CLINICAL ARTICLE

J Neurosurg 128:756–767, 2018

Reoperation and readmission after clipping of an

unruptured intracranial aneurysm: a National Surgical
Quality Improvement Program analysis
Hormuzdiyar H. Dasenbrock, MD, Timothy R. Smith, MD, PhD, MPH, Robert F. Rudy, BS,
William B. Gormley, MD, MPH, MBA, M. Ali Aziz-Sultan, MD, and Rose Du, MD, PhD
Department of Neurosurgery, Cushing Neurosurgical Outcomes Center, Brigham and Women’s Hospital, Harvard Medical
School, Boston, Massachusetts

OBJECTIVE  Although reoperation and readmission have been used as quality metrics, there are limited data evaluating
the rate of, reasons for, and predictors of reoperation and readmission after microsurgical clipping of unruptured aneu-
rysms.
METHODS  Adult patients who underwent craniotomy for clipping of an unruptured aneurysm electively were extracted
from the prospective National Surgical Quality Improvement Program registry (2011–2014). Multivariable logistic re-
gression and recursive partitioning analysis evaluated the independent predictors of nonroutine hospital discharge,
unplanned 30-day reoperation, and readmission. Predictors screened included patient age, sex, comorbidities, American
Society of Anesthesiologists (ASA) classification, functional status, aneurysm location, preoperative laboratory values,
operative time, and postoperative complications.
RESULTS  Among the 460 patients evaluated, 4.2% underwent any reoperation at a median of 7 days (interquartile
range [IQR] 2–17 days) postoperatively, and 1.1% required a cranial reoperation. The most common reoperation was
ventricular shunt placement (23.5%); other reoperations were tracheostomy, craniotomy for hematoma evacuation, and
decompressive hemicraniectomy. Independent predictors of any unplanned reoperation were age greater than 51 years
and longer operative time (p ≤ 0.04). Readmission occurred in 6.3% of patients at a median of 6 days (IQR 5–13 days)
after discharge from the surgical hospitalization; 59.1% of patients were readmitted within 1 week and 86.4% within
2 weeks of discharge. The most common reason for readmission was seizure (26.7%); other causes of readmission
included hydrocephalus, cerebrovascular accidents, and headache. Unplanned readmission was independently associ-
ated with age greater than 65 years, Class II or III obesity (body mass index > 35 kg/m2), preoperative hyponatremia, and
preoperative anemia (p ≤ 0.04). Readmission was not associated with operative time, complications during the surgical
hospitalization, length of stay, or discharge disposition. Recursive partitioning analysis identified the same 4 variables,
as well as ASA classification, as associated with unplanned readmission. The most potent predictors of nonroutine
hospital discharge (16.7%) were postoperative neurological and cardiopulmonary complications; other predictors were
age greater than 51 years, preoperative hyponatremia, African American and Asian race, and a complex vertebrobasilar
circulation aneurysm.
CONCLUSIONS  In this national analysis, patient age greater than 65 years, Class II or III obesity, preoperative hypona-
tremia, and anemia were associated with adverse events, highlighting patients who may be at risk for complications after
clipping of unruptured cerebral aneurysms. The preponderance of early readmissions highlights the importance of early
surveillance and follow-up after discharge; the frequency of readmission for seizure emphasizes the need for additional
data evaluating the utility and duration of postcraniotomy seizure prophylaxis. Moreover, readmission was primarily as-
sociated with preoperative characteristics rather than metrics of perioperative care, suggesting that readmission may be
a suboptimal indicator of the quality of care received during the surgical hospitalization in this patient population.
https://thejns.org/doi/abs/10.3171/2016.10.JNS161810
KEY WORDS  cerebral aneurysm; craniotomy; National Surgical Quality Improvement Program; NSQIP; readmission;
reoperation; vascular disorders

ABBREVIATIONS  ACS = American College of Surgeons; ASA = American Society of Anesthesiologists; BMI = body mass index; CI = confidence interval; COPD = chronic
obstructive pulmonary disease; CPT = Current Procedural Terminology; IQR = interquartile range; NSQIP = National Surgical Quality Improvement Program; OR = odds
ratio; ROC = receiver operating characteristic; SSI = surgical site infection; TIA = transient ischemic attack; UTI = urinary tract infection; WBC = white blood cell.
SUBMITTED  July 22, 2016.  ACCEPTED  October 31, 2016.
INCLUDE WHEN CITING  Published online April 7, 2017; DOI: 10.3171/2016.10.JNS161810.

