgent trans-sphenoidal decompression,as bromocriptine can hardly be expected toshrink a haemorrhagic infarct. On the otherhand, there is one case report of a third nervepalsy due to apoplexy, treated with bromocrip-tine to good eVect and chiasmal compressionsecondary to prolactinomas is certainly revers-ible with medical therapy.99 Brain MRI indicatesthat a significant number of patients have smallhaemorrhages into the adenoma which are usu-ally asymptomatic but may form the basis forsubsequent fibrosis. There is evidence that bro-mocriptine treatment for 6–10 weeks increasesthe degree of fibrosis, leading to a greater surgi-cal morbidity thereafter.100 101 Therefore, earlypituitary surgery should be considered inpatients with visual or cranial nerve problems ifthere is either no improvement or an increase intumour size despite maximal dopamine agonisttherapy, or in the few patients who are intolerantor non-compliant with drug therapy.102

Trans-sphenoidal microsurgery (see later) isthe primary therapeutic modality of choice inCushing’s disease and specific medical man-agement is very limited. It is useful to treatpatients preoperatively with metyrapone,which blocks 11â-hydroxylase, to reduce thesurgical morbidity resulting from the eVects ofhigh cortisol concentrations on tissue healing,blood pressure, and immunity.103 Metyrapone,ketoconazole (inhibits steroidogenesis), andmitotane (cytotoxic to adrenal cortical cells)cannot be used long term because of their sideeVects and centrally acting drugs (bromocrip-tine, cyproheptadine, sodium valproate) areseldom eVective. Longterm cure rates of up to80% are being reported after surgery,3 104

reflecting several improvements in manage-ment: firstly, inferior petrosal sinus samplinghas proved particularly useful in making a firmpreoperative diagnosis—particularly in difficultcases; secondly, recent studies indicate that anundetectable morning serum cortisol postop-eratively correlates most closely with the likeli-hood of long term cure. A non-suppressedpostoperative cortisol (>150 nmol/l) maytherefore be an indication for an early secondexploration with more radical pituitary surgery.Ectopic adenomas in the suprasellar part of thepituitary stalk105 or multiple areas of cortico-trophic hyperplasia may contribute to surgicalfailure. Failed pituitary surgery usually leads tolocal radiotherapy, particularly as cortico-trophic adenomas may be radiosensitive and itis essential to protect against the developmentof Nelson’s syndrome. Radiation treatmenttakes many years to exert its eVect, necessitat-ing the use of bilateral adrenalectomy toachieve biochemical remission of Cushing’sdisease in the interim. The recent advent of

laparoscopic techniques for adrenalectomymay reduce the morbidity associated with thisoperation.

Trans-sphenoidal hypophysectomy is alsothe treatment of choice in acromegaly.106 Thelikelihood of a surgical response seems propor-tional to tumour mass and extrasellar spread,with a 70% cure rate for adenomas associatedwith circulating growth hormone concentra-tions of 100 mU/l but only a 40% cure rate withconcentrations greater than l00 mU/l.3 107 Onceagain, radiotherapy is used postoperativelywhen surgery has failed, but can take 10–20years to achieve biochemical control. More-over, it is now recognised that the increasedgrowth hormone concentrations associatedwith active acromegaly are associated withconsiderable excess cardiorespiratory mortalityand morbidity and possibly a risk of colonicneoplasia.108 Accordingly, medical therapy isused to reduce mean growth hormone to safeconcentrations of less than 5 mU/l. A minority(10%) of acromegalic tumours, especiallythose that also secrete prolactin, respond tobromocriptine.109 For the remainder, octreotide(a somatostatin analogue),110 is administered bythrice daily injection, but newer highly effectivefortnightly or monthly depot preparations arenow available. Octreotide treatment can also beused preoperatively to shrink extrasellartumour extension, but it is not known whetherthis improves the surgical outcome. Octreotidemay also have a role as the primary longtermtreatment modality, despite the risk of gallstones and the expense, in an older patient withmild acromegaly, a circumscribed tumourwithout local compression, perhaps with addedfactors (for example, cardiac failure) whichconstitute a poor operative risk. Finally, octre-otide is also highly eVective in controllinghyperthyroidism due to rare pituitary thyro-trophin secreting adenomas which have notbeen controlled surgically.

Although many of the non-functioningadenomas are derived from the gonadotrophlineage, they are generally unresponsive tomedical therapies including dopamine ago-nists, growth hormone, and somatostatinanalogues. In this context, radiotherapy playsan important part in arresting growth of thepostoperative tumour remnant but this needscareful follow up with serial MRI as there areno circulating hormonal markers which can bemeasured to monitor residual tumour activity.

Surgical managementSurgical approaches for “sellar” lesions haveevolved since Sir Victor Horsley’s first crani-otomy in 1889 and range from the straightfor-ward transphenoidal operation to the complexmultidisciplinary anterior skull base andtransventricular/transcallosal approaches (fig9). Subspecialisation is required if the patient isto be oVered safe, appropriate treatment.Modern transphenoidal surgery, of which thereare many excellent accounts,2 3 111 is very safe,with a mortality in well trained hands of0.2%-1% depending on the case mix, and issuitable for the great majority of pituitarytumours. It is also inexpensive when compared




j.com/

J Neurol N


ownloaded from


with drug therapy. Transnasal, sublabial tran-septal, and transethmoidal approaches eachhave their advocates— it is useful to be able touse each route in selected cases particularly forreoperations. Transnasal endoscopy112 has notyet been widely adopted except for repair ofsome CSF leaks. Refinements such as intraop-erative hormone assays are not generally avail-able.

Contraindications to the transphenoidalapproach include dumb-bell tumours (tightdiaphragma prevents access to the suprasellarextension from below), inaccessible extrasellarextensions (subfrontal, subtemporal,ndretrosellar—where a combined or staged ap-proach is preferable—see later). and intrasellarvascular lesions including aneurysms and“kissing” carotids. With the use of an imageintensifier and the awareness that this ensuressafety only in the anteroposterior plane, it ispossible and reasonably safe to drill through anincompletely pneumatised sphenoid air sinus.Chronic sinusitis and polyps should be rigor-ously dealt with by an ear, nose, and throat col-league before transphenoidal surgery. Particu-lar care needs to be taken with reoperationsespecially if the first operation has been under-taken elsewhere and the midline anatomicallandmarks have been removed, circumstancesin which the information from preoperative CTwith bone windows is useful.

