- Email: [email protected]
Thirty-day non-seizure outcomes following temporal lobectomy for adult epilepsy
Clinical Neurology and Neurosurgery 160 (2017) 12–18
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Thirty-day non-seizure outcomes following temporal lobectomy for adult epilepsy Brandon A. Sherrod, Matthew C. Davis, Kristen O. Riley
MARK
⁎
Department of Neurological Surgery, University of Alabama at Birmingham, Birmingham, AL, USA
A R T I C L E I N F O
A B S T R A C T
Keywords: Epilepsy Temporal lobectomy NSQIP Morbidity Mortality Adverse event Complication
Objective: Multi-institutional rates of acute adverse outcomes other than seizures after temporal lobectomy (TL) are not well understood. Here we analyzed short-term morbidity and mortality following TL using a validated national database. Patients and methods: The multi-institutional American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was queried by Current Procedural Terminology (CPT) code for TL procedures performed for adult patients with diagnoses related to epilepsy from 2008 to 2014. Patient demographics, operative variables, hospital variables, preoperative laboratory values, and preexisting comorbidities were analyzed using univariate and multivariate techniques to determine associations with 30-day postoperative morbidity and mortality. Results: A total of 202 TL procedures were analyzed, 80 (39.6%) with intraoperative electrocorticography (ECOG) and 122 (60.4%) without ECOG. Mean age was 40.4 ± 13.7 years, and 47.5% of patients were male. Overall morbidity and mortality were 11.4% and 2.0%, respectively. The most common adverse outcomes were reoperation (5.4%), stroke with residual deficit (2.5%), failure to wean from ventilator (2.0%), and surgical site infection (2.0%). Adverse event rates were not significantly different between TLs with and without ECOG (13.1% vs. 8.8%, p = 0.375). Independent predictors of adverse events included prior stroke (OR 7.60, 95% CI 1.22–47.17, p = 0.029) and chronic steroid use (OR 10.90, 95% CI 1.03–115.79, p = 0.048). Diabetes mellitus (p = 0.078) and older age (p = 0.145) approached but did not reach significance in the multivariate model. Conclusions: We report rates of acute morbidity and mortality following TL procedures using a national database. These findings can be used both to assist with patient selection as well as patient counseling prior to surgery.
1. Introduction
medical complications, major medical complications, and mortality following TL was 5.4%, 1.6%, and 0.4%, respectively [16]. However, Hader et al. (2013) recognized the limitations of lacking prospectively collected, standardized complication data. Furthermore, follow-up period requirements were not standardized across studies in the review. An analysis of the Nationwide Inpatient Sample (NIS) dataset found that morbidity after anterior TL was 10.8%, with no mortality [17]. However, the NIS is an administrative, not clinical, dataset, and the International Classification of Diseases, Ninth Revision [ICD-9] codes used in the NIS study are not specific to TL [18]. Most other studies on TL outcomes focus on seizure outcomes and are limited to single-center experiences with relatively small case series [8,10,19]. Furthermore, many of the previously cited studies lack data on complications considered to be “minor” such as urinary tract infection, wound disruption, etc. Minor non-seizure complications still have significant potential to negatively affect quality of life, length of hospital stay, and cost of care
Refractory temporal lobe epilepsy (TLE) is the most common cause of pharmacoresistant seizures [1,2]. Temporal lobectomy (TL) is performed after exhausting most other efforts to control refractory TLE. Given that seizure control is the ultimate goal of TL, numerous studies have investigated seizure outcomes after TL in adults [3–11] and children [12–15]. However, few studies have investigated other (non-seizure) postoperative outcomes after TL, especially using multicenter data. Due to the recognized underutilization of TL for refractory TLE, understanding the full extent of TL complications is paramount in making informed decisions prior to TL surgery [1,2]. A review of the literature on major and minor complications following TL is displayed in Table 1. Hader et al. (2013) reviewed rates of major and minor complications following procedures for invasive seizure monitoring or resective surgery, finding that the rates of minor ⁎
Corresponding author at: Department of Neurosurgery, University of Alabama at Birmingham, 1720 2nd Avenue S., FOT 1008, Birmingham, AL 35294-3410, USA. E-mail address: [email protected] (K.O. Riley).
http://dx.doi.org/10.1016/j.clineuro.2017.05.027 Received 16 January 2017; Received in revised form 3 May 2017; Accepted 29 May 2017 Available online 08 June 2017 0303-8467/ © 2017 Elsevier B.V. All rights reserved.
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
excluded from the analysis if the NSQIP postoperative diagnosis was not epilepsy-related. Trauma cases (e.g., hematoma evacuation) and patients under the age of 18 are automatically excluded from the NSQIP dataset and were not included for this analysis. The University of Alabama at Birmingham Institutional Review Board (IRB) does not require IRB approval of NSQIP studies because NSQIP data are de-identified and considered a public dataset [30].
Table 1 Literature reports of major and minor complications after TL surgery in adults. Major Complicationsb
Minor Complicationsb
Complication
Rates, %
Citation
Complication
Rates, %
Citation
Death
0.4
[16]
Memory impairment (mild)
3.2
[16]
0.0 3.0
[17,19] [10]
1.2
[17]
Hemorrhage or hematoma
2.0
[16]
Infectionc
0.8 1.9 2.0 2.9
[17] [16] [10] [19]
Visual field defect
1.1 3.0
[17] [10]
Hydrocephalus
1.3 0.4
[16] [17]
DVT or PE
0.0 0.7
[17] [16]
2.2. Variables of interest Psychiatrica
5.8
[16]
Dysphagia
2.1 5.5
[16] [10]
Patient demographics included age, gender, race, and body mass index (BMI). Patient comorbidities included obesity, diabetes, tobacco use within the previous year, hypertension requiring medication, chronic steroid use, previous stroke with/without deficit, and preoperative sepsis. Preoperative laboratory values included hypernatremia (Na > 145 mEq/L), hyponatremia (Na < 135 mEq/L), thrombocytopenia (< 150k platelets/μL), leukocytosis (> 11k WBC/ μL), elevated creatinine (> 1.2 mg/dL), and anemia (hematocrit < 36% for females; < 40% for males). Lab values were only included if at least 85% of cases had recorded values (e.g., AST and ALT laboratory values are available in NSQIP, but less than 85% of patients in our cohort had these lab values recorded). Postoperative diagnoses were determined by ICD-9 codes. Hospital variables included length of stay, time from operation to discharge, time from admission to operation, and prior operation within 30 days. Operative variables included operative time, emergent operation status, American Society of Anesthesiologist (ASA) class, perioperative blood transfusion, and wound classification. Complications are tracked by NSQIP for up to 30 days postoperatively; follow-up beyond the 30-day window is not recorded. Complications for this study included reoperation, death, bleeding requiring transfusion, subdural hemorrhage, intracerebral hemorrhage, surgical site infection (superficial incisional, deep incisional, or organ/ space [includes osteomyelitis, ventriculitis, meningitis, and intracranial abscesses]), pneumonia, failure to wean from ventilator, urinary tract infection (UTI), deep venous thrombosis (DVT), pulmonary embolism (PE), sepsis/septic shock, and stroke with resultant neurological deficit. NSQIP complication coding follows strict criteria for entry and are coded by trained data abstractors with minimal (< 2%) inter-rater disagreement; full definitions for complication coding criteria are provided in the referenced ACS NSQIP User Guides [20–25].
DVT = deep venous thrombosis; PE = pulmonary embolism. a Psychiatric includes personality changes, etc. otherwise not attributable to cognitive decline or memory changes alone. b Definitions of “major” and minor vary by study. Here, we have classified major and minor complication by relative severity, not acuity or duration. Time to complications, whether acute or chronic, was not distinguished.. c Minor wound infections were excluded.
even if long term health is not affected. There is a need for a nationwide, multicenter study on rates of complications following TL surgery. The purpose of the current study was to investigate non-seizure postoperative complications following TL surgery using a nationwide, multi-institutional clinical database.
2. Material and methods 2.1. Data acquisition The American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) was initiated in 2005 to identify areas of quality improvement in surgery. The NSQIP database contains millions of procedures across hundreds of hospitals primarily located in the United States and Canada, including academic and non-academic centers [20–25]. The ACS trains clinical data abstractors at each NSQIP site and employs strict data collection and case exclusion criteria, with inter-rater reliability rates at 98% or greater [20–26]. Researchers and clinicians at NSQIP participating hospitals may access the entire dataset, which is stripped of any patient or hospital identifying information. The ACS NSQIP adult dataset was obtained for years 2008–2014. The number of hospitals participating in each database year are: 211 hospitals (2008), 237 hospitals (2009), 258 hospitals (2010), 315 hospitals (2011), 374 hospitals (2012), 435 hospitals (2013), and 517 hospitals (2014) [20–25]. The accuracy (in comparison to chart review), follow up, ability to improve outcomes for participating hospitals, and inter-rater reliability of this database have been validated by several independent studies and internal ACS audits [26–29]. Procedures performed by a neurosurgeon were filtered from the main dataset by querying the surgical subspecialty variable for “Neurosurgery”. Temporal lobectomy procedures were filtered from the neurosurgeryonly dataset by Current Procedural Terminology (CPT) codes: 61537 (craniotomy; temporal lobectomy, without intraoperative ECOG) and 61538 (craniotomy; temporal lobectomy, with ECOG). Cases were
2.3. Statistical analysis Univariate analyses of the association between overall morbidity and patient demographics, preexisting comorbidities, abnormal lab values, and operative factors were performed using Chi-square test, Fisher’s exact test, binary logistic regression, or independent sample Student’s t-test where appropriate. The alpha value for significance was set at 0.05. Variables that reached significance in the univariate analyses were then entered into a multivariate logistic regression model. An area under the curve (AUC) analysis for a receiver operating characteristic (ROC) curve was performed to determine predictive accuracy of the multivariate model. All statistical analyses were performed using SPSS Version 23.0 (IBM Corp, Armonk, NY, 2015). 3. Results 3.1. Patient demographics, comorbidities, and laboratory values A total of 202 procedures were performed, 80 (39.6%) with intraoperative ECOG and 122 (60.4%) without intraoperative ECOG. Patient demographics, comorbidities, and laboratory values are shown in Table 2. Age ranged from 18 to 78 years, with 13 patients age 65 or older in this study (6.4% of total). There was essentially an even gender distribution (47.5% male). The most common comorbidities were 13
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Table 2 Patient demographics, preexisting comorbidities, and preoperative laboratory values. Parameter
Mean ± SD
Age (years) BMI (kg/m2)a
40.4 ± 13.7 29.2 ± 7.2
Parameter
Table 3 Hospital stay and operative variables.