756 J Neurosurg  Volume 128 • March 2018 ©AANS 2018, except where prohibited by US copyright law

A
lthough unplanned 30-day readmission and reop-
eration have been increasingly used as metrics to
assess the quality of care physicians and hospitals
provide, their utility as quality indicators—particularly in
surgical patients—is contentious.28 Hospitals now incur
penalties for high readmission rates in specified circum-
stances, including pneumonia.16,28 Medical patients often
require readmission for exacerbation of the same condi-
tion treated during their initial hospital stay (known as
the index hospitalization), an argument used to support
the use of readmission to grade hospitals.21 However, the
degree to which readmission is an appropriate quality in-
dicator in surgical patients is debated.28,48
Prior studies evaluating readmission after neurosurgi-
cal intervention have arrived at divergent conclusions re-
garding its utility as a quality metric.13,39,53 Some authors,
particularly those evaluating patients treated for ventricu-
lar shunt revision, have argued that readmission and reop-
eration are preventable and thus appropriate quality met-
rics.9,17,47 However, others have suggested that readmission
typically occurs in the absence of a breach of appropriate
care, and therefore may be a suboptimal metric.13,19,33
The National Surgical Quality Improvement Program
(NSQIP) is a large nationwide prospective surgical regis-
try. However, few studies have used the NSQIP to evaluate
patients undergoing treatment of unruptured intracranial
aneurysms.41 The goal of this study was to use the NSQIP
to 1) evaluate the rate of, reasons for, and predictors of
readmission and reoperation after elective clipping of un-
ruptured cerebral aneurysms; and 2) compare the predic-
tors of these metrics with nonroutine hospital discharge, to
identify patients who may be at higher risk of postopera-
tive adverse events.
Methods
Data Source
Data were extracted from the prospective NSQIP reg-
istry from 2011 to 2014. The NSQIP is a multiinstitutional
program where trained surgical reviewers use a uniform
protocol to collect data at more than 400 academic and
private hospitals. Enrolled patients are followed longitu-
dinally for 30 days, and all complications are recorded
whether they occurred during the surgical hospitalization
or after discharge. The American College of Surgeons
(ACS) routinely audits data from contributing institutions
and prior publications have validated the accuracy of data
from the NSQIP.44 This data source has been previously
used to evaluate patients undergoing neurosurgical inter-
vention,1–3,5–8,11,12,20,23,24,27,35,37,38,41–43 including craniotomy
for clipping of cerebral aneurysms.41 The 1st year included
was 2011 as this is when the NSQIP began collecting data
on readmissions. Individual studies using the NSQIP have
been designated as exempt from review by our institution-
al review board.
Inclusion Criteria
To identify those who underwent elective clipping of an
unruptured intracranial aneurysm, patients were included
who met the following criteria: 1) age ≥ 18 years; 2) an
ICD-9-CM diagnosis code of an unruptured cerebral an-
Readmission and unruptured aneurysms
eurysm (437.3); 3) underwent microsurgical clipping of a
simple (61700) or complex (61697, 61703, or 61705) ca-
rotid or simple (61702) or complex (61698) vertebrobasi-
lar circulation aneurysm, as indicated by the documented
Current Procedural Terminology (CPT) code; and 4) the
recorded attending surgeon’s specialty was neurosurgery.
The NSQIP does not systematically collect data on pa-
tients undergoing endovascular intracranial interventions.
Patients were excluded if 1) the recorded admission diag-
nosis was subarachnoid hemorrhage (430) or intracranial
hemorrhage (431, 432.x); 2) the hospital admission was
designated as nonelective; or 3) the documented case ur-
gency was emergent.
Patient Stratification
Patients were stratified by reoperation and unplanned
readmission, both of which are collected by the NSQIP
regardless of whether it occurred at the same hospital as
the initial surgery or at a different institution. External
validation of the NSQIP’s collection of rates and reasons
for readmission has reported concordance (kappa) val-
ues of 0.98 and 0.75, respectively, compared with physi-
cian review.44 Unplanned readmission was only evaluated
among patients who were discharged from the surgical
hospitalization alive, and therefore at risk for readmission.
In 2012, the NSQIP began collecting data on the reasons
for reoperation and readmission. Reasons for reoperation
were encoded by the operation performed (using CPT
codes); indications for readmission are collected by diag-
nosis (some categories including infections are recorded
directly, while any other diagnoses are recorded through
ICD-9-CM codes).
To compare risk factors for readmission and reopera-
tion with other adverse events, hospital discharge disposi-
tion was evaluated. A nonroutine hospital discharge was
defined as any other than to home. Prior research has sug-
gested that discharge disposition is correlated with func-
tional neurological deficits (including modified Rankin
Scale score) at hospital discharge in patients with cerebral
aneurysms.50
Predictor Variables
Pertinent predictor variables collected by the NSQIP
were extracted, including age, sex, race or ethnicity,
American Society of Anesthesiologists (ASA) physical
classification designation, and baseline functional status.
Additionally, patients were stratified by aneurysm type:
carotid versus vertebrobasilar circulation location, and
simple versus complex operation.
Comorbidities collected by the NSQIP and present in
at least 20 patients were evaluated individually: smoking
(within the past year), alcohol abuse (at least 2 drinks per
day), prior transient ischemic attack (TIA), preoperative
neurological deficits, hypertension requiring medication,
cardiac disease (recent angina, myocardial infarction, or
congestive heart failure exacerbation), pulmonary dis-
ease (chronic obstructive pulmonary disease [COPD] or
dyspnea), diabetes, or a bleeding disorder (vitamin K de-
ficiency, hemophilia, thrombocytopenia, and long-term
anticoagulant use). The designation of a bleeding disor-
J Neurosurg  Volume 128 • March 2018 757
H. H. Dasenbrock et al.
der did not include aspirin use. Patient height and weight
were used to calculate body mass index (BMI) and body
habitus was classified as nonobese, Class I obese (BMI
30–34.99 kg/m2), Class II or III obese (BMI > 35 kg/m2),
or missing.10
Pertinent preoperative laboratory values available were
extracted and evaluated categorically to appropriately ac-
count for missing data. These values and their stratifica-
tion were sodium (by 135 meq/L), creatinine (by 1.4 mg/
dl), albumin (by 3.5 g/dl), white blood cell [WBC] count
(by 12,000 ml), hematocrit (by 36%) and platelet count (by
150,000/ ml). Operative time was evaluated categorically
as < 240, 240–330, and > 330 minutes, as these are the
30-minute intervals closest to the median and upper-quar-
tile of the entire population.
Postoperative complications during the surgical hospi-
talization, length of hospital stay, and discharge disposition
were also evaluated as potential predictors of readmission.
Complications that occurred during the surgical hospital-
ization were determined based on the complication tim-
ing and length of hospital stay. Postoperative cerebrovas-
cular accidents are defined by the NSQIP as an embolic,
thrombotic, or hemorrhagic accident with motor, sensory,
or cognitive dysfunction that persists for at least 24 hours;
data are also collected on new postoperative coma. Medi-
cal complications extracted were cardiopulmonary (cardi-
ac arrest, acute myocardial infarction, extubation failure,
and mechanical ventilation for greater than 48 hours); ve-
nous thromboembolic (symptomatic pulmonary embolus
or deep venous thrombosis); and infectious (surgical site
infection [SSI], pneumonia, urinary tract infection [UTI],
and sepsis). Length of hospital stay was evaluated both
continuously and dichotomously as the proportion of pa-
tients with a hospitalization longer than the upper quartile
of the entire population.4
Missing Data
The NSQIP explicitly denotes if data are missing on
any given parameter. Data were missing on outcomes in a
small proportion of patients—reoperation in 1.3% (n = 6)
and readmission in 2.4% (n = 11)—and these patients were
excluded from such analyses. Some covariates had miss-
ing data in a proportion of patients; for these variables,
patients with missing data were categorized into a dif-
ferent stratum, to compare patients with known presence
and absence of that parameter in multivariable regression,
without decreasing the size of the study population.
Statistical Analysis
Descriptive statistics and regression were performed
in STATA (version 13, STATACorp.). Preoperative and
operative variables were screened for association with re-
admission, reoperation, or nonroutine hospital discharge
using univariable logistic regression. Variables with a p
value less than 0.20 in any strata were thereafter entered
into a multivariable model, and the final model was con-
structed using stepwise backward elimination. The receiv-
er operating characteristic (ROC) assessed the discrimi-
natory capacity of models, where a model with a value
of 0.5 has no greater discrimination than chance alone,
758 J Neurosurg  Volume 128 • March 2018
and a value of 1.0 leads to complete separation of cases.
The Hosmer-Lemeshow test evaluated the calibration of
regression models, which tests the null hypothesis that the
observed and expected proportion in each decile is equal,
and a model that fails to reject the null hypothesis (value
> 0.05) is accepted. A p value less than 0.05 was defined
as significant.
Additionally, recursive partitioning analysis was used
as a complementary form of model building using JMP
Pro (version 12; SAS): this algorithm creates decision trees
to identify cases by splitting the population into subpopu-
lations based on critical values of independent variables.
Some advantages of recursive partitioning analysis com-
pared with regression analysis include superior evaluation
of associations that are not linear or monotonic. Classifi-
cation trees were constructed using a randomly selected
70% of the population as the training set and 30% as the
validation set. The predicted probability of each terminal
leaf, the ROC in the training and validation sets, and the
column contributions from included variables were re-
corded.
Results
Study Demographics
A total of 460 patients who underwent elective clip-
ping of an unruptured intracranial aneurysm were in-
cluded, and their baseline demographics are compared
by unplanned readmission or reoperation in Table 1. All
patients were admitted directly from home, and no patient
was transferred from another hospital or admitted from
the emergency department. Only 1 patient required me-
chanical ventilation preoperatively, and that patient did
not undergo a subsequent reoperation. Length of stay was
not associated with unplanned hospital readmission when
evaluated as a continuous variable (p = 0.44) or dichoto-
mously as the proportion of patients with a hospitalization
longer than the upper quartile of the total study population
(p = 0.53).
Outcomes
The in-hospital mortality rate was 1.1% (n = 5), the me-
dian length of stay was 4 days (interquartile range [IQR]
3–6 days), and the hospital discharge disposition was
nonroutine in 16.7% (n = 77). Data on reoperation were
available for 454 patients: the total 30-day reoperation rate
was 4.2% (n = 19) and the cranial reoperation rate (ex-
cluding CSF diversion procedures) was 1.1% (n = 5). The
median time to a cranial reoperation was 1 day (IQR 1–2
days) and to any reoperation was 7 days (IQR 2–17 days)
postoperatively (Fig. 1D). Additionally, data on unplanned
hospital readmission were available for 444 patients, and
the 30-day readmission rate was 6.3% (n = 28). The me-
dian time from discharge from the surgical hospitaliza-
tion to readmission was 6 days (IQR 5–13 days); 59.1% of
readmissions occurred within 1 week and 86.4% within 2
weeks of discharge. The median time to readmission from
aneurysm clipping was 11 days (IQR 9–17 days; Fig. 1C).
Reasons for Reoperation and Readmission
The procedure code associated with the reoperation
 Readmission and unruptured aneurysms