Patients should be counselled that there is asmall risk of infection of 1%-2% (meningitis;sinusitis; septal or pituitary abscess) and thatthere is no evidence that prophylactic antibiot-ics are of any value except perhaps in patientswith Cushing’s disease or recent sinusitis. Sep-tal perforation, numb teeth, and a saddle noseare potential late complications but much lessimportant than those of subfrontal retractionvia a craniotomy. Significant haemorrhage isunusual (<1%). Damage to the internal carotidartery or rupture of an unsuspected intrasellaraneurysm can be readily controlled with amuscle pack. When intraoperative arterialinjury is suspected, postoperative angiographyis essential to forestall any delayed, lifethreatening epistaxis.113 The usual cerebrovas-cular techniques may be required includingballoon occlusion, clipping or coiling of ananeurysm, and embolisation of a torn sphen-opalatine artery. Arterial injuries should be-come even rarer with routine preoperativeMRI. Some tumours continue to ooze and thepatient may awake with a fine drain issuingfrom one nostril, the other end of which iswithin the tumour bed, in addition to the usualself expanding nasal packs. Suprasellar exten-sions may be encouraged to descend with theinjection of Ringer’s lactate (10 ml-30 ml) orair into a lumbar CSF catheter. If a suprasellarextension does not descend, there is a risk ofhaemorrhagic infarction in the remnant neces-sitating urgent return to the operating theatre.If there is any doubt about the state of the rem-nant or about haemostasis, early CT is helpful.Postoperative CSF leaks are uncommon (1%-6%) when there is an aggressive policy for seal-ing the pituitary fossa and sphenoid sinus ifthere is the slightest hint of CSF intraopera-

tively. A sandwich technique of muscle or fatwith tissue (triple fibrinogen) glue, combinedwith lumbar CSF drainage for a few days,works well. Patients should be warned that theymay wake up with a lumbar catheter and mayhave a low pressure headache for a few days.Visual deterioration is rare after transphenoidalsurgery (<0.5%) and immediate CT should beperformed to exclude a remediable cause.Curettage of lateral extensions may provoke athird or sixth nerve palsy but such palsies areusually temporary.

Postoperative diabetes insipidus is overdiag-nosed and overtreated! It is not unknown forelderly patients to be discharged on deamino-D-arginine vasopressin only to return inhyponatraemic coma. Except in the case ofchildren after craniotomy for extensive crani-opharyngiomas, when the thirst mechanismand hypothalamus may be defective, it is saferto follow the fluid balance chart and reassurethe patient—nasal packs make mouth breath-ing inevitable and patients want to keep theirmouths moist. The diagnosis of diabetes insip-idus should be based not just on urinary output(>2500 ml/day) but also on an increasedplasma osmolarity (>300 mosmol/kg) anddilute urine (<300 mosmol/kg). Sleep depriva-tion should be avoided by introducingdeamino-D-arginine vasopressin first at night.

Delayed hyponatraemia may occur sponta-neously up to 2 weeks postoperatively in about2% of patients.114 115 The original concept wasthat this reflected release of antidiuretichormone from degenerating posterior pituitaryneurosecretory terminals (SIADH) or was theresult of overenthusiastic cortisol withdrawalparticularly in patients with Cushing’s diseasewhose receptors had become adjusted to a veryhigh cortisol concentration. However, antidiu-retic hormone concentrations may not beraised. Treatment is by mild fluid restrictionand salt replacement.

Patients should be warned that they mayneed second operations, long term follow upand long term hormone replacement as well asadjuvant therapy (for example, bromocriptine,somatostatin, radiotherapy). It is diYcult togive a precise risk of postoperative hypopitui-tarism as it depends on the preoperative statusand tumour type and morphology101 but anoverall figure of up 10% is sensible. There is thepotential for improvement in pituitary functionin a minority of patients after transphenoidaldecompression of non-functioning pituitaryadenomas (16%-57% depending on the seriesand pituitary sector investigated).116 117 It wouldbe unusual for there to be any deterioration inpituitary function after transphenoidal marsu-pialisation of an intrasellar cyst in a young per-son. By contrast, it would be intended thatreoperation for refractory Cushing’s or Nel-son’s disease would result in hypopituitarism.

In addition to trans-sphenoidal surgery,more complex tumours including craniophar-yngiomas require various approaches (fig 9)that may carry very significant risks. Thejudgement and experience of a multidiscipli-nary team is invaluable.2 118–121 There is a temp-tation with multilobular suprasellar extensions




j.com/

J Neurol N


ownloaded from


of pituitary adenomas, for example, to chaseevery lobule via pterional, subfrontal,translaminar terminalis and subtemporalroutes but the morbidity and mortality may beunacceptable to the patient. Great care has tobe taken when dissecting around the chiasm. Amore conservative transphenoidal decompres-sion plus a restrained craniotomy followed byradiotherapy may achieve a much better qualityof life for the patient even if the postoperativescans are less impressive. Lateral extensionsinvading the cavernous sinus and engulfing thecarotid artery are seldom amenable to any sur-gical route and are the province of theradiotherapist.

Current issues in radiotherapy treatmentof pituitary tumoursThere is a consensus that radiotherapy shouldnot be given routinely postoperatively and eachpatient must be considered individually. Withthe hypersecreting adenomas, there is a tumourmarker to guide adjuvant therapy. With the“non-secretors” in which there has been com-plete tumour removal at operation as judged onpostoperative imaging, detectable recurrenceoccurs in 16% but further therapy was onlyrequired in 6%.117 If a small residual tumour isseen on postoperative imaging, radiation maybe withheld until a change indicating evidentgrowth is seen or repeat surgery oVeredwithout radiotherapy.122 123 Postoperative imag-ing can be deceptive and not all residual tissueis a tumour remnant. Radiotherapy is useful forthe very occasional patient who is too frail forsurgery.

