- Email: [email protected]
Outcomes in octogenarians with subdural hematomas
Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
Contents lists available at SciVerse ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Outcomes in octogenarians with subdural hematomas Julie Pilitsis a,b,∗ , Britney Atwater a,b , Daniel Warden b , Gina Deck b , James Carroll c , Jillian Smith c , Sing Chau Ng c , Jennifer Tseng c a b c
Division of Neurosurgery, Albany Medical College, Albany, NY 12208, USA Department of Neurosurgery, University of Massachusetts Medical Center, Worcester, MA 01655, USA Surgical Outcomes Analysis & Research (SOAR), Department of Surgery, University of Massachusetts Medical Center, Worcester, MA 01655, USA
a r t i c l e
i n f o
Article history: Received 11 July 2012 Received in revised form 10 December 2012 Accepted 20 January 2013 Available online 18 February 2013 Keywords: Elderly National database NIS Subdural hematoma
a b s t r a c t Objective: In the majority of literature concerning age in TBI, specifically in subdural hematomas (SDH), the mean age of patients considered elderly is 55–65. Limited data in SDH patients > 75 years suggest an increased mortality rate. The impact of medical decision making on these data is not well-documented. Patients/Methods: We use the Nationwide Inpatient Sample (NIS) database to compare outcomes between SDH patients 60–79 and ≥80. As administrative databases have some shortcomings, i.e. in-hospital data only, acute and chronic SDHs listed together, we examined institutional data to evaluate the impact of these factors on medical decision making which may falsely elevate mortality rates. Results: In-hospital mortality was increased in NIS patients > 80 treated both surgically and non-surgically (P < 0.05). Our institutional data confirmed higher in-hospital mortality rates in patients > 80 with SDHs as a group. However, the SDH patients > 80 who underwent surgery at our institution had much lower mortality rates. We found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. Type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on surgical decision making. Conclusion: This study examines mortality rates in patients > 80 with SDHs who are managed surgically and non-surgically using a large administrative database and institutional data. It provides preliminary insight into medical decision making which make affect mortality rates of the very elderly. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Traumatic brain injuries (TBI) and their after-effects account for approximately $25 billion in healthcare costs annually in the United States, with an estimated 50–99% of moderate to severe head injuries resulting in long term sequelae [1]. Most clinicians are well aware of the peak in TBI that occur in young adult males, but are less familiar with the additional TBI peak occurring in the elderly, even though TBI is seen more frequently in older female patients than younger ones. In the majority of literature concerning age in TBI, specifically in subdural hematomas (SDH), the mean age of patients considered elderly is 55 to 65. There are three series we are aware of that examine mortality rates in patients > 75; one examines traumatic acute intracerebral hematoma (ICH) and two look at outcome following chronic
∗ Corresponding author at: Albany Medical College, Division of Neurosurgery MC10, 47 New Scotland Ave, Albany, NY 12208, USA. Tel.: +1 518 262 0739; fax: +1 518 262 5400. E-mail addresses: [email protected], [email protected] (J. Pilitsis). 0303-8467/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.clineuro.2013.01.017
subdural hematoma (cSDH). Patients > 75 year with an acute ICH have a significantly higher mortality rate [2]. Mortality rates in two studies on cSDH in very elderly patients appear higher [3,4]. In the most recent series of patients with chronic SDHs (cSDH) whose mean age was 80.6 years, it was clear that patients with SDHs were significantly more likely to die within 2 years than those without cSDHs, regardless of surgical intervention or anticoagulant use [3]. Likewise, Jones and Kafetz found that 31% of patients presenting with cSDH died within 6 months [4]. Both allude to the fact that cSDH may be a marker of chronic disease and impaired health status, rather than a cause of death. In order to better evaluate this hypothesis, we examine the impact of age and comorbidities on outcomes in patients > 80 with SDH as compared to a cohort of patients 60–79. We use the Nationwide Inpatient Sample (NIS) database to examine this data on a large scale. As administrative databases have some shortcomings, i.e. in-hospital data only, acute and chronic SDHs listed together, baseline status unknown, we examined institutional data to evaluate the impact of these factors on medical decision making which may falsely elevate mortality rates.
