Intracranial Aneurysms and Subarachnoid Hemorrhage Report on a Randomized Treatment Study

Transcription

1 Intracranial Aneurysms and Subarachnoid Hemorrhage Report on a Randomized Treatment Study IV-A. Regulated Bed Rest DONALD W. NIBBELINK, M.D., JAMES C. TORNER, M.S., AND WILLIAM G. HENDERSON, PH.D. SUMMARY Three weeks of regulated bed rest was one of four treatments evaluated in the Cooperative Aneurysm Study. A total of patients with a recently ruptured intracranial aneurysm had subarachnoid hemorrhage confirmed by lumbar puncture. A group of patients were assigned to treatment within days after the bleed, between and days, and between days and days. During the mean follow-up interval of. years, mortality was.%. A proved rebleed was the cause of death in.%, progressive deterioration from aneurysm rupture in.%, and a suspected rebleed in.%. A total of.% died of causes related directly to the cerebral effects of the ruptured aneurysm. Downloaded from by on November, McKISSOCK AND COLLABORATORS ' s were the first investigators to stress the importance of controlled clinical trials for evaluation of certain treatments in patients with a ruptured intracranial aneurysm. In the present study regulated bed rest was recommended as one of four selected treatments to be evaluated in patients with a recently ruptured intracranial aneurysm. A specific format of therapy was followed during the -day period of bed rest. The remaining three treatment categories included drug-induced hypotension with bed rest, ipsilateral common carotid ligation and bed rest, and intracranial surgery. The results from patients allocated to each treatment category are discussed in separate sections in a series of reports. The introductory paper presents a general outline of the rationale and purpose of the study. The second report* includes the list of participants and the design and objectives of this aneurysm study. The third section' deals with the results of those patients who were allocated to intracranial surgery. This section includes the medical evaluation of patients allocated to regulated bed rest. The next publication will include the statistical analysis (Section IV-B). Clinical Material Between June, and February,, data forms from patients who received bed rest treatment were submitted to the Central Registry for analysis. All but eight patients were allowed to complete their designated course of treatment. Among these eight patients two experienced repetitive seizure activity, one was discharged nine days after randomization, two were restless and uncooperative, and three failed to be admitted to the reporting hospital within five weeks after the last bleed. The distribution of the major aneurysm locations is shown in table. Further details on specific aneurysm locations appear in table. As of June,, the allocation of patients with a vertebrobasilar aneurysm was suspended because an in- From the Department of Neurology, University Hospitals, Iowa City, Iowa. Reprint requests: Aneurysm Registry, Department of Neurology, University of Iowa, Iowa City, Iowa S. Dr. Nibbclink's present address is Director, Clinical Research, Merck Sharp & Dohme, West Point, Pa.. sufficient number of patient data forms were submitted to the Central Registry for statistical purposes. The seven patients allocated before the suspension date were in good neurological condition (grade ; see table for definition). One was in poor medical condition with severe pneumonia, but the remaining six were in good or fair condition (defined in section on "Age and Medical Condition"). Five were women, and two were men. Three died following a proved rebleed at day, days, and days, respectively, after the day of allocation to treatment. These seven patients with a vertebrobasilar aneurysm will be eliminated from further analyses. The remaining patients are included for detailed evaluation. The distribution of the interval from the last bleed to randomization is shown in table. Neurological Condition and Interval Between Last Bleed and Allocation to Treatment The distribution of patients in each neurological condition was analyzed with respect to various intervals from last bleed to day of randomization. For patients randomized within days after the hemorrhage,.% ( of ) were in good condition (grades and, table ). For those patients randomized during the through -day interval,.% were in good condition, and in the group allocated to treatment during the through -day interval,.% were in such condition. Only patients were included in the latter interval. The average distribution of patients in good neurological condition for the entire group was.% ( of ). Definition of Regulated Bed Rest The following criteria were included in the regulated bed rest treatment program: () hospitalization for three weeks; () subdued lighting for three weeks; () head of bed elevated no more than ; () patient allowed to turn but not to sit up or feed himself; () proper regulation of bowel movements with commonly used stool softeners; () after the first week, the patient could be helped to a bedside commode; () maintenance of adequate airway and oxygenation as necessary; () prophylaxis of seizures by administration of diphenylhydantoin, mg tid, for at least four weeks; () maintenance of, to, ml of fluids per hours by

2 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. Downloaded from by on November, TABLE Distribution of Aneurysm Site Site IC MC AC VB No IC = internal carotid; MC = middle cerebral; AC = anterior cerebral; VB = vertebrobasilar. mouth, nasogastric tube, or intravenously; () maintenance of, calories of nutrition per day, either by oral diet or tube feeding formula; () if sedation was required for restlessness, paraldehyde, phenobarbital, or chloral hydrate was administered in usual dosage; () condom drainage for men and catheter drainage for women was ordered for patients with incontinence; () proper analgesia was maintained by administration of acetylsalicylic acid, codeine, or meperidine hydrochloride; () every effort was made by each contributing center to provide a sustained high standard of nursing care encompassing frequent checking of vital signs, level of responsiveness, neurological assessment of verbal requests, pupil size, reactivity to light, ability to move extremities, and relative strength of grip; and () no agents such as osmotic diuretics, corticosteroids, or similar medications were permitted. TABLE Distribution of Patients with Specific Intracranial Aneurysm Location Major distribution Specific site Right Center Left INTERNAL CAROTID Subclinoid Supraclinoid, ophthalmic region Supraclinoid, posterior communicating junction Supraclinoid, at bifurcation Posterior communicating (distinct from internal carotid junction) Subtotal MIDDLE CEREBRAL Proximal to first main branching At main branching Distal to main branching _ Subtotal ANTERIOR CEREBRAL - ANTERIOR COMMUNICATING Proximal to anterior communicating At anterior communicating junction Distal to anterior communicating At anterior communicating Subtotal VERTEBRAL - BASILAR Basilar artery termination Posterior inferior cerebellar artery Basillar, trunk Vertebral Subtotal ~ _ TABLE Definition of Neurological Condition Grade Definition Symptom-free Minor symptoms (headache, meningeal irritation, diplopia) Major neurological deficit but fully responsive Impaired state of alertness but capable of protective or other adaptive responses to noxious stimuli Poorly responsive but with stable vital signs No response to address or shaking, nonadaptive response to noxious stimuli, and progressive instability of vital signs Results Mortality was.% among the patients treated with regulated bed rest during a mean follow-up interval of. years. The highest mortality (.%) occurred in patients allocated to treatment within days after the last bleed (table ). Mortality among patients who were allocated to treatment during the to -day interval and to -day interval was considerably less (.% and.%, respectively). Mortality among patients with an internal carotid aneurysm was.%; for those with a middle cerebral aneurysm it was.%, and in the group with an aneurysm on the anterior cerebral complex it was.% (table ). These figures are not significantly different. Among the group of patients who were hospitalized and who had their diagnostic workup completed within one week after onset of symptoms, mortality was appreciably higher. Several clinical parameters were analyzed with respect to potential influence on mortality: age, sex, aneurysm site, neurological and medical conditions, systolic and mean blood pressures, interval between last bleed and randomization, cerebral vasospasm, size of aneurysm, cause of death, and associated medical disorders. In the discussion to follow, each parameter is discussed with respect to mortality and rebleed rate. Age, sex distribution, and aneurysm site are displayed in tables and. Age and Neurological Condition The severity of each neurological condition was correlated with respect to age. The age group to had the highest number of patients in each neurological condition (table ). As expected, mortality was lowest in grade patients (.%) and highest in grade (%). With each increase in severity of neurological deficit, a proportionate rise in mortality was noted (table ). For the group with relatively TABLE Distribution of Interval from Last Bleed to Randomization Days* No. %.... From last bleed to randomization.