n (%)
a
Gender Male Female
96 (47.5) 105 (52.0)
Race White Black Asian Native American Unknown
164 (81.2) 9 (4.5) 6 (3.0) 3 (1.5) 20 (9.9)
Body weight Underweight (BMI < 18.5) Class I obesity (BMI 30–34.9) Class II obesity (BMI 35–39.9) Class III obesity (BMI ≥ 40)
1 (0.5) 43 (21.3) 16 (7.9) 18 (8.9)
Diabetes History of tobacco use in previous year Hypertension requiring medication Chronic steroid use Bleeding disorder Previous stroke with deficitb Previous stoke without deficitb Preoperative sepsis Hypernatremia (Na > 145 mEq/L) Hyponatremia (Na < 135 mEq/L) Thrombocytopenia (< 150,000 platelets/μL) Leukocytosis (> 11,000 WBC/μL) Elevated serum creatinine (> 1.2 mg/dL) Anemia (< 36% Hct for females; < 40% Hct for males)
4 (2.0) 42 (20.8) 33 (16.3) 4 (2.0) 4 (2.0) 6 (3.0) 1 (0.5) 7 (3.5) 3 (1.5) 16 (7.9) 2 (1.0) 18 (8.9) 5 (2.5) 30 (14.9)
Parameter
n (%) or mean ± SD
Total Length of stay (days) Total length of stay > 30 daysa Time from operation to discharge (days) Time from admission to operation (days) Length of operation (minutes)
5.4 ± 5.1 2 (0.8) 4.1 ± 3.5 1.3 ± 3.7 247.5 ± 83.1
ASA Classification, n (%) ASA Class I ASA Class II ASA Class III ASA Class IV ASA Class V Preoperative blood transfusion, n (%)
0 (0.0) 100 (47.5) 95 (47.0) 6 (3.0) 1 (0.5) 2 (0.8)
Wound classification, n (%) Clean Clean-contaminated Contaminated Dirty/infected
198 (98.0) 3 (1.5) 1 (0.5) 0 (0.0)
ASA = American Society of Anesthesiologists. a Two patients had total length of stay (including preoperative and postoperative hospital stay) greater than 30 days (32 and 35 days), but no patient had a postoperative length of stay greater than 30 days. Table 4 Postoperative 30-day temporal lobectomy complications.
SD = standard deviation; IQR = interquartile range; BMI = body mass index; WBC = white blood cell; Hct = hematocrit. a Data missing for 1 case. b Stroke history unknown for 80 cases.
hypertension, obesity, and tobacco use. Although NSQIP does not include a thorough neurological comorbidity assessment, variables such as prior stroke are included. The most common abnormal preoperative laboratory values were anemia, leukocytosis, and hyponatremia. Elevated serum creatinine, thrombocytopenia, and hypernatremia were all relatively rare (occurring in less than 3% of cases). Importantly, not all patients had preoperative laboratory values measured or recorded; however, lab values were only included if at least 85% of cases had recorded values.
Outcome
n (%)
Time to Complication in days, median (range)a
Any adverse outcome Reoperation Stroke with deficit Ventilator requirement > 48 h Death Organ/space surgical site infection Urinary tract infection Sepsis Subarachnoid hemorrhage Pneumonia Bleeding requiring transfusion Deep venous thrombosis Wound disruption Intracerebral hemorrhage Deep incisional surgical site infection Pulmonary embolism Superficial incisional surgical site infection
23 (11.4) 11 (5.4) 5 (2.5) 4 (2.0) 4 (2.0) 4 (2.0) 3 (1.5) 3 (1.5) 2 (1.0) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5)
4 (0–28) 1 (0–9) 1 (0–5) 4.5 (1–9) 9 (3–19) 5 (4–28) 10 (4–23) 4.5 (4–5) Unknown 3 (3–3) 2 (2–2) 6 (6–6) 15 (15–15) Unknown 15 (15–15)
0 (0.0) 0 (0.0)
– –
a For the complications with days to event data. Not all reoperations had days to event data, and none of the hemorrhages had days to event data.
reoperation (5.4%), stroke with deficit (2.5%), organ space/surgical site infection (2.0%), and intracranial hemorrhage (1.5%). The most common medical complications were failure to wean from ventilator (2.0%), UTI (1.5%), sepsis (1.5%), and pneumonia (0.5%). Complications occurred at varying postoperative times, with a range from 0 to 28 days overall. It is important to recognize that the NSQIP reoperation variable does not capture whether the reoperation was related to the index procedure. Therefore, these results should be interpreted with caution.
3.2. Hospital and operative variables Hospital and operative factors may be seen in Table 3. Substantial variation was observed for hospital length of stay, time from operation to discharge, time from admission to operation, and length of operation. Since NSQIP does not track outcomes after 30-days, length of hospital stay results are likely missing data on patients who had substantially longer and complicated hospital stays. 3.3. Thirty-day complication rates
3.4. Predictors of thirty-day complications
Thirty-day complications are shown in Table 4. Adverse events occurred after 23 cases (11.4%).Ooverall mortality frequency was 4 (2.0%). Complication rates were not significantly different between TLs with and without ECOG (13.1% without ECOG vs. 8.8% with ECOG, p = 0.375). The most common neurosurgical complications were
Predictors of 30-day complications via univariate and multivariate analyses are displayed in Table 5. Independent risk factors for any adverse event included history of prior stroke and chronic steroid use. Older age (p = 0.145) and diabetes (p = 0.078) approached but did not reach independent significance in the multivariate model. Variables 14
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Table 5 Predictors of 30-day adverse eventsb following TL. Variables in bold represent significance (p < 0.05). Parameter
Univariate OR (95% CI)
Procedure TL without EEG TL with EEG
Ref 0.635 (0.249–1.621)
Adjusted OR (95% CI)
P
Age (years) Gender Male Female
1.053(1.019–1.087)a
1.034 (0.989–1.080)
0.145
2.269 (0.916–5.620) Ref
2.144 (0.747–6.158)
0.156
Race Minority Non-minority
0.898 (0.287–2.810) Ref
ASA Classification ASA Class I–II ASA Class III–V
Ref 2.472 (0.970–6.298)
1.642 (0.540–5.000)
0.382
Body Weight Class I obesity (BMI 30–34.9) Class II obesity (BMI 35–39.9) Class III obesity (BMI ≥ 40)
1.031 (0.359–2.956) 1.122 (0.238–5.286) 2.481 (0.741–8.307)
1.019 (0.166–6.251)
0.983
1.728 (0.471–6.333) 10.896 (1.025–115.893)
0.409 0.048
9.108 (0.783–105.983) 7.598 (1.224–47.165)
0.078 0.029
4.917 (0.427–56.696)
0.202
Hypertension Chronic steroid use Smoker in last year Diabetes Prior stroke Preoperative sepsis Hypernatremia (Na > 145 mEq/L) Hyponatremia (Na < 135 mEq/L) Thrombocytopenia (< 150,000 platelets/μL) Leukocytosis (> 11,000 WBC/μL) Elevated serum creatinine (> 1.2 mg/dL) Anemia (< 36% Hct for females; < 40% Hct for males) Operative time (hours; interval) 0–3 h 3–6 h 6+ hours Operative time (hours; continuous) Preoperative length of stay in days
4.152 (1.620–10.642) 8.429 (1.128–62.999) 1.800 (0.688–4.710) 8.429 (1.128–62.999) 12.351 (2.570–59.366) 3.314 (0.605–18.164) 4.023 (0.350–46.199) 0.497 (0.063–3.949) 8.091 (0.489–133.983) 1.640 (0.437–6.162) 13.275 (2.092–84.255) 0.514 (0.114–2.314) Ref 0.391 0.408 0.755 0.989
(0.149–1.024) (0.078–2.127) (0.534–1.067)a (0.872–1.122)
TL = temporal lobectomy; EEG = electrocorticography; ASA = American Society of Anesthesiologists; BMI = body mass index; WBC = white blood cell; Hct = hematocrit. a Per unit increase. b Includes all complications and mortality.
significant in the univariate analysis but not the multivariate analysis included male gender, elevated preoperative serum creatinine, diabetes, hypertension, class III obesity, ASA class 3 or greater, and older age. The results of the multivariate model ROC curve yielded an AUC, or C-statistic, of 0.797 (95% CI 0.69–0.90, p < 0.001). Fig. 1 displays the multivariate model ROC curve. 4. Discussion Temporal lobectomy is a well-established treatment option for patients with TLE. However, national rates of postoperative TL complications unrelated to seizures have not been well reported using standardized clinical data. Here, we report recent, standardized, nationwide data from the ACS NSQIP dataset on acute non-seizure complications following TL. 4.1. Complications The overall 30-day morbidity following TL was 11.4%, and the overall 30-day mortality following TL was 2.0%. The reported morbidity rate includes all complications, including relatively minor complications such as UTI and superficial wound infection. Furthermore, the rates of serious complications (e.g., death, organ/space SSI, intracranial hemorrhage) were relatively low − rates of these types of complications were less than 2%. These findings validate previous studies demonstrating relatively low rates of adverse events following
Fig. 1. Receiver Operating Characteristic (ROC) curve for multivariate model. The diagonal line represents the predictive power of chance alone. Area under the curve (AUC) was 0.797 (95% CI = 0.693–0.900, p < 0.001).
