Endovascular techniques are now used twice as often as traditional open craniotomy and clipping to treat unruptured intracranial aneurysms and may have more favorable outcomes.1–3 Previous studies of patients undergoing elective craniotomy indicate that routine intensive care unit (ICU) admission may not be necessary.4–6 Although endovascular techniques are less invasive than open craniotomy, the potential for major cerebrovascular and other complications remains.7,8 Currently, it is not clear whether postoperative ICU admission should occur for all, none, or some of these patients.9
Accordingly, in this single-center retrospective case series, we reviewed the outcomes of patients who underwent elective endovascular treatment of unruptured intracranial aneurysms. During a 3-year period, we progressively changed the (operating room [OR]) managerial practice of admitting all such patients directly to the ICU in lieu of admission to the postanesthesia care unit (PACU). We evaluated whether neurologic outcomes (and/or other major adverse events) differed between patients who were directly admitted to the ICU following their procedure versus patients who were admitted to the PACU. We also reviewed our data with regard to anesthesia duration for (1) case scheduling and (2) day of surgery prediction of ICU admission.
This study was approved by the IRB of the University of Iowa (#201207734), and the need for individual patient consent was waived. Following Strengthening the Reporting of Observational Studies in Epidemiology guidelines for a cohort study,a we retrospectively reviewed the electronic medical records of 162 patients who underwent a total of 170 consecutive elective endovascular procedures to treat ≥1 unruptured intracranial aneurysms between July 20, 2009 and September 28, 2012. Each procedure was considered to be an independent event. We did not perform an a priori power analysis because we studied the entire sample of patients undergoing these procedures. However, we knew before analysis that our sample size would be greater than, or equivalent to, several previous studies reporting neurologic outcomes and adverse events of patients undergoing elective endovascular procedures to treat unruptured aneurysms.10–15
Postoperative Care Pathways and Protocols
All patients received postoperative care in ≥2 of the following 3 locations: (1) the ICU, (2) the PACU, and (3) the neurosurgical ward (after initial ICU or PACU admission) (Table 1). Postoperative care protocols for patients undergoing these procedures in each of these locations are summarized in Table 1. For each patient, the decision regarding immediate postoperative care location was made jointly by the neurosurgeon and the anesthesiologist assigned to the patient that day. Patients were considered appropriate for PACU admission and subsequent routine neurosurgical ward care if they had not experienced a recognized intraoperative event with the potential for ongoing cerebral ischemia, such as aneurysm rupture, in-stent thrombosis, coil migration, persistent coil herniation, or cerebrovascular dissection or occlusion. In all patients, regardless of initial postoperative care location (ICU or PACU), postoperative systolic arterial blood pressure values that prompted neurosurgery team notification were those >160 mm Hg or <90 mm Hg.
Patients assigned to PACU admission had femoral arterial access sheaths removed in the OR before extubation. With 1 exception, all PACU patients were tracheally extubated in the OR before PACU arrival. If a PACU patient was receiving IV tirofiban on arrival, the infusion was discontinued at 2 hours (total infusion duration) unless a new neurologic complication was noted. Unless conditions requiring ICU care or other intervention were identified in the PACU (see Results), PACU patients were subsequently transferred to the neurosurgical ward. Transfer from the PACU to the neurosurgical ward required that the patients have (1) stable arterial blood pressure, heart rate, and respiratory rate within target ranges for ≥3 consecutive measurements ≥15 minutes apart (systolic pressure 90–200 mm Hg, diastolic pressure 60–100 mm Hg, heart rate 50–120 beats/min, and respiratory rate 8–29 breaths/min); (2) no respiratory distress or obstruction with pulse oximetry values ≥92%, with or without supplemental oxygen; (3) a core temperature 36.0 to 38.5°C; (4) a level of consciousness sufficient to either verbally call for assistance and/or use a call light; and (5) control of pain and nausea with last administration of IV agents ≥15 and ≥30 minutes before transfer, respectively. The minimal PACU stay was 30 minutes. There was no maximal PACU stay, but patients who did not meet criteria for transfer to the neurosurgical ward within 2 hours were assessed for possible ICU transfer.
Neurosurgical ward patients who developed a need for intubation, intra-arterial pressure monitoring, continuous IV vasopressors, or any other form of intensive care were transferred to the ICU (not to the PACU).