3 STROKE VOL, No, MARCH-APRIL Downloaded from by on November, TABLE Distribution of Patients in Various Neurological Conditions for Each Interval from Last Bleed to Randomization Days* l Neurological conditiont From last bleed to randomization. tsee table for definition. a minor symptoms (grades and ), died (.% of ), while of those in poor condition (grades,, and S), died (.% of ). This difference was highly significant (P <.). For patients years of age or over who were in poor neurological condition, the chance of survival was less than %. For all aneurysm sites combined, a proved rebleed was the cause of death in.% ( of ) of all patients, or.% of the deaths. Progressive decline in clinical condition following the direct cerebral effects of the original bleed was reported in.% ( of ) or.% of deaths. Death following a suspected rebleed was reported in nine patients. Therefore,.% ( of ) of deaths were directly related to the cerebral effects of the ruptured aneurysm. Age and Medical Condition Some medical conditions were present prior to the onset of symptoms of subarachnoid hemorrhage, while other conditions appeared during the initial phases of the illness. As the specific medical condition was recognized, it was regarded as an associated condition necessary for consideration as the course of treatment was followed. At the time of treatment allocation, each patient was appraised as to the degree of risk for a major surgical procedure, using the rating scale of good, fair, poor, and prohibitive. This classification was unrelated to the aneurysm per se. A good condition patient was alert and normotensive. He had no more than a degree of fever and was considered to be in good health prior to his subarachnoid hemorrhage. In general, patients in fair condition had a single underlying medical problem such as hypertension, cirrhosis of the liver, mild to moderate anemia, poor nutrition, generalized atherosclerosis, or chronic pulmonary disease. A patient with more than one adverse medical factor or a single condition which was moderately severe (e.g., recent myocardial infarction, electrolyte abnormalities, malignant hypertension, advanced atherosclerotic disease) was regarded as being in poor condition. Patients in the TABLE Age and Sex Distribution with Number of Deaths Among All Sites Combined Age Men Women % Deaths - - ()* () () () - () () - () () - () () - - () () () () () () Unknown KD () % Deaths.. *Number in parenthesis denotes deaths. () () () () () () () () () () prohibitive category had a complicated and serious illness with one or more medical disorders. Among patients, (.%) were in good condition, (.%) in fair condition, and (.%) in poor condition; in patients (.%) the medical condition was categorized as prohibitive (table ). With an overall mortality of.%, good condition patients alone had a mortality of.%, whereas in fair condition patients it was.%. For patients in poor condition, mortality was.%. All patients with a prohibitive medical status died. Clearly, medical disorders associated with subarachnoid hemorrhage had considerable influence on mortality. This pattern was consistent for all age groups. The number of deaths below age as compared with those age or above revealed the death rate to be significantly higher in the older group. Age was not a significant factor after analyzing the number of deaths above ( of or.%) or below ( of or.%) age in good medical condition patients alone. However, for patients in the fair, poor, and prohibitive medical conditions combined,.% ( of ) died in the group below age, whereas.% ( of ) died in the group age or above (P =.). Patients in good and fair medical conditions had a higher proportion of deaths after a rebleed (.%, or of, table ) as compared with patients in poor and prohibitive conditions (.%, or of ). Comparison of the distribution of deaths which followed a proved rebleed among various medical conditions shows no significant difference among all patients in good, fair, or poor medical condition. Cause of Death The major cause of death for patients with an aneurysm at various sites was a proved rebleed. Table shows the distribution of the cause of death for each aneurysm site. TABLE Mortality Distribution for Each Aneurysm Site and Interval to Randomization Days* IC Site MC AC % Deaths - ()t () () (). - () () () (). - () () () ().. () () () () % Deaths.... IC = internal carotid; MC = middle cerebral; AC anterior cerebral. From last bleed to randomization. fnumber in parenthesis denotes deaths. TABLE Sex Distribution by Site of Aneurysm Aneurysm site Sex IC MC AC Males ()* () () ().%.%.%.% Females () () () ().%.%.%.% () () () () % Deaths.%.%.%.% IC = internal carotid; MC = middle cerebral; AC anterior cerebral. 'Number in parentheses denotes deaths.

4 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. TABLE Deaths by Age and Severity of Neurological Condition with All Aneurysm Sites Combined Age Unknown % deaths l K)t () () () () ().% Defined in table. fnumber in parentheses denotes deaths. () () () () () () () ().% Neurological condition* () () () () () () ().% () () () () () () ().% () () () () KD () % () () () () () () () () () ().% Downloaded from by on November, Evaluation of the interval to death following allocation to formal bed rest therapy revealed that the major proportion of patients died during the three-week period of bed rest. During the first days after randomization, died (.% of ) following a proved rebleed; died (.% of ) within days (table ). From day and following, there were more deaths which followed a recurrent hemorrhage (.% of ). Analysis of the remaining causes of death reveals that among the patients who died following a suspected rebleed, succumbed within days after randomization and the remaining after the -day interval. Among patients who died following progressive deterioration in clinical condition, deaths were reported within days, and the remaining died after the -day interval. Two of these developed hydrocephalus (one required a ventriculopleural shunt), and remained poorly responsive following the last prerandomized bleed. Eight patients died from unrelated causes. Five of these occurred after the one year interval following randomization. Mortality and Blood Pressure on Day of Randomization The level of blood pressure as measured within days after the last bleed was another variable which was related to differences in mortality. The average value of four systolic blood pressures taken at intervals of four hours or more were calculated at the Central Registry. For this analysis patients were subdivided into two groups. One group had average TABLE Distribution by Age and Medical Condition with All Aneurysm Sites Combined Medical condition* Age Good Fair Poor Prohibitive io Unknown (l)t () () () () () () () () () () () () () KD () () KD () () () () () KD () KD () % deaths.%.%.% % *See text for definition. fnumber in parentheses denotes deaths. () () () () () () () () () ().% systolic blood pressures of mm Hg or below, and the other had average values above that level. Table shows that the mortality was.% in patients with average systolic values of mm Hg or below, while in the group with levels above mm Hg, it was.%. This difference was highly significant (P <.). Rebleed Rate, Neurological Condition, and Blood Pressure The highest percentage of patients who rebled were those allocated to treatment within days and whose average systolic blood pressures were over mm Hg (table ). However, these percentages were not different statistically from the group whose blood pressure was mm Hg or below. By comparison with the mortality table (table ), a proved rebleed as a cause of death remained the nearest association for patients who were in good neurological condition no matter what the interval following randomization. If systolic blood pressure alone had more direct influence on rebleeding, one would have expected a more direct relationship between mortality and rebleeding. Recurrent Hemorrhage Following Randomization The frequency distribution of recurrent hemorrhages as associated with mortality was presented in the previous discussion. Whenever deterioration in mental status and/or neurological condition was noted, a lumbar puncture was performed as conditions permitted. When significant increase in red cell count was noted in the cerebrospinal fluid or recent gross hemorrhage was observed at postmortem ex- TABLE Distribution of Cause of Death by Severity of Medical Condition Medical condition* Cause of death Good Fair Poor Prohibitive Proved rebleed Progressive decline Suspected rebleed Unrelated cause Unknown _ " " ~ patients each condition See text for definition.