15
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
substantially different age groups. Previous studies have reported higher rates of complications in older patients after TL [11]. Older patients should be thoroughly assessed prior to TL surgery and counseled adequately on the increased risk of acute postoperative complications. The authors recognize that a separate morbidity and mortality analysis would be useful. The statistical significance of a model predicting an event (death) that only occurred in 4 patients would be questionable however. We are specifically concerned about type II statistical error, or failure to report a truly significant variable due to underpowered sampling. As a result, we elected to combine all adverse events into a single outcome measure. The low frequency of “minor” complications (such as UTI and wound disruption, which only occurred after four combined cases out of 23 with postop adverse events) relative to the rates of death, stroke, failure to wean from ventilator, reoperation, and intracranial hemorrhage indicates that our model is weighted heavily towards more “major” complications. Furthermore, organ/space SSI was the most frequent SSI subtype, which is not a minor complication per se in the authors’ opinions given that osteomyelitis, ventriculitis, meningitis, and intracranial abscesses are included in the organ/space SSI definition.
TL [8,10,17,19,31,32], highlighting the likely underutilization of TL for refractory TLE given that TLE has been shown to be superior to medical therapy alone in randomized trials [32]. Implementation of measures to prevent complications after TL based on these data could reduce the rate of postoperative adverse events. We observed that prior history of stroke and chronic steroid use were independent risk factors for developing any adverse event within 30 days of TL surgery. Diabetes mellitus approached but did not reach independent significance (p = 0.078) as well. Therefore, proper risk stratification and counseling for patients with prior history of stroke, chronic steroid use, and diabetes is warranted. Furthermore, controlling modifiable risk factors (e.g., glucose control or hemoglobin A1C for diabetics, evaluating whether steroid tapering for steroid chronic steroid users is possible, and assessing anticoagulation/antiplatelet therapy modification for patients with history of stroke) could lead to lower rates of complications after TL surgery in patients with these risk factors. The rate of infectious complications was found to be very low at 2%. As with all neurosurgical cases, attention to protocols to minimize postoperative infection is warranted. Finally, an inverse relationship between hospital volume and TL complication rates has been reported [31]. Therefore, referring physicians should refer TL candidates to high volume neurosurgical centers if possible. Despite the relatively low rates of adverse events after TL, neurosurgeons should counsel patients that TL can result in both serious neurologic and medical/surgical complications. Since providers and patients may be more likely to emphasize seizure outcomes after TL surgery, one must remain cognizant of the possibility of aforementioned non-seizure outcomes as well. These acute non-seizure complications still have significant potential to negatively affect quality of life, length of hospital stay, and cost of care even if long term health is not affected.
4.3. Prior literature There have been relatively few studies reporting non-seizure complications following TL. One systemic review reported rates of major and minor complications following procedures for invasive seizure monitoring or resective surgery, finding that the rates of minor medical complications, major medical complications, and mortality following TL were 5.4%,1.6%, and 0.4%, respectively [16]. An analysis of the Nationwide Inpatient Sample (NIS) dataset by McClelland et al. (2011) found that morbidity after anterior TL was 10.8%; there was no mortality [17]. Another analysis of the NIS database found that hospital volume influences TL complication rates, with low, medium, and high volume centers experiencing complications in 12.4%, 10.0%, and 6.0% of cases, respectively [31]. The limitations of these studies, particularly in regards to the NIS database and the systemic review by Hader et al., are well established and discussed above. A single center series of 200 TL cases reported 6 late deaths with no 30-day deaths, major (permanent and/or severe) neurological complication rates of 8%, and minor (temporary or not severe) complication rates of 45.5% [10]. Another single center study of 140 TL cases reported three major non-seizure complications (2.1%) and 15 minor non-seizure complications (10.7%) [8]. Finally, a single center study of 58 patients reported non-seizure complications in 7 (12%) of patients, with infection being the most common complication (2.9%) [19]. In the only randomized controlled trial comparing TL vs. medical therapy for mesial temporal lobe epilepsy, TL was demonstrated to be superior in regards to seizure outcome, with a 10% rate of adverse effects in the surgical arm [32]. An often overlooked outcome after surgery is patient satisfaction; one study found that both side of resection and degree of seizure control correlated with patient satisfaction [37]. In another study of 57 patients who underwent TL, 84% of patients stated that they would undergo TL again at 17 year follow up [38]. Future study on patient satisfaction with TL in the context of non-seizure adverse effects may provide valuable information on how secondary surgical outcomes influence patient satisfaction. A systemic review on surgical candidacy for TL recommended that patients be evaluated for surgery as soon as they have failed two AEDs [39], based on the principle that earlier intervention provides better outcomes for TL, and delaying surgical candidacy evaluation until further signs or symptoms develop could lead to underutilization of TL. Based on the recommendations of the aforementioned studies and the aggregate data on low rates of morbidity and mortality, TL surgery is very likely underutilized for refractory epilepsy, especially given the relatively high rates of patient satisfaction.
4.2. Independent risk factors for morbidity and mortality The results from our multivariate analysis found prior history of stroke and chronic steroid use as risk factors for postoperative complications after TL. As indicated by the ROC analysis, the model achieved good predictive accuracy (AUC = 0.797). Surgeons, referring physicians, and prospective TL candidate patients should be aware of these postoperative morbidity risk factors prior to surgery. Prior history of stroke before surgery was an independent risk factor for postoperative adverse events. Of note, history of prior stroke was unknown for some patients (see footnote in Table 2). Patients with a known previous stroke may be more likely to be on anticoagulation or antiplatelet agents. However, medication data and data on the exact nature of prior stroke are not available in NSQIP, limiting a more thorough analysis of stroke history, etiology, and risk. Chronic steroid use was the only other independent risk factor for any adverse outcome. Steroid use has previously been associated with adverse outcomes in neurosurgery after spine and intracranial procedures [33,34]. Steroid use itself may not necessarily predispose to complications; rather, the underlying pathology requiring steroid use could be responsible. Nevertheless, patients who require chronic steroids may benefit from complication risk stratification prior to TL. Diabetes mellitus approached independent significance as a risk factor for adverse events following TL (p = 0.078). Diabetes is a well-recognized indicator for poor surgical candidacy and a known risk factor for postoperative complications following neurosurgery [35,36]. Although NSQIP does not track blood glucose or diabetes disease severity, diabetic TL surgical candidates may benefit from specific attention to perioperative glucose management. Older age also approached, but did not reach, independent significance as a risk factor for complication development. Complications such as UTI, pneumonia, and stroke, all of which are more likely to develop with older age regardless. There were 13 patients age 65 or older in this study (6.4% of total). Importantly, we analyzed age as a continuous variable rather than as an interval or dichotomous variable, which may lead to false negatives for 16
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Conflict of interest
4.4. Limitations
The authors have no relevant financial disclosures to report.
The limitations of the NSQIP dataset are multiple, particularly with regard to neurosurgery-specific and epilepsy-specific data. Furthermore, the lack of follow-up beyond 30 days likely results in a significant underestimation of longer-term TL surgery complications. The NSQIP dataset does not contain data on whether complications are related to the index procedure or not, which likely results in a significant overestimation of what we define as “complications” (i.e., that the adverse outcomes are in fact a result of the surgical procedure). Additionally, there are multiple neurologic outcomes which NSQIP does not capture, including cranial nerve palsies, seizure outcomes, memory impairment, subtle language difficulties and visual field deficits, which likely results in an underestimation of TL-specific complications. Psychological outcomes after TL such as IQ reduction have been reported previously [15] but are not recorded in the NSQIP database, further limiting morbidity assessment. The nature of many categorical variables in NSQIP limits a more thorough analysis of comorbidities and complications. For instance, the “Previous stroke” variables do not provide information on type of stroke, area of infarct, or exact nature of the resulting deficit (if present). Furthermore, stroke history was unknown for 80 of the 202 cases, potentially confounding our finding that prior stroke was an independent risk factor for complications. Specific procedural information such as side of surgery, operative technique, operative approach, etc. is not included in the NSQIP database. The postoperative diagnosis variables in NSQIP are coded using ICD-9 codes, which are often vague and do not provide more insight into underlying pathology or surgical indication. Preoperative laboratory values were included for this analysis if available, but the underlying cause of the laboratory abnormalities is not always clear, particularly due to the lack of medication information in NSQIP. Cases from NSQIP are sampled randomly, thus our analysis may not be representative of the complete TL patient population. Additionally, NSQIP does not contain hospital-identifying information, meaning that outlier institutions with extremely high or low rates of complications may have a significant but unappreciable effect on our analysis. This limitation is important given the previous literature highlighting the inverse relationship between hospital volume and TL complication rate [31]. However, a sample size above 200 cases for a surgical series on TL outcomes is rare, and will hopefully provide more robust population-based data than previously published single-institution experiences. Finally, there is risk of type I error due to the relatively small number of postoperative events relative to the number of predictor variables analyzed. This limitation is made evident by the large confidence interval ranges for some variables.