TABLE 1. Demographics of the study population, stratified by an unplanned readmission or reoperation

Readmission Reoperation
Variable Total* Yes No p Value Yes No p Value
No. of patients 444.0 28 416 19 435.0
Age (yrs)
 18–50 27.3 17.9 27.9 Ref 5.3 27.8 Ref
 51–65 47.8 39.3 48.3 0.67 36.8 48.5 0.20
 >65 25.0 42.9 23.8 0.06 57.9 23.7 0.02
Female sex 73.0 82.1 72.4 0.27 57.9 73.8 0.13
Race or ethnicity
  White, not Hispanic 63.1 64.3 63.0 Ref 68.4 63.0 Ref
  African American 10.6 17.9 10.1 0.30 21.1 10.1 0.27
 Hispanic 6.3 7.1 6.3 0.88 0 6.4 —
 Asian 5.0 3.6 5.1 0.73 5.3 5.1 0.97
 Unknown 15.1 7.1 15.6 0.29 5.3 15.4 0.27
Smoking 40.8 25.0 41.8 0.09 42.1 40.9 0.92
Alcohol use
 No 32.9 32.1 32.9 Ref 26.3 33.6 Ref
 Yes 1.4 0 1.4 — 5.3 1.2 0.14
 Missing 65.8 67.9 65.6 0.89 68.4 65.3 0.59
Prior TIA
 No 28.6 28.6 28.6 Ref 21.1 29.2 Ref
 Yes 5.6 3.6 5.8 0.66 10.5 5.5 0.28
 Missing 65.8 67.9 65.6 0.94 68.4 65.3 0.52
Neurological deficits
 No 28.6 28.6 28.6 Ref 21.1 29.2 Ref
 Yes 5.6 3.6 5.8 0.66 10.5 5.5 0.28
 Missing 65.8 67.9 65.6 0.94 68.4 68.3 0.94
Hypertension 57.9 82.1 56.3 0.01 68.4 57.5 0.35
Cardiac disease
 No 32.2 32.1 32.2 Ref 31.6 32.2 Ref
 Yes 2.7 0 2.9 — 0 3.2 —
 Missing 65.1 67.9 64.9 0.91 68.4 64.6 0.88
Pulmonary disease 14.2 14.3 14.2 0.99 15.8 14.3 0.85
Bleeding disorder 1.8 7.1 1.4 0.05 5.3 1.6 0.26
Diabetes mellitus 11.7 17.9 11.3 0.30 10.5 11.5 0.90
Body habitus
 Nonobese 62.6 50.0 63.5 Ref 52.6 62.8 Ref
 Obese 21.2 17.9 21.4 0.91 15.8 21.6 0.84
  Morbidly obese 15.5 32.1 14.4 0.02 31.6 14.9 0.08
 Missing 0.7 0 0.7 — 0 0.69 —
ASA Classification
 I–II 30.4 25.0 30.8 Ref 15.8 30.6 Ref
 III 59.9 67.9 59.4 0.45 73.7 59.3 0.17
  IV & V 9.7 7.1 9.9 0.89 10.5 10.1 0.45
Dependent functional status 1.8 3.6 1.7 0.48 10.5 1.8 0.03
Operative type
  Carotid simple 39.2 50.0 38.5 Ref 21.1 39.5 Ref
  Carotid complex 50.5 42.9 51.0 0.29 63.2 50.1 0.14
  VB simple 2.3 0 2.4 — 0 2.3 —
  VB complex 8.1 7.1 8.2 0.61 15.8 8.1 0.10
Preop sodium (mEq/L)
  ≥135 90.3 85.7 90.6 Ref 94.7 90.3 Ref
 <135 2.9 10.7 2.4 0.03 5.3 2.8 0.58
 Missing 6.8 3.6 7.0 0.56 0 6.9 —
CONTINUED ON PAGE 760 »