Radiotherapy has a very high therapeuticratio, defined by the balance of tumour controlrates which are very high, and normal tissuecomplication rates which are extremely low.For non-functioning adenomas, tumour con-trol is defined as lack of progression of macro-scopic disease. For hormone secreting adeno-mas, reduction of hormone production is alsorequired. After radiotherapy, rates of freedomfrom progression are extremely high, and

should be around 95% at 10 years,124–128 and88% at 20 years.124 Secreting tumours faremarginally worse in some series, 89% versus97% at 10 years and 69% versus 96% at 20years,124 though others record equal controlrates of over 90%.125 126 Survival rates are excel-lent, although there is a slightly increased rela-tive risk of death in patients with pituitarytumours requiring radiotherapy.124

Despite the excellent rates of tumour con-trol, control of hormone secretion is poorer,with up to 20%-40% of patients failing to gaincontrol of hypersecretion,126 128 129 even withprotracted follow up. These rates of hormonecontrol can be improved by a further 20% bythe addition of medical treatment.126 TheeVects of radiotherapy on tumour hormoneproduction may take several years to becomemaximal, so the timing of assessment of thisend point must be chosen carefully. Ashormone concentrations are reduced by abouta first order reaction, the initial hormone con-centration will in part determine the timerequired for normalisation, if control isachieved.129 130

Radiotherapy doseDoses of 45–50 Gy in 25–30 fractions are typi-cal, well tolerated, and highly eVective. The fewstudies which report results of a wider range ofradiotherapy doses show a dose response, withlower rates of control associated with lowerdoses.128 131 The use of lower doses for patientswith acromegaly has been investigated as a wayto reduce the risk of pituitary insuYciency.132 Adose of 20 Gy in eight fractions seems toachieve relatively good control rates,132 thoughthis has not been formally compared with con-ventional larger doses, and the long termoutcome is not certain.

It is known that individual diVerences existin normal tissue radiosensitivity in some tissuesand it is very likely that this applies to the CNSalso, but it is not yet possible to predictprospectively133 Avoidance of damage to anypatient thus requires the choice of dose which

Figure 9 (A and B) Preoperative and postoperative transcallosal excision of a third ventricular craniopharyngioma(coronal T1 weighted images). Note the small residual tumour in the floor of the third ventricle which was subsequentlytreated with radiotherapy.




j.com/

J Neurol N


ownloaded from


is safe for even the most sensitive patient. Thisvariation in individual sensitivity is probablythe major reason that higher doses have beengiven to some patients without inflicting dam-age.

Complications of radiotherapyComplications of radiotherapy can be dividedinto “acute”, occurring during or just after thecourse of radiotherapy, or “late”, occurring 6months up to many years after treatment.Acute side eVects are rarely serious and includea mild skin reaction, loss of temporal hair, andsecretory otitis media, due to inflammation ofthe anterior part of the eustachian tube. Thisnormally settles spontaneously but requires amyringotomy in 2%-3%.125 126 However, it isthe late complications which determine safedoses and define the therapeutic ratio.

LATE COMPLICATIONS

Visual impairmentIn a series of 411 patients treated at the RoyalMarsden Hospital with 45 50 Gy in 25–30fractions, 305 patients had a detectable visualdefect before treatment, and 55% improvedafter radiotherapy. Only two (0.5%) developedlate deterioration of vision presumed to be dueto radiotherapy, and both actually retainedvision.124 In a large series from Princess Marga-ret Hospital in Toronto, out of 160 patientswith non-functioning adenomas treated with amedian dose of 45 Gy, there were no cases ofvisual damage, and of the 145 cases withsecreting tumours treated with a median doseof 50 Gy one patient (who received 42.5 Gy in22 fractions) developed deterioration in thevision of the left eye only.125 126 Surgicalexploration of this patient showed densefibrotic tissue adherent to the left optic nerveand a postfixed chiasm, with herniation of theleft optic nerve into the empty sella. This wasconsidered to be due to the combined eVects ofsurgery and radiotherapy,126 and gives anincidence of visual impairment of 0.3%.

Brain necrosisRadiation necrosis has been described inpatients having pituitary irradiation, but mostcases occurred many decades ago, with whatwould now be considered poor techniques,often with low voltage machines (for example,250–300 kV, rather than 6 MV typically usednow), and with higher doses than currentlyprescribed.134 Provided the dose is 50 Gy orless, and the fraction size is 2 Gy or less, the riskof necrosis should be extremely small orabsent.134 135

Hypothalamic pituitary axis dysfunction isdependent on radiotherapy dose: and also on doseper fractionWith clinical doses, failure occurs in manypatients after pituitary radiotherapy, and is cer-tainly dose dependent.136 As noted above,smaller doses (20 Gy in eight fractions) havebeen used to treat acromegaly,132 on the basisthat the risk of dysfunction of the hypothalamicpituitary axis will be reduced and its onsetdelayed relative to “conventional” doses. Thiswarrants further investigation in a clinical trialsetting. It may be that a similar dose but givenwith a lower dose per fraction will furtherreduce the risk of hypothalamic pituitary axisinsuYciency. As well as total dose, the dose perfraction has an additional eVect,136 such thatlarger doses per fraction, especially doses over 2Gy per fraction, produce a disproportionateeVect on the axis. Even with dose fractionationschedules which relatively spare the axis, thereis a high incidence of failure, which rises withtime and probably has no plateau (fig 10), evenbeyond 15 years. Some 50%-70% of patientsrequire replacement of gonadotrophins, gluco-corticoids, and thyroxine.124 125 137 This highincidence is the product of initial tumour dam-age and surgery, as well as the eVects of radio-therapy which probably accounts for 10%-20%of patients who develop it.125 126 Associated withdysfunction of the hypothalamic pituitary axis

1.0B

0.5

10

Time (y)

TSH

ACTH

FSH/LH

Pro

bab

ility

of

no

hyp

op

itu

itar

ism

0 987654321

1.0A

0.5

10

Time (y)

TSH

ACTHFSH/LH

GH

Pro

bab

ility

of

no

rmal

hyp

oth

alam

ic p

itu

itar

y ax

is

0 987654321

Figure 10 Probability of hypothalamic-pituitary axis dysfunction after treatment forpituitary adenoma. A (redrawn from Littley et al137) shows data from 165 adults treatedwith radiotherapy doses of 37.5-42.5 Gy in 15-16 fractions with a median follow up of 4years, which is relatively short. B (redrawn from Tsang et al125) shows data for 160 patientswith a median follow up of 8.5 years. Radiotherapy doses were typically 45 Gy in 22-23fractions for minimal residual disease (in 95), 50 Gy in 25 fractions for gross residualdisease (in 41), and 40-42 Gy in 25 fractions for unspecified reasons. There are obviousdiVerences between the two series which may reflect the diVerent dose fractionation schedulesused.




j.com/

J Neurol N


ownloaded from


is hyperprolactinaemia, although this is usuallynot permanent or extreme.137 Also striking isthe complete loss of growth hormone produc-tion by 5 years. It is virtually unknown forpatients to develop diabetes insipidus as aresult of radiotherapy for pituitary tumour.138

Hypopituitarism in adults does carry a slightsurvival disadvantage, which is not fully elimi-nated by replacement with cortisol, thyroxine,and sex hormones, and it is possible thatgrowth hormone deficiency is a factor.139

Vascular damageDamage to the large vessels around thepituitary gland has not been noted with radio-therapy, but such damage at other sites isdescribed, although typically with higher dosesthan standard pituitary treatment.140 Occlusionof one or both carotid arteries might have cata-strophic consequences, and is one reasonagainst the use of stereotactic radiosurgery forthe initial treatment of large target volumes orfor retreatment for tumours close to these largevessels.