Acknowledgements The American College of Surgeons National Surgical Quality Improvement Program 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. Dr. Davis completed this work as a Women’s Leadership Council Clinical Scholar in the Department of Neurosurgery at the University of Alabama at Birmingham, supported by the Kaul Foundation. References [1] J. Engel Jr., Surgical treatment for epilepsy: too little, too late? JAMA 300 (2008) 2548–2550. [2] B.C. Jobst, G.D. Cascino, Resective epilepsy surgery for drug-resistant focal epilepsy: a review, JAMA 313 (2015) 285–293. [3] K.A. Bujarski, F. Hirashima, D.W. Roberts, et al., Long-term seizure, cognitive, and psychiatric outcome following trans-middle temporal gyrus amygdalohippocampectomy and standard temporal lobectomy, J. Neurosurg. 119 (2013) 16–23. [4] D.J. Englot, A.T. Lee, C. Tsai, et al., Seizure types and frequency in patients who fail temporal lobectomy for intractable epilepsy, Neurosurgery 73 (2013) 838–844. [5] A.A. Cohen-Gadol, B.G. Wilhelmi, F. Collignon, et al., Long-term outcome of epilepsy surgery among 399 patients with nonlesional seizure foci including mesial temporal lobe sclerosis, J. Neurosurg. 104 (2006) 513–524. [6] R.E. Elliott, R.J. Bollo, J.L. Berliner, et al., Anterior temporal lobectomy with amygdalohippocampectomy for mesial temporal sclerosis: predictors of long-term seizure control, J. Neurosurg. 119 (2) (2013) 261–272. [7] L.E. Jehi, D.C. Silveira, W. Bingaman, I. Najm, Temporal lobe epilepsy surgery failures: predictors of seizure recurrence, yield of reevaluation, and outcome following reoperation, J. Neurosurg. 113 (2010) 1186–1194. [8] L. Jutila, A. Immonen, E. Mervaala, et al., Long term outcome of temporal lobe epilepsy surgery: analyses of 140 consecutive patients, J. Neurol. Neurosurg. Psychiatry 73 (2002) 486–494. [9] S.A. Lee, S.B. Yim, Y.M. Lim, J.K. Kang, J.K. Lee, Factors predicting seizure outcome of anterior temporal lobectomy for patients with mesial temporal sclerosis, Seizure 15 (2006) 397–404. [10] E.A. Popovic, G.C. Fabinyi, G.A. Brazenor, S.F. Berkovic, P.F. Bladin, Temporal lobectomy for epilepsy –complications in 200 patients, J. Clin. Neurosci. 2 (1995) 238–244. [11] T. Srikijvilaikul, S. Lerdlum, S. Tepmongkol, S. Shuangshoti, C. Locharernkul, Outcome of temporal lobectomy for hippocampal sclerosis in older patients, Seizure 20 (2011) 276–279. [12] D.J. Englot, S.J. Han, J.D. Rolston, et al., Epilepsy surgery failure in children: a quantitative and qualitative analysis, J. Neurosurg. Pediatr. 14 (2014) 386–395. [13] D.J. Englot, J.D. Rolston, D.D. Wang, P.P. Sun, E.F. Chang, K.I. Auguste, Seizure outcomes after temporal lobectomy in pediatric patients, J. Neurosurg. Pediatr. 12 (August) (2013) 134–141. [14] D.J. Englot, M.J. Rutkowski, M.E. Ivan, et al., Effects of temporal lobectomy on consciousness-impairing and consciousness-sparing seizures in children, Childs Nerv. Syst. 29 (10) (2013) 1915–1922. [15] S. Vadera, V.R. Kshettry, P. Klaas, W. Bingaman, Seizure-free and neuropsychological outcomes after temporal lobectomy with amygdalohippocampectomy in pediatric patients with hippocampal sclerosis, J. Neurosurg. Pediatr. 10 (2012) 103–107. [16] W.J. Hader, J. Tellez-Zenteno, A. Metcalfe, et al., Complications of epilepsy surgery: a systematic review of focal surgical resections and invasive EEG monitoring, Epilepsia 54 (2013) 840–847. [17] S. McClelland 3rd, H. Guo, K.S. Okuyemi, Population-based analysis of morbidity and mortality following surgery for intractable temporal lobe epilepsy in the United States, Arch. Neurol. 68 (2011) 725–729. [18] K. Kaiboriboon, N. Schiltz, S.M. Koroukian, S.D. Lhatoo, M.Z. Koubeissi, Limitations of NIS database in evaluation of epilepsy surgery morbidity and mortality, Arch. Neurol. 68 (November) (2011) 1483. [19] H.I. Ipekdal, O. Karadas, E. Erdogan, Z. Gokcil, Spectrum of surgical complications of temporal lobe epilepsy surgery: a single − center study, Turk. Neurosurg. 21 (2011) 147–151. [20] American College of Surgeons, User Guide for the 2008 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug08. ashx . (Accessed June 14, 2016). [21] American College of Surgeons, User Guide for the 2009 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug09. ashx . (Accessed June 14, 2016). [22] American College of Surgeons, User Guide for the 2010 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug10. ashx . (Accessed June 14, 2016).
5. Conclusions We report rates of acute non-seizure morbidity and mortality following TL procedures using a validated national database. The rate of morbidity and mortality following temporal lobectomy is relatively low, especially considering the comparative severity of the complications reported here. The data from this analysis are largely consistent with complication rates previously reported in the literature. We have identified prior history of stroke and chronic steroid use as independent risk factors for the development of acute postoperative complications after TL. Other risk factors approaching, but not reaching, independent significance included diabetes and older age. Surgeons and referring physicians should be aware of these risk factors in order to appropriately counsel patients about surgical options.
Funding The authors have no financial disclosures to report. 17
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
perioperative events at US epilepsy centers, J. Neurosurg. 118 (2013) 169–174. [32] S. Wiebe, W.T. Blume, J.P. Girvin, M. Eliasziw, Effectiveness and efficiency of surgery for temporal lobe epilepsy study group. A randomized, controlled trial of surgery for temporal-lobe epilepsy, N. Engl. J. Med. 345 (2001) 311–318. [33] S.E. Emery, S. Akhavan, P. Miller, et al., Steroids and risk factors for airway compromise in multilevel cervical corpectomy patients: a prospective, randomized, double-blind study, Spine (Phila Pa 1976) 34 (2009) 229–232. [34] N. Alan, A. Seicean, S. Seicean, D. Neuhauser, E.C. Benzel, R.J. Weil, Preoperative steroid use and the incidence of perioperative complications in patients undergoing craniotomy for definitive resection of a malignant brain tumor, J. Clin. Neurosci. 22 (2015) 1413–1419. [35] L.B. Chambless, S.L. Parker, L. Hassam-Malani, M.J. McGirt, R.C. Thompson, Type 2 diabetes mellitus and obesity are independent risk factors for poor outcome in patients with high-grade glioma, J. Neurooncol. 106 (2012) 383–389. [36] E.T. Martin, K.S. Kaye, C. Knott, et al., Diabetes and risk of surgical site infection: a systematic review and meta-analysis, Infect. Control Hosp. Epidemiol. 37 (2016) 88–99. [37] R.E. Iachinski, M.S. de Meneses, C.A. Simão, S.F. da Rocha, F. de Oliveira Braga, P.A. Kowacs, Patient satisfaction with temporal lobectomy/selective amygdalohippocampectomy for temporal lobe epilepsy and its relationship with Engel classification and the side of lobectomy, Epilepsy Behav. 31 (2014) 377–380. [38] J.E. Jones, J.B. Blocher, D.C. Jackson, Life outcomes of anterior temporal lobectomy: serial long-term follow-up evaluations, Neurosurgery 73 (2013) 1018–1025. [39] C.S. Kwon, J. Neal, J. Telléz-Zenteno, et al., CASES Investigators. Resective focal epilepsy surgery − Has selection of candidates changed? A systematic review, Epilepsy Res. 122 (2016) 37–43.