1430
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
Table 1 Mortality rates in patients with SDH both who were managed medically and surgically. NIS surgical Overall No. (%) Total No. Patients age 60–79 ≥80 * **
NIS non-surgical Mortality (%)
2596 1669 (64.3) 927 (35.7)
Overall No. (%)
Institutional surgical Mortality (%)
49570 145 (8.6) 112 (12.1)*
27249 (55.0) 22321 (45.0)
Overall No. (%)
Mortality (%)
29 3770 (13.7) 3612 (16.2)**
21 (72.5) 8 (27.5)
Institutional non-surgical Overall No. (%)
Mortality (%)
31 2 (9.5) 0 (0)**
10 (32.3) 21 (67.7)
3 (33) 10 (48)**
P < 0.01. P < 0.001.
2. Clinical material and methods 2.1. NIS database 2.1.1. Data source Data for the years 1998–2007 inclusive was gathered from the Nationwide Inpatient Sample (NIS), the largest all-payer inpatient database in the United States. This database is a representative sample of 20% of US hospitals and is maintained by the Agency for Healthcare Research and Quality. It contains data on over 8 million hospital stays from over 1000 hospitals [5]. This approximates a 20% stratified sample of US community hospitals. Unweighted data from this 20% sample was used throughout this study. 2.1.2. Inclusion criteria To identify the appropriate patient population, we used diagnosis and procedure codes from the International Classifications of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) [6]. Patients from 1998 to 2007 with a diagnosis of SDH (ICD-9-CM 432.1, 852.2) both who underwent and did not undergo evacuation (ICD-9-CM 01.24) were included in the analysis. Patient gender, race, and co-morbidities were analyzed. 2.1.3. Outcome variables The primary outcome was in-hospital mortality. Mortality was defined as death from any cause prior to discharge from the hospital. The secondary outcome was the presence of in-hospital post-operative complications. 2.1.4. Statistical analysis All data are presented in unweighted raw values and were analyzed using SAS version 9.2 (SAS Institute, Inc, Cary, NC). Univariate analysis was performed using 2 tests, with P < 0.05 as statistically significant. Multivariate logistic regression was created with outcome measures as dependent variables. Independent variables included patient age (grouped 60–79 and ≥80 years) and various comorbidities (including congestive heart failure, liver disease, renal failure, diabetes mellitus, and obesity). In order to adjust for comorbidities, we used the NIS Comorbidity Software, Version 3.0, to identify the conditions from the Elixhauser index that are relevant to surgical patients [5,7]. Adjusted odds ratios (ORs) were calculated in order to determine the effects of the identified covariates on outcome measures. 2.1.5. Single institution experience Eighty-four patients over 60 years old were admitted to a single institution with a primary diagnosis of SDH over a 24 month period from 2008–2010. Of these patients, 60 patients had SDHs without a confounding intracranial diagnosis. Patients with significant intracerebral hemorrhage, subarachnoid hemorrhage, intraventricular hemorrhage, and ventriculoperitoneal shunts were excluded. Forty-four of these isolated SDHs had surgical lesions, which were defined as SDHs ≥1 cm with mass effect
and midline shift. We compared outcomes across patients treated surgically and conservatively across age groups. Patient gender, race, comorbidities, use of anticoagulation, and type of SDH (acute v. chronic) were analyzed. The primary outcome was in-hospital all-cause death. We examined one year survival rates through the social security database. We obtained IRB approval for this study. 2.1.6. Statistical analysis All data was analyzed using Student’s t-tests/Chi Square were performed with P < 0.05 as statistically significant. Patients were analyzed based on age, gender, anticoagulant use, type of subdural, Glasgow coma score and baseline cognitive status. 3. Results 3.1. NIS database Between 1998 and 2007, 52,166 people over 60 years old in the NIS database were discharged with a primary diagnosis of SDH and had valid mortality information. Patients between 80 and 95 made up 42.7% (22,321) of this group. Of those, 3612 died during their hospital stay (16.1% of patients) which was the highest rate of mortality of all age groups. Of the patients who underwent SDH evacuation (n = 2596), 927 patients (35.7%) were ≥80. Patient gender, race, and hospital teaching status did not affect mortality rates. A significant difference in mortality rate did exist between different age groups (Table 1). Mortality was significantly increased across all groups by congestive heart failure (P < 0.0001), liver disease (P < 0.01), and renal failure (P < 0.0001). Octogenarians with ≥3 comorbidities had an inhospital mortality rate of 36.6% as compared to 11.8% in the same age group without any comorbidities. 3.2. Multivariate analysis Using multivariate logistic regression, we assessed the impact of age v. comorbidities on in-hospital mortality and complications on SDH surgical patients in NIS (Table 2). Age appeared to increase the risk of in-hospital mortality, but not complications, while comorbidities increased both rates. Table 2 Multivariate analysis of NIS data showing effect of patient age and comorbidities on mortality and complications after subdural hematoma evacuation. Variable
Mortality Odds ratio (95% C.I.)