All data were collected (SHE) using a standardized data collection instrument that included 128 items. Variables included date of procedure, patient demographics, and coexisting nonneurologic conditions, including hypertension, smoking, and alcohol dependence.16b Comorbidities, including cardiovascular, peripheral vascular, respiratory, renal, and liver disease, as well as information regarding malignancy and acquired immune deficiency syndrome, were recorded. Neurologic history, previous neurologic interventions, and preoperative functional status (Rankin score17) were also recorded. With the use of these data, the Charlson Index (comorbidity) score18,19 was calculated. These data are provided in Table 2.
Aneurysm characteristics recorded included aneurysm discovery, size, location, and the presence of parent vessel abnormalities (Table 3).
Intraoperative Treatments and Events
All procedures were conducted under general anesthesia with a volatile anesthetic, with (n = 23, 14%) or without nitrous oxide, supplemented with continuous IV infusions of remifentanil (n = 23, 14%) and/or propofol (n = 3, 2%) (Table 4). All patients were endotracheally intubated. In addition to standard noninvasive hemodynamic, respiratory, and neuromuscular blockade monitoring, all patients had continuous intraarterial blood pressure monitoring. There was no predetermined target range for intraoperative arterial blood pressure, but 69 of 170 (41%) patients received a continuous IV vasopressor (phenylephrine) infusion intraoperatively. No patient received a continuous IV vasodilator intraoperatively. Intraoperative neuroelectrophysiological monitoring was not used. Anesthesia records were reviewed for ASA physical status score as assigned by the anesthesia care team, the presence of sustained hypotension (mean arterial pressure <60 mm Hg for ≥5 minutes [no events]) or hypoxemia (pulse oximetry values <80% for ≥1 minute [no events]) or any other anesthesia-related problems.
All procedures were performed by one faculty neurosurgeon who is also a neurointerventionalist (DMH). Anesthesia duration was defined as the interval between anesthesia start and anesthesia stop. Operative notes and radiographic images were reviewed to characterize the number of aneurysms treated, endovascular techniques used, total number of coils deployed, and intraoperative events such as aneurysm rupture, vessel dissection or perforation, coil herniation or embolization, vasospasm, local or distal thrombosis, and embolism. More than one aneurysm was treated in 11 of 170 procedures. For statistical analysis of these 11 procedures, aneurysm size and location were referenced to the largest aneurysm treated. These data are summarized in Table 4.
Preoperative and intraoperative anticoagulation medications and greatest intraoperative activated clotting time were recorded. Patients who received intraoperative tirofiban (n = 52) were managed 2 ways. In the absence of recognized intraoperative thrombotic or embolic complications, intraoperative tirofiban infusions (0.1 μg/kg/min) were discontinued after an infusion duration of 2 hours. In patients who had recognized intraoperative thrombus formation, coil migration or herniation, in-stent thrombosis, or compromised arterial flow, intraoperative tirofiban infusions (with or without a bolus of 0.4 μg/kg/min for 30 minutes) were continued for a maximum of 24 hours after the procedure. Tirofiban infusions were discontinued before 24 hours if follow-up angiograms demonstrated thrombus resolution or if there was evidence of intracranial hemorrhage.
Data were recorded regarding initial postoperative care location (ICU versus PACU), blood pressure on arrival, and if the patient was intubated and/or receiving vasoactive medications on arrival. We recorded the occurrence of all adverse events within the first 96 hours after the procedure (neurologic, peripheral vascular, cardiovascular, respiratory, etc.) and the date and time of first diagnosis/recognition of the event. Neurologic adverse events were defined as any new motor or sensory sign or symptom, any change in level of consciousness, or any new symptom referable to the central nervous system (e.g., severe headache or dizziness). Also recorded were the occurrence of any new postoperative intracranial intervention, neurologic status at discharge (or death) compared with preoperative status, and level of care needed at discharge (or death) compared with preoperative status were recorded.