[23] American College of Surgeons, User Guide for the 2011 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug11. ashx . (Accessed June 14, 2016). [24] American College of Surgeons, User Guide for the 2012 ACS NSQIP Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/ nsqip/ug12.ashx . (Accessed June 14, 2016). [25] American College of Surgeons, User Guide for the 2013 ACS NSQIP Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/ nsqip/2013_acs_nsqip_puf_user_guide.ashx . Accessed (June 14, 2016). [26] M. Shiloach, S.K. Frencher Jr., J.E. Steeger, et al., Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program, J. Am. Coll. Surg. 210 (2010) 6–16. [27] B.L. Hall, B.H. Hamilton, K. Richards, K.Y. Bilimoria, M.E. Cohen, C.Y. Ko, Does surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals, Ann. Surg. 250 (2009) 363–376. [28] M.M. Sellers, R.P. Merkow, A. Halverson, et al., Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program, J. Am. Coll. Surg. 216 (2013) 420–427. [29] A. Weiss, J.E. Anderson, D.C. Chang, Comparing the National Surgical Quality Improvement Program with the Nationwide Inpatient Sample Database, JAMA Surg. 150 (2015) 815–816. [30] University of Alabama at Birmingham Institutional Review Board, Is IRB Review Required for Use of Public Datasets? UAB Research, (2017) http://www.uab.edu/ research/administration/offices/IRB/FAQs/Pages/PublicDatasets.aspx . Accessed (October 1, 2015).. [31] D.J. Englot, D. Ouyang, D.D. Wang, J.D. Rolston, P.A. Garcia, E.F. Chang, Relationship between hospital surgical volume, lobectomy rates, and adverse
18
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Thirty-day non-seizure outcomes following temporal lobectomy for adult epilepsy Brandon A. Sherrod, Matthew C. Davis, Kristen O. Riley
MARK
⁎
Department of Neurological Surgery, University of Alabama at Birmingham, Birmingham, AL, USA
A R T I C L E I N F O
A B S T R A C T
Keywords: Epilepsy Temporal lobectomy NSQIP Morbidity Mortality Adverse event Complication
Objective: Multi-institutional rates of acute adverse outcomes other than seizures after temporal lobectomy (TL) are not well understood. Here we analyzed short-term morbidity and mortality following TL using a validated national database. Patients and methods: The multi-institutional American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was queried by Current Procedural Terminology (CPT) code for TL procedures performed for adult patients with diagnoses related to epilepsy from 2008 to 2014. Patient demographics, operative variables, hospital variables, preoperative laboratory values, and preexisting comorbidities were analyzed using univariate and multivariate techniques to determine associations with 30-day postoperative morbidity and mortality. Results: A total of 202 TL procedures were analyzed, 80 (39.6%) with intraoperative electrocorticography (ECOG) and 122 (60.4%) without ECOG. Mean age was 40.4 ± 13.7 years, and 47.5% of patients were male. Overall morbidity and mortality were 11.4% and 2.0%, respectively. The most common adverse outcomes were reoperation (5.4%), stroke with residual deficit (2.5%), failure to wean from ventilator (2.0%), and surgical site infection (2.0%). Adverse event rates were not significantly different between TLs with and without ECOG (13.1% vs. 8.8%, p = 0.375). Independent predictors of adverse events included prior stroke (OR 7.60, 95% CI 1.22–47.17, p = 0.029) and chronic steroid use (OR 10.90, 95% CI 1.03–115.79, p = 0.048). Diabetes mellitus (p = 0.078) and older age (p = 0.145) approached but did not reach significance in the multivariate model. Conclusions: We report rates of acute morbidity and mortality following TL procedures using a national database. These findings can be used both to assist with patient selection as well as patient counseling prior to surgery.
1. Introduction
medical complications, major medical complications, and mortality following TL was 5.4%, 1.6%, and 0.4%, respectively [16]. However, Hader et al. (2013) recognized the limitations of lacking prospectively collected, standardized complication data. Furthermore, follow-up period requirements were not standardized across studies in the review. An analysis of the Nationwide Inpatient Sample (NIS) dataset found that morbidity after anterior TL was 10.8%, with no mortality [17]. However, the NIS is an administrative, not clinical, dataset, and the International Classification of Diseases, Ninth Revision [ICD-9] codes used in the NIS study are not specific to TL [18]. Most other studies on TL outcomes focus on seizure outcomes and are limited to single-center experiences with relatively small case series [8,10,19]. Furthermore, many of the previously cited studies lack data on complications considered to be “minor” such as urinary tract infection, wound disruption, etc. Minor non-seizure complications still have significant potential to negatively affect quality of life, length of hospital stay, and cost of care
Refractory temporal lobe epilepsy (TLE) is the most common cause of pharmacoresistant seizures [1,2]. Temporal lobectomy (TL) is performed after exhausting most other efforts to control refractory TLE. Given that seizure control is the ultimate goal of TL, numerous studies have investigated seizure outcomes after TL in adults [3–11] and children [12–15]. However, few studies have investigated other (non-seizure) postoperative outcomes after TL, especially using multicenter data. Due to the recognized underutilization of TL for refractory TLE, understanding the full extent of TL complications is paramount in making informed decisions prior to TL surgery [1,2]. A review of the literature on major and minor complications following TL is displayed in Table 1. Hader et al. (2013) reviewed rates of major and minor complications following procedures for invasive seizure monitoring or resective surgery, finding that the rates of minor ⁎
Corresponding author at: Department of Neurosurgery, University of Alabama at Birmingham, 1720 2nd Avenue S., FOT 1008, Birmingham, AL 35294-3410, USA. E-mail address: [email protected] (K.O. Riley).
http://dx.doi.org/10.1016/j.clineuro.2017.05.027 Received 16 January 2017; Received in revised form 3 May 2017; Accepted 29 May 2017 Available online 08 June 2017 0303-8467/ © 2017 Elsevier B.V. All rights reserved.
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
excluded from the analysis if the NSQIP postoperative diagnosis was not epilepsy-related. Trauma cases (e.g., hematoma evacuation) and patients under the age of 18 are automatically excluded from the NSQIP dataset and were not included for this analysis. The University of Alabama at Birmingham Institutional Review Board (IRB) does not require IRB approval of NSQIP studies because NSQIP data are de-identified and considered a public dataset [30].
Table 1 Literature reports of major and minor complications after TL surgery in adults. Major Complicationsb
Minor Complicationsb
Complication
Rates, %
Citation
Complication
Rates, %
Citation
Death
0.4
[16]
Memory impairment (mild)
3.2
[16]
0.0 3.0
[17,19] [10]
1.2
[17]
Hemorrhage or hematoma
2.0
[16]
Infectionc
0.8 1.9 2.0 2.9
[17] [16] [10] [19]
Visual field defect
1.1 3.0
[17] [10]
Hydrocephalus
1.3 0.4
[16] [17]
DVT or PE
0.0 0.7
[17] [16]
2.2. Variables of interest Psychiatrica
5.8
[16]
Dysphagia
2.1 5.5
[16] [10]
Patient demographics included age, gender, race, and body mass index (BMI). Patient comorbidities included obesity, diabetes, tobacco use within the previous year, hypertension requiring medication, chronic steroid use, previous stroke with/without deficit, and preoperative sepsis. Preoperative laboratory values included hypernatremia (Na > 145 mEq/L), hyponatremia (Na < 135 mEq/L), thrombocytopenia (< 150k platelets/μL), leukocytosis (> 11k WBC/ μL), elevated creatinine (> 1.2 mg/dL), and anemia (hematocrit < 36% for females; < 40% for males). Lab values were only included if at least 85% of cases had recorded values (e.g., AST and ALT laboratory values are available in NSQIP, but less than 85% of patients in our cohort had these lab values recorded). Postoperative diagnoses were determined by ICD-9 codes. Hospital variables included length of stay, time from operation to discharge, time from admission to operation, and prior operation within 30 days. Operative variables included operative time, emergent operation status, American Society of Anesthesiologist (ASA) class, perioperative blood transfusion, and wound classification. Complications are tracked by NSQIP for up to 30 days postoperatively; follow-up beyond the 30-day window is not recorded. Complications for this study included reoperation, death, bleeding requiring transfusion, subdural hemorrhage, intracerebral hemorrhage, surgical site infection (superficial incisional, deep incisional, or organ/ space [includes osteomyelitis, ventriculitis, meningitis, and intracranial abscesses]), pneumonia, failure to wean from ventilator, urinary tract infection (UTI), deep venous thrombosis (DVT), pulmonary embolism (PE), sepsis/septic shock, and stroke with resultant neurological deficit. NSQIP complication coding follows strict criteria for entry and are coded by trained data abstractors with minimal (< 2%) inter-rater disagreement; full definitions for complication coding criteria are provided in the referenced ACS NSQIP User Guides [20–25].
DVT = deep venous thrombosis; PE = pulmonary embolism. a Psychiatric includes personality changes, etc. otherwise not attributable to cognitive decline or memory changes alone. b Definitions of “major” and minor vary by study. Here, we have classified major and minor complication by relative severity, not acuity or duration. Time to complications, whether acute or chronic, was not distinguished.. c Minor wound infections were excluded.
even if long term health is not affected. There is a need for a nationwide, multicenter study on rates of complications following TL surgery. The purpose of the current study was to investigate non-seizure postoperative complications following TL surgery using a nationwide, multi-institutional clinical database.
2. Material and methods 2.1. Data acquisition The American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) was initiated in 2005 to identify areas of quality improvement in surgery. The NSQIP database contains millions of procedures across hundreds of hospitals primarily located in the United States and Canada, including academic and non-academic centers [20–25]. The ACS trains clinical data abstractors at each NSQIP site and employs strict data collection and case exclusion criteria, with inter-rater reliability rates at 98% or greater [20–26]. Researchers and clinicians at NSQIP participating hospitals may access the entire dataset, which is stripped of any patient or hospital identifying information. The ACS NSQIP adult dataset was obtained for years 2008–2014. The number of hospitals participating in each database year are: 211 hospitals (2008), 237 hospitals (2009), 258 hospitals (2010), 315 hospitals (2011), 374 hospitals (2012), 435 hospitals (2013), and 517 hospitals (2014) [20–25]. The accuracy (in comparison to chart review), follow up, ability to improve outcomes for participating hospitals, and inter-rater reliability of this database have been validated by several independent studies and internal ACS audits [26–29]. Procedures performed by a neurosurgeon were filtered from the main dataset by querying the surgical subspecialty variable for “Neurosurgery”. Temporal lobectomy procedures were filtered from the neurosurgeryonly dataset by Current Procedural Terminology (CPT) codes: 61537 (craniotomy; temporal lobectomy, without intraoperative ECOG) and 61538 (craniotomy; temporal lobectomy, with ECOG). Cases were
2.3. Statistical analysis Univariate analyses of the association between overall morbidity and patient demographics, preexisting comorbidities, abnormal lab values, and operative factors were performed using Chi-square test, Fisher’s exact test, binary logistic regression, or independent sample Student’s t-test where appropriate. The alpha value for significance was set at 0.05. Variables that reached significance in the univariate analyses were then entered into a multivariate logistic regression model. An area under the curve (AUC) analysis for a receiver operating characteristic (ROC) curve was performed to determine predictive accuracy of the multivariate model. All statistical analyses were performed using SPSS Version 23.0 (IBM Corp, Armonk, NY, 2015). 3. Results 3.1. Patient demographics, comorbidities, and laboratory values A total of 202 procedures were performed, 80 (39.6%) with intraoperative ECOG and 122 (60.4%) without intraoperative ECOG. Patient demographics, comorbidities, and laboratory values are shown in Table 2. Age ranged from 18 to 78 years, with 13 patients age 65 or older in this study (6.4% of total). There was essentially an even gender distribution (47.5% male). The most common comorbidities were 13
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Table 2 Patient demographics, preexisting comorbidities, and preoperative laboratory values. Parameter
Mean ± SD
Age (years) BMI (kg/m2)a
40.4 ± 13.7 29.2 ± 7.2
Parameter
Table 3 Hospital stay and operative variables.