Cognitive dysfunctionImpairment of memory and executive functionhas been found in patients with pituitarytumours. One recent study suggested that nei-ther surgery nor radiotherapy alone aVectedcognitive function, but that combined treat-ment caused some dysfunction.127 However,other detailed psychometric studies suggestthat pituitary radiotherapy does not lead toimpairment of cognitive function.141 142

Risk of second tumourExposure to radiation, including therapeuticradiation, carries a small but significant risk ofdeveloping a second tumour, extending for atleast two decades after exposure. In one seriesof 334 patients treated with conservativesurgery and radiotherapy and followed up for amedian of 11 years, a total of five secondtumours occurred, although only three of thesewere malignant. This gives a cumulative risk ofsecond tumour of 1.3% in the first 10 yearsafter radiotherapy and 1.9% over 20 years.143

Summing the cases in the relevant literature, atotal of 1510 patients followed up for an aver-age of 10 years developed 13 new secondtumours, giving an approximate incidence of0.86%.143–147 The absolute risk may in fact beslightly lower than this, as developments inradiotherapy techniques and equipment havealmost certainly reduced the risk below theseconcentrations.148

Newer developments in radiotherapySTEREOTACTIC RADIOSURGERY

It is diYcult to see how this technique couldimprove the outcome for pituitary tumours inmost circumstances, especially as standardexternal beam radiotherapy gives such goodresults.152 Stereotactic radiosurgery is charac-terised by the use of a large dose ofradiotherapy given as a single fraction anddelivered with high accuracy. The use of a largesingle fraction may be eVective against tu-mours. Indeed, the technique was originally

developed as a method of giving a dosedesigned to cause necrosis of the target, and itstherapeutic ratio is provided by the rapid dropfrom high to low dose (called the penumbra)which occurs over a distance of just a few mil-limetres, and which allows avoidance of criticalstructures. However, large single fractions arethe antithesis of fractionated treatment and canbe extremely damaging to normal tissues ifthey are inadvertently treated. Thus thetechnique, even with its high precision, is farfrom ideal for treatment in which the target hasto include tumour extension onto or aroundthe critical normal tissues. Relatively fewpatients with pituitary disease have tumourswhich are truly confined to the pituitary fossawithout some encroachment laterally or supe-riorly. The optic nerves and chiasm, the nervesassociated with the cavernous sinus, and possi-bly the carotid arteries, are highly susceptibleto radiation damage, particularly when thedose per fraction is large, as in stereotacticradiosurgery. If a tumour abuts the opticchiasm, for example, it is possible to limit thedose at the edge of the tumour to safe levels,but this is at the expense of lowering the dose topart of the tumour. Thus, stereotactic radiosur-gery is unlikely to improve tumour control overconventional radiotherapy.

Many patients with pituitary adenomas haveapparently been successfully treated usingstereotactic radiosurgery.149 150 However, as yetthere is no evidence that it produces superiorresults. Proton and heavy ion beams which candeliver relatively focused radiotherapy havealso been employed, though likewise do notnecessarily confer any particular advantage.130

Despite some enthusiastic reports, catastrophicdamage has been described after stereotacticradiosurgery. Rocher et al151 treated 135patients with stereotactic radiosurgery, includ-ing 36 with pituitary adenomas. Over a shortperiod of follow up, the rate of pituitarytumour control was comparable with conven-tional external beam radiotherapy. The compli-cation rate, however, was excessive. Twelvepatients (33%) had serious visual complica-tions, including two who developed bilateralblindness. These complications were directlyattributable to the proximity of the tumour,and hence the high dose volume to the chiasm.This group has now abandoned the use ofstereotactic radiosurgery for pituitary tumours.

CONFORMAL RADIOTHERAPY

Conformal radiotherapy uses advanced radio-therapy treatment planning in three dimen-sions to achieve better conformation of treat-ment volume to target volume. This allowsreduction in the volume of normal tissueirradiated around the target. Conventionaldose fractionation is normally used withconformal radiotherapy. As conventional exter-nal beam radiotherapy is so safe and eVective,conformal radiotherapy would not be consid-ered necessary in most cases. However, it islikely that the indications will expand withtime. It is possible that the technique mightlower doses to the hypothalamus, althoughthis requires formal evaluation. Conformal




j.com/

J Neurol N


ownloaded from


radiotherapy is likely to have an advantage fortreating children, by reducing cognitive dys-function. Currently, its place may be intreating particularly extensive tumours, forretreatments, or for treating children.

1 Royal College of Physicians Working Party. Pituitarytumours: recommendations for service provision and guidelinesfor management of patients. London: Royal College of Physi-cians of London, 1997:1–39.

2 Powell M, Lightman SL. The management of pituitarytumours: a handbook. Edinburgh: Churchill Livingstone,1997:1–238.

3 Landolt AM, Vance ML, Reilly PL. Pituitary adenomas. NewYork: Churchill Livingstone, 1996:1–540.

4 Pituitary tumour network association. Web site www.pitui-tary.com

5 Reincke M, Allolio B, Saegar W, et al. The incidentaloma ofthe pituitary gland. JAMA 1990;263:2772–6.

6 Nishizawa S, Ohta S, Yokoyama T, et al. Therapeutic strat-egy for incidentally found pituitary tumours (pituitary inci-dentalomas). Neurosurgery 1998;43:1344–50.

7 Hall WA, Luciano MG, Doppman JL, et al. Pituitarymagnetic resonance signaling in normal human volunteers:occult adenomas in the general population. Ann Intern Med1994;120:817–20.

8 Teramoto A, Hirakawa K, Sanno N, et al. Incidentalpituitary lesions in 1000 unselected autopsy specimens.Radiology 1994;193:161–4.