Initially, it was our institution’s practice for all patients to be admitted directly to the ICU. However, the incidence of direct ICU admission decreased over time (see Results). For analysis of ICU admission rates, we batched the data to create 6 sequential periods, each with approximately 7 patients admitted to the ICU. As shown in Table 5, the first period began from the start of data collection (July 20, 2009) through the 6th ICU admission (6 ICU admissions among 6 patients); the second period began with the next day that a patient underwent a procedure through the date of the 13th ICU admission (7 additional ICU admissions among the next 17 patients); the third period began with the next day that a patient underwent a procedure through the date of the 20th ICU admission (7 additional ICU admissions among the next 17 patients), and so forth through the last date of this case series (September 28, 2012). Among the 6 sequential periods, we analyzed the percentage of patients admitted to the ICU versus PACU by using the Cochrane-Armitage trend test. Because the periods were chosen based on the data, this approach makes the P value for the change during time in ICU admission potentially inaccurate. However, given that it is statistically implausible for trend to have been absent (i.e., P < 0.0001 by any method of analysis), this approach seemed reasonable. The important question was not whether there was an increase in the percentage of patients admitted to the PACU (that was obvious), but rather whether there were differences in rates of neurologic events among statistically independent periods. The incidences of perioperative neurologic events (present versus absent) were compared among the same 6 sequential periods, using the Cochrane-Armitage trend test.
To estimate effect sizes, the first 3 periods were compared with the last 3 periods. Relative risks of ICU admission and neurologic events were calculated, including exact 95% confidence intervals (CIs).
Correlations between the incidence of direct ICU admission and neurologic events were tested by using 2 different methods. The first method treated all events as Bernoulli (i.e., statistically independent). For each of the 6 sequential periods, there was a corresponding 2 × 2 table of counts and an odds ratio that ICU admission was associated with neurologic event. The odds ratio being significant within periods is unimportant because that merely confirms recognition of intraoperative procedural complications and admitting the patient to the ICU. The focus was on assessing heterogeneity of the odds ratio among periods because the incidence of ICU admission was markedly changed. If there were no change in the association between ICU admission and neurologic events among periods, the Breslow and Day test for the homogeneity of the odds ratios among periods would not be significant. A limitation of this first method is that the neurologic events within periods may be correlated because the decision for ICU admission was made by clinicians who likely knew of previous patients’ neurologic events. The second method treated the observed proportions within each of the 6 sequential periods as summary statistics. The correlation coefficient between the n = 6 ICU admission rates and neurologic event rates was calculated by using Kendall τ. This second method avoided the limitation of the first, but with the consequence of having wide CIs. Because it was (essentially) impossible for ICU admission decisions for all 60 patients in the 4th interval to be correlated to one another, the actual CIs likely are somewhere between these 2 extremes.
Consistent with there being (unmeasured) correlations among patients, a CI for the incidence of postoperative neurologic events was calculated by using the raw incidence from each period. After Freeman-Tukey transformation, the Student t distribution was applied to calculate the 95% 1-group CI in the transformed dimension.20 The inverse of the transform was then calculated as derived by Miller, using the harmonic mean sample size among periods.21
Associations between all recorded patient, aneurysm, and procedural characteristics and neurologic events (present versus absent) were tested by using the Fisher exact test for categorical variables and Wilcoxon-Mann-Whitney for continuous variables. Associations between preoperative variables likely to be known at the time of case scheduling and anesthesia duration were tested by using Kendall τ for continuous and ranked variables and Kruskall-Wallis test for categorical variables. Statistical software used included StatXact-9 (Cytel, Inc., Cambridge, MA). All P values are exact and 2 sided, without adjustments for multiple comparisons.
In this case series of 170 patients over n = 6 sequential periods, 16 (9%) patients experienced a perioperative neurologic event (95% CI, 3.1% to 14.7%). The incidence of ICU admission decreased progressively among periods (P < 0.0001; Fig. 1) but not the incidence of perioperative neurologic events (P = 0.79). The relative risk of ICU admission during the latter 3 periods versus the first 3 periods was 0.59 (95% CI, 0.40 to 0.78), whereas that for a neurologic event was 0.96 (95% CI, 0.80 to 1.07). The Kendall τ correlation between period and neurologic events was −0.20 (P = 0.72; asymptotic 95% CI, −0.99 to +0.71) (i.e., very wide, but negative trend).