n (%)
a
Gender Male Female
96 (47.5) 105 (52.0)
Race White Black Asian Native American Unknown
164 (81.2) 9 (4.5) 6 (3.0) 3 (1.5) 20 (9.9)
Body weight Underweight (BMI < 18.5) Class I obesity (BMI 30–34.9) Class II obesity (BMI 35–39.9) Class III obesity (BMI ≥ 40)
1 (0.5) 43 (21.3) 16 (7.9) 18 (8.9)
Diabetes History of tobacco use in previous year Hypertension requiring medication Chronic steroid use Bleeding disorder Previous stroke with deficitb Previous stoke without deficitb Preoperative sepsis Hypernatremia (Na > 145 mEq/L) Hyponatremia (Na < 135 mEq/L) Thrombocytopenia (< 150,000 platelets/μL) Leukocytosis (> 11,000 WBC/μL) Elevated serum creatinine (> 1.2 mg/dL) Anemia (< 36% Hct for females; < 40% Hct for males)
4 (2.0) 42 (20.8) 33 (16.3) 4 (2.0) 4 (2.0) 6 (3.0) 1 (0.5) 7 (3.5) 3 (1.5) 16 (7.9) 2 (1.0) 18 (8.9) 5 (2.5) 30 (14.9)
Parameter
n (%) or mean ± SD
Total Length of stay (days) Total length of stay > 30 daysa Time from operation to discharge (days) Time from admission to operation (days) Length of operation (minutes)
5.4 ± 5.1 2 (0.8) 4.1 ± 3.5 1.3 ± 3.7 247.5 ± 83.1
ASA Classification, n (%) ASA Class I ASA Class II ASA Class III ASA Class IV ASA Class V Preoperative blood transfusion, n (%)
0 (0.0) 100 (47.5) 95 (47.0) 6 (3.0) 1 (0.5) 2 (0.8)
Wound classification, n (%) Clean Clean-contaminated Contaminated Dirty/infected
198 (98.0) 3 (1.5) 1 (0.5) 0 (0.0)
ASA = American Society of Anesthesiologists. a Two patients had total length of stay (including preoperative and postoperative hospital stay) greater than 30 days (32 and 35 days), but no patient had a postoperative length of stay greater than 30 days. Table 4 Postoperative 30-day temporal lobectomy complications.
SD = standard deviation; IQR = interquartile range; BMI = body mass index; WBC = white blood cell; Hct = hematocrit. a Data missing for 1 case. b Stroke history unknown for 80 cases.
hypertension, obesity, and tobacco use. Although NSQIP does not include a thorough neurological comorbidity assessment, variables such as prior stroke are included. The most common abnormal preoperative laboratory values were anemia, leukocytosis, and hyponatremia. Elevated serum creatinine, thrombocytopenia, and hypernatremia were all relatively rare (occurring in less than 3% of cases). Importantly, not all patients had preoperative laboratory values measured or recorded; however, lab values were only included if at least 85% of cases had recorded values.
Outcome
n (%)
Time to Complication in days, median (range)a
Any adverse outcome Reoperation Stroke with deficit Ventilator requirement > 48 h Death Organ/space surgical site infection Urinary tract infection Sepsis Subarachnoid hemorrhage Pneumonia Bleeding requiring transfusion Deep venous thrombosis Wound disruption Intracerebral hemorrhage Deep incisional surgical site infection Pulmonary embolism Superficial incisional surgical site infection
23 (11.4) 11 (5.4) 5 (2.5) 4 (2.0) 4 (2.0) 4 (2.0) 3 (1.5) 3 (1.5) 2 (1.0) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5)
4 (0–28) 1 (0–9) 1 (0–5) 4.5 (1–9) 9 (3–19) 5 (4–28) 10 (4–23) 4.5 (4–5) Unknown 3 (3–3) 2 (2–2) 6 (6–6) 15 (15–15) Unknown 15 (15–15)
0 (0.0) 0 (0.0)
– –
a For the complications with days to event data. Not all reoperations had days to event data, and none of the hemorrhages had days to event data.
reoperation (5.4%), stroke with deficit (2.5%), organ space/surgical site infection (2.0%), and intracranial hemorrhage (1.5%). The most common medical complications were failure to wean from ventilator (2.0%), UTI (1.5%), sepsis (1.5%), and pneumonia (0.5%). Complications occurred at varying postoperative times, with a range from 0 to 28 days overall. It is important to recognize that the NSQIP reoperation variable does not capture whether the reoperation was related to the index procedure. Therefore, these results should be interpreted with caution.
3.2. Hospital and operative variables Hospital and operative factors may be seen in Table 3. Substantial variation was observed for hospital length of stay, time from operation to discharge, time from admission to operation, and length of operation. Since NSQIP does not track outcomes after 30-days, length of hospital stay results are likely missing data on patients who had substantially longer and complicated hospital stays. 3.3. Thirty-day complication rates
3.4. Predictors of thirty-day complications
Thirty-day complications are shown in Table 4. Adverse events occurred after 23 cases (11.4%).Ooverall mortality frequency was 4 (2.0%). Complication rates were not significantly different between TLs with and without ECOG (13.1% without ECOG vs. 8.8% with ECOG, p = 0.375). The most common neurosurgical complications were
Predictors of 30-day complications via univariate and multivariate analyses are displayed in Table 5. Independent risk factors for any adverse event included history of prior stroke and chronic steroid use. Older age (p = 0.145) and diabetes (p = 0.078) approached but did not reach independent significance in the multivariate model. Variables 14
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Table 5 Predictors of 30-day adverse eventsb following TL. Variables in bold represent significance (p < 0.05). Parameter
Univariate OR (95% CI)
Procedure TL without EEG TL with EEG
Ref 0.635 (0.249–1.621)
Adjusted OR (95% CI)
P
Age (years) Gender Male Female
1.053(1.019–1.087)a
1.034 (0.989–1.080)
0.145
2.269 (0.916–5.620) Ref
2.144 (0.747–6.158)
0.156
Race Minority Non-minority
0.898 (0.287–2.810) Ref
ASA Classification ASA Class I–II ASA Class III–V
Ref 2.472 (0.970–6.298)
1.642 (0.540–5.000)
0.382
Body Weight Class I obesity (BMI 30–34.9) Class II obesity (BMI 35–39.9) Class III obesity (BMI ≥ 40)
1.031 (0.359–2.956) 1.122 (0.238–5.286) 2.481 (0.741–8.307)
1.019 (0.166–6.251)
0.983
1.728 (0.471–6.333) 10.896 (1.025–115.893)
0.409 0.048
9.108 (0.783–105.983) 7.598 (1.224–47.165)
0.078 0.029
4.917 (0.427–56.696)
0.202
Hypertension Chronic steroid use Smoker in last year Diabetes Prior stroke Preoperative sepsis Hypernatremia (Na > 145 mEq/L) Hyponatremia (Na < 135 mEq/L) Thrombocytopenia (< 150,000 platelets/μL) Leukocytosis (> 11,000 WBC/μL) Elevated serum creatinine (> 1.2 mg/dL) Anemia (< 36% Hct for females; < 40% Hct for males) Operative time (hours; interval) 0–3 h 3–6 h 6+ hours Operative time (hours; continuous) Preoperative length of stay in days
4.152 (1.620–10.642) 8.429 (1.128–62.999) 1.800 (0.688–4.710) 8.429 (1.128–62.999) 12.351 (2.570–59.366) 3.314 (0.605–18.164) 4.023 (0.350–46.199) 0.497 (0.063–3.949) 8.091 (0.489–133.983) 1.640 (0.437–6.162) 13.275 (2.092–84.255) 0.514 (0.114–2.314) Ref 0.391 0.408 0.755 0.989
(0.149–1.024) (0.078–2.127) (0.534–1.067)a (0.872–1.122)
TL = temporal lobectomy; EEG = electrocorticography; ASA = American Society of Anesthesiologists; BMI = body mass index; WBC = white blood cell; Hct = hematocrit. a Per unit increase. b Includes all complications and mortality.
significant in the univariate analysis but not the multivariate analysis included male gender, elevated preoperative serum creatinine, diabetes, hypertension, class III obesity, ASA class 3 or greater, and older age. The results of the multivariate model ROC curve yielded an AUC, or C-statistic, of 0.797 (95% CI 0.69–0.90, p < 0.001). Fig. 1 displays the multivariate model ROC curve. 4. Discussion Temporal lobectomy is a well-established treatment option for patients with TLE. However, national rates of postoperative TL complications unrelated to seizures have not been well reported using standardized clinical data. Here, we report recent, standardized, nationwide data from the ACS NSQIP dataset on acute non-seizure complications following TL. 4.1. Complications The overall 30-day morbidity following TL was 11.4%, and the overall 30-day mortality following TL was 2.0%. The reported morbidity rate includes all complications, including relatively minor complications such as UTI and superficial wound infection. Furthermore, the rates of serious complications (e.g., death, organ/space SSI, intracranial hemorrhage) were relatively low − rates of these types of complications were less than 2%. These findings validate previous studies demonstrating relatively low rates of adverse events following
Fig. 1. Receiver Operating Characteristic (ROC) curve for multivariate model. The diagonal line represents the predictive power of chance alone. Area under the curve (AUC) was 0.797 (95% CI = 0.693–0.900, p < 0.001).