5 STROKE VOL, No, MARCH-APRIL TABLE Distribution of Cause of Death for Each Aneurysm Site Aneurysm Bite Internal carotid ()* Middle cerebral () Anterior cerebral () () Proved rebleed (.%) (.%) (.%) (.%) Suspected rebleed (.%) (.%) (.%) (.%) Cause of death Progressive decline (.%) (.%) (.%) (.%) Unrelated cause (.%) (.%) (.%) (.%) Unknown (.%) (%) (.%) (.%) (.%) (.%) (.%) (.%) " patients allocated to bed rest for each aneurysm site. Downloaded from by on November, amination, these findings were designated as a proved rebleed. Whenever deterioration in clinical condition was not followed by a lumbar puncture or postmortem examination, a suspected rebleed was reported. The foregoing discussion is an analysis of proved rebleeds which followed the last prerandomized bleed responsible for hospitalization. Neurological condition, aneurysm site, and interval following last bleed were the most important parameters related to recurrent hemorrhage. Inspection of table reveals that the number of rebleeds for all aneurysm sites was.% ( of ). The highest frequency occurred during the through -day interval when (.% of ) patients rebled during this period alone. The frequency distribution gradually diminished in subsequent intervals. The majority (.%, or of ) occurred within days after the last bleed. The percentage of patients who rebled was.% for the internal carotid group,.% for the middle cerebral group, and.% for the anterior cerebral complex. Rebleed rate in grade patients was.% for the internal carotid group,.% for the middle cerebral group, and.% for the anterior cerebral complex. The higher percentage in the middle cerebral group may be influenced by the fewer number of patients for this site. The overall rebleed rate for good condition patients (grades and ) was.% ( of ). Comparison of rebleed rate among poor condition patients (grades,, and ) reveals that.% ( of ) rebled among the internal carotid group,.% ( of ) in the middle cerebral group, and.% ( of ) among patients with an anterior cerebral-anterior communicating aneurysm for an overall rate of.% ( of ). Therefore, no significant difference in rebleed rate occurred among good and poor neurological condition patients (.% vs..%). Mortality after recurrent hemorrhage was high for all aneurysm sites. Seven of patients died from the cerebra effects of the rebleed in the internal carotid group, of patients in the middle cerebral group, and among patients in the anterior cerebral group, for a mortality of.? ( of rebleeds). These deaths followed a proved rebleed with an additional deaths which followed s suspected rebleed, for an overall mortality of.% ( ol ). For comparison with the gross mortality of.! among patients in the entire bed rest treatment category, th( difference of.% includes such causes as progressive deterioration from the direct cerebral effects of the preran domized bleed and from unrelated and unknown factors. Interval to Randomization and Recurrent Hemorrhages Rebleeding was the major cause of morbidity and mor tality during the bed rest treatment period. For all aneurysn sites the group of patients allocated to bed rest within day after last bleed had the highest mortality and rebleed rate i the follow-up period. Figure presents schematically th rebleed rate for each aneurysm site among patients allocate to treatment within seven days. The group of patients wit an aneurysm on the anterior cerebral complex had th slowest rise during the through -day interval followin the last bleed. The curves in figure represent the data i table for patients randomized within seven days. Thi table contains the percentages in proved rebleed rate durin the days after the last prerandomized bleed for patient randomized within each interval. Among patients rar domized within days in the internal carotid group, f rebled; % rebled by day. No rebleeds were noted amon the six patients allocated to treatment during the to -da interval or in the four patients allocated during the to % day interval. Therefore, those percentages are not included i TABLE Interval to Death Following Randomization for Each Cause of Death with All Sites Combined Days from randomisation to death yr Over yr Proved rebleed " Suspected rebleed Progressive course IE Unrelated cause oo Unknown T "

6 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. TABLE Distribution of Death Rates Tabulated with Respect to Interval After Last Bleed, Neurological Condition, and Blood Pressure Systolic blood (mm Hg) Allocation to treatmijnt after last bleed - days Neurological condition* - days Neurological condition* Good Poor Subtotal Good Poor Subtotal or below i _^ _^ _ ^ (.%) (.%) (.%) (.%) (.%) (.%) (.%) Over (.%) (.%) (.%) (.%) (.%) (.%) (.%) j JLL _JL J IPjL (.%) (.%) (.%) (.%) (.%) (.%) (.%) See text for definition. Downloaded from by on November, table. Although the number of patients with a middle cerebral aneurysm was small, a relatively high proportion rebled. Of the patients randomized within the -day interval, five rebled, with one additional patient who rebled at days (not shown in table ). Three of seven patients rebled who were randomized within the to -day interval. Of the three patients allocated to treatment during the to -day interval, one rebled four years later. Therefore, a total of.% ( of ) rebled in the middle cerebral group. Patients with an aneurysm on the anterior cerebral complex who were randomized within days had a high rebleed rate. Forty-one per cent rebled during the -day interval following the prerandomized bleed (table ). Over % rebled within days. Patients allocated to bed rest during the to -day interval had a lower rebleed rate. Only % rebled by the fifty-ninth day following last bleed. None of the seven patients rebled who were allocated to bed rest during the to -day interval. Therefore, this interval was not included in table. For all sites combined, for patients randomized following the twenty-first day (when all patients randomized within this interval were allocated to treatment), % rebled between day and day (table ). Following the initial two-month interval, more patients with an aneurysm on the anterior cerebral-anterior communicating complex rebled than those with an aneurysm at the remaining two major sites. From the third through forty-eighth month following last bleed, nine patients rebled who were allocated to treatment within seven days, and four patients rebled in the group originally allocated to treatment during the to -day interval (table ). No late rebleeds were reported among patients allocated to treatment within the to -day interval. As mentioned previously for the middle cerebral group, one patient rebled on the seventy-fifth day, and one rebled four years later. No other late rebleeds occurred. Interval to Randomization and Subsequent Mortality The mortality distribution was variable for specific intervals following last bleed at each aneurysm site. Twenty-five per cent of patients with a middle cerebral aneurysm allocated to bed rest within days actually died within days;.% died in the anterior cerebral group (table ). Thirty-six per cent of patients with an internal carotid aneurysm died during that interval. By the twentieth day following last bleed,.% of patients died in the anterior cerebral group. Among those with an internal carotid and middle cerebral aneurysm, the mortality was.% and.%, respectively (table ). After the twentieth day, mortality continued to rise gradually among the anterior cerebral group, whereas mortality among patients with an internal carotid and middle cerebral aneurysm remained similar through the sixty-second day. Figure represents schematically the death rate for patients randomized within seven days (table ). Mortality between the fifth and twentieth day was most frequent among patients with an internal carotid and middle cerebral aneurysm. However, patients with an anterior cerebral aneurysm had a more gradual rise which did not reach its maximum until the sixty-second day. Table includes a summary distribution of deaths at TABLE Distribution of Rebleed Rates Tabulated with Respect to Interval After Last Bleed, Neurological Condition, and Blood Pressure Systolic blood (mm Hg) or below Over Allocation to treatment after last bleed - days Neurological condition* - days Neurological condition* Good Poor Subtotal Good Poor (.%) (.%) (.%) See text for definition. (.%) (.%) (.%) (.%) (.%) (.%) (.%) (.%) (.%) (.%) (.%) (.%) Subtotal (.%) (.%) (.%) (.%) (.%) (.%)

7 STROKE VOL, No, MARCH-APRIL TABLE Distribution of Patients with a Proved Rebleed for Each Neurological Condition and Aneurysm Site with Respect to Interval After Last Prerandomized Bleed Days after last bleed yr Over yr Internal carotid Neurological condition* See table for definition. Middle cerebral Neurological condition Anterior cerebrali Neurological condition Downloaded from by on November, monthly intervals for all patients with an aneurysm at each location and interval following last bleed. The group of patients who were randomized in the to -day and to -day intervals with an internal carotid or middle cerebral aneurysm was too small for secondary statistical analysis. However, % ( of ) died within two months and % within four months. A lower rebleed rate was the reason for the lower mortality two months following last bleed in the group of patients with an anterior cerebral aneurysm allocated to treatment during the to -day interval. After the sixty-second day mortality stabilized for all groups with the exception of patients allocated to treatment within days with an anterior cerebral-anterior communicating aneurysm. For patients with this aneurysm site, survived the initial days. Between the third and sixtieth month, additional patients (.%) died following a rebleed. These data suggest that during the initial eight-week period, patients with an anterior cerebral-anterior communicating aneurysm had a slower rebleed rate than those with an internal carotid or middle cerebral aneurysm. After this eight-week interval patients with an aneurysm on the anterior cerebral complex TABLE Cumulative Rebleeds After Allocation to Bed Rest Treatment Days after last bleed Internal carotid ()* No. -t % No. () - % Middle cerebral " patients allocated to specific interval to randomization. TDays to randomization after last bleed. No. had comparatively more recurrent hemorrhages than those with aneurysms at the remaining two sites. Overall Medical Condition of Patients with Ruptured Intracranial Aneurysm Each medical disorder was reported which existed prior to treatment or which was acquired during treatment. In an earlier section the relationship of mortality to overall medical condition was analyzed. The foregoing analysis includes data on various medical conditions as they relate to age, neurological condition, recurrent hemorrhage, and cause of death. Those conditions recognized prior to or at the time of randomization will be discussed individually from those conditions which developed during the bed rest treatment program. Medical Disorders at Randomization Pre-existing hypertension was the most frequently reported medical condition prior to allocation to treatment. () - %.. " No. Anterior cerebral () - % No. % () No. %