The association among periods between direct ICU admission (39 of 170) and perioperative neurologic events (16 of 170) was first tested assuming statistical independence of successive ICU admissions and neurologic events within periods. The odds ratios among periods were not significantly heterogeneous (P = 0.097), indicating no change in the association between ICU admission and a neurologic event. The second analysis approach used rank correlation between each period’s point estimates of direct ICU admission and neurologic events (i.e., n = 6 pairwise numbers). There also was no significant correlation (P = 0.99), but logically a wide CI (Kendall τ = 0.00; 95% CI, −0.81 to +0.81).
Neurologic events were not associated with type of volatile anesthetic used (P = 0.18), use of nitrous oxide (P = 0.99), intraoperative vasopressor administration (P = 0.79), or anesthesia duration (P = 0.37). The factors associated with neurologic events were (1) preoperative systolic blood pressure (no neurologic event: 137 mm Hg [25th to 75th quartiles: 125–148 mm Hg] versus neurologic event 148 mm Hg [135–159 mm Hg], P = 0.0179) and (2) any intraoperative procedural complication (no neurologic event: 29 of 154 [19%] versus neurologic event: 8 of 16 [50%], P = 0.0084).
As summarized in Table 6, 8 of 131 PACU patients had neurologic events. Four events were first recognized while patients were still in the PACU: 0.0, 0.4, 1.0, and 1.0 hours after completion of the procedure. Of these 4 patients, 2 were immediately transferred to the ICU and 2 were transferred to the ICU after an intracranial intervention. Two additional PACU patients were transferred from the PACU to the ICU for nonneurologic causes (reintubation for respiratory failure and continued intubation for epistaxis with anticoagulation). Among the other 125 PACU patients (95%), discharge to the neurosurgical ward occurred after a median PACU stay of 1.3 hours (25th to 75th quartiles = 1.1–1.7 hours). In comparison, 8 of 39 ICU patients had neurologic events, 5 while patients were still in the ICU: 0.0, 0.0, 0.4, 1.0, and 1.0 hours after completion of the procedure.
Among patients with neurologic events (n = 16), symptoms resolved before discharge in 10 patients (90% within 24 hours). Neurologic symptoms did not resolve before discharge or the patient died for 3 of 8 patients initially admitted to the ICU versus 3 of 8 initially admitted to the PACU (Table 6). Four patients had new neurologic symptoms that did not fully resolve by their discharge on postoperative days 1, 5, 5, and 13. Two other patients died (on postoperative days 3 and 11) as the result of acute subarachnoid hemorrhage first recognized approximately 1 hour postoperatively; both patients were receiving tirofiban infusions.
As summarized in Table 5, mean anesthetic duration was greater in patients directly admitted to the ICU (P = 0.0060). With the use of a receiver operating curve (dependent variable ICU admission, independent variable anesthetic duration), the area under the curve (0.83) was not sufficient for decision making for individual cases on the day of surgery (see Discussion). Therefore, we considered the information known before the procedure started. Because the mean ± SD of anesthesia duration was 2.43 ± 0.76 hours (median, 2.3 hours; 25th to 75th quartiles, 1.9–2.7 hours), these procedures would be scheduled for 2 hours, 30 minutes (2:30). Although 2 preoperative variables were associated with anesthetic duration (ASA physical class and aneurysm dome size), neither association was large enough to be managerially useful (i.e., for preoperative case scheduling; Table 7; again, see Discussion).
This management case series considers the postoperative outcome of patients as our institution/hospital started to admit patients to the PACU rather than ICU after elective endovascular treatment of unruptured intracranial aneurysm. To our knowledge, only one other report describes the postoperative management of these patients (ICU versus PACU) and does not provide details regarding outcomes.9 The direct ICU admission rate in our case series (approximately 15% in the latter 3 periods) appears to be less than reported by Burrows et al. (35%).9
In our case series, of the 8 PACU patients (8 of 131 = 6%) who experienced new postoperative neurologic events, 4 (50%) were recognized while the patients were still in the PACU (<1 hour postoperatively) and were promptly treated. Four other PACU patients developed new neurologic findings after transfer from the PACU to the neurosurgical ward (4 of 125 = 3%), one of whom (1 of 125 ≤ 1%) was eventually transferred to the ICU. Our review provided no indication that the diagnosis of new postoperative neurologic events in patients who had first been admitted to the PACU was substantively delayed or that their outcome was adversely affected. In comparison, 3 of 8 (38%) neurologic events that occurred among patients admitted to the ICU had their onset after transfer from the ICU to the neurosurgical ward. Accordingly, even if every patient had been directly admitted to the ICU, it appears likely that a substantive percentage of patients would have experienced neurologic events after discharge from the ICU.