J Neurosurg  Volume 128 • March 2018 759

H. H. Dasenbrock et al.

» CONTINUED FROM PAGE 759

TABLE 1. Demographics of the study population, stratified by an unplanned readmission or reoperation
Readmission Reoperation
Variable Total* Yes No p Value Yes No p Value
Preop creatinine (mg/dl)
 <1.4 90.5 85.7 90.9 Ref 84.2 91.0 Ref
  ≥1.4 5.2 10.7 4.8 0.19 15.8 4.6 0.05
 Missing 4.3 3.6 4.3 0.90 0 4.4 —
Preop albumin (g/dl)
  ≥3.5 29.7 32.1 29.6 Ref 36.8 29.4 Ref
 <3.5 4.1 7.1 3.9 0.52 5.3 3.9 0.95
 Missing 66.2 60.7 66.6 0.68 57.9 66.7 0.46
Preop WBC (/µl)
  ≤12,000 88.5 82.1 88.9 Ref 100.0 88.1 Ref
 >12,000 4.3 10.7 3.9 0.10 0 4.6 —
 Missing 7.2 7.1 7.2 0.90 0 7.4 —
Preop hematocrit (%)
  ≥36 81.3 67.9 82.2 Ref 79.0 81.2 Ref
 <36 11.5 25.0 10.6 0.03 21.1 11.5 0.28
 Missing 7.2 7.1 7.2 0.80 0 7.4 —
Preop platelets (/µl)
 >150,000 89.9 92.9 89.7 Ref 100.0 89.7 Ref
  ≤150,000 3.4 0 3.6 — 0 3.5 —
 Missing 6.8 7.1 6.7 0.97 0 6.9 —
Op time (mins)
 <240 45.7 42.9 45.9 Ref 10.5 46.7 Ref
 240–330 31.3 42.9 30.5 0.34 47.4 30.6 0.02
 >330 23.0 14.3 23.6 0.47 42.1 22.8 0.009
Complications during surgical hospitalization
 Neurological 6.5 7.1 6.5 0.89 — — —
  Cardiac or pulmonary 5.4 7.1 5.3 0.68 — — —
 VTE 1.13 0 1.2 — — — —
 Infectious 4.3 3.6 4.3 0.85 — — —
Length of hospital stay ≥7 days† 22.8 17.9 23.1 0.53 — — —
Nonroutine discharge disposition 16.1 16.7 16.1 0.95 — — —
Ref = reference; VB = vertebrobasilar; VTE = venous thrombotic event.
All data are given as percentages. Boldface type indicates statistical significance in the univariable logistic regression analysis.
*  For whom there are data available on readmission.
†  Length of hospital stay of at least 7 days was the upper quartile of the IQR for the study population.

was available for 89.5% (n = 17) of patients who under-
went a reoperation, and the most frequent operations were
ventricular shunt placement, tracheostomy, craniotomy for
hematoma evacuation, and decompressive craniectomy
(Fig. 1B). The median times to ventricular shunt place-
ment and tracheostomy were 20 days (IQR 10.5–25 days)
and 8 days (IQR 7–10 days) postoperatively, respectively.
All cranial reoperations were a craniotomy for evacua-
tion of an extraaxial or intraparenchymal hematoma or
decompressive craniectomy. There were no reoperations
for SSIs.
The reasons for readmission were available for 53.6%
(n = 15) of readmitted patients during the years these
data were collected by the NSQIP (2012–2014). The most
common reason for readmission was seizure (26.7%);
other indications for readmission were hydrocephalus,
headache, stroke, gastrointestinal and cardiac complica-
tions, altered mental status, and intracranial hemorrhage
(Fig. 1A). The median time to readmission for patients
returning to acute care with seizure was 7.5 days (IQR
6–9 days) after surgery and 4 days (IQR 3–5.5 days) after
hospital discharge.
Regression Modeling
Multivariable logistic regression models were con-
structed to identify predictors of readmission, any reop-
eration, and nonroutine hospital discharge. Age greater
than 65 years, Class II or III obesity, and preoperative hy-
ponatremia and anemia were independent predictors of an
unplanned readmission (Table 2, Fig. 2). Notably, neither
the development of complications during the surgical hos-
pitalization nor discharge disposition was associated with
readmission. Age greater than 50 years and operative time
longer than 240 minutes, as well as dependent functional

760 J Neurosurg  Volume 128 • March 2018


FIG. 1. Pie charts depict the most common reasons for readmission (A) and reoperation (B). Kaplan-Meier curves show the time
to readmission (C) and reoperation (D) postoperatively. CVA = cerebrovascular accident; GI = gastrointestinal; ICH = intracranial
hemorrhage.
status (of borderline statistical significance) were associ-
ated with any reoperation (Table 2). Two different models
evaluated predictors of nonroutine hospital discharge: the
first was exclusively based on preoperative and operative
characteristics, while the second also included complica-
tions that developed during the surgical hospitalization
(Table 3).
Recursive Partitioning Analysis
Thereafter, recursive partitioning analysis was used
to evaluate for congruence with the results of multivari-
able regression analyses, and performed for unplanned
hospital readmission and nonroutine hospital discharge.
Partition trees of unplanned hospital readmission (Fig. 3
upper) revealed similar variables from regression analy-
sis, and the column contributions were from patient age
(32.9%), body habitus (28.1%), preoperative sodium level
(16.5%), ASA classification (12.8%), and preoperative he-
matocrit level (9.7%). Additional partition trees evaluated
nonroutine hospital discharge, the first of which only used
preoperative and operative variables, and the second of
which included postoperative complications. The parti-
tion tree based on preoperative and operative variables
(data not shown) had good discrimination (ROC = 0.76
in the training and 0.74 in the validation sets). Column
contributions were from patient age (34.1%), body habitus
(15.7%), operative time (13.5%), preoperative creatinine
(12.2%), TIA (10.8%), hypertension (4.7%), preoperative
Readmission and unruptured aneurysms
hematocrit level (4.7%), and neurological deficits (4.4%).
A subsequent partition tree included postoperative com-
plications (Fig. 3 lower), and column contributions were
from neurological complications (35.4%), patient age
(17.7%), cardiopulmonary complications (16.4%), ASA
classification (10.3%), aneurysm location (9.0%), race/
ethnicity (5.0%), operative time (3.8%), and preoperative
hematocrit level (2.3%).
Discussion
The benefit of treating unruptured intracranial an-
eurysms depends on weighing the risk of sustaining an-
eurysmal subarachnoid hemorrhage from an untreated
aneurysm against the risk of developing a perioperative
complication from treatment of the aneurysm. To that end,
several studies have attempted to risk-stratify patients un-
dergoing microsurgical clipping of unruptured cerebral
aneurysms, and patient age, aneurysm size, and location
have consistently emerged as predictors of untoward out-
comes.22,30,34,52 The recent unruptured intracranial aneu-
rysm treatment score was developed as multidisciplinary
consensus among cerebrovascular specialists, which as-
signs points based on known predictors of aneurysmal
rupture and perioperative complications.18 While an im-
portant risk-stratification tool, the only comorbidities in-
cluded in the score are neurocognitive, psychiatric, and
hematological. Additionally, Newman et al.32 used admin-
J Neurosurg  Volume 128 • March 2018 761
H. H. Dasenbrock et al.

