Semenyutin V.B., Vaynshenker Yu.I., Melkishev V.F., Alexeev V.V., Kamalova G.M., Zhinzhina I.V.
Russian Polenov Neurosurgical Institute
Pituitary apoplexy syndrome (PAS) comprises certain clinical manifestations resulting from sudden enlargement of pituitary adenoma (PA) and adjacent hypophysis due to necrosis and/or hemorrhage into tumor tissue. Recently there appeared a tendency to consider PAS as hemorrhage/necrosis into PA with a variable clinical course (asymptomatic to apoplectiform) [11, 14]. Hemorrhage into PA is observed in 9.6 to 17% of cases, ischemic infarction (necrosis) and degenerative cyst development within PA have been described in 7 to 25% . Hemorrhage into PA is 5.4 times as frequent as into other intracranial neoplasms . PAS can be observed both before and after an operation, it can be spontaneous or provoked by some factors [5–7, 12–15, 20]. PAS is to be differentiated from intracranial aneurysm rupture, craniocerebral injury, meningoencephalitis, internal carotid (ICA) occlusion [6, 7, 12, 19, 20].
Chiasmal-sellar localization of PA close to major arteries (their compression or coating by tumor, altered hormonal levels, an effect on diencephalic structures and rostral areas of the brainstem with considerable supra- and retrosellar tumor growth are possible causes of impaired cerebral blood flow in patients with PA [2, 4, 5, 6]. Impaired CSF dynamics in marked suprasellar PA growth may result in intracranial hypertension (ICH). PAS seems to considerably increase this probability, for, in sharp tumor enlargement there is an additional mechanic effect on major vessels, cavernous blood outflow is disturbed, hypothalamic dysfunction increases, and, if PA capsule is ruptured, blood and tumor masses enter the basal cistern [5, 6, 12, 19]. Besides, biologically active substances resorbed from the hemorrhagic or ischemic focus affect the state of cerebral vessels and blood-brain barrier [17, 22, 26].
The studies of cerebral hemodynamic disturbances in PAS patients were of empiric character up to now [4, 5, 6, 19, 20], though PAS is often a fatal PA complication [5, 6, 13]. Discrepancies between neurologic symptoms and direction of tumor growth were often observed [6, 7, 19]. The possibility of vasospasm (VS) in PAS accompanied by subarachnoid hemorrhage (SAH) was shown [6, 12]. Carotid angiographic (CAG) findings of marked narrowing of supraclinoid ICA and anterior cerebral artery (ACA) were considered to be a specific sign of the syndrome . CAG is an invasive method and therefore may be counter-indicated in grave PA patients. Besides, CAG is often a provoking factor in PAS development [6, 19]. Transcranial Dopplerography (TCD) is a non-invasive, harmless and reliable method to record linear blood flow velocity (BFV) of major cerebral vessels [9, 10]. TCD in patients with large and giant PA revealed: increased BFV in C-1 segment of ICA, M-1 segment of MCA, and A-1 segment of ACA and the orbit venous outflow with abnormal anterograde blood flow through orbital veins [1, 16]. At the same time, PA in these studies was considered with no special attention to PAS and the data obtained gave no possibility to evaluate BFV. Thus, cerebral hemodynamic disturbance in patients with PAS-complicated PA has not been studied in detail yet.
The goal of the present study was to compare clinical symptoms of PAS and major hemodynamic cerebrovascular changes in patients with PAS-complicated PA in order to evaluate their effect on the results of treatment.
MATERIAL AND METHODS
Retrospective analysis of the results of examination and treatment of 136 PAS patients was carried out; it made 31.6% of all the PA patients (430 cases) treated at the Russian Polenov Neurosurgical Institute during a 10-year period of 1989-1999. The patients with PAS were classified into two groups. Group I comprised the patients with gross PA hemorrhage with eventual necrotic foci, and signs of hemorrhage in the walls and cystic fluid (85 cases – 19.8%). Group II comprised the patients with gross PA necrosis with eventual small hemorrhages and cysts without signs of hemorrhage in the cystic fluid and the walls (51 cases – 11.9%). A control group of 50 patients with PA and absence of gross PA hemorrhage/necrosis was to help comparative evaluation of the results. Hemorrhage/necrosis was diagnosed with the help of CT, MRT or intraoperatively and verified histologically.