15
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
substantially different age groups. Previous studies have reported higher rates of complications in older patients after TL [11]. Older patients should be thoroughly assessed prior to TL surgery and counseled adequately on the increased risk of acute postoperative complications. The authors recognize that a separate morbidity and mortality analysis would be useful. The statistical significance of a model predicting an event (death) that only occurred in 4 patients would be questionable however. We are specifically concerned about type II statistical error, or failure to report a truly significant variable due to underpowered sampling. As a result, we elected to combine all adverse events into a single outcome measure. The low frequency of “minor” complications (such as UTI and wound disruption, which only occurred after four combined cases out of 23 with postop adverse events) relative to the rates of death, stroke, failure to wean from ventilator, reoperation, and intracranial hemorrhage indicates that our model is weighted heavily towards more “major” complications. Furthermore, organ/space SSI was the most frequent SSI subtype, which is not a minor complication per se in the authors’ opinions given that osteomyelitis, ventriculitis, meningitis, and intracranial abscesses are included in the organ/space SSI definition.
TL [8,10,17,19,31,32], highlighting the likely underutilization of TL for refractory TLE given that TLE has been shown to be superior to medical therapy alone in randomized trials [32]. Implementation of measures to prevent complications after TL based on these data could reduce the rate of postoperative adverse events. We observed that prior history of stroke and chronic steroid use were independent risk factors for developing any adverse event within 30 days of TL surgery. Diabetes mellitus approached but did not reach independent significance (p = 0.078) as well. Therefore, proper risk stratification and counseling for patients with prior history of stroke, chronic steroid use, and diabetes is warranted. Furthermore, controlling modifiable risk factors (e.g., glucose control or hemoglobin A1C for diabetics, evaluating whether steroid tapering for steroid chronic steroid users is possible, and assessing anticoagulation/antiplatelet therapy modification for patients with history of stroke) could lead to lower rates of complications after TL surgery in patients with these risk factors. The rate of infectious complications was found to be very low at 2%. As with all neurosurgical cases, attention to protocols to minimize postoperative infection is warranted. Finally, an inverse relationship between hospital volume and TL complication rates has been reported [31]. Therefore, referring physicians should refer TL candidates to high volume neurosurgical centers if possible. Despite the relatively low rates of adverse events after TL, neurosurgeons should counsel patients that TL can result in both serious neurologic and medical/surgical complications. Since providers and patients may be more likely to emphasize seizure outcomes after TL surgery, one must remain cognizant of the possibility of aforementioned non-seizure outcomes as well. These acute non-seizure complications still have significant potential to negatively affect quality of life, length of hospital stay, and cost of care even if long term health is not affected.
4.3. Prior literature There have been relatively few studies reporting non-seizure complications following TL. One systemic review reported rates of major and minor complications following procedures for invasive seizure monitoring or resective surgery, finding that the rates of minor medical complications, major medical complications, and mortality following TL were 5.4%,1.6%, and 0.4%, respectively [16]. An analysis of the Nationwide Inpatient Sample (NIS) dataset by McClelland et al. (2011) found that morbidity after anterior TL was 10.8%; there was no mortality [17]. Another analysis of the NIS database found that hospital volume influences TL complication rates, with low, medium, and high volume centers experiencing complications in 12.4%, 10.0%, and 6.0% of cases, respectively [31]. The limitations of these studies, particularly in regards to the NIS database and the systemic review by Hader et al., are well established and discussed above. A single center series of 200 TL cases reported 6 late deaths with no 30-day deaths, major (permanent and/or severe) neurological complication rates of 8%, and minor (temporary or not severe) complication rates of 45.5% [10]. Another single center study of 140 TL cases reported three major non-seizure complications (2.1%) and 15 minor non-seizure complications (10.7%) [8]. Finally, a single center study of 58 patients reported non-seizure complications in 7 (12%) of patients, with infection being the most common complication (2.9%) [19]. In the only randomized controlled trial comparing TL vs. medical therapy for mesial temporal lobe epilepsy, TL was demonstrated to be superior in regards to seizure outcome, with a 10% rate of adverse effects in the surgical arm [32]. An often overlooked outcome after surgery is patient satisfaction; one study found that both side of resection and degree of seizure control correlated with patient satisfaction [37]. In another study of 57 patients who underwent TL, 84% of patients stated that they would undergo TL again at 17 year follow up [38]. Future study on patient satisfaction with TL in the context of non-seizure adverse effects may provide valuable information on how secondary surgical outcomes influence patient satisfaction. A systemic review on surgical candidacy for TL recommended that patients be evaluated for surgery as soon as they have failed two AEDs [39], based on the principle that earlier intervention provides better outcomes for TL, and delaying surgical candidacy evaluation until further signs or symptoms develop could lead to underutilization of TL. Based on the recommendations of the aforementioned studies and the aggregate data on low rates of morbidity and mortality, TL surgery is very likely underutilized for refractory epilepsy, especially given the relatively high rates of patient satisfaction.
4.2. Independent risk factors for morbidity and mortality The results from our multivariate analysis found prior history of stroke and chronic steroid use as risk factors for postoperative complications after TL. As indicated by the ROC analysis, the model achieved good predictive accuracy (AUC = 0.797). Surgeons, referring physicians, and prospective TL candidate patients should be aware of these postoperative morbidity risk factors prior to surgery. Prior history of stroke before surgery was an independent risk factor for postoperative adverse events. Of note, history of prior stroke was unknown for some patients (see footnote in Table 2). Patients with a known previous stroke may be more likely to be on anticoagulation or antiplatelet agents. However, medication data and data on the exact nature of prior stroke are not available in NSQIP, limiting a more thorough analysis of stroke history, etiology, and risk. Chronic steroid use was the only other independent risk factor for any adverse outcome. Steroid use has previously been associated with adverse outcomes in neurosurgery after spine and intracranial procedures [33,34]. Steroid use itself may not necessarily predispose to complications; rather, the underlying pathology requiring steroid use could be responsible. Nevertheless, patients who require chronic steroids may benefit from complication risk stratification prior to TL. Diabetes mellitus approached independent significance as a risk factor for adverse events following TL (p = 0.078). Diabetes is a well-recognized indicator for poor surgical candidacy and a known risk factor for postoperative complications following neurosurgery [35,36]. Although NSQIP does not track blood glucose or diabetes disease severity, diabetic TL surgical candidates may benefit from specific attention to perioperative glucose management. Older age also approached, but did not reach, independent significance as a risk factor for complication development. Complications such as UTI, pneumonia, and stroke, all of which are more likely to develop with older age regardless. There were 13 patients age 65 or older in this study (6.4% of total). Importantly, we analyzed age as a continuous variable rather than as an interval or dichotomous variable, which may lead to false negatives for 16
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
Conflict of interest
4.4. Limitations
The authors have no relevant financial disclosures to report.
The limitations of the NSQIP dataset are multiple, particularly with regard to neurosurgery-specific and epilepsy-specific data. Furthermore, the lack of follow-up beyond 30 days likely results in a significant underestimation of longer-term TL surgery complications. The NSQIP dataset does not contain data on whether complications are related to the index procedure or not, which likely results in a significant overestimation of what we define as “complications” (i.e., that the adverse outcomes are in fact a result of the surgical procedure). Additionally, there are multiple neurologic outcomes which NSQIP does not capture, including cranial nerve palsies, seizure outcomes, memory impairment, subtle language difficulties and visual field deficits, which likely results in an underestimation of TL-specific complications. Psychological outcomes after TL such as IQ reduction have been reported previously [15] but are not recorded in the NSQIP database, further limiting morbidity assessment. The nature of many categorical variables in NSQIP limits a more thorough analysis of comorbidities and complications. For instance, the “Previous stroke” variables do not provide information on type of stroke, area of infarct, or exact nature of the resulting deficit (if present). Furthermore, stroke history was unknown for 80 of the 202 cases, potentially confounding our finding that prior stroke was an independent risk factor for complications. Specific procedural information such as side of surgery, operative technique, operative approach, etc. is not included in the NSQIP database. The postoperative diagnosis variables in NSQIP are coded using ICD-9 codes, which are often vague and do not provide more insight into underlying pathology or surgical indication. Preoperative laboratory values were included for this analysis if available, but the underlying cause of the laboratory abnormalities is not always clear, particularly due to the lack of medication information in NSQIP. Cases from NSQIP are sampled randomly, thus our analysis may not be representative of the complete TL patient population. Additionally, NSQIP does not contain hospital-identifying information, meaning that outlier institutions with extremely high or low rates of complications may have a significant but unappreciable effect on our analysis. This limitation is important given the previous literature highlighting the inverse relationship between hospital volume and TL complication rate [31]. However, a sample size above 200 cases for a surgical series on TL outcomes is rare, and will hopefully provide more robust population-based data than previously published single-institution experiences. Finally, there is risk of type I error due to the relatively small number of postoperative events relative to the number of predictor variables analyzed. This limitation is made evident by the large confidence interval ranges for some variables.