TABLE 2. Multivariable logistic regression evaluating the predictors of readmission and reoperation
Unplanned Readmission* Reoperation†
Variable OR 95% CI p Value OR 95% CI p Value
Age (yrs)
 18–50 Ref — — Ref — —
 51–65 1.40 0.46–4.26 0.55 3.77 1.21–11.79 0.02
 >65 3.31 1.08–10.14 0.04 5.89 1.56–22.17 0.009
Body habitus
 Nonobese Ref — — — — —
 Obese 1.09 0.36–3.31 0.89
  Morbidly obese 4.02 1.55–10.42 0.004
Dependent functional status — — — 5.93 0.99–35.40 0.05
Preop sodium (mEq/L)
  ≥135 Ref — — — — —
 <135 5.73 1.32–24.89 0.02
Preop hematocrit (%)
  ≥36 Ref — — — — —
 <36 3.74 1.38–10.13 0.01
Op time (mins)
 <240 — — — Ref — —
 240–330 6.45 1.34–31.05 0.02
 >330 7.59 1.55–37.21 0.012
Boldface type indicates statistical significance.
*  C-statistic = 0.76, Hosmer-Lemeshow test = 0.65.
†  C-statistic = 0.78, Hosmer-Lemeshow test = 0.77.

istrative claims data to propose a comorbidity index for
patients with unruptured intracranial aneurysms; however,
due to the limitations of claims data, it is difficult to ascer-
tain if many of the most important comorbidities (includ-
ing neurological disorder, renal insufficiency, and gastro-
intestinal hemorrhage) were truly present upon admission
or misclassified postoperative complications. Comorbidi-
ties are known to impact postoperative outcomes, and
additional data highlighting patients who are at risk for
adverse postoperative outcomes are needed to optimally
risk-stratify patients.
In this NSQIP analysis, 460 adult patients who under-
went craniotomy for microsurgical clipping of a cerebral
aneurysm electively were included to evaluate reoperation
and readmission. There are many advantages to using the
NSQIP to evaluate these quality metrics. As data are pro-
spectively collected by surgical reviewers, comorbidities
are distinguished from postoperative complications, and
the accuracy of rates of and reasons for readmission has
been shown to be higher than with administrative data-
bases.44 Additionally, the NSQIP collects readmission data
regardless of whether the patient was readmitted to the
same hospital as the index surgery or a different institu-
tion, which is determined through direct communication
with the patient and his or her health care providers. Using
the results of the NISQIP analysis, further elaboration re-
garding 3 questions of clinical interest is merited.
Which Patients Have the Greatest Odds of Adverse
Events?
Multivariable logistic regression models and partition
trees were constructed of readmission, reoperation, and
762 J Neurosurg  Volume 128 • March 2018
nonroutine hospital discharge, to identify patients with the
greatest odds of an adverse event. This may augment pre-
operative risk-stratification and identify patient character-
istics to be included in future risk-benefit ratio schematics
evaluating repair of unruptured aneurysms. Models of read-
mission had good discrimination (ROCs between 0.67 and
0.82) and nonroutine discharge excellent discrimination
(ROC range 0.80–0.92), indicating their ability to distin-
guish cases from noncases. Four preoperative factors—age
greater than 65 years, Class II or III obesity, preoperative
hyponatremia, and preoperative anemia—were consistent
predictors of adverse events, in analyses using both regres-
sion and recursive partitioning. These variables were the
only independent predictors of unplanned hospital read-
mission, and entered at least 1 model of nonroutine hospital
discharge. While patient age is a well-known risk factor for
complications after clipping of unruptured aneurysms and
a recent publication emphasized the importance of anemia
in all patients with intracranial aneurysms,41 these models
highlight the importance of other characteristic that have
not been highlighted in this patient population.
Can Readmission Rates be Decreased?
In this study, the 30-day rate of reoperation was 4.2%
and readmission was 6.3%. The most common indication
for reoperation was ventricular shunt placement and for
readmission was seizures, although hydrocephalus, head-
ache, and stroke were other diagnoses associated with
readmission. Although the readmission rate is lower than
several other studies evaluating cranial neurosurgical pa-
tients (ranging from 7% to 24%),9,26,29,45,47 the indications
for readmission are in concordance with those of Singh et

FIG. 2. Independent predictors of readmission (A) and reoperation (B), as well as preoperative (C) and postoperative (complica-
tions; (D) predictors of a nonroutine hospital discharge from multivariable logistic regression.
al.46 and Marcus et al.,26 who have found similar associ-
ated diagnoses in the subarachnoid hemorrhage and brain
tumor populations, respectively. These indications for re-
admission can shape patient surveillance in readmission
reduction programs.16,28
While it is difficult to determine from diagnosis alone
which readmissions may have been preventable, the analy-
sis emphasizes some important determinants of readmis-
sion. The utility, selection, and duration of antiepileptic
prophylaxis after craniotomy remains debated,31,36,40,​51,54
and because seizure was the most common reason for re-
admission, further research on this topic is needed. More-
over, the frequency of early readmission underscores the
importance of early surveillance and follow-up after hos-
pital discharge.
The most common reason for reoperation was ven-
tricular shunt placement. As patients may present with
symptoms from mass effect from a large cerebral aneu-
rysm including obstructive hydrocephalus, the optimal
treatment of such patients is particularly challenging.14,15
In these cases, ventricular shunt placement would not be
an unplanned reoperation, but related to disease severity,
as a consequence of aneurysm size and location. However,
such patients are unlikely to undergo surgery electively
and are more likely to present acutely; in this patient popu-
lation, all admissions were directly from home. Neverthe-
less, prior research has primarily evaluated delayed hydro-
Readmission and unruptured aneurysms
cephalus after aneurysmal subarachnoid hemorrhage,25,49
and additional data on predictors of delayed hydrocepha-
lus in patients with unruptured aneurysms are needed.
Is Readmission an Appropriate Quality Indicator?
Notably, postoperative complications from the index
hospitalization, length of hospital stay, and discharge dis-
position did not meet entry criteria for regression models
or recursive partitioning analysis evaluating readmission,
and only preoperative characteristics were predictive of
readmission. This is in contrast to nonroutine hospital dis-
charge, which was largely predicted by postoperative com-
plications, particularly neurological or cardiopulmonary.
Therefore, the quality of care provided postoperatively
may be less important than baseline characteristics in
predicting readmission after cerebral aneurysm clipping.
Thus, readmission may be a suboptimal quality indicator
in this population. Nevertheless, preoperative optimization
of comorbidities and appropriate patient selection is criti-
cal to postoperative outcomes, and readmission rates may
be one metric to indirectly measure such care, given the
association of comorbidities and laboratory abnormalities
with readmission.
Limitations and Advantages to the Study Design
There are many notable limitations to this study. One
J Neurosurg  Volume 128 • March 2018 763
H. H. Dasenbrock et al.