Complications Odds ratio (95% C.I.)
0.606 (0.421–0.873) 0.755 (0.566–1.007)
0.959 (0.754–1.219) 0.940 (0.766–1.153)
0.859 (0.485–1.523) 1.832 (1.095–3.065) 2.550 (1.564–4.159)
1.858 (1.245–2.771) 3.249 (2.212–4.770) 4.460 (3.076–6.466)
Age 60–69 vs ≥ 80 70–79 vs ≥ 80 Comorbidities 1 vs 0 2 vs 0 ≥3 vs 0
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
1431
Table 3 Institutional mortality rates for patients with subdural hematoma by age. Factor
Total No. Patients age 60–79 ≥80
All cases
Patients undergoing SDH evacuation In-hospital mortality (%)
One year mortality (%)
60
15
12
31 29
5 (16.1) 10 (34.5)
6 (19.4) 6 (20.7)
In-hospital mortality (%)
One year mortality (%)
29
2
6
21 8
2 (9.5) 0
5 (23.8) 1 (12.5)
Table 4 Impact of age on patients with surgical lesions on progression to surgery.
Number of patients with surgical lesions proceeding to surgery Number of patients with surgical acute SDHs proceeding to surgery Number of patients with surgical chronic SDHs proceeding to surgery Number of patients with surgical lesions who did not undergo surgery and died in-hospital Number of patients with surgical lesions who did undergo surgery and died in-hospital Number of patients with surgical lesions who did undergo surgery and died by 1 year
60–79 year
≥80
21 of 23 10 of 12 11 of 11 1 of 2 2 of 21 5 of 19
8 of 21 4 of 14 4 of 6 8 of 13 0 of 8 1 of 8
3.3. Retrospective study
4. Discussion
Our institutional data confirmed higher in-hospital mortality rates in patients ≥ 80 with SDHs as a group. (Table 3). In-hospital complication rates were equivalent across age groups (35% in patients < 80 and 38% in patients ≥ 80). Patients ≥ 80 with surgical lesions that did not have surgery had an in-hospital mortality rate of 61.5%, while none of the 8 patients ≥ 80 that underwent surgery died in the hospital; 7 were discharged to rehab and 1 went home. According to the social security database, only one patient in this cohort had died at one year follow-up. Patients with another neurological condition including previous stroke, Parkinson’s, Alzheimers, and epilepsy had a higher one year mortality (P < 0.05). Other medical comorbidities did not correlate with one year outcome. We were able to quantitate factors that related to which patients were offered surgery (Table 4) and found that 87% of the patients that had surgical lesions that did not have operations were ≥80. In patients over 80, type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on decision making (Table 5).
The term “elderly” has been too vaguely and inconsistently defined to be helpful in decision making for patients with TBI. Many studies report conflicting age-related mortality data based on various types of TBI, such as a study of acute SDHs in 28 patients > 65 with a mortality rate of 82%, and only 5% having a functional survival [8]. Amacher and Bybee showed 12% mortality in patients > 80 after head injury, even if minor [9], whereas a case-study reports a 102 year old woman surviving two acute SDHs with craniotomies in 12 months [10]. While these studies offer important information, it is difficult to interpret the data because the studies have different parameters and exclude details on decision making, which is no doubt important in the extremes of age. NIS data shows that patients ≥ 80 both who undergo surgery and do not undergo surgery for any reason (i.e. medically unstable, SDH does not warrant surgery, advanced directives) have significantly increased mortality rates. In sicker patients, i.e. those who have 3 or more comorbidities, mortality rates are nearly equivalent between groups. The impact of other variables is difficult to assess in NIS or any administrative database [11,12]. In order to gain a better
Table 5 Decision making in patients over 80 with SDHs.