9 Ezzatt S. Living with acromegaly. Endocrinol Metab ClinNorth Am 1992;21:753–60.

10 Koibuchi N, Kagegawa T, Suzuki M. Electrical stimulationof the basolateral amygdala elicits only growth hormonesecretion among six anterior pituitary hormones in thephenobarbital-anaesthetized male rat. J Neuroendocrinol1991;3:685–7

11 Pickett JBE, Layzer RB, Levin SR. Neuromuscular compli-cations of acromegaly. Neurology 1975;25:638.

12 Sussman JD, Sarkies N, Pickard JD. Benign intracranialhypertension. Adv Tech Stand Neurosurg 1998;24:261–305.

13 Chatterjee VKK, Beck-Peccoz P. Thyroid hormone resist-ance. Baillieres Clin Endocrinol Metab 1994;8:267–83.

14 Heyburn PJ, Gibby OM, Hourihan M. Primary hypo-thyroidism presenting as amenorrhoea and galactorrhoeawith hyperprolactinaemia and pituitary enlargement. BMJ1986;292:1660–3.

15 Smith MV, Laws ER. Magnetic resonance imaging meas-urements of pituitary stalk compression and deviation inpatients with non-prolactin secreting intrasellar andparasellar tumours: lack of correlation with serum prolactinlevels. Neurosurgery 1994;34:834–9.

16 Lees PD, Pickard JD. Hyperprolactinaemia, intrasellarpituitary tissue pressure and the pituitary stalk compres-sion syndrome. J Neurosurg 1987;67:192–6.

17 Tien RD. Sequence of enhancement of various portions ofthe pituitary gland on gadolinium: enhanced MR images.Correlation with regional blood supply. AJR Am J Roentge-nol 1992;158:651–4.

18 Bevan JS, Burke CW, Esiri MM, et al. Misinterpretation ofprolactin levels leading to management errors in patientswith sellar enlargement. Am J Med 1987;82:29–32.

19 Honnegger J, Fahlbusch R, Bornemann A. Lymphocyticand granulomatous hypophysitis: experience with ninecases. Neurosurgery 1997;40:713–22.

20 Thodou E, Asa S, Kontoeorgos G, et al. Clinical Care Semi-nar - lymphocytic hypophysitis: clinicopathological find-ings. J Clin Endocrinol Metab 1995;80:2302–11.

21 Cardoso ER, Peterson EW. Pituitary apoplexy: a review.Neurosurgery 1984;44:363–73.

22 Kruse A, Astrup J, Cold CF, et al. Pressure and blood flowin pituitary adenomas measured during transphenoidalsurgery. Br J Neurosurg 1992;6:333–42.

23 Lees PD, Lynch DT, Richards HK, et al. Blood in portalsystems with special reference to the rat pituitary gland. JCBF Metab 1992;12:128–38.

24 McFadzean RM, Doyle D, Rampling R, et al. Pituitary apo-plexy and its eVect on vision. Neurosurgery 1991;29:669–675.

25 Yousem DM, Arrington JA, Ziureich SJ, et al. Pituitaryadenomas: possible role of bromocriptine in intratumoralhaemorrhage. Radiology 1989;170:239–243.

26 Lundin Per, Bergstrom K, Nyman R, et al.Macroprolactinomas: serial MR imaging in long-term Bro-mocriptine therapy. AJNR Am J Neuroradiol 1992;13:1279–1.

27 Maccagnan P, Maceda CL, Kayath HJ, et al. Conservativemanagement of pituitary apoplexy: a prospective study. JClin Endocrinol Metab 1995;80:2190–7.

28 Zucchini S, Ambrosetto P, Carla G, et al. Primary emptysella: diVerences and similarities between children andadults. Acta Paediatr 1995;84:1382–5.

29 Akculin S, Ocal G, Berberoglu M, et al. Association ofempty sella and neuroendocrine disorders in childhood.Acta Paediatr 1995;37:347–51.

30 Buchfelder M, Brockmeier S, Pichl J, et al. Results ofdynamic endocrine testing of hypothalmic pituitary func-tion in patients with primary ‘empty’ sella syndrome.HormMetabRes 1989;21:573–576.

31 Weisberg LA, Housepian EM, Saur DP. Empty sellasyndrome as a complication of benign intracranial hyper-tension. J Neurosurg 1975;43:177–80.

32 Komatsu M, Kondo T, Yamauchi K, et al. Anti-pituitaryantibodies in patients with the primary empty sellasyndrome. J Clin Endocrinol Metab 1988;67:633–8.

33 Applebaum EL, Desai NM. Primary empty sella syndromewith CSF rhinorrhoea. JAMA 1980;244:1606–8.

34 Kaufman B, Tomsik RL, Kaufman BA. Herniation of thesuprasellar visual system and third ventricle into emptysellae: morphologic and clinical considerations. Am J Neu-roradiology 1989;10:65.

35 Cybulsky GR, Stone JL, Geremia G, et al. Intrasellar ballooninflation for treatment of symptomatic empty sellasyndrome. Neurosurgery 1989;24:105–9.

36 Moses AM, Clayton B, Hochhauser L. Use of T1 weightedMR imaging to diVerentiate between primary polydipsiaand central diabetes insipidus. AJNR Am J Neuroradiol1992;13:1273–7.

37 Terano T, Seya A, Tamura Y, et al. Characteristics of thepituitary gland in elderly subjects from magnetic resonanceimages: relationship to pituitary hormone secretion. ClinEndocrinol 1996;45:273–9.

38 El Gammel T, Brooks BS, HoVman WH. MR imaging ofthe ectopic bright signal of posterior pituitary regeneration.Am J Neuroradiol 1989;10:323.

39 Kelly WM, Jucharczyk LV, Kurcharcyzk J. Posteriorpituitary ectopia: an MR feature of pituitary dwarfism. AmJ Neuroradiol 1988;9:453–60.

40 Aaberg TM Jr, Kay M, Mracek Z, et al. Metatastic tumoursof the pituitary. Am J Ophthalmol 1995;119:779–85.

41 Hoyt WF. Correlative functional anatomy of the opicchiasm: 1969. Clin Neurosurg 1970;17:189–208.

42 Rucker CW. The concept of a semi-decussation of the opticnerves. Arch Ophthalmol 1978;59:159–71.

43 Bergland RM, Ray BS, Torack RM. Anatomical variationsin the pituitary gland and adjacent structure in 225 autopsycases. J Neurosurg 1968;28:93.

44 Wilbrand H, Saenger A. Neurologie des Auges. Vol 6.Wiesbaden: Verlag Von JF Bergmann, 1915.

45 Horton JC. Wilbrand’s knee of the primate optic chiasm isan artefact of monocular enucleation. Trans Am OphthalmolSoc 1997;95:579–609.

46 Kirkham TH. The occular symptomatology of pituitarytumours. Proceedings of the Royal Society of Medicine1972;65:517–8.