TABLE 3. Multivariable logistic regression evaluating predictors of a nonroutine hospital discharge, based on
preoperative and operative characteristics (Model 1), and including complications during the surgical hospitalization
(Model 2)
Model 1* Model 2†
Variable OR 95% CI p Value OR 95% CI p Value
Age (yrs)
 18–50 Ref — — Ref — —
 51–65 6.34 2.28–17.64 <0.001 5.43 1.86–15.88 0.002
 >65 10.83 3.65–32.12 <0.001 5.13 1.58–16.67 0.007
Race or ethnicity
  White, not Hispanic Ref — — Ref — —
  African American 4.00 1.81–8.85 0.001 5.19 2.10–12.85 <0.001
 Hispanic 1.33 0.35–5.02 0.67 1.33 0.28–6.25 0.72
 Asian 2.76 0.94–8.08 0.06 3.56 1.04–12.09 0.04
 Unknown 0.59 0.21–1.71 0.33 0.62 0.18–2.19 0.46
Prior TIA 3.17 1.15–8.69 0.03 5.41 1.70–17.24 0.004
Body habitus
 Nonobese Ref — — Ref — —
 Obese 1.23 0.59–2.53 0.58 1.64 0.69–3.87 0.26
  Morbidly obese 2.05 0.98–4.29 0.06 2.02 0.85–4.80 0.11
Op type
  Carotid simple — — — Ref — —
  Carotid complex 1.32 0.63–2.74 0.46
  VB simple 4.07 0.60–27.79 0.15
  VB complex 3.48 1.05–11.47 0.04
Preop sodium (mEq/L)
  ≥135 Ref — — Ref — —
 <135 2.51 0.66–9.52 0.18 5.07 1.13–22.62 0.03
Preop creatinine (mg/dl)
 <1.4 — — — Ref — —
  ≥1.4 3.22 0.93–11.18 0.07
Preop hematocrit (%)
  ≥36 Ref — — Ref — —
 <36 2.31 1.07–4.98 0.03 2.27 0.92–5.65 0.08
Op time (mins)
 <240 Ref — — — — —
 240–330 1.68 0.84–3.33 0.14
 >330 3.28 1.61–6.69 0.001
Complications during the surgical hospitalization
 Neurological — — — 23.37 7.18–76.11 ≤0.001
 Cardiac/pulmonary — — — 7.99 2.30–27.79 0.001
 Infectious — — — 4.02 0.98–16.61 0.05
VB = vertebrobasilar.
Boldface type indicates statistical significance.
*  C-statistic = 0.80, Hosmer-Lemeshow test = 0.90.
†  C-statistic = 0.89, Hosmer-Lemeshow test = 0.31.

limitation to the use of the NSQIP to evaluate readmission
rates after neurosurgery is that patients are followed for
30 days postoperatively, and not within a specified time
period after discharge, and therefore rates of readmission
may be slightly underestimated by the NSQIP. Moreover,
while the NSQIP collects data evaluating many pertinent
preoperative characteristics and medical complications,
rates of hydrocephalus and seizures during the surgical
hospitalization, as well as the use of antiepileptic drugs,
could not be ascertained. Additionally, data are not sys-
tematically collected on endovascular intervention by the
764 J Neurosurg  Volume 128 • March 2018
NSQIP, and therefore readmission rates could not be com-
pared by treatment modality selected for aneurysm repair.
Data are not available on surgeon or institution character-
istics, including provider volume. Nonetheless, the NSQIP
provides a multiinstitutional perspective with a nationally
accrued patient population, increasing the external gener-
alizability compared with single-center studies. The data
collected by the NSQIP allow for a detailed analysis of
the rate, predictors, and reasons for readmission and re-
operation after craniotomy for microsurgical clipping of
unruptured aneurysms.
 Readmission and unruptured aneurysms

FIG. 3. Partition trees evaluating unplanned hospital readmission (upper) and nonroutine hospital discharge (lower), which iden-
tify the predictive probability of each event based on a series of dichotomous partitionings of the population into subpopulations.

Conclusions for, and predictors of 30-day readmission and reoperation.