8 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. Uj p -O.O -O O-- O -Q'---O O"...Q-.-CJ o D- a MCI) / d i a & o a $ if Downloaded from by on November, DAYS FOLLOWING LAST BLEED FIGURE. Graphic comparison of rebleed rates at specific intervals following last bleed for each aneurysm site in patients randomized within seven days after last bleed. MC = middle cerebral, AC = anterior cerebral-anterior communicating, IC = internal carotid. No information was supplied with respect to duration or type of antihypertensive medication prescribed. Internal Carotid Aneurysm In these patients (.%) had known pre-existing hypertension, and in a definite history of hypertension was negative. No information was available in three. Among the nine patients with pre-existing hypertension, two had systolic levels in the range of to mm Hg, two in the range of to mm Hg, and two over mm Hg. In three patients this information was incomplete. Average age in these nine patients was. years (range to ). There were two males and seven females. Four patients were in grade neurological condition, two in grade, and three in grade. One patient was in good medical condition, four were in fair condition, and four in poor condition. Average systolic pressure prior to randomization was mm Hg systolic (range to ) and mm Hg diastolic (range to ). Average mean blood pressure was mm Hg. All nine patients had one bleed prior to hospitalization. Eight of the nine patients died: two followed a proved rebleed at and days after the last bleed, respectively; one followed a suspected rebleed; two followed progressive deterioration in clinical condition; and one each died with pneumonia, pulmonary edema, and from unknown cause. Middle Cerebral Aneurysm Three of patients (.%) had pre-existing hypertension; in patients there was no history of hypertension. No information was available in three. Among the three patients with pre-existing hypertension, one had known systolic levels in the range of to mm Hg, one in the range of to mm Hg, and in one the range was unknown. Average age was. years (range to ), and all were females. Two were in grade neurological condition, and one was in grade. One patient was in good medical condition, one in fair condition, and one in poor condition. Average systolic blood pressure at randomization was mm Hg (range to ), and average diastolic blood pressure was mm Hg (range to ). Average mean blood pressure was mm Hg. Two patients had had one previous bleed, and one patient had two bleeds. Two patients died " fc " DAYS FOLLOWING LAST BLEED FIGURE. Distribution of deaths at specific intervals following last bleed for each aneurysm site in patients randomized within seven days after last bleed. MC = middle cerebral, IC = internal carotid, AC = anterior cerebral-anterior communicating. following a proved rebleed and days after their last prerandomized bleed. One patient remained alive. Anterior Cerebral Aneurysm Among patients, (.%) had known pre-existing hypertension. Eighty-eight patients had no history of hypertension. No information was reported in. Among the patients with pre-existing hypertension, five had systolic levels between and mm Hg, one between and mm Hg, and one between and mm Hg. In seven patients systolic levels were over mm Hg. In ten patients the levels were unknown. Average age was. years (range to ). Nineteen were females; five were males. One patient was in grade neurological condition, in grade, two in grade, five in grade, and one in grade. Eleven patients were in good medical condition, eight in fair condition, four in poor condition, and one was in prohibitive condition. Average systolic pressure at randomization was mm Hg (range to ), and average diastolic pressure was mm Hg (range to ). Average mean blood pressure was mm Hg. Three patients had two previous bleeds, and had one bleed. Ten of the had a rebleed following randomization; two rebled five days after last bleed, one at nine days, one at ten days, two at days, and one each at days, days, days, and days. Seven of the ten patients died following their proved rebleed. In addition, two died following progressive clinical deterioration, TABLE Frequency Distribution of Proved Rebleeds for Patients Randomized at Specified Intervals Anterior Cerebral Aneurysms Only Months after ' last bleed Interval to randomization after last bleed (days) - -

9 STROKE VOL, No, MARCH-APRIL TABLE Cumulative Deaths After Allocation to Bed Rest Treatment DayB after hut bleed Internal carotid ()* -J No. % O No Middle cerebral ()* ()* -J % *Number of patients allocated to interval of randomization. t patients for all aneurysm sites and treatment intervals. (Interval to randomization after last bleed. No -J % Anterior cerebral ()* -J ()* -* No. % No % ()t No. % Downloaded from by on November, one following a suspected rebleed, and two from unknown cause, for a total of deaths (% of the patients with pre-existing hypertension). Summary This analysis reveals that among a total of patients with known pre-existing hypertension (table ), (.%) were in good neurological condition at the time of initial evaluation, and most patients ( of or.%) were in fair or poor medical condition. There were deaths. This number of deaths (. % of ) was not significantly greater than the.% for the entire treatment group. Therefore, pre-existing hypertension prior to treatment did not significantly influence mortality over the mean.-year interval. Medical Disorders Acquired During Bed Rest Treatment For all aneurysm locations.% ( of ) acquired various medical disorders during the bed rest treatment program. A genitourinary tract infection was reported in patients, pulmonary infection in, and a lower frequency of several other conditions for each aneurysm site. An acquired infectious disorder was the usual medical condition which complicated this course of treatment. Cerebral Angiography The technique of cerebral angiography was not specified according to protocol. Some institutions pursued the angiographic study within several hours of admission to the hospital, while others followed routine scheduling. In spite of these variables, several important factors relating to these procedures were analyzed. These analyses include number of procedures, complications, associated abnormalities (whether or not related to the aneurysm), vasospasm, interval after last bleed to first angiographic procedure, interval after last bleed to randomization, neurological condition, systolic and mean blood pressures, and size of the aneurysm. TABLE Frequency Distribution of Deaths After Last Bleed for Patients Randomized During Specified Intervals Internal carotid Middle cerebral Anterior cerebral Months after ()* () () () () () last bleed -t - - $ () Over? patients analyzed during each interval to randomization, tlnterval to randomisation.

10 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. TABLE Distribution of Clinical Characteristics in Patients vrith Pre-existing Hypertension Internal carotid () Middle cerebral () Anterior cerebralanterior communicating () () Average age (yr)... Good = grades and. tpoor - grades,,. See table for definition. JSee text for definition. Neurological condition Good* Poorf Medical condition} Good Fair Poor Average blood pressure Systolic Diastolic (mm Hg) (mm Hg) deaths Downloaded from by on November, These analyses are based on the sample of patients. This number includes the patients who completed their course of treatment, plus one patient who was discharged nine days after randomization. Occasionally, a result will be based on a total of patients because insufficient angiographic information was reported on two patients. Multiple angiographic procedures (more than one) were performed on of patients (.%) allocated to bed rest for a number of reasons. These procedures were distributed in similar proportions among the three aneurysm sites. In some institutions each carotid system was studied at separate intervals. Some patients were re-evaluated after deterioration in neurological condition associated with a rebleed or ischemic neurological deficit. Other patients had their vertebral-basilar system visualized angiographically during later procedures. Although the protocol suggested complete cerebral angiographic surveys, vertebral-basilar visualization was completed in only of patients (.%). Therefore, the presence of a single aneurysm was certain in these patients only. As each aneurysm was discovered, a complete angiographic survey was not always pursued when the clinical presentation was consistent with the angiographic findings. The number of multiple angiographic procedures (.%) may have been one factor responsible for delay in formal allocation to treatment. Another possible reason for delay was poor neurological and/or medical condition prior to transfer from the local hospital to the participating institution. The distribution was similar in patients allocated to irug-induced hypotension, carotid ligation, and intracranial surgery. The interval from the first angiographic procedure :o the day of randomization (the formal beginning of bed est) was no more than two days in.%, five days in.%, ind eight days in.%. Assuming that the first angiographic procedure was pursued in a uniform manner at each participating institution, the delay in randomization of more han five days in.% was not unusual. The reasons for lelay in those instances were not clarified. Post-treatment Angiography Among patients who had post-treatment angiographic ;urveys for determination of the volumetric size of the meurysm, the fundus portion of the aneurysm enlarged in.%), the dimensions remained unchanged in (.%), ind the aneurysm was smaller in (.%). Information on he patient data form was insufficient for three patients. Although patients who died prior to post-treatment angiogaphy did not have follow-up studies, among the survivors.% ( patients) had angiographic evidence of increased volume in the fundus portion of the aneurysm. Morbidity Associated with Angiography Unexpected deterioration in neurological condition was regarded as a complication when it occurred during or within hours after the angiographic procedure. Among patients subjected to procedures, (.% of ) had unexpected complications (table ). Among patients with an internal carotid aneurysm, two acquired hemiplegia and aphasia. In one patient the condition cleared eventually, and in the other it remained permanent and the patient died. One additional patient had transient hemiparesis which cleared in three hours. No patient rebled or had recognized hypotensive episodes. In the middle cerebral group one patient had temporary hemiplegia and aphasia which cleared in three hours, one acquired permanent hemiparesis, one became poorly responsive without focal manifestations and died, and one had a proved rebleed hours after cerebral angiography. In the anterior cerebral group four patients developed hemiparesis alone; in two the hemiparesis remained permanent, and in two it cleared within three hours. Three patients acquired temporary, dysphasia which cleared in three to six hours. One patient acquired permanent hemiparesis and aphasia, one was lethargic following a left pneumothorax and died four days later, and one had temporary diplopia and increased headache. Three patients rebled, one fatally, and the other two became poorly responsive. One of these two patients died five days later, and the other patient died five months later. One additional patient developed convulsions during the procedure. TABLE Distribution of Patients with Complications Which Occurred During or Within Hours After Angiographic Procedure Complicating event Hemiplegia and aphasia Hemiparesis alone Hemiparesis and aphasia Dysphasia alone Rebleed (remained alive) Rebleed (fatal) Convulsions Poorly responsive Left pneumothorax and poorly responsive Diplopia and headache Internal carotid Aneurysm site Middle cerebral Anterior cerebral