47 Miller NM. Walsh and Hoyt’s clinical neuro-ophthalmology.4th ed. Baltimore: Williams and Wilkins, 1982–1994;1:119–27.

48 Plant GT. Residual visual function in chiasmal compressionin man. J Physiol 1985;360:22P

49 Grochowski M, Vighetto A, Berquet S, et al. Contrast sensi-tivity function and pituitary adenoma: a study of 40 cases.Br J Ophthalmol 1990;74:358.

50 Gutowski NJ, Heron JR, Scase MO. Early impairment offoveal magno- and parvocellular pathways in juxta chiasmaltumours. Vision Res 1997;37:1401–8.

51 Trobe JD, Tao AH, Schuster JJ. Perichiasmal tumours: diag-nostic and prognostic features. Neurosurgery 1984;15:391.

52 Reese BE, Cowey A. The neurologic consequences of a sub-chiasmal tumour on the retino-geniculo-striate pathway ofa macaque monkey. Clin Vision Sci 1989;4:341–56.

53 Burke W, Cottee LJ, Garvey J, et al. Selective degenerationof optic nerve fibres in the cat produced by a pressureblock. J Physiol 1986;376:461–76.

54 Glaser JS. Topical diagnosis: the optic chiasm. In: Glaser JS,ed. Neuro-ophthalmology. Philadelphia: JB Lippincott,1990:171–212.

55 Powell M. Recovery of vision following transphenoidsurgery for pituitary adenomas. Br J Neurosurg 1995;9:367–73.

56 Tawana LK, Sedgwick EM, Pickard JD, et al. Early recoveryof vision and visual evoked potentials following surgicaldecompression of the optic chiasm. In: Papakostopoulos D,Butter S, Martin I, eds. Clinical and experimental neuropsy-chophysiology. London: Croom Helm, 1984:105–31.

57 Masson EA, Atkin SL, Diver M, et al. Pituitary apoplexy andsudden blindness following the administration of gonado-trophic releasing hormone. Clin Endocrinol 1993;38:109–10.

58 Oldfield EH, Doppman JL, Nieman LK. Petrosal sinussampling with and without corticotrophic-releasing hor-mone for the diVerential diagnosis of Cushing’s syndrome.N Engl J Med 1991;325:897–905.

59 Zimmerman RA. Imaging of intrasellar, suprasellar andparasellar tumours. Semin Roentgenol 1990;25:174–97.

60 Elster AD. Imaging of the sella: anatomy and pathology.Semin Ultrasound CT MR 1993;14:182–94.

61 Suzuki M, Takashima T, Kadoya M, et al. Height of normalpituitary gland on MR imaging: age and sex differentiation.J Comput Assist Tomogr 1990;14:36–9.

62 Elster AD, Sanders TG, Vines FS, et al. Size and shape ofthe pituitary gland during pregnancy and post partum:measurement with MR imaging. Radiology 1991;181:531–5.

63 Faglia G. Epidemiology and pathogenesis of pituitaryadenoma. Acta Endocrinol 1993;129(suppl):1–5.

64 Stadmik T, Sprunyt D, Van Binst A, et al. Pituitarymicroadenomas: diagnosis with dynamic serial CT, con-ventional CT, and T1-weighted MR imaging before andafter injection of gadolinium. Eur J Radiol 1994;18:191–8.

65 Wu W, Thomas KA. Pituitary microadenomas: MR appear-ance and correlation with CT. Acta Radiol 1995;36:529–35.

66 Colombo N, Loli P, Vignati F, et al. MR of corticotrophin-secreting pituitary microadenomas. Am J Neuroradiol 1994;15:1591–5.




j.com/

J Neurol N


ownloaded from


67 Lopez J, Barielo B, Lucas T, et al. Petrosal sinus sampling fordiagnosis of Cushing’s disease: evidence of false negativeresults. Clin Endocrinol 1996:45:147–56.

68 Newton DR, Dillon WP, Norma D, et al. Gd-DTPAenhanced MR imaging of pituitary adenomas. Am J Neuro-radiol 1989;10:949–54.

69 Hamon-Kerautret M, Leclerc X, Devailly D, et al. Pituitarymicroadenomas: experience with Gd-DPTA enhanced MRimaging at 0.5 Tesla. Eur J Radiol 1994;18:185–90.

70 Bartynski W, Lin L. Dynamic and conventional spin echoMR of pituitary microlesions. Am J Neuroradiol 1997;18:965–72.

71 Yuh WT, Fisher DJ, Nguyen HD, et al. Sequential MRenhancement pattern in normal pituitary gland and inpituitary adenoma. Am J Neuroradiol 1994;15:101–8.

72 Donovan JL, Nesbit GM. Dysfunction of masses involvingthe sella and suprasellar space: specificity of imagingfeatures. Review article. AJR Am J Roentgenol 1996;167:596–603.

73 Ikeda H, Yoshimoto T. Visual disturbances in patients withpituitary adenoma. Acta Neurol Scand 1995;92:157–60.

74 Kremer P, Forsling M, Hamer J, et al. MR imaging ofresidual tumour tissue after trans-sphenoidal surgery ofhormone inactive pituitary macroadenomas: a prospectivestudy. Acta Neurochir Wien 1996;(suppl 65):27–30.

75 Kleinschmidt De Masters BK, Lillehei KO, et al. Thepathologic, surgical and MR spectrum of Rathke’s cleftcysts. Surg Neurol 1995;44:19–26.

76 Sumida M, Mozumi T, Mukada K, et al. Rathke’s cleft cysts:correlation of enhanced MR and surgical findings. Am JNeuroradiol 1994;15:525–32.

77 Ross DA, Norman D, Wilson CB. Radiologic characteristicsand results of surgical management of Rathke’s cysts in 43patients. Neurosurgery 1992;30:173–8.

78 Burger P, Scheithauer B. Craniopharyngiomas. Tumours ofthe central nervous system. In: Armed Forces Institute ofPathology. Atlas of tumour pathology. Washington DC:Armed Forces Institute of Pathology, Fascicle, 1994;10:349–54.

79 Crotty TB, Scheithamer BW, Young WFJ, et al. Papillarycraniopharyngioma: a clinicopathological study of 48cases. J Neurosurg 1995;83:216–14.

80 Sartoretti-Schefer S, Wichmann W, Aguzzi A, et al. MR dif-ferentiation of adamantinous and squamous papillarycraniopharyngioma. Am J Neuroradiol 1997;18:77–87.

81 Hold KJ, Eldevik OP, Quint DJ, et al. Pre and post operativeMR imaging of craniopharyngiomas. Acta Radiol 1996;37:806–12.