Of the patients who experienced neurologic events (ICU [n = 8] and PACU [n = 8]), symptoms did not resolve before discharge or resulted in death in 3 of 8 (38%) and 3 of 8 (38%), respectively. With these estimates, we can design a randomized trial of patients admitted to the ICU or PACU to detect a large (50%) relative difference in failure to recover from neurologic deficits between groups (e.g., 38% in PACU patients versus 19% in ICU patients). A study population of 218 patients with postoperative neurologic events would be required to demonstrate the difference (α = 0.05, β = 0.20). If 8.5% of all patients who undergo elective endovascular treatment of unruptured aneurysms experience any type of postoperative neurologic event, then a total study population of 2580 patients undergoing these procedures would be required to establish that neurologic outcomes are improved with routine ICU admission. This estimate presupposes that all neurologic events would have their onset in the ICU, which our observations suggest would not occur without prolonged ICU stays.
Our management case series shows that, in the absence of such a randomized trial, among patients without intraoperative events with the potential for ongoing cerebral ischemia, admission to the PACU is reasonable. This conclusion is compatible with previous studies of ICU needs of patients undergoing elective craniotomy.4–6 However, unqualified generalization of these observations to every patient at every medical center would be imprudent. Our conclusion depends on the care provided by PACU and neurosurgery ward nursing personnel, who are familiar with these patients and who frequently assess patient neurologic status. Likewise, this conclusion likely depends on a sufficient period of postoperative PACU monitoring before discharge to the neurosurgical ward and continued patient assessment on the neurosurgery ward. In our series, median PACU stay of patients transferred to the neurosurgical ward (n = 125) was 1.3 hours. Medical centers that have PACU policies and procedures that would not allow for postprocedure patient monitoring at similar levels or durations might consider routine ICU admission to be advisable.
In this case series, intraoperative procedural complications were associated with postoperative neurologic events, increasing approximately 3- to 4-fold. This is consistent with previous reports that indicate neurologic complications associated with endovascular treatment of cerebral aneurysms are the consequence of direct cerebrovascular injuries (e.g., rupture, perforation, dissection),8,15,22,23 and/or complications from coils (e.g., protrusion, embolization, thrombosis)7,15,22–24 or stents (e.g., migration, thrombosis),8,15,25 although fewer than half of such events result in symptomatic neurologic deficits.7,8,15,22–25
We also observed an association between preoperative systolic hypertension and postoperative neurologic events. Because of the multiple statistical comparisons made in this study, it is possible that this apparent association is a type I error. In addition, our case series cannot determine whether systolic hypertension contributed to more procedural complications or, alternatively, to a decreased capacity for neurologic recovery when ischemic events occur. However, hypertension causes widespread changes in cerebrovascular structure and function that impair postischemic compensatory mechanisms and may lead to less favorable outcomes.26 Although our routine postoperative threshold value for systolic blood pressure that required notification of a physician was 90 mm Hg, greater values might be advisable, particularly in patients with preoperative hypertension.
The findings of our study provide additional insight managerially. We plan our anesthesia (OR) time months in advance using a forecast of the total workload. Cases are then scheduled by using durations that are an unbiased estimate of the contribution of each case to the total workload (i.e., the mean of historical data).27–29 In our case series, mean case duration was 2.43 hours, with a mean absolute percentage error of 22.1% ± 1.7% (SE). This error is indistinguishable from that reported from a previous series of patients undergoing neurointerventional procedures in our institution, but with a different interventionalist (24% ± 1% SE).30 This suggests that our anesthetic duration data likely are generalizable. In our case series, the actual anesthetic duration was not meaningfully associated with information available at the time of scheduling (see Table 7). Therefore, our case series confirms that the scheduling of elective endovascular treatment of unruptured intracranial aneurysms based on the pooled mean anesthetic duration is logical and practical. The anesthetic duration was greater in patients directly admitted to the ICU versus those admitted to the PACU. However, the difference between ICU patients and PACU patients in anesthetic duration was insufficient to determine whether, on the day of surgery, a patient who had anesthetic duration exceeding the historical average would be reliably predicted to be admitted to the ICU.31,32 Decisions to admit patients directly to the ICU after these procedures, in general, were made on the basis of intraoperative events not readily predicted preoperatively (e.g., intraoperative coagulation management and procedural complications).