In this NSQIP analysis, 460 patients who underwent
craniotomy for microsurgical clipping of an unruptured
aneurysm were evaluated to analyze the rate of, reasons
The strongest predictors of adverse events were patient
age greater than 65 years, Class II or III obesity, preopera-
tive hyponatremia, and preoperative anemia, highlighting
baseline characteristics that are key in the outcomes of pa-
J Neurosurg  Volume 128 • March 2018 765
H. H. Dasenbrock et al.
tients with unruptured intracranial aneurysms. The most
common reasons for readmission and reoperation were
seizures and ventricular shunt placement, respectively,
emphasizing the need for additional data on the utility and
duration of postcraniotomy seizure prophylaxis. Given
that 59.1% of readmissions were within the 1st week and
86.4% within 2 weeks of hospital discharge, early post-
operative surveillance and follow-up may be important.
Nevertheless, as readmission was related to baseline pa-
tient characteristics, and not differential quality of care
received during the surgical hospitalization, its utility as
a quality indicator may be suboptimal. However, readmis-
sion may be an indirect proxy for preoperative optimiza-
tion of comorbidities, correction of laboratory values, and
appropriate patient selection.
References
  1. Abt NB, Bydon M, De la Garza-Ramos R, McGovern K, Ol-
ivi A, Huang J, et al: Concurrent neoadjuvant chemotherapy
is an independent risk factor of stroke, all-cause morbidity,
and mortality in patients undergoing brain tumor resection. J
Clin Neurosci 21:1895–1900, 2014
  2. Alan N, Seicean A, Seicean S, Neuhauser D, Benzel EC, Weil
RJ: Preoperative steroid use and the incidence of periopera-
tive complications in patients undergoing craniotomy for
definitive resection of a malignant brain tumor. J Clin Neu-
rosci 22:1413–1419, 2015
  3. Alan N, Seicean A, Seicean S, Neuhauser D, Weil RJ: Impact
of preoperative anemia on outcomes in patients undergoing
elective cranial surgery. J Neurosurg 120:764–772, 2014
  4. Basques BA, Fu MC, Buerba RA, Bohl DD, Golinvaux NS,
Grauer JN: Using the ACS-NSQIP to identify factors affect-
ing hospital length of stay after elective posterior lumbar fu-
sion. Spine (Phila Pa 1976) 39:497–502, 2014
  5. Basques BA, Varthi AG, Golinvaux NS, Bohl DD, Grauer
JN: Patient characteristics associated with increased post-
operative length of stay and readmission after elective lami-
nectomy for lumbar spinal stenosis. Spine (Phila Pa 1976)
39:833–840, 2014
  6. Bekelis K, Bakhoum SF, Desai A, Mackenzie TA, Goodney
P, Labropoulos N: A risk factor-based predictive model of
outcomes in carotid endarterectomy: the National Surgical
Quality Improvement Program 2005–2010. Stroke 44:1085–
1090, 2013
  7. Bekelis K, Bakhoum SF, Desai A, Mackenzie TA, Roberts
DW: Outcome prediction in intracranial tumor surgery: the
National Surgical Quality Improvement Program 2005–2010.
J Neurooncol 113:57–64, 2013
  8. Bekelis K, Desai A, Bakhoum SF, Missios S: A predictive
model of complications after spine surgery: the National
Surgical Quality Improvement Program (NSQIP) 2005–2010.
Spine J 14:1247–1255, 2014
  9. Buchanan CC, Hernandez EA, Anderson JM, Dye JA, Leung
M, Buxey F, et al: Analysis of 30-day readmissions among
neurosurgical patients: surgical complication avoidance as
key to quality improvement. J Neurosurg 121:170–175, 2014
10. Buerba RA, Fu MC, Gruskay JA, Long WD III, Grauer JN:
Obese Class III patients at significantly greater risk of multi-
ple complications after lumbar surgery: an analysis of 10,387
patients in the ACS NSQIP database. Spine J 14:2008–2018,
2014
11. Bydon M, Abt NB, De la Garza-Ramos R, Macki M, Witham
TF, Gokaslan ZL, et al: Impact of resident participation on
morbidity and mortality in neurosurgical procedures: an
analysis of 16,098 patients. J Neurosurg 122:955–961, 2015
12. Bydon M, Abt NB, Macki M, Brem H, Huang J, Bydon A, et
766 J Neurosurg  Volume 128 • March 2018
al: Preoperative anemia increases postoperative morbidity in
elective cranial neurosurgery. Surg Neurol Int 5:156, 2014
13. Chern JJ, Bookland M, Tejedor-Sojo J, Riley J, Shoja MM,
Tubbs RS, et al: Return to system within 30 days of discharge
following pediatric shunt surgery. J Neurosurg Pediatr
13:525–531, 2014
14. Dengler J, Maldaner N, Bijlenga P, Burkhardt JK, Graewe
A, Guhl S, et al: Perianeurysmal edema in giant intracranial
aneurysms in relation to aneurysm location, size, and partial
thrombosis. J Neurosurg 123:446–452, 2015
15. Dengler J, Maldaner N, Bijlenga P, Burkhardt JK, Graewe A,
Guhl S, et al: Quantifying unruptured giant intracranial an-
eurysms by measuring diameter and volume—a comparative
analysis of 69 cases. Acta Neurochir (Wien) 157:361–368,
2015
16. Dharmarajan K, Hsieh AF, Lin Z, Bueno H, Ross JS, Horwitz
LI, et al: Hospital readmission performance and patterns of
readmission: retrospective cohort study of Medicare admis-
sions. BMJ 347:f6571, 2013
17. Dickinson H, Carico C, Nuño M, Mukherjee D, Ortega A,
Black KL, et al: Unplanned readmissions and survival fol-
lowing brain tumor surgery. J Neurosurg 122:61–68, 2015
18. Etminan N, Brown RD Jr, Beseoglu K, Juvela S, Raymond
J, Morita A, et al: The unruptured intracranial aneurysm
treatment score: a multidisciplinary consensus. Neurology
85:881–889, 2015
19. Greenberg JK, Washington CW, Guniganti R, Dacey RG
Jr, Derdeyn CP, Zipfel GJ: Causes of 30-day readmission
after aneurysmal subarachnoid hemorrhage. J Neurosurg
124:743–749, 2016
20. Gruskay JA, Fu M, Basques B, Bohl DD, Buerba R, Webb
ML, et al: Factors affecting length of stay and complications
following elective anterior cervical discectomy and fusion:
a study of 2164 patients from the American College of Sur-
geons National Surgical Quality Improvement Project Data-
base (ACS NSQIP). Clin Spine Surg 29:E34–E42, 2016
21. Houkin K, Baba T, Minamida Y, Nonaka T, Koyanagi I,
Iiboshi S: Quantitative analysis of adverse events in neurosur-
gery. Neurosurgery 65:587–594, 2009
22. Khanna RK, Malik GM, Qureshi N: Predicting outcome
following surgical treatment of unruptured intracranial aneu-
rysms: a proposed grading system. J Neurosurg 84:49–54,
1996
23. Kimmell KT, Jahromi BS: Clinical factors associated with
venous thromboembolism risk in patients undergoing crani-
otomy. J Neurosurg 122:1004–1011, 2015
24. Kimmell KT, Walter KA: Risk factors for venous thrombo-
embolism in patients undergoing craniotomy for neoplastic
disease. J Neurooncol 120:567–573, 2014
25. Lewis A, Irvine H, Ogilvy C, Kimberly WT: Predictors for
delayed ventriculoperitoneal shunt placement after external
ventricular drain removal in patients with subarachnoid hem-
orrhage. Br J Neurosurg 29:219–224, 2015
26. Marcus LP, McCutcheon BA, Noorbakhsh A, Parina RP,
Gonda DD, Chen C, et al: Incidence and predictors of 30-day
readmission for patients discharged home after craniotomy
for malignant supratentorial tumors in California (1995–
2010). J Neurosurg 120:1201–1211, 2014
27. McCutcheon BA, Ciacci JD, Marcus LP, Noorbakhsh A,
Gonda DD, McCafferty R, et al: Thirty-day perioperative
outcomes in spinal fusion by specialty within the NSQIP
database. Spine (Phila Pa 1976) 40:1122–1131, 2015
28. Merkow RP, Ju MH, Chung JW, Hall BL, Cohen ME, Wil-
liams MV, et al: Underlying reasons associated with hospital
readmission following surgery in the United States. JAMA
313:483–495, 2015
29. Moghavem N, Morrison D, Ratliff JK, Hernandez-Boussard
T: Cranial neurosurgical 30-day readmissions by clinical
indication. J Neurosurg 123:189–197, 2015