82 Kinjo T, Al-Mefty O, Ciric I. Diaphragma sella meningi-oma. Neurosurgery 1995;36:1082–92.

83 Sumida M, Uozumi K, Kiyak, et al. MRI of intracranialgerm cell tumours. Neuroradiology 1995;37:32–7.

84 Hardy J. Transphenoidal microsurgery of the normal andpathological pituitary. Clin Neurosurg 1969;10:185–217.

85 Teasdale G. Surgical management of pituitary adenoma.Clin Endocrinol Metab 1983;12:789–823.

86 Nistor R, Fahlbusch R. Buchfelder M, et al. Magnetic reso-nance imaging of parasellar developed pituitary adenomas:new consequences for pituitary surgery. In: Samii M, ed.Surgery of the sellar region and paranasal sinuses. Berlin:Springer-Verlag, 1991:199–204.

87 Horvath E, Scheithauer B, Kovacs K, et al. Regionalneuropathology: hypothalamus and pituitary. In: GrahamD, Lantos P, eds. Greenfield’s neuropathology. 6th ed.London: Arnold, 1997;1:1007–94.

88 Aylwin SJB, King A, Blenke A, et al. Free á-subunit andintact TSH secretion in vitro are closely associated inhuman somatotroph adenomas. Eur J Endocrinol 1998;139:378–86.

89 Hsu DW, Hakim F, Biller BM, et al. Significance of prolifer-ating cell nuclear antigen index in predicting pituitaryadenoma recurrence. J Neurosurg 1993;78:753–61.

90 Burger P, Scheithauer B, Vogel F. Pituitary neoplasia. Surgi-cal pathology of the nervous system and its coverings. 3rd ed.New York: Churchill Livingstone, 1991:503–68.

91 Puchner M, Ludecke D, Saeger W, et al. Gangliocytomas ofthe sellar region: a review. Exp Clin Endocrinol Diabetes1995;103:129–49.

92 Nagaya T, Seo H, Kuwayama A, et al. Pro-opiomelanocortingene expression in silent corticotroph-cell adenoma andCushing’s disease. J Neurosurg 1990;72:262–7.

93 Webster J, Peters JR, John R, et al. Pituitary stone: two casesof densely calcified thyrotrophin-secreting pituitary adeno-mas. Clin Endocrinol (Oxf) 1994;40:137–43.

94 Ho DM, Hsu CY, Ting LT, et al. The clinicopathologicalcharacteristics of gonadotroph cell adenoma: a study of 118cases. Hum Pathol 1997;28:905–11.

95 Nishioka H, Ito H, Hirano A, et al. Immunocytochemicalstudy of pituitary oncocytic adenomas. Acta Neuropathol(Berl) 1997;94:42–7.

96 Buchfelder M, Fahlbusch R, Adams EF, et al. Proliferationparameters for pituitary adenomas. Acta Neurochir (Wien)1996;65(suppl):18–21.

97 Pericone P, Scheiithauer B, Sebo T, et al. Pituitarycarcinoma: a clinicopathologic study of 15 cases. Cancer1996;79:804–12.

98 Pei L, Melmed S, Scheithauer B, et al. H-ras mutations inhuman pituitary carcinoma metastases. J Clin EndocrinolMetab 1994;78:842–6.

99 Brisman MH, Katz G, Post KD. Symptoms of pituitaryapoplexy rapidly reversed with bromocriptine. J Neurosurg1996;85:1153–5.

100 Landolt AM, Osterwalder V. Perivascular fibrosis in pitui-tary adenomas: is it increased by bromocriptine? J ClinEndocrinol Metab 1984;58:1179–83.

101 Bevan JS, Adams CBT, Burke CW. Factors in the outcomeof transphenoidal surgery for prolactinoma and non-functioning tumour including preoperative bromocriptivetherapy. Clin Endocrinol 1987;26:541–56.

102 Thomson JA, Davies DL, McLaren EH, et al. Ten yearfollow-up of microprolactinoma treated by transphenoidalsurgery. BMJ 1994;309:1409–10.

103 Child DF, Burke CW, Burley DM, et al. Drug control ofCushing’s syndrome: combined aminoglutethamide andmetyrapore therapy. Acta Endocrinol 1979;82:330–41.

104 Luedecke DK, Chrousos GP, Tolis G, eds. ACTH,Cushing’s syndrome and other hypercortisolemic states. NewYork: Raven, 1990.

105 Mason RB, Nieman LK, Doppman JL, et al. Selectiveexcision of adenomas originating in or extending into thepituitary stalk with preservation of pituitary function. JNeurosurg. 1997;87:343–51.

106 Wass JAH, ed. Treating acromegaly: 100 years on. Bristol:Society for Endocrinology, 1994.

107 Freda PV, Wardlaw SL, Post KD. Long term endocrino-logical follow-up evaluation in 115 patients who underwenttransphenoidal surgery for acromegaly. J Neurosurg 1998;89:353–8.

108 Ezzat S, Melmed S. Are patients with acromegaly atincreased risk for neoplasia? J Clin Endocrinol Metab 1991;72:245–9.

109 Besser GM, Wass JAH, Thorner MO. Acromegaly: resultsof long-term treatment with bromocriptine. Acta Endocrinol1978;88(suppl 216):187–98.

110 Ezzat S, Synder PJ, Young WF. Octreotide treatment ofacromegaly: a randomised multicenter study. Ann InternMed 1992;117:711–8.

111 Hardy J. Atlas of transphenoidal microsurgery in pituitarytumours. New York: Ikagu-Shoin Medical, 1991.

112 Jho H-D, Carrau RL, Ico Y, et al. Endoscopic pituitarysurgery: an early experience. Surg Neurol 1997;47:213–23.

113 Raymond J, Hardy J, Czepto R, et al. Arterial injuries intransphenoidal surgery for pituitary adenoma, the role ofangiography and endovascular treatment. Am J Neuroradiol1997;18:655–65.

114 Whitaker SJ, Meanock CI, Turner GF, et al. Fluid balanceand secretion of antidiuretic hormone following transphe-noidal pituitary surgery. A preliminary series. J Neurosurg1985;63:404–12.

115 Taylor SL, Tyrrell JB, Wilson CB. Delayed onset ofhyponatraemia after transphenoidal surgery for pituitaryadenomas. Neurosurgery 1995;37:649–53.

116 Arafah BM, Broadkey JS, Manni A. Reversible hypopitui-tarism in patients with large non-functioning pituitaryadenomas. J Clin Endocrinol Metab 1986;62:1173–9.