The primary limitation of this case series is that it is a retrospective review of procedures from a single center and a single surgeon and may not be generalizable or reproducible in other centers.33 It is a management case series. Nevertheless, our patients’ characteristics (age, sex, and Charlson comorbidity score) and in-hospital mortality (2 of 170 = 1.2%; 95% CI, 0.1%–4.2%) were similar to those reported in the Nationwide Inpatient Sample of patients undergoing endovascular treatment of unruptured cerebral aneurysms between 1998 and 2007 (mortality = 1.2%).3 Likewise, presenting symptoms7,22,34 and aneurysm locations7,25,34,35 are similar to those reported by other centers performing endovascular procedures to treat unruptured intracranial aneurysms. In our case series, median aneurysm dome size (10 mm) was greater than that reported from many other centers.7,22,25,34,35
Because endovascular technologies are rapidly evolving, several techniques are currently in use to treat cerebral aneurysms: (1) coiling alone; (2) balloon-assisted coiling; (3) stent-assisted coiling; and (4) stent-based flow diversion or disruption.36 Each technique has advantages and disadvantages, and the superiority of one technique over another is not certain.36 Nevertheless, outcomes and complications may differ among techniques, and the advisability of routine postoperative ICU care may differ in kind. In our case series, most procedures (139 of 170 = 82%) consisted of stent-assisted coiling. Outcomes in our series are comparable to those reported in 2 recent large reviews. In 2010, Naggara et al.34 analyzed 71 reports of endovascular treatment of unruptured aneurysms. The overall rates of intraoperative thromboembolism (7.6%), permanent neurologic injury and/or death (4.3%), and mortality (1.2%) reported by Naggara et al.34 are indistinguishable from those reported in our case series: 8 of 170 (4.7%), 6 of 170 (3.5%), and 2 of 170 (1.2%), respectively. In 2012, Shapiro et al.37 analyzed 39 reports specific to stent-assisted aneurysm coiling, largely (78%) taking place in patients with unruptured aneurysms. The overall rates of intraoperative thromboembolism (10%) and mortality (2.1%) reported in this study are also comparable to those in our case series.
Based on our case series, in the absence of intraoperative events with the potential for ongoing cerebral ischemia (e.g., aneurysm rupture, thrombosis, coil migration, persistent coil herniation, or cerebrovascular dissection or occlusion), we suggest that the great majority of patients undergoing elective endovascular treatment of unruptured intracranial aneurysms can be appropriately managed postoperatively without direct ICU admission. This conclusion is based on the patient-to-nurse ratios and monitoring practices of our institution as summarized in Table 1 (PACU [1:2 nurse-to-patient ratio]; neurosurgical ward [1:3 nurse-to-patient ratio]; and ICU [1:1 nurse-to-patient ratio]). These procedures should be scheduled based on historical averages of anesthesia duration. On the day of surgery, prolonged case duration was not helpful in determining whether an individual patient is more likely to be admitted to the ICU.
Dr. Franklin Dexter is the Statistical Editor and Section Editor for Economics, Education, and Policy for Anesthesia & Analgesia. This manuscript was handled by Dr. Gregory J. Crosby, Section Editor for Neuroscience in Anesthesiology and Perioperative Medicine and Pediatric Neuroscience for the Journal, and Dr. Dexter was not involved in any way with the editorial process or decision.
Name: Sarah H. Eisen, MD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Sarah H. Eisen has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Bradley J. Hindman, MD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Bradley J. Hindman has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Emine O. Bayman, PhD.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Emine O. Bayman has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Franklin Dexter, MD, PhD.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Franklin Dexter has analyzed the data and approved the final manuscript.
Name: David M. Hasan, MD.
Contribution: This author helped design the study, conduct the study, and write the manuscript.
Attestation: David M. Hasan has seen the original study data and approved the final manuscript.
a Available at: http://www.strobe-statement.org/?id=available-checklists. Accessed November 7, 2013.
b The reviewer (SHE) determined whether diagnostic criteria for alcohol dependence were likely satisfied preoperatively based upon the information available in the medical record with dates preceding the procedure.