11 STROKE VOL, No, MARCH-APRIL TABLE Each Site Site Internal carotid ()* Middle cerebral () Anterior cerebral () () Distribution of Associated Lesions in Addition to a Single Aneurysm at Angiographic Survey for Angiographic abnormality Stretching Extracerebral Intracerebral of anterior mass mass cerebral n Irregular intima (atherosclerosis) Infundibular dilatation Number of patients allocated to each site. tnine patients had more than one angiographic abnormality. _ {Additional patients with miscellaneous findings included four patients with nonfilling of one anterior cerebral artery, one had displacement of midline vessel*, one an occluded right internal carotid artery in the neck, and one a superficial temporal artery a- neurysm. f Downloaded from by on November, In summary, five patients (.%) died after their rebleed or ischemic deficit. Morbidity was.%. It must be emphasized that the mortality and morbidity were associated with multiple angiographic procedures in acutely ill patients with a hemorrhagic cerebral vascular disorder. Many patients with subarachnoid hemorrhage are also at greater risk than those with an ischemic vascular disorder or with normal intracranial pressure. Angiographic Abnormalities Associated with the Aneurysm Abnormalities associated with single ruptured intracranial aneurysms were noted in patients. A total of lesions were reported among all aneurysm sites (table ). Intracerebral mass lesions were noted in.% ( of ). Most lesions were consistent with a hematoma and/or localized area of cerebral swelling. Sixteen patients (.%) had increased sweep of the anterior cerebral vessels; (.%) had evidence of an extracerebral mass lesion. Only two patients had an irregular intima consistent with intracranial atherosclerosis, and one had an infundibular dilatation. For each aneurysm location the most frequent abnormality was an intracerebral mass lesion as determined by angiographic examination. Correlation of intracerebral mass lesions (), extracerebral mass lesions (), and stretching of the anterior cerebral vessels () revealed that these lesions occurred times in patients. Such lesions are commonly associated with poor neurological condition. Therefore, one would expect increased mortality with progressive clinical deterioration from direct cerebral effects of the bleed. Among eight patients with an internal carotid aneurysm, six died: four after progressive deterioration in neurological condition and one each from unrelated and unknown cause. Among six patients with a middle cerebral aneurysm, five died: three after a proved rebleed and one each after a suspected rebleed and unrelated cause. Among patients with an anterior cerebral aneurysm, died: after a proved rebleed, three after a suspected rebleed, two from unrelated causes, and one after progressive clinical deterioration. In summary, deaths (.% of patients) were associated with these angiographic findings. Sixteen of these deaths followed a proved rebleed, five followed progressive deterioration, four followed a suspected rebleed, and four were from unrelated and one from unknown causes. Comparison of the overall mortality of.% in the bed rest treatment group with the.% figure cannot be regarded as a statistically significant increase. Most deaths ( of ) followed a proved rebleed. Cerebral Vasospasm The angiographic demonstration of decreased intraarterial size is usually regarded as cerebral vasospasm. Although the primary objective of this study was not the evaluation of vasospasm, the degree of arterial narrowing as interpreted on the cerebral angiogram was reported to the Central Registry. We are cognizant of the pitfalls in the analysis of the data submitted due to variation in the interpretation between one observer and another as to what actually constitutes vasospasm. Cerebral vasospasm was categorized on each angiogram as one of three degrees ol severity: no vasospasm; localized vasospasm (narrowing ol intra-arterial diameter over to -cm distances on the proximal portions of the intracranial portion of the internal carotid artery, middle cerebral, anterior cerebral, oi vertebral-basilar artery); and diffuse vasospasm (narrowed diameter of the intracranial internal carotid artery and all the main ipsilateral branches). Various degrees of vasospasm were correlated with age, sex, neurological condition systolic and mean blood pressures, interval from last bleed tc angiography, size of the aneurysm (mm*), and mortality. Age The frequency of vasospasm among all aneurysm sites anc age groups was.% ( of ). The proportion of patient: with vasospasm increased proportionately with age. A Iov figure of.% was observed in the age group to, witl TABLE Age Distribution with Respect to Each Category o Vasospasm Age No data Cerebral vasospiasm None Localized* Diffuse* *See text for definition. % with vasospasm

12 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. Downloaded from by on November, the highest incidence of.% in the age group to (table ). Among the seven patients in the age group to, only two had localized involvement. Analysis of the entire group below age reveals.% ( of ) had some degree of vasospasm, whereas in those patients who were years of age or over,.% had this condition. This difference was not statistically significant. However, vasospasm was present in all age groups, with a higher frequency among the older groups, especially among females (see Discussion, next section). Sex Distribution Evaluation of sex distribution with each aneurysm site revealed that patients with an internal carotid aneurysm included males and females. In contrast, the proportion of males and females in the middle cerebral and anterior cerebral groups was very similar (table ). The distribution of patients with various degrees of vasospasm was proportional among each aneurysm location with the exception of females in the middle cerebral group where four patients had no vasospasm, three had localized vasospasm, and four had diffuse vasospasm. For all aneurysm sites combined, % ( of ) of men had variable degrees of vasospasm in comparison with.% ( of ) of women. Localized spasm must be clarified, particularly in relation to the A, segment of the anterior cerebral arteries. Differentiation between congenital narrowing of this segment and localized vasospasm was difficult in some instances. Analysis of the patient report forms reveals that unilateral narrowing of the A! segment proximal to the anterior communicating artery was observed in of patients with an aneurysm on the anterior cerebral complex. In two additional instances localized spasm was bilateral, and in one patient with an internal carotid aneurysm it was also bilateral. No instances of selected A, narrowing were reported in the middle cerebral group. Therefore, a total of patients with "localized vasospasm" conceivably had congenital narrowing of the Ai segment. This would reduce the total number of patients with localized and diffuse vasospasm to (.% of ) instead of (.%), a difference which is not significant. Therefore, this observation can be disregarded in various correlative analyses. TABLE Sex Distribution for Aneurysm Site and Degree of Vasospasm Cerebral vasospasm Aneurysm site None Localized Diffuse * Internal carotid Middle cerebral Anterior cerebral MALES FEMALES Internal carotid Middle cerebral Anterior cerebral There was insufficient information on one patient, making a total of. oo co co Interval from Last Bleed to Initial Angiography Analysis of the number of patients with and without vasospasm at specific intervals following last bleed revealed that % had their initial angiographic procedure within three days after the bleed. During this interval patients (.% of ) had no vasospasm, (.%) had localized vasospasm, and (.%) had diffuse involvement (table ). By comparison, among the % who had their initial angiographic procedure after the three-day interval, (.% of ) had no vasospasm, (.%) had localized spasm, and (.%) had diffuse vasospasm. Thus, a total of patients (.%) had localized or diffuse vasospasm when the initial angiogram was performed four or more days after the last bleed. For all sites combined, this percentage was a.% increase. Particularly during the to -day interval (fig. ), the percentage of patients with vasospasm increased appreciably. The appearance of intra-arterial narrowing, as demonstrated by cerebral angiography, had its peak incidence during that interval. Neurological Condition Correlation of neurological deficit with specific degree of vasospasm reveals that patients in poor neurological condition (grades,, and ) had a greater frequency of localized and diffuse arterial narrowing. Among various neurological conditions.% of patients in grade had localized or severe vasospasm,.% in grade, and.% in grade (table ). Comparison of good condition patients (grades and ) with those in poor condition (grades,, and ) reveals that in the former group.% ( of ) had vasospasm (mostly localized), whereas.% ( of ) of those in the latter group had localized or diffuse vasospasm in similar proportions (table ). The frequency and severity of arterial narrowing were significantly increased (P <.) in poor condition patients. TABLE Distribution of Patients with Respect to Interval from Last Bleed to First Angiographic Procedure Tabulated for Various Degrees of Vasospasm Among All Aneurysm Sites Interval* Over % Cerebral vasospasm None Localized Diffuse (.%) (.%) (.%) f % initial angiogram completed Days from last bleed to first cerebral angiogram. tno information was available on one patient, making a total of.