Total No. patients Gender Male Female Comorbidities 0 1 2 Anticoagulation Yes No Mechanism of injury Fall Motor vehicle crash Other high energy injury GCS at presentation GCS 3–8 GCS 9–11 GCS 12–15 Baseline cognitive status Good Marginal Poor a
Total Number of patients with SDH diagnosis (%)
Surgery (%)
Patients with surgical lesions who did not undergo surgery (%)
n = 29 n = 29 13 (44.8) 16 (55.2) n = 25 13 (52.0) 8 (32.0) 4 (16.0) n = 27 20 (74.1) 7 (25.9) n = 27 25 (92.6) 1 (3.7) 1 (3.7) n = 26 5 (19.2) 3 (11.5) 18 (69.2) n = 27 15 (55.6) 5 (18.5) 7 (25.9)
n=8 n=8 6 (75.0) 2 (25.0) n=8 5 (62.5) 2 (25.0) 1 (12.5) n=8 6 (75.0) 2 (25.0) n=8 6 (75) 1 (12.5) 1 (12.5) n=8 1 (12.5) 0 7 (87.5) n=8 8 (100.0)a 0 0
n = 13 n = 13 6 (46.2) 7 (53.8) n = 10 4 (40.0) 5 (50.0) 1 (10.0) n = 10 8 (80.0) 2 (20.0) n = 11 11 (100.0) 0 0 n = 10 2 (20.0) 3 (30.0) 5 (50.0) n = 11 4 (36.4) 2 (18.2) 5 (45.5)
Patients > 80 who underwent surgery presented with significantly better baseline cognitive statuses than patients who had surgical lesions but did not have surgery (P < 0.05).
1432
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
understanding of these factors including mechanism of injury, baseline cognitive status, type of subdural, admission GCS and presence of anticoagulation, we utilized institutional data. Acute SDH cases accounted for 90% of deaths in our institutional patients ≥ 80 and 66% of complications. High rates of mortality and complications were also found in patients ≥ 80 on anticoagulant treatment (62.5%). Many studies have shown an increase in complications associated with anticoagulant treatments [8,9,13–15]. One study showed that 1 in 10 older patients (≥75) on anticoagulant treatment had bleeding complications and the rates of complication rose with increasing age [14]. In our study, anticoagulation use did not significantly influence decision making, while admission GCS, age, and baseline cognitive status significantly did. Our data found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. The mortality rate for this group was 61.5%, while none of the surgical patients ≥ 80 died in the hospital. These rates were no doubt influenced by medical decision making taking into account the surgeon’s assessment of survivability and disability from the lesion as well as the patient’s comorbid status. Interestingly, the presence of an underlying neurological disorder but no other medical comorbidity (renal, cardiac, pulmonary, diabetes, ophthalmological) led to a significantly higher mortality at one year. We hypothesize that this is due to a continued high fall risk in this group. In the elderly, falls have been shown to be the most common cause of traumatic brain injury, especially SDHs [2,4,8,9]. Falls can also lead to serious injuries like hip fractures. Prevention of falls in the home can be as simple as installing grab bars in bathrooms and anti-slip appliques in bath tubs, in addition to repairs of any environmental hazards and maintenance of walking aids. Another useful intervention is exercise, which can be focused to improve gait, balance and strength. Lastly, providers should closely monitor medications which may cause dizziness, unsteadiness and hypotension as these also may lead to a rise in falls. One study documented a significant decrease in fall risk after a 3 month intervention period in both high and low risk patients as well as a greater time until first fall [16]. As neurosurgeons caring for these patients, it is essential that we provide education on fall prevention in conjunction with the patient’s primary doctor. 5. Conclusion Our review of NIS data shows that patients ≥ 80 both who undergo and do not undergo surgery have an increased in-hospital mortality rate as compared to patients 60–79. In patients with 3 or more comorbidities, mortality rates are nearly equivalent between groups. Our institutional data showed an increased mortality rate in those ≥80 who did not undergo surgery, but not in those who did.