1. Alshekhlee A, Mehta S, Edgell RC, Vora N, Feen E, Mohammadi A, Kale SP, Cruz-Flores S. Hospital mortality and complications of electively clipped or coiled unruptured intracranial aneurysm. Stroke. 2010;41:1471–6
2. Brinjikji W, Rabinstein AA, Nasr DM, Lanzino G, Kallmes DF, Cloft HJ. Better outcomes with treatment by coiling relative to clipping of unruptured intracranial aneurysms in the United States, 2001–2008. Am J Neuroradiol. 2011;32:1071–5
3. Lin N, Cahill KS, Frerichs KU, Friedlander RM, Claus EB. Treatment of ruptured and unruptured cerebral aneurysms in the USA: a paradigm shift. J Neurointerv Surg. 2012;4:182–9
4. Beauregard CL, Friedman WA. Routine use of postoperative ICU care for elective craniotomy: a cost-benefit analysis. Surg Neurol. 2003;60:483–9
5. Bui JQ, Mendis RL, van Gelder JM, Sheridan MM, Wright KM, Jaeger M. Is postoperative intensive care unit admission a prerequisite for elective craniotomy? J Neurosurg. 2011;115:1236–41
6. Rhondali O, Genty C, Halle C, Gardellin M, Ollinet C, Oddoux M, Carcey J, Francony G, Fauvage B, Gay E, Bosson JL, Payen JF. Do patients still require admission to an intensive care unit after elective craniotomy for brain surgery? J Neurosurg Anesthesiol. 2011;23:118–23
7. Khosla A, Brinjikji W, Cloft H, Lanzino G, Kallmes DF. Age-related complications following endovascular treatment of unruptured intracranial aneurysms. Am J Neuroradiol. 2012;33:953–7
8. Lee SJ, Cho YD, Kang HS, Kim JE, Han MH. Coil embolization using the self-expandable closed-cell stent for intracranial saccular aneurysm: a single-center experience of 289 consecutive aneurysms. Clin Radiol. 2013;68:256–63
9. Burrows AM, Rabinstein AA, Cloft HJ, Kallmes DF, Lanzino G. Are routine intensive care admissions needed after endovascular treatment of unruptured aneurysms? Am J Neuroradiol. 2013;34:2199–201
10. van Rooij WJ, Sluzewski M. Procedural morbidity and mortality of elective coil treatment of unruptured intracranial aneurysms. Am J Neuroradiol. 2006;27:1678–80
11. Gizewski ER, Göricke S, Wolf A, Schoch B, Stolke D, Forsting M, Wanke I. Endovascular treatment of intracranial aneurysms in patients 65 years or older: clinical outcomes. Am J Neuroradiol. 2008;29:1575–80
12. Mocco J, Snyder KV, Albuquerque FC, Bendok BR, Alan S B, Carpenter JS, Fiorella DJ, Hoh BL, Howington JU, Jankowitz BT, Liebman KM, Rai AT, Rodriguez-Mercado R, Siddiqui AH, Veznedaroglu E, Hopkins LN, Levy EI. Treatment of intracranial aneurysms with the Enterprise stent: a multicenter registry. J Neurosurg. 2009;110:35–9
13. Piotin M, Blanc R, Spelle L, Mounayer C, Piantino R, Schmidt PJ, Moret J. Stent-assisted coiling of intracranial aneurysms: clinical and angiographic results in 216 consecutive aneurysms. Stroke. 2010;41:110–5
14. Hwang SK, Hwang G, Oh CW, Jin SC, Park H, Bang JS, Kwon OK. Endovascular treatment for unruptured intracranial aneurysms in elderly patients: single-center report. Am J Neuroradiol. 2011;32:1087–90
15. Gao X, Liang G, Li Z, Wei X, Hong Q. Complications and adverse events associated with Neuroform stent-assisted coiling of wide-neck intracranial aneurysms. Neurol Res. 2011;33:841–52
16. American Psychiatric Association.. Substance-related disorders. In: Diagnostic and Statistical Manual of Mental Disorders. 20004th ed., Text Revision Washington, DC American Psychiatric Association:197–214
17. Rankin J. Cerebral vascular accidents in patients over the age of 60. II. Prognosis. Scott Med J. 1957;2:200–15
18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83
19. Reponen E, Tuominen H, Korja M. Evidence for the use of preoperative risk assessment scores in elective cranial neurosurgery: a systematic review of the literature. Anesth Analg. 2014;119:420–32