117 Ebersold MJ, Quast LM, Laws ER Jnr, et al. Long-termresults in transphenoidal removal of non-functioning pitui-tary adenomas. J Neurosurg 1986;64:713–19.

118 Dolenc VV. Transcranial epidural approach to pituitarytumours extending beyond the sella. Neurosurgery 1997;41:542–50.

119 Takakura K, Teramoto A. Management of huge pituitaryadenomas. Acta Neurochir (Wien) 1996;(suppl 65):13–25.

120 Goel A, Nadkami T, Kobayashi S. Surgical management ofgiant pituitary tumours. In: Kobayashi S, Goel A, Hono K,eds. Neurosurgery of complex tumours and vascular lesions.New York: Churchill Livingstone, 1997:259–76.

121 Apuzzo MLJ, ed. Surgery of the third ventricle. 2nd ed.Baltimore: Williams and Wilkins, 1998.

122 Lillehei KO, Kirschman DL, Kleinschmidt-De MastersBK, et al. Reassessment of the role of radiation therapy inthe treatment of endocrine-inactive pituitary macroadeno-mas. Neurosurgery 1998;43:432–9.

123 Post KD. Comments on Lillehei et al122. Neurosurgery1998;43:438.

124 Brada M, Rajan B, Traish D, et al. The long term eYcacyof conservative surgery and radiotherapy in the control ofpituitary adenomas. Clin Endocrinol 1993;38:571–8.

125 Tsang RW, Brierley JD, Panzarella T, et al. Radiation therapyfor pituitary adenoma: treatment outcome and prognosticfactors. Int J Radiat Oncol Biol Phys 1994;30:557–65.

126 Tsang RW, Brierley JD, Panzarella T, et al. Role ofradiation therapy in clinical hormonally-active pituitaryadenomas. Radiother Oncol 1996;41:45–53.

127 McCollough WM, Marcus RB, Rhoton AL, et al.Long-term follow-up of radiotherapy for pituitaryadenoma: the absence of late recurrence after > 4500 cGy.Int J Radiat Oncol Biol Phys 1991;21:607–14.

128 Zierhut D, Flentje M, Adolph J, et al. External radio-therapy of pituitary adenomas. Int J Radiat Oncol Biol Phys1995;33:307–14.

129 Tsagarakis S, Grossman A, Plowman PN, et al. Megavoltagepituitary irradiation in the management of prolactinomas:longterm follow-up. Clin Endocrinol 1991;34:399–406.

130 Andrews DW. Pituitary adenomas. Curr Opin Oncol 1997;9:55–60.

131 Grigsby PW, Simpson JR, Emarni BN, et al. Prognosticfactors and results of surgery and postoperative irradiationin the management of pituitary adenomas. Int J RadiatOncol Biol Phys 1989;16:1411–7.

132 Littley MD, Shalet SM, Swindell R, et al. Low-dose pitui-tary irradiation for acromegaly. Clin Endocrinol 1990;32:261–70.

133 Bumet NG, Wurm R, Nyman J, et al. Normal tissueradiosensitivity: how important is it? Clin Oncol 1996;8:25–34.




j.com/

J Neurol N


ownloaded from


134 Sheline GE, Wara WM, Smith V. Therapeutic irradiationand brain injury. Int J Radiat Oncol Biol Phys 1980;6:1215–28.

135 Grattan-Smith PJ, Morris JG, Langlands AO. Delayedradiation necrosis of the central nervous system in patientsirradiated for pituitary tumours. J Neurol NeurosurgPsychiatry 1992;55:949–55.

136 Littley MD, Shalet SM, Beardwell CG, et al. Radiation-induced hypopituitarism is dose-dependent. Clin Endocri-nol 1989;31:363–73.

137 Littley MD, Shalet SM, Beardwell CG, et al. Hypopituitar-ism following external radiotherapy for pituitary tumoursin adults. Q J Med1989;70:145–60.

138 O’Halloran DJ, Shalet SM. Radiotherapy for pituitaryadenomas: an endocrinologist’s perspective. Clin Oncol1996;8:79–84.

139 Hadden DR. Growth hormone for adults. J R Soc Med1997;90:122–3.

140 Schultz-Hector S, Kallfass E, Sund M. Radiation sequelaein the large arteries. A review of clinical and experimentaldata. Strahlenther Onkol 1995;171:427–36.

141 Grattan-Smith PJ, Morris JG, Shores EA, et al. Neuro-psychological abnormalities in patients with pituitarytumours. Acta Neurol Scand 1992;86:626–31.

142 Peace KA, Orme SM, Thompson AR, et al. Cognitive dys-function in patients treated for pituitary tumours. J ClinExp Neuropsychol 1997;19:1–6.

143 Brada M, Ford D, Ashley S, et al. Risk of second braintumour after conservative surgery and radiotherapy forpituitary adenoma. BMJ 1992;304:1343–7.

144 Jones A. Radiation oncogenesis in relation to the treatmentof pituitary tumours. Clin Endocrinol 1991;35:379–97.

145 Tsang RW, Laperriere NJ, Simpson WJ, et al. Glioma aris-ing after radiation therapy for pituitary adenoma. A reportof 4 patients and estimation of risk. Cancer 1993;72:2227–33.

146 Bliss P, Kerr GR, Gregor A. Incidence of second braintumours after pituitary irradiation in Edinburgh 1962–90.Clin Oncol 1994;6:361–3.

147 Hughes NM, Llamas KJ, Yeland ME, et al. Pituitaryadenomas: long term results for long term radiotherapyalone and post operative radiotherapy. Int J Radiat OncolBiol Phys 1993;27:1035–43.

148 Feigen M. Should cancer survivors fear radiation inducedsarcomas? Sarcoma 1997;1:5–15.

149 Degerblad M, Rahn T, Bergstrand G, et al. Long termresults of stereotactic radiosurgery to the pituitary gland inCushings disease. Acta Endocrinol 1986:112;310–14.

150 Thoren M, Rahn T, Guo WY, et al. Stereotactic radiosur-gery with the cobalt-60 gamma unit in the treatment ofgrowth hormone-producing pituitary tumours. Neurosurg1991;29:663–8.

151 Rocher FP, Sentenac I, Berger C, et al. Stereotacticradiosurgery: the Lyon experience. Acta Neurochir (Wien)1995;(suppl 63):09–114.

152 Brada M, Cruickshank G. Radiosurgery for brain tumours.Triumph of marketing over evidence based medicine. BMJ1999;318:411−12.




j.com/

J Neurol N


ownloaded from


neurology of the pituitary gland - jnnp’s ambition is to...

Documents