13 STROKE VOL, No, MARCH-APRIL Downloaded from by on November, r ^ K & DAYS FOLLOWING LAST BLEED FIGURE. Per cent of patients with localized and diffuse vasospasm at specific intervals after last bleed. The percentages plotted in the figure were calculated from the data in table. Systolic Blood Pressure Four representative blood pressures, taken at more than four-hour intervals between hospital admission and randomization, were recorded by the usual manometric technique. These four readings were averaged at the Central Registry. Average systolic values at various increments were tabulated with various degrees of cerebral vasospasm for all sites combined. Most patients ( of or.%) had systolic values between and mm Hg (table ). Ninety-four patients (.% of ) had systolic values of mm Hg or below, (.%) between and mm Hg, and (.%) had systolic values over mm Hg. Therefore, nearly one-half of all patients had systolic values of mm Hg or below. The percentage of patients with vasospasm at various increments varied between.% and.% (table ). No association could be established whereby elevated systolic blood pressure suggested increased incidence of vasospasm. TABLE Distribution of Neurological Condition with Respect to Degree of Vasospasm for All Aneurysm Sites Neurological grade* Cerebral vasospasm None Localized Diffuse Tl f % with vasospasm..... "See table for definition. fthere was no information on one patient, making a total of. TABLE Distribution of Patients with Various Degrees of Vasospasm Tabulated with Respect to Designated Levels of Systolic Blood Pressure for All Aneurysm Sites Systolic blood pressure (mm Hg) Below Over Cerebral vasospasm None Localized Diffuse No information on one, making a total of. Mean Blood Pressure * % with vasospasm Mean blood pressures were calculated, using the data compiled for systolic and diastolic determinations. In order to determine the relationship of vasospasm to systolic and diastolic values, mean blood pressure distribution was compared with various degrees of vasospasm. Forty-nine patients (.% of ) had mean blood pressures varying from to mm Hg (table ). Most patients (.% or of ) had mean pressures between and mm Hg. In this group.% ( of ) had localized or diffuse vasospasm. The ratio of patients with vasospasm versus those without vasospasm varied among various values of mean blood pressure. No evidence of an association was observed between elevated mean pressure and severity of vasospasm. Size of Aneurysm Measurements of the intraluminal diameter of the aneurysm on the cerebral angiogram were recorded at each participating institution. We are cognizant of variations in measurements taken from one institution to another. Each dimension was reported to the nearest fraction of a millimeter. All volumetric determinations were calculated at the Central Registry. The median volumetric size was mm*. That group with an aneurysmal volume below mm* had a.% incidence of vasospasm (table ). The group of patients with volumetric measurements above mm' had a similar frequency of vasospasm (.% or of ). Therefore, aneurysm size was unrelated to the incidence of vasospasm. Mortality Increased mortality and degree of vasospasm were related to poor neurological condition. The relationship of mortality to various degrees of vasospasm was also analyzed. In patients with no vasospasm,.% died among all aneurysm sites combined. This figure compares with.% and.% among those patients with localized and diffuse vasospasm, respectively (table ). Therefore, vasospasm itself was unrelated to mortality on a statistical basis. Summary of Vasospasm Data The role of vasospasm in the production of neurological dysfunction in patients with subarachnoid hemorrhage has not been fully investigated in this study, and only some

14 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et ai Downloaded from by on November, TABLE Dislribution of Mean Blood Pressure Tabulated with Respect to Various Degrees of Vasospasm for All Aneurysm sues Mean Cerebral vasaspasm % with blood pressure Nont Localized Diffuse vasospasm Over *No information on one, making a total of. * sn n. general statements can be made. Neither does this analysis fully contribute to a clear understanding of the problem. However, these data suggest that poor neurological condition has a close relationship to vasospasm in comparison with other variables under consideration. The severity of neurological deficit correlated highly with the presence of some degree of vasospasm. Analysis of sex distribution shows that females with an internal carotid aneurysm had a higher incidence of vasospasm than males. Too few patients with a middle cerebral aneurysm were collected for making a definite conclusion. There was no correlation between males and females with a ruptured aneurysm on the anterior cerebral complex. The parameter of age and the severity of vasospasm revealed no statistical relationship, but the average percentage of patients with vasospasm was % higher in those above years of age versus those in the fifth decade or below. No relationship to vasospasm was observed for variables such as average systolic or mean blood pressures, aneurysm size, or mortality. Discussion Patients allocated to regulated bed rest provided a baseline and control for comparison of effectiveness of the other treatment modalities. Every effort was made by each participating institution to conduct this therapeutic program in as uniform a manner as possible. The adequacy of randomization was evaluated with respect to age, sex, aneurysm site, neurological and medical conditions, and vasospasm. TABLE Distribution of Aneurysm Size Tabulated with Respect to Degree of Vasospasm for All Sites Combined Size (mm )* Over Cerebral vasoepaam None Localized Diffuse TI " *See text for method of calculation. tlnsumcient information in one, making a total of. Tf TABLE Distribution of Degree of Vasospasm with Respect to Each Aneurysm Site site carotid Middle cerebral Anterior cerebral Per cent Cerebral vasospasm None Localized Diffuse ()» () () () () () () ().. () () () (). *Number in parentheses denotes deaths. t patients with treatment followed according to protocol. () () () f (). These parameters were all found to be in proper distribution for all aneurysm sites. Patient distribution in the bed rest treatment group was.% for those in the age group to,.% in the age group to,.% in the age group to, and.% in the age group to. This distribution was similar to that reported by Pakarinen' (.%,.%,.%, and.%, respectively), and it also concurs with his review of several published reports. Comparison of the results of this study with nonrandomized published reports is somewhat hazardous because conservatively treated patients usually form a highly selected portion of the representative sample undergoing analysis. Therefore, the results for morbidity and mortality tend to be inordinately high in such groups. Patients in Logue's series who were treated conservatively and followed months to years had a % ( of ) recurrent rebleed rate with a mortality of % from such recurrences. In a series of patients treated conservatively and followed an average of years (varying from Vh to years), % ( of ) had a recurrent bleed. Forty-eight of these (%) were fatal. In Parkarinen's series* the patients treated conservatively (i.e., one month of bed rest) included a very selected group. Sixty-four deaths occurred before angiographic demonstration of the aneurysm. Furthermore, patients were more than years of age. Among the deaths occurred during the first week after onset of symptoms. Recurrent bleeding episodes for the total series was.%, with an incidence of.% at eight weeks. Cumulative mortality two weeks after last bleed was.%, at the fourth week.%, and at the sixth week.%. These figures did not include the preoperative patients. By comparison, the cumulative mortality in the present study was.% at the end of the second week,.% at the fourth week, and.% at the sixth week. However, comparison of all conservatively treated patients in Pakarinen's series (including all those prior to operation) reveals that.% died within two weeks,.% within four weeks, and.% within six weeks after the last bleed. These percentages, interestingly, show a striking similarity to those of the present study. The values presented in this study probably represent a mortality which is nearest to the actual natural course following subarachnoid hemorrhage from a ruptured intracranial aneurysm in patients admitted to referral centers. After this six-week interval mortality rises slowly. In a randomized trial* stratified with respect to age, hypertension, clinical condition, and interval following last