TCD was used in 23 PAS patients and 8 control group patients treated at the Institute during the period of 1997 to 1999. To evaluate cerebral hemodynamic disturbances, BFV of MCA, ACA, ICA extracranial portions and siphons were dynamically measured bilaterally using Multi Dop X (Germany). PAS was diagnosed before PA removal in 17, intraoperatively in 2 (to remove PA, bilateral subfrontal approach with preservation of transitional veins was used), and postoperatively in 4 patients.
TCD was performed in 12 patients of Group I, 8 males and 4 females aged 16 to 53. PA size was classified according to B.A.Kadashev : microadenoma – 1, small PA – 2, moderate PA – 1, large PA – 2 and giant PA – 6 cases. Classification based on endocrine activity was as follows: endocrine-inactive PA – 8, growth hormone (GH)-secreting PA – 2, prolactinoma – 1, PA with mixed hormonal activity – 1. PA was removed using bilateral subfrontal approach in 10 patients; 1 patient was treated therapeutically; 1 patient, operated on previously, was drug-treated for PAS.
Eleven patients of Group II underwent TCD examination, 5 males and 6 females aged 17 to 64. PA size was small in 2 cases, moderate in 1, large in 2, and giant in 6 cases. Endocrine-inactive PA, GH-secreting PA and prolactinoma were watched in 8, 1 and 3 cases respectively. PA was removed in 10 patients, 1 patient was treated therapeutically.
Eight patients of the control group were selected, 3 males and 5 females aged 38 to 51. PA size was small in 2 cases, moderate in 1, large in 2, and giant in 3. Distribution according to endocrine activity was as follows: endocrine-inactive PA – 3, GH-secreting PA – 3, prolactinoma – 1 and PA with mixed hormonal activity – 1. PA was removed in 7 patients, 1 was treated therapeutically.
TCD was performed from the very first day of PAS development in 19 patients and in 1–2 months after PAS onset in 4 cases (Group I – 1 and Group II - 3 patients). PAS was of a typical (including apoplectiform) course in 12 patients (Group I – 7 cases, in 2 of them hemorrhage into PA was accompanied by verified SAH and Group II – 5 patients). Pituitary apoplexy was characterized by scantiness of symptoms in 11 patients (Group I – 4 cases and Group II – 7 cases). Evaluation of results was based on the analysis of fatal outcomes and restoration of visual functions.
TCD studies revealed two types of hemodynamic changes, i.e. VS and changes being an indirect indication of intracranial hypertension (ICH). Increase of the Lindegaard index (LI) exceeding 3 served the criterion for VS and increase of pulsatile index (PI) in all major intracranial arteries up to 1.0 and more – for ICH.
All the 19 cases subject to examination from the very beginning of PAS development demonstrated VS or ICH. Three types of cerebral hemodynamic disturbances were observed during the first two days of PAS development, i.e.:
– VS + ICH (42.1% of cases, 4 from Group I, and 4 from Group II). In the ICA siphon BFV and PI were 140.0± 23.6 cm/sec and 1.13 ± 0.10 respectively; these indices were equal to 120.0± 23.0 cm/sec and 1.09 ± 0.15 for the MCA and to 129.8 ± 32.5 cm/sec and 1.12 ± 0.18 for the ACA.
– ICH without VS (47.4% of cases, 4 from Group I and 5 from Group II). In the ICA siphon BFV and PI were 44.5 ± 10.1 cm/sec and 1.29 ± 0.21 respectively; these indices were equal to 56.5± 9.1 cm/sec and 1.15± 0.17 for the MCA and to 48.2 ± 17.8 cm/sec and 1.12 ± 0.13 for the ACA.
– VS without ICH (10.5% of cases, 1 from Group I and 1 from Group II). As for the first case BFV, PI and LI in the ICA siphon were 106 cm/sec, 0.65 and 3.1 respectively; in the second case these indices in the MCA were 82 cm/sec, 0.62 and 3.5 respectively.
BFV values in VS + ICH differed reliably from those in ICH (P<0.05), whereas no reliable difference was obtained for PI. As VS without ICH was observed in two cases only, statistical difference was not calculated. In VS + ICH patients VS in the ICA, MCA, ACA was watched in 3 cases; in the ICA, MCA – 1 case; in the MCA, ACA – 2 cases; in the MCA – 1 case and in the ACA – 1 case; VS was bilateral in a half of the cases.