20. Dexter F, Marcon E, Epstein RH, Ledolter J. Validation of statistical methods to compare cancellation rates on the day of surgery. Anesth Analg. 2005;101:465–73 erratum 2002:114:693
21. Miller JJ. The inverse of the Freeman-Tukey double arcsine transformation. Am Stat. 1978;32:138
22. Gallas S, Drouineau J, Gabrillargues J, Pasco A, Cognard C, Pierot L, Herbreteau D. Feasibility, procedural morbidity and mortality, and long-term follow-up of endovascular treatment of 321 unruptured aneurysms. Am J Neuroradiol. 2008;29:63–8
23. Im SH, Han MH, Kwon OK, Kwon BJ, Kim SH, Kim JE, Oh CW. Endovascular coil embolization of 435 small asymptomatic unruptured intracranial aneurysms: procedural morbidity and patient outcome. Am J Neuroradiol. 2009;30:79–84
24. Hwang G, Jung C, Park SQ, Kang HS, Lee SH, Oh CW, Chung YS, Han MH, Kwon OK. Thromboembolic complications of elective coil embolization of unruptured aneurysms: the effect of oral antiplatelet preparation on periprocedural thromboembolic complication. Neurosurgery. 2010;67:743–8
25. Gentric JC, Biondi A, Piotin M, Mounayer C, Lobotesis K, Bonafé A, Costalat VFrench SENAT Investigators. . Safety and efficacy of neuroform for treatment of intracranial aneurysms: a prospective, consecutive, French multicentric study. Am J Neuroradiol. 2013;34:1203–8
26. Pires PW, Dams Ramos CM, Matin N, Dorrance AM. The effects of hypertension on the cerebral circulation. Am J Physiol Heart Circ Physiol. 2013;304:H1598–614
27. Dexter F, Macario A, Ledolter J. Identification of systematic underestimation (bias) of case durations during case scheduling would not markedly reduce overutilized operating room time. J Clin Anesth. 2007;19:198–203
28. Dexter F, Dexter EU, Ledolter J. Influence of procedure classification on process variability and parameter uncertainty of surgical case durations. Anesth Analg. 2010;110:1155–63
29. Dexter F, Epstein RH, Bayman EO, Ledolter J. Estimating surgical case durations and making comparisons among facilities: identifying facilities with lower anesthesia professional fees. Anesth Analg. 2013;116:1103–15
30. Dexter F, Yue JC, Dow AJ. Predicting anesthesia times for diagnostic and interventional radiological procedures. Anesth Analg. 2006;102:1491–500
31. Tiwari V, Dexter F, Rothman BS, Ehrenfeld JM, Epstein RH. Explanation for the near-constant mean time remaining in surgical cases exceeding their estimated duration, necessary for appropriate display on electronic white boards. Anesth Analg. 2013;117:487–93
32. Dexter F, Epstein RH, Lee JD, Ledolter J. Automatic updating of times remaining in surgical cases using bayesian analysis of historical case duration data and “instant messaging” updates from anesthesia providers. Anesth Analg. 2009;108:929–40
33. Shafer SL, Dexter F. Publication bias, retrospective bias, and reproducibility of significant results in observational studies. Anesth Analg. 2012;114:931–2
34. Naggara ON, White PM, Guilbert F, Roy D, Weill A, Raymond J. Endovascular treatment of intracranial unruptured aneurysms: systematic review and meta-analysis of the literature on safety and efficacy. Radiology. 2010;256:887–97
35. Oishi H, Yamamoto M, Shimizu T, Yoshida K, Arai H. Endovascular therapy of 500 small asymptomatic unruptured intracranial aneurysms. Am J Neuroradiol. 2012;33:958–64
36. Pierot L, Wakhloo AK. Endovascular treatment of intracranial aneurysms: current status. Stroke. 2013;44:2046–54
37. Shapiro M, Becske T, Sahlein D, Babb J, Nelson PK. Stent-supported aneurysm coiling: a literature survey of treatment and follow-up. Am J Neuroradiol. 2012;33:159–63