15 STROKE VOL, No, MARCH-APRIL Downloaded from by on November, bleed to the beginning of treatment, patients with an aneurysm at the junction of the internal carotid and posterior communicating artery were treated with six weeks of bed rest. Six months after the bleed,.% had died. In the present study.% ( of ) died during the six-month interval following last bleed. With regard to patients with a middle cerebral aneurysm,.% died within six months after the last bleed. In this study.% ( of ) died during that interval. Mortality in patients with an anterior cerebral-anterior communicating aneurysm in the randomized trial of McKissock et al. was.% ( of ) at six months (.% died following a rebleed within days after the last bleed). In this study aneurysms on the anterior cerebralanterior communicating complex had a cumulative mortality of.% at four weeks,.% at the end of six weeks, and.% ( of ) at six months. These figures compare favorably with the series from London for each aneurysm site. Recurrent hemorrhage within two months after the bleed has been the major cause of death in many published reports of unoperated patients. In undifferentiated subarachnoid hemorrhage reviewed by Pakarinen," an average of.% rebled within eight weeks after their initial bleed, the major proportion occurring in the second and third weeks (.% and.%, respectively). In Pakarinen's series of patients who had a proved aneurysm (), including preoperative recurrences in patients eventually treated operatively, patients (.%) had a recurrent bleed within eight weeks. Fourteen (.%) occurred during the first week, (.%) during the second, (.%) during the third, and (.%) during the fourth. Therefore, a total of.% rebled within four weeks, with an additional.% in the following two weeks. This high rebleed rate shows the importance of differentiating subarachnoid hemorrhage due to ruptured aneurysm alone, rather than including all cases in an undifferentiated series. In addition, the influence of patient selection (nonsurgical patients) was also apparent. Furthermore, aneurysm location can be influential. In the randomized series of conservatively treated patients with an internal carotid aneurysm, (%) rebled within six weeks, and of these died (.%, of ). Most of them () occurred during the four to six-week interval. In the present study five patients (.%) with an internal carotid aneurysm rebled within two weeks, six (.%) in four weeks, and nine (.%) in six weeks, with no further rebleeds to the end of the eight-week interval. Essentially, these percentages relate the number of rebleeds which occurred during treatment with bed rest. Twenty-six patients began their official period of bed rest within seven days after last bleed. However, when consideration was given to the number of rebleeds which occurred during the interval prior to hospitalization, eight patients had clinical neurological symptoms or signs of an initial bleed during a -day interval preceding the bleed prompting admission to the reporting hospital. Among all patients with an internal carotid aneurysm, eight had two bleeds prior to official commencement of bed rest. Nine additional patients rebled within six weeks after entry into the study. Therefore, a total of patients (.%) with recurrent hemorrhage were reported over a period of weeks. Analysis of those patients from the study who started their bed rest program within seven days after the last bleed showed that five of patients (.%) rebled in days, six (.%) rebled within four weeks, and a total of nine (.%) in eight weeks. The rebleed rate among patients with an internal carotid aneurysm in the Cooperative Study" (nonrandomized) was.% ( of ) at days and.% ( of ) at the end of days (four weeks). The rebleed rate at the six-week interval was not reported. In the randomized trial of McKissock et al.,' of (.%) conservatively treated patients with a middle cerebral aneurysm rebled during the initial two-week interval,.% ( of ) during the three-week interval, and.% ( of ) during the four-week interval with no recurrent hemorrhages to the end of six weeks. In the present study.% ( of ) rebled during the initial -day period,.% ( of ) at days, with no further rebleeds to the end of six weeks. In the Cooperative Study series" (nonrandomized),.% ( of ) rebled during the -day interval, and.% ( of ) by the end of four weeks. The data at six weeks were not reported The randomized series of anterior communicating aneurysms treated conservatively had a rebleed rate of.% ( of ) during the first days, with.% ( of ) having a rebleed during the four-week interval following the bleed prior to hospitalization. In the present study.% ( of ) rebled during the first two weeks following last bleed and a total of.% ( of ) within days. In the previous Cooperative Study,".% ( of ) rebled within days and.% ( of ) within four weeks. No statistical differences were observed among the three studies at these intervals following the bleed. Mortality among patients following recurrent hemorrhages in the previous Cooperative Study" was % for internal carotid aneurysms, % for middle cerebral aneurysms, and % for anterior cerebral-anterior communicating aneurysms. In the present study death due to rebleeding at days was.% ( of ) for patients with an internal carotid aneurysm,.% ( of ) for those with a middle cerebral aneurysm, and.% ( of ) for patients with an anterior cerebral-anterior communicating aneurysm. In Pakarinen's series".% died following recurrent hemorrhage. Mortality during the six-week interval in the conservatively treated patients in the randomized trial ' * among those who had a second bleed was.% ( of, category B) for patients with an internal carotid aneurysm,.% ( of ) for those with a middle cerebral aneurysm, and.% ( of ) for the anterior cerebralanterior communicating group. In general, mortality of nonsurgically treated patients is lower in studies conducted with a design of randomization in comparison with the number of deaths reported from a nonrandomized selection. The wide spectrum of clinical manifestations of subarachnoid hemorrhage following ruptured intracranial aneurysm was confirmed in this investigation. Most patients were in good neurological condition after their initial bleed. Sixtyseven per cent were in grades and,.% ( of ) were in grade, and.% ( of ) were lethargic with varying degrees of neurological deficit (grades and ). The number of patients with severe central nervous system deficits from the initial rupture was small. More patients experience minimal symptoms frequently, which is consistent