TCD signs of ICH were seen from the very first hours of PAS development. Further they were either stable or diminished in parallel with BFV increase. There was a progradient increase of VS with the maximum on the 5th-7th day. VS was characterized by a long-lasting, sometimes wave-like course of 3 and more weeks. Dynamics of BFV and PI changes in the MCA, ACA and ICA siphon during 11 days after PAS development is shown in Fig. 1. BFV and PI values in patients with VS and absence of ICH signs patients are not shown in this figure due to a small number of cases, still their dynamics corresponded to that of BFV and PI in VS + ICH. VS and/or ICH were not diagnosed in 8 patients of the control group. BFV and PI in the ICA siphon were 51.5 ± 19.1 cm/sec and 0.68 ± 0.12 respectively, these indices were equal to 61.2± 8.2 cm/sec and 0.75 ± 0.07 for the MCA and to 62.6 ± 11.4 cm/sec and 0.76 ± 0.03 for the ACA. In 4 patients examined in a post-apoplexy period (1-2 months) BFV and PI in the ICA siphon were 34.0 ± 2.0 cm/sec and 0.83 ± 0.05 respectively; these indices were equal to 58.5 ± 6.6 cm/sec and 0.74 ± 0.11 for the MCA and to 57.5 ± 1.0 cm/sec and 0.79 ± 0.16 for the ACA.
There was no reliable difference of BFV and PI indices in these patients as compared to the control group (P>0.10). In 2 patients VS was diagnosed immediately after PAS development by CAG, MRT-AG just on their admission to a hospital. They were admitted to the Polenov Neurosurgical Institute 1 or 2 months after PAS development and TCD examination failed to demonstrate any signs of VS or ICH. As two patients were not examined during the first month of PAS development, they were likely to have had no VS and ICH; however, their regress is a more probable phenomenon.
According to TCD findings with taking into consideration dynamics of VS and indirect ICH signs development after PAS there were 4 periods of hemodynamic changes: (a) a super-acute period (the first 6-12 hours after the onset of clinical signs of PAS) with absence of VS signs and appearance of indirect manifestations of ICH; (b) an acute period (6-12 hours to 2 days), when the 3 types of BFV and PI changes were seen; (c) A period of complications (3 days to 3 weeks), when either VS or indirect ICH signs were found: (d) sequelae period (after 3 weeks), when VS or indirect ICH signs regressed.
As BFV and PI values in the control group patients and in the patients examined a month after PAS did not differ significantly, they were combined into one group, i.e. with no changes. Fig.2 demonstrates the results of treatment of PA patients in different types of cerebral hemodynamic changes.
Visual deterioration after the treatment was observed in 9 and lethal outcome in 3 patients with detected hemodynamic disturbances. Combination of VS and ICH was seen in 5 cases, ICH solely in 7 cases. In VS without ICH there were no unfavourable outcomes; but the number of cases is too small to make any conclusions. The best results were seen in cases without hemodynamic disturbances (patients of the control group and cases from Groups I and II operated on at the sequelae period): visual amelioration was achieved in 10 patients, no visual deterioration or lethal outcomes were seen.
The operation was made during the period of complications in 9 patients with VS and/or ICH and unfavourable surgical outcomes. It was a period when 2 ICH patients, previously treated therapeutically, died (1 from Group I and 1 from Group II). Three patients with CBF disorders had visual amelioration after the operation performed at the sequelae or acute periods. To understand clinical significance of TCD results, retrospective analysis of 136 cases with PAS was carried out. During the first two days a clinical picture of PAS was characterized by acute hypothalamic-pituitary dysfunction, direction of sudden tumor enlargement and possible rupture of its capsule. Then it could be aggravated by hypertension syndrome and ischemic and brainstem disturbances.
Diencephalic syndrome was observed in PAS patients before PA began to affect visual pathways; it was especially typical of gradual onset of the disease. Other characteristic symptoms of PAS were sharp visual deterioration to the minimum particularly in cases without an effect of PA on visual pathways, papilledema (14.8% in Group I, and 6.7% in Group II), brainstem and hemispheric symptoms with no tumor effect on appropriate cerebral structures; diabetes insipidus; decompensation of insulin-resistant hyperglycaemia.
Before the operation VS (diffuse and segmental) differentiated by CAG was diagnosed only in PAS patients (22.2% for Group I, 11.1% for Group II). CT and MRI findings in PAS patients were indicative of ventricular enlargement irrespective of a degree of suprasellar growth of PA (41.9% for Group I and 35.5% for Group II); as for the control group it was typical only of cases with compression of the III ventricle by tumor (25.0%) (P<0.05).