30. Morgan MK, Wiedmann M, Assaad NN, Heller GZ: Com-
plication-effectiveness analysis for unruptured intracranial
aneurysm surgery: a prospective cohort study. Neurosurgery
78:648–659, 2106
31. Naidech AM, Garg RK, Liebling S, Levasseur K, Macken
MP, Schuele SU, et al: Anticonvulsant use and outcomes af-
ter intracerebral hemorrhage. Stroke 40:3810–3815, 2009
32. Newman WC, Neal DW, Hoh BL: A new comorbidities index
for risk stratification for treatment of unruptured cerebral
aneurysms. J Neurosurg 125:713–719, 2016
33. Nuño M, Ly D, Mukherjee D, Ortega A, Black KL, Patil CG:
Quality of surgical care and readmission in elderly glioblas-
toma patients. Neurooncol Pract 1:33–39, 2014
34. Ogilvy CS, Carter BS: A proposed comprehensive grading
system to predict outcome for surgical management of intra-
cranial aneurysms. Neurosurgery 42:959–970, 1998
35. Rambachan A, Smith TR, Saha S, Eskandari MK, Bendok
BR, Kim JY: Reasons for readmission after carotid endarter-
ectomy. World Neurosurg 82:e771–e776, 2014
36. Raper DM, Starke RM, Komotar RJ, Allan R, Connolly ES
Jr: Seizures after aneurysmal subarachnoid hemorrhage: a
systematic review of outcomes. World Neurosurg 79:682–
690, 2013
37. Rolston JD, Han SJ, Bloch O, Parsa AT: What clinical fac-
tors predict the incidence of deep venous thrombosis and
pulmonary embolism in neurosurgical patients? J Neurosurg
121:908–918, 2014
38. Rolston JD, Han SJ, Lau CY, Berger MS, Parsa AT: Frequen-
cy and predictors of complications in neurological surgery:
national trends from 2006 to 2011. J Neurosurg 120:736–
745, 2014
39. Sarda S, Bookland M, Chu J, Shoja MM, Miller MP, Reisner
SB, et al: Return to system within 30 days of discharge fol-
lowing pediatric non-shunt surgery. J Neurosurg Pediatr
14:654–661, 2014
40. Sayegh ET, Fakurnejad S, Oh T, Bloch O, Parsa AT: Anticon-
vulsant prophylaxis for brain tumor surgery: determining the
current best available evidence. J Neurosurg 121:1139–1147,
2014
41. Seicean A, Alan N, Seicean S, Neuhauser D, Selman WR,
Bambakidis NC: Risks associated with preoperative anemia
and perioperative blood transfusion in open surgery for intra-
cranial aneurysms. J Neurosurg 123:91–100, 2015
42. Seicean A, Alan N, Seicean S, Worwag M, Neuhauser D,
Benzel EC, et al: Impact of increased body mass index on
outcomes of elective spinal surgery. Spine (Phila Pa 1976)
39:1520–1530, 2014
43. Seicean A, Seicean S, Schiltz NK, Alan N, Jones PK, Neu-
hauser D, et al: Short-term outcomes of craniotomy for ma-
lignant brain tumors in the elderly. Cancer 119:1058–1064,
2013
44. Sellers MM, Merkow RP, Halverson A, Hinami K, Kelz RR,
Bentrem DJ, et al: Validation of new readmission data in the
American College of Surgeons National Surgical Quality
Improvement Program. J Am Coll Surg 216:420–427, 2013
45. Shah MN, Stoev IT, Sanford DE, Gao F, Santiago P, Jaques
DP, et al: Are readmission rates on a neurosurgical service
indicators of quality of care? J Neurosurg 119:1043–1049,
2013
46. Singh M, Guth JC, Liotta E, Kosteva AR, Bauer RM, Prab-
hakaran S, et al: Predictors of 30-day readmission after sub-
arachnoid hemorrhage. Neurocrit Care 19:306–310, 2013
Readmission and unruptured aneurysms
47. Taylor BE, Youngerman BE, Goldstein H, Kabat DH, Ap-
pelboom G, Gold WE, et al: Causes and timing of unplanned
early readmission after neurosurgery. Neurosurgery 79:356–
369, 2016
48. Tsai TC, Joynt KE, Orav EJ, Gawande AA, Jha AK: Varia-
tion in surgical-readmission rates and quality of hospital
care. N Engl J Med 369:1134–1142, 2013
49. Walcott BP, Iorgulescu JB, Stapleton CJ, Kamel H: Incidence,
timing, and predictors of delayed shunting for hydrocephalus
after aneurysmal subarachnoid hemorrhage. Neurocrit Care
23:54–58, 2015
50. Washington CW, Derdeyn CP, Dacey RG Jr, Dhar R, Zipfel
GJ: Analysis of subarachnoid hemorrhage using the Nation-
wide Inpatient Sample: the NIS-SAH Severity Score and
Outcome Measure. J Neurosurg 121:482–489, 2014
51. Weston J, Greenhalgh J, Marson AG: Antiepileptic drugs as
prophylaxis for post-craniotomy seizures. Cochrane Data-
base Syst Rev 3:CD007286, 2015
52. Wiebers DO, Whisnant JP, Huston J III, Meissner I, Brown
RD Jr, Piepgras DG, et al: Unruptured intracranial aneu-
rysms: natural history, clinical outcome, and risks of surgical
and endovascular treatment. Lancet 362:103–110, 2003
53. Wrubel DM, Riemenschneider KJ, Braender C, Miller BA,
Hirsh DA, Reisner A, et al: Return to system within 30 days
of pediatric neurosurgery. J Neurosurg Pediatr 13:216–221,
2014
54. Wu AS, Trinh VT, Suki D, Graham S, Forman A, Weinberg
JS, et al: A prospective randomized trial of perioperative
seizure prophylaxis in patients with intraparenchymal brain
tumors. J Neurosurg 118:873–883, 2013
Disclosures
Dr. Aziz-Sultan has served as a proctor for Covidien and
Medtronic. Dr. Gormley served as a proctor for Codman. The
ACS NSQIP and the hospitals participating in the ACS NSQIP
are the source of the data used herein; they have not verified and
are not responsible for the statistical validity of the data analysis
or the conclusions derived by the authors.
Author Contributions
Conception and design: Du, Dasenbrock. Acquisition of data:
Dasenbrock. Analysis and interpretation of data: Dasenbrock,
Rudy. Drafting the article: Dasenbrock. Critically revising the
article: all authors. Reviewed submitted version of manuscript: all
authors. Approved the final version of the manuscript on behalf of
all authors: Du. Statistical analysis: Dasenbrock. Study supervi-
sion: Du.
Supplemental Information
Previous Presentations
The abstract of this manuscript was presented as a podium pre-
sentation at the 84th AANS Annual Meeting in Chicago, Illinois,
April 30–May 4, 2016.
Correspondence
Rose Du, Department of Neurosurgery, Brigham and Women’s
Hospital and Harvard Medical School, 75 Francis St., Boston, MA
02115. email: rdu@bwh.harvard.edu.
J Neurosurg  Volume 128 • March 2018 767