16 ANEURYSM/HEMORRHAGE TREATMENT IV-A BED REST/Nibbelink et al. Downloaded from by on November, with minor bleeding episodes. A recent report" suggests that such symptoms occur in.% (/) as a warning sign. In the present study.% (/) had a recent history of symptoms characterized by headache or neck or back pain as manifestations of meningeal irritation. In at least threefourths of patients with initial rupture of an intracranial aneurysm, symptoms and signs are relatively minor. By far, the most typical presentation was that complex of symptoms shown by patients in grade, namely, severe headache, meningeal signs, and varying degrees of nausea and vomiting, with a transient episode of amnesia in some instances. Sex distribution has been of particular interest among patients with an internal carotid-posterior communicating aneurysm. For this treatment category were males, and were females. This distribution was similar among all four treatment categories. Sex distribution among patients with a middle cerebral and anterior cerebral-anterior communicating aneurysm was nearly equal. The reason for this peculiar distribution of more females among patients with rupture of an internal carotid aneurysm remains unexplained. Among patients who rebled, died. Twenty-eight patients (.%, table ) rebled during the interval from the fifth through the fourteenth day following last bleed. The most crucial interval was this period. A unique feature was the distribution of rebleeds among patients with an anterior cerebral-anterior communicating aneurysm. These patients continued to have the most prolonged interval to recurrent hemorrhage, an observation which remains unexplained. This study appraises and confirms the variation in intervals to recurrent hemorrhages among aneurysm sites. At one end of this spectrum were the so-called "minor bleeds" which may herald a warning sign an average of two weeks prior to the more severe ictus. At the opposite end were those who had one severe hemorrhagic catastrophe with devastating effects to the central nervous system at the initial rupture. Between these two extremes was the large group of patients with symptoms characterized by headache, nausea, vomiting, and usually transient loss of consciousness which prompted medical attention. This group comprised nearly % of the entire patient population. Particular attention must be focused on those patients who had their workup completed and were allocated to therapy within seven days following subarachnoid hemorrhage. This group probably provides the most important information with respect to the number of concrete observations that are associated with subarachnoid hemorrhage following a ruptured intracranial aneurysm. In the group with minor symptoms, too many factors may not appear clinically distinct from other common disorders. In the severe catastrophic lesions, the central nervous system was too extensively and too rapidly damaged to make a definite course of therapy of any value. Patients allocated to treatment within seven days after last bleed had the highest death rate during the fifth through twentieth day (fig. ). During this interval rebleeding was the most frequent cause of death for all aneurysm sites. Patients allocated to treatment after the seven-day interval were distributed similarly among the various neurological conditions as compared with those patients allocated to treatment within seven days. With % of all patients allocated to treatment during the seven-day interval, the risk of rebleeding was most substantial among patients with an internal carotid aneurysm, followed by those with a middle cerebral aneurysm. Patients with an anterior cerebralanterior communicating aneurysm had the slowest rate of recurrent hemorrhage. Various medical conditions which were present prior to rupture were relatively infrequent. The most frequent condition prior to rupture was pre-existing hypertension. Only % had known hypertension, a frequency which would not lead one to designate hypertension as an underlying cause for an aneurysm to rupture. Mortality in this group was.%, a percentage not significantly different from the gross mortality of.%. Conclusions. During the interval from June, through February,, patients with a single intracranial aneurysm were randomly allocated to regulated bed rest. After the last bleed were allocated to treatment within days, during the to -day interval, and during the to -day interval.. Thirty-seven patients had a single aneurysm on the intracranial portion of the internal carotid system, had an aneurysm on the middle cerebral distribution, on the anterior cerebral-anterior communicating complex, and on the vertebral-basilar system. There were men and women.. Eight patients failed to receive their designated treatment. Subtracting these eight patients in addition to the seven with a vertebral-basilar aneurysm allowed a remainder of for correlative analysis.. During the mean follow-up interval of. years, mortality was.%. Cumulative mortality was.% for those patients randomized within days,.% for those randomized during the to -day interval, and.% for those allocated to treatment during the through -day interval.. Eleven patients in grade neurological condition had a.% mortality. Among patients in grade,.% died; among patients in grade and patients in grade,.% died, and among patients in grade, all died.. For all aneurysm sites a proved rebleed was the cause of death in.%, progressive deterioration from aneurysm rupture in.%, and a suspected rebleed in.%. A total of. % died from causes related directly to the cerebral effects of the ruptured aneurysm.. Mortality in good medical condition patients was not significantly different in the groups below or above age. Among patients in fair and poor medical conditions, mortality above age was significantly higher than for those below that age.. For all aneurysm sites patients allocated to bed rest within seven days after last bleed had the highest mortality and rebleed rate.. In.% of patients several medical disorders needed medical attention during the treatment period. The most frequent conditions were pulmonary or genitourinary tract infections.. Unexpected deterioration in neurological condition occurred during or within hours after angiography in.%. Mortality associated with this procedure was.%.. The incidence of vasospasm as demonstrated by

17 STROKE VOL, No, MARCH-APRIL cerebral angiography was.%. In patients below age,.% had definite evidence of vasospasm; for those age or above,.% had this condition.. For patients in good neurological condition (grades and ),.% had vasospasm, whereas in poor condition patients (grades,, ),.% had some degree of vasospasm.. Systolic and mean blood pressures were statistically unrelated to the frequency or severity of vasospasm.. Aneurysm size was unrelated statistically to the incidence of vasospasm.. Mortality among patients with no vasospasm was.%. In the group of patients with localized vasospasm,.% died. In patients with diffuse vasospasm, the mortality was.%. Acknowledgment The authors are indebted to each of the following for reviewing this manuscript: Mark L. Dyken, M.D., A. L. Sahs, M.D., and S. C. Strickland, M.D. References. McKissock W, Paine K, Walsh L: Further observations on subarachnoid haemorrhage. J Neurol Neurosurg Psychiat : -,. McKissock W, Richardson A, Walsh L: "Posterior-communicating" aneurysms: Controlled trial of the conservative and surgical treatment of ruptured aneurysms of the internal carotid artery at or near the point of origin of the posterior communicating artery. Lancet : -,. Sahs AL: Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: Report on a randomized treatment study: I. Introduction. Stroke : -,. Nibbelink DW, Knowler LA: Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: Report on a randomized treatment study: II. Objectives and design of randomized aneurysm study. Stroke S: -,. Graf CJ, Nibbelink DW: Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: Report on a randomized treatment study: III. Intracranial surgery. Stroke : -,. Logue V: Surgery in spontaneous subarachnoid haemorrhage: Operative treatment of aneurysms on the anterior cerebral and anterior communicating artery. Brit Med J : -,. Tappura M: Prognosis of subarachnoid haemorrhage: Study of patients with unoperated intracranial arterial aneurysms and patients without vascular lesions demonstrable in bilateral carotid angiograms. Acta Med Scand Suppl, pp -,. Pakarinen S: Incidence, aetiology, and prognosis of primary subarachnoid haemorrhage. Acta Neurol Scand Suppl, pp -,. McKissock W, Richardson A, Walsh L: "Middle cerebral" aneurysms: Further results in the controlled trial of conservative and surgical treatment of ruptured intracranial aneurysms. Lancet : -,. McKissock W, Richardson A, Walsh L: "Anterior communicating" aneurysms: Trial of conservative and surgical treatment. Lancet : -,. Sahs AL, Perret GE, Locksley HB, et al: Intracranial Aneurysms and Subarachnoid Hemorrhage. A Cooperative Study. Philadelphia, JB Lippincott Co,. Okawara S: Warning signs prior to rupture of an intracranial aneurysm. J Neurosurg : -, Downloaded from by on November, Central Nervous System Angioendotheliosis SUMMARY CNS neoplastic angioendotheliosis is a treatable primary proliferative disorder of the endothelial cells of blood vessels characterized by a clinical neurological picture of multiple infarct dementia and an inordinate amount of local cerebral edema, so striking that it may simulate primary or metastatic central nervous NEOPLASTIC ANGIOENDOTHELIOSIS is a rare disorder of blood vessels characterized by a bizarre array of neurological symptoms associated with dementia, strokelike syndrome, and dermatological involvement. There is widespread vascular endothelial cell proliferation contained within the blood vessels throughout the body. This condition has been referred to under various designations as "angioendotheliomastosis proliferans, systemistra" and "diffuse malignant proliferation of the vascular endothelium." From the Departments of Neurology, Pathology (Division of Neuropathology), and Radiology (Nuclear Medicine Section), Indiana University School of Medicine, Indianapolis, Indiana. This investigation was supported by CVD Research Grant NS-- USPHS. Dr. Reinglass is now at the Northeastern Ohio University College of Medicine. Reprint requests to Dr. Reinglass, Harrison Avenue N.W., Canton, Ohio. A Treatable Multiple Infarct Dementia JAMES L. REINGLASS, M.D., JANS MULLER, M.D., STANLEY WISSMAN, M.D., AND HENRY WELLMAN, M.D. system tumor. The malignant cells remain within the lumen of the vessels and rarely if ever metastasize or occur in peripheral blood. There is remarkable improvement in symptoms by treating with high dose steroids. Antimetabolites and irradiation are suggested means of additional treatment. The purpose of this communication is to describe the remarkable response of the neurological symptoms to treatment with steroids and to add an additional case report of this rare disease to the literature. Case History This -year-old, white, right-handed woman was referred to the Department of Neurology of the Indiana University Medical Center because of progressive dementia and left hemiparesis. She had been well until two months previously when she developed the sudden onset of aphasia and rightsided weakness. She was admitted to her local hospital, and a diagnosis of stroke was made. In the following weeks, however, she suffered several episodes of incoherent speech and was readmitted to the original hospital where she had a normal brain scan and left carotid arteriogram. Lethargy, dress-