The nearest postoperative period in the PAS patients was more severe compared to the control group; there were persistent headache and dyscirculatory disturbances. Relapses of PAS were observed on the 2nd -8th day (more often on the 4th-8th day). The results of treatment of patients with PAS were worse than those of other PA patients. Total mortality in patients with PA was found to be 5.8%: Group I – 3.7%, Group II – 1.2%, patients without PAS – 0.9%. Visual function deteriorated postoperatively in 35% of PAS patients. Unfavourable results were seen in the patients operated on at the periods of 6-12 hours to 3 weeks after apoplectiform PAS accompanied by consciousness disturbances and in 3 days to 3 weeks after PAS with a less grave course. These periods are peculiar for high frequency of recurrence, deterioration of aggravation of blood flow disorders at a diencephalic area and in the basin of major cerebral arteries. It caused neurologic deterioration and high mortality.
Retrospective analysis of 136 PAS cases revealed certain clinical peculiarities. Discrepancy between neuroophthalmologic and other focal neurologic signs and direction of PA growth is probably due to VS, for the latter was revealed by preoperative CAG in PAS patients only. Papilledema is a sign of ICH. E.Zh.Tron  had demonstrated papilledema to be an extremely rare sign in PA patients (1.2%) without taking PAS into account. We found it to be a more frequent sign, but in PAS patients only. Ventricular enlargement may serve another sign of ICH, independent of the degree of suprasellar tumor growth in PAS patients.
CBF studies using TCD in PA patients revealed both cerebral arterial VS and ICH to be etiologically connected with PAS development.
Unfavourable results of treatment depended upon cerebral hemodynamic disturbances at the time of PA removal. The 4 periods determined on the basis of BFV and PI changes (super-acute, acute, complications and sequelae periods) correspond to clinical dynamics of PAS. The results of the operation performed at the complication period were reliably worse than those obtained at earlier or later periods, it was due to marked VS or persistent indirect signs of ICH observed by that time. Poor results of the operation performed at an acute period of apoplectiform PAS were probably due to earlier development of major cerebral hemodynamic disturbances.
Taking into consideration gradual development of VS at an acute period and associated PI decrease which was not accompanied by clinical ICH decrease, one may consider VS absence after PAS to be a sign of major disorder of CBF autoregulation which is to be studied. Transcranial operation is a considerable damaging factor and may result in postoperative VS development and aggravation of hypothalamo-pituitary disturbances [5, 10, 21]. The possibility of unfavourable results after the operation performed against a background of VS or vascular atonia is extremely high.
Impaired venous outflow into cavernous sinus developing acutely in sudden, primarily intrasellar, tumor enlargement seems to be the most probable cause of ICH development in PAS. Due to this fact ICH may develop independently of the tumor size. Some other mechanisms may contribute to ICH genesis: SAH, altered water-electrolyte balance due to an effect of tumor on neurohypophysis, CBF dynamic disorders, especially in patients with evident suprasellar growth of PA.
There exist many theories explaining VS origin, but all of them consider cerebrovascular pathology [17, 22, 26]. Still, one cannot exclude an effect of vasoactive substances released with tumor necrosis , as well as hormonal vasculopathy [2, 4]. The latter is responsible for the fact that intracranial aneurysms are frequently (up to 7.4%) combined with PA . The relationship between a degree to which VS was marked and severity of hypothalamic involvement was found in the patients suffering from cerebrovascular pathology. It demonstrated a possible impact of hypothalamic lesions on VS development. According to the suggested theory pathologic impulses from the posterior hypothalamus ascend to cervical sympathetic ganglia that mediate development of VS of cerebral arteries . Diencephalic syndrome was observed before PAS in our series and increased considerably during an acute period. Thus, neurohumoral mechanism of VS in PAS seems to be very convincing, for it explains VS development in the absence of SAH and presence of hypothalamo-pituitary dysfunction.
It is probable that apoplexy is immediately followed by VS of tumor vessels; hypothalamic-pituitary dysfunction becomes more marked and VS acquires a diffused character especially in case of hormonal vasculopathy. At the same time, one cannot say for sure that VS develops primarily after PAS for circulatory changes demonstrated by TCD concern only major cerebral arteries, whereas PAS is preceded by hypothalamic dysfunction, as it has been shown above. There exists an opinion that hypothalamic dysfunction stimulates PA growth . It is quite probable that VS develops primarily within the inferior hypophyseal artery resulting in PAS, and then it involves major arteries of the brain base.