Acknowledgements The American College of Surgeons National Surgical Quality Improvement Program 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. Dr. Davis completed this work as a Women’s Leadership Council Clinical Scholar in the Department of Neurosurgery at the University of Alabama at Birmingham, supported by the Kaul Foundation. References [1] J. Engel Jr., Surgical treatment for epilepsy: too little, too late? JAMA 300 (2008) 2548–2550. [2] B.C. Jobst, G.D. Cascino, Resective epilepsy surgery for drug-resistant focal epilepsy: a review, JAMA 313 (2015) 285–293. [3] K.A. Bujarski, F. Hirashima, D.W. Roberts, et al., Long-term seizure, cognitive, and psychiatric outcome following trans-middle temporal gyrus amygdalohippocampectomy and standard temporal lobectomy, J. Neurosurg. 119 (2013) 16–23. [4] D.J. Englot, A.T. Lee, C. Tsai, et al., Seizure types and frequency in patients who fail temporal lobectomy for intractable epilepsy, Neurosurgery 73 (2013) 838–844. [5] A.A. Cohen-Gadol, B.G. Wilhelmi, F. Collignon, et al., Long-term outcome of epilepsy surgery among 399 patients with nonlesional seizure foci including mesial temporal lobe sclerosis, J. Neurosurg. 104 (2006) 513–524. [6] R.E. Elliott, R.J. Bollo, J.L. Berliner, et al., Anterior temporal lobectomy with amygdalohippocampectomy for mesial temporal sclerosis: predictors of long-term seizure control, J. Neurosurg. 119 (2) (2013) 261–272. [7] L.E. Jehi, D.C. Silveira, W. Bingaman, I. Najm, Temporal lobe epilepsy surgery failures: predictors of seizure recurrence, yield of reevaluation, and outcome following reoperation, J. Neurosurg. 113 (2010) 1186–1194. [8] L. Jutila, A. Immonen, E. Mervaala, et al., Long term outcome of temporal lobe epilepsy surgery: analyses of 140 consecutive patients, J. Neurol. Neurosurg. Psychiatry 73 (2002) 486–494. [9] S.A. Lee, S.B. Yim, Y.M. Lim, J.K. Kang, J.K. Lee, Factors predicting seizure outcome of anterior temporal lobectomy for patients with mesial temporal sclerosis, Seizure 15 (2006) 397–404. [10] E.A. Popovic, G.C. Fabinyi, G.A. Brazenor, S.F. Berkovic, P.F. Bladin, Temporal lobectomy for epilepsy –complications in 200 patients, J. Clin. Neurosci. 2 (1995) 238–244. [11] T. Srikijvilaikul, S. Lerdlum, S. Tepmongkol, S. Shuangshoti, C. Locharernkul, Outcome of temporal lobectomy for hippocampal sclerosis in older patients, Seizure 20 (2011) 276–279. [12] D.J. Englot, S.J. Han, J.D. Rolston, et al., Epilepsy surgery failure in children: a quantitative and qualitative analysis, J. Neurosurg. Pediatr. 14 (2014) 386–395. [13] D.J. Englot, J.D. Rolston, D.D. Wang, P.P. Sun, E.F. Chang, K.I. Auguste, Seizure outcomes after temporal lobectomy in pediatric patients, J. Neurosurg. Pediatr. 12 (August) (2013) 134–141. [14] D.J. Englot, M.J. Rutkowski, M.E. Ivan, et al., Effects of temporal lobectomy on consciousness-impairing and consciousness-sparing seizures in children, Childs Nerv. Syst. 29 (10) (2013) 1915–1922. [15] S. Vadera, V.R. Kshettry, P. Klaas, W. Bingaman, Seizure-free and neuropsychological outcomes after temporal lobectomy with amygdalohippocampectomy in pediatric patients with hippocampal sclerosis, J. Neurosurg. Pediatr. 10 (2012) 103–107. [16] W.J. Hader, J. Tellez-Zenteno, A. Metcalfe, et al., Complications of epilepsy surgery: a systematic review of focal surgical resections and invasive EEG monitoring, Epilepsia 54 (2013) 840–847. [17] S. McClelland 3rd, H. Guo, K.S. Okuyemi, Population-based analysis of morbidity and mortality following surgery for intractable temporal lobe epilepsy in the United States, Arch. Neurol. 68 (2011) 725–729. [18] K. Kaiboriboon, N. Schiltz, S.M. Koroukian, S.D. Lhatoo, M.Z. Koubeissi, Limitations of NIS database in evaluation of epilepsy surgery morbidity and mortality, Arch. Neurol. 68 (November) (2011) 1483. [19] H.I. Ipekdal, O. Karadas, E. Erdogan, Z. Gokcil, Spectrum of surgical complications of temporal lobe epilepsy surgery: a single − center study, Turk. Neurosurg. 21 (2011) 147–151. [20] American College of Surgeons, User Guide for the 2008 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug08. ashx . (Accessed June 14, 2016). [21] American College of Surgeons, User Guide for the 2009 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug09. ashx . (Accessed June 14, 2016). [22] American College of Surgeons, User Guide for the 2010 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug10. ashx . (Accessed June 14, 2016).
5. Conclusions We report rates of acute non-seizure morbidity and mortality following TL procedures using a validated national database. The rate of morbidity and mortality following temporal lobectomy is relatively low, especially considering the comparative severity of the complications reported here. The data from this analysis are largely consistent with complication rates previously reported in the literature. We have identified prior history of stroke and chronic steroid use as independent risk factors for the development of acute postoperative complications after TL. Other risk factors approaching, but not reaching, independent significance included diabetes and older age. Surgeons and referring physicians should be aware of these risk factors in order to appropriately counsel patients about surgical options.
Funding The authors have no financial disclosures to report. 17
Clinical Neurology and Neurosurgery 160 (2017) 12–18
B.A. Sherrod et al.
perioperative events at US epilepsy centers, J. Neurosurg. 118 (2013) 169–174. [32] S. Wiebe, W.T. Blume, J.P. Girvin, M. Eliasziw, Effectiveness and efficiency of surgery for temporal lobe epilepsy study group. A randomized, controlled trial of surgery for temporal-lobe epilepsy, N. Engl. J. Med. 345 (2001) 311–318. [33] S.E. Emery, S. Akhavan, P. Miller, et al., Steroids and risk factors for airway compromise in multilevel cervical corpectomy patients: a prospective, randomized, double-blind study, Spine (Phila Pa 1976) 34 (2009) 229–232. [34] N. Alan, A. Seicean, S. Seicean, D. Neuhauser, E.C. Benzel, R.J. Weil, Preoperative steroid use and the incidence of perioperative complications in patients undergoing craniotomy for definitive resection of a malignant brain tumor, J. Clin. Neurosci. 22 (2015) 1413–1419. [35] L.B. Chambless, S.L. Parker, L. Hassam-Malani, M.J. McGirt, R.C. Thompson, Type 2 diabetes mellitus and obesity are independent risk factors for poor outcome in patients with high-grade glioma, J. Neurooncol. 106 (2012) 383–389. [36] E.T. Martin, K.S. Kaye, C. Knott, et al., Diabetes and risk of surgical site infection: a systematic review and meta-analysis, Infect. Control Hosp. Epidemiol. 37 (2016) 88–99. [37] R.E. Iachinski, M.S. de Meneses, C.A. Simão, S.F. da Rocha, F. de Oliveira Braga, P.A. Kowacs, Patient satisfaction with temporal lobectomy/selective amygdalohippocampectomy for temporal lobe epilepsy and its relationship with Engel classification and the side of lobectomy, Epilepsy Behav. 31 (2014) 377–380. [38] J.E. Jones, J.B. Blocher, D.C. Jackson, Life outcomes of anterior temporal lobectomy: serial long-term follow-up evaluations, Neurosurgery 73 (2013) 1018–1025. [39] C.S. Kwon, J. Neal, J. Telléz-Zenteno, et al., CASES Investigators. Resective focal epilepsy surgery − Has selection of candidates changed? A systematic review, Epilepsy Res. 122 (2016) 37–43.
[23] American College of Surgeons, User Guide for the 2011 Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/nsqip/ug11. ashx . (Accessed June 14, 2016). [24] American College of Surgeons, User Guide for the 2012 ACS NSQIP Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/ nsqip/ug12.ashx . (Accessed June 14, 2016). [25] American College of Surgeons, User Guide for the 2013 ACS NSQIP Participant Use Data File, (2017) https://www.facs.org/∼/media/files/quality%20programs/ nsqip/2013_acs_nsqip_puf_user_guide.ashx . Accessed (June 14, 2016). [26] M. Shiloach, S.K. Frencher Jr., J.E. Steeger, et al., Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program, J. Am. Coll. Surg. 210 (2010) 6–16. [27] B.L. Hall, B.H. Hamilton, K. Richards, K.Y. Bilimoria, M.E. Cohen, C.Y. Ko, Does surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals, Ann. Surg. 250 (2009) 363–376. [28] M.M. Sellers, R.P. Merkow, A. Halverson, et al., Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program, J. Am. Coll. Surg. 216 (2013) 420–427. [29] A. Weiss, J.E. Anderson, D.C. Chang, Comparing the National Surgical Quality Improvement Program with the Nationwide Inpatient Sample Database, JAMA Surg. 150 (2015) 815–816. [30] University of Alabama at Birmingham Institutional Review Board, Is IRB Review Required for Use of Public Datasets? UAB Research, (2017) http://www.uab.edu/ research/administration/offices/IRB/FAQs/Pages/PublicDatasets.aspx . Accessed (October 1, 2015).. [31] D.J. Englot, D. Ouyang, D.D. Wang, J.D. Rolston, P.A. Garcia, E.F. Chang, Relationship between hospital surgical volume, lobectomy rates, and adverse
18