This difference is likely secondary to the impact on medical decision making on which patients to take to surgery. We found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. Type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on surgical decision making. Conflict of interest The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. References [1] Tallon JM, Acokroyd-Stolarz S, Karim SA, Clarke DB. The epidemiology of surgically treated acute subdural and epidural hematomas in patients with head injuries: a population-based study. Canadian Journal of Surgery 2008;51(5):339–45. [2] Jamjoom A, Nelson R, Stranjalis G, Wood S, Chissell H, Kane N, et al. Outcome following surgical evacuation of traumatic intracranial haematomas in the elderly. Journal of Neurosurgery 1992;6(1):27–32. [3] Miranda LB, Braxton E, Hobbs J, Quigley MR. Chronic subdural hematoma in the elderly: not a benign disease. Journal of Neurosurgery 2011;114(January (1)):72–6. [4] Jones S, Kafetz K. A prospective study of chronic subdural haematomas in elderly patients. Age Ageing 1999;28(October (6)):519–21. [5] HCUP Nationwide Inpatient Sample (NIS). Healthcare cost and utilization project (HCUP). Rockville, MD: Agency for healthcare research and quality; 1997–2007 www.hcup-us.ahrq.gov/nisoverview.jsp [6] International classifications of diseases, ninth revision, clinical modification. Washington, DC: Public health service, US dept of health and human services; 1998. [7] Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Medical Care 1998;36:8–27. [8] Wilberger JE, Harris M, Diamond D. Acute subdural hematoma: morbidity, mortality, and operative timing. Journal of Neurosurgery 1991;74:212–8. [9] Amacher AL, Bybee DE. Toleration of head injury by the elderly. Neurosurgery 1987;20:954–8. [10] Vyas N, Chicoine M. Extended survival after evacuation of subdural hematoma in a 102-year-old patient: case report and review of the literature. Surgical Neurology 2007;67:314–7. [11] Iezzoni L, Foley SM, Daley J, Hughes J, Fisher ES, Heeren T. Comorbidities, complications, and coding bias: does the number of diagnosis codes matter in predicting in-hospital mortality? JAMA 1992;267:2197–203. [12] Jencks SF, Williams DK, Kay TL. Assessing hospital-associated deaths from discharge data: the role of length of stay and comorbidities. JAMA 1988;260:2240–6. [13] Bershad EM, Farhadi S, Suri FK, Feen ES, Hernandez OH, Selman WR, et al. Coagulopathy and inhospital deaths in patients with acute subdural hematoma. Journal of Neurosurgery 2008;109:664–9. [14] Palareti G, Neali N, Coccheri S, Poggi M, Manotti C, D’Angelo A, et al. Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaboration study (ISCOAT). Lancet 1996;348:423–8. [15] Flaherty ML, Tao H, Haverbusch M, Sekar P, Kleindorfer D, Kissela B, et al. Warfarin use leads to larger intracerebral hematomas. Neurology 2008;71:1084–9. [16] Jensen J, Nyberg L, Gustafson Y, Lundin-Olsson L. Fall and injury prevention in residential care – effects in residents with higher and lower levels of cognition. Journal of the American Geriatrics Society 2003 May;51(5):627–35.
Contents lists available at SciVerse ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Outcomes in octogenarians with subdural hematomas Julie Pilitsis a,b,∗ , Britney Atwater a,b , Daniel Warden b , Gina Deck b , James Carroll c , Jillian Smith c , Sing Chau Ng c , Jennifer Tseng c a b c
Division of Neurosurgery, Albany Medical College, Albany, NY 12208, USA Department of Neurosurgery, University of Massachusetts Medical Center, Worcester, MA 01655, USA Surgical Outcomes Analysis & Research (SOAR), Department of Surgery, University of Massachusetts Medical Center, Worcester, MA 01655, USA
a r t i c l e
i n f o
Article history: Received 11 July 2012 Received in revised form 10 December 2012 Accepted 20 January 2013 Available online 18 February 2013 Keywords: Elderly National database NIS Subdural hematoma
a b s t r a c t Objective: In the majority of literature concerning age in TBI, specifically in subdural hematomas (SDH), the mean age of patients considered elderly is 55–65. Limited data in SDH patients > 75 years suggest an increased mortality rate. The impact of medical decision making on these data is not well-documented. Patients/Methods: We use the Nationwide Inpatient Sample (NIS) database to compare outcomes between SDH patients 60–79 and ≥80. As administrative databases have some shortcomings, i.e. in-hospital data only, acute and chronic SDHs listed together, we examined institutional data to evaluate the impact of these factors on medical decision making which may falsely elevate mortality rates. Results: In-hospital mortality was increased in NIS patients > 80 treated both surgically and non-surgically (P < 0.05). Our institutional data confirmed higher in-hospital mortality rates in patients > 80 with SDHs as a group. However, the SDH patients > 80 who underwent surgery at our institution had much lower mortality rates. We found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. Type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on surgical decision making. Conclusion: This study examines mortality rates in patients > 80 with SDHs who are managed surgically and non-surgically using a large administrative database and institutional data. It provides preliminary insight into medical decision making which make affect mortality rates of the very elderly. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Traumatic brain injuries (TBI) and their after-effects account for approximately $25 billion in healthcare costs annually in the United States, with an estimated 50–99% of moderate to severe head injuries resulting in long term sequelae [1]. Most clinicians are well aware of the peak in TBI that occur in young adult males, but are less familiar with the additional TBI peak occurring in the elderly, even though TBI is seen more frequently in older female patients than younger ones. In the majority of literature concerning age in TBI, specifically in subdural hematomas (SDH), the mean age of patients considered elderly is 55 to 65. There are three series we are aware of that examine mortality rates in patients > 75; one examines traumatic acute intracerebral hematoma (ICH) and two look at outcome following chronic
∗ Corresponding author at: Albany Medical College, Division of Neurosurgery MC10, 47 New Scotland Ave, Albany, NY 12208, USA. Tel.: +1 518 262 0739; fax: +1 518 262 5400. E-mail addresses: [email protected], [email protected] (J. Pilitsis). 0303-8467/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.clineuro.2013.01.017
subdural hematoma (cSDH). Patients > 75 year with an acute ICH have a significantly higher mortality rate [2]. Mortality rates in two studies on cSDH in very elderly patients appear higher [3,4]. In the most recent series of patients with chronic SDHs (cSDH) whose mean age was 80.6 years, it was clear that patients with SDHs were significantly more likely to die within 2 years than those without cSDHs, regardless of surgical intervention or anticoagulant use [3]. Likewise, Jones and Kafetz found that 31% of patients presenting with cSDH died within 6 months [4]. Both allude to the fact that cSDH may be a marker of chronic disease and impaired health status, rather than a cause of death. In order to better evaluate this hypothesis, we examine the impact of age and comorbidities on outcomes in patients > 80 with SDH as compared to a cohort of patients 60–79. We use the Nationwide Inpatient Sample (NIS) database to examine this data on a large scale. As administrative databases have some shortcomings, i.e. in-hospital data only, acute and chronic SDHs listed together, baseline status unknown, we examined institutional data to evaluate the impact of these factors on medical decision making which may falsely elevate mortality rates.
1430
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
Table 1 Mortality rates in patients with SDH both who were managed medically and surgically. NIS surgical Overall No. (%) Total No. Patients age 60–79 ≥80 * **
NIS non-surgical Mortality (%)
2596 1669 (64.3) 927 (35.7)
Overall No. (%)
Institutional surgical Mortality (%)
49570 145 (8.6) 112 (12.1)*
27249 (55.0) 22321 (45.0)
Overall No. (%)
Mortality (%)
29 3770 (13.7) 3612 (16.2)**
21 (72.5) 8 (27.5)
Institutional non-surgical Overall No. (%)
Mortality (%)
31 2 (9.5) 0 (0)**
10 (32.3) 21 (67.7)
3 (33) 10 (48)**
P < 0.01. P < 0.001.
2. Clinical material and methods 2.1. NIS database 2.1.1. Data source Data for the years 1998–2007 inclusive was gathered from the Nationwide Inpatient Sample (NIS), the largest all-payer inpatient database in the United States. This database is a representative sample of 20% of US hospitals and is maintained by the Agency for Healthcare Research and Quality. It contains data on over 8 million hospital stays from over 1000 hospitals [5]. This approximates a 20% stratified sample of US community hospitals. Unweighted data from this 20% sample was used throughout this study. 2.1.2. Inclusion criteria To identify the appropriate patient population, we used diagnosis and procedure codes from the International Classifications of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) [6]. Patients from 1998 to 2007 with a diagnosis of SDH (ICD-9-CM 432.1, 852.2) both who underwent and did not undergo evacuation (ICD-9-CM 01.24) were included in the analysis. Patient gender, race, and co-morbidities were analyzed. 2.1.3. Outcome variables The primary outcome was in-hospital mortality. Mortality was defined as death from any cause prior to discharge from the hospital. The secondary outcome was the presence of in-hospital post-operative complications. 2.1.4. Statistical analysis All data are presented in unweighted raw values and were analyzed using SAS version 9.2 (SAS Institute, Inc, Cary, NC). Univariate analysis was performed using 2 tests, with P < 0.05 as statistically significant. Multivariate logistic regression was created with outcome measures as dependent variables. Independent variables included patient age (grouped 60–79 and ≥80 years) and various comorbidities (including congestive heart failure, liver disease, renal failure, diabetes mellitus, and obesity). In order to adjust for comorbidities, we used the NIS Comorbidity Software, Version 3.0, to identify the conditions from the Elixhauser index that are relevant to surgical patients [5,7]. Adjusted odds ratios (ORs) were calculated in order to determine the effects of the identified covariates on outcome measures. 2.1.5. Single institution experience Eighty-four patients over 60 years old were admitted to a single institution with a primary diagnosis of SDH over a 24 month period from 2008–2010. Of these patients, 60 patients had SDHs without a confounding intracranial diagnosis. Patients with significant intracerebral hemorrhage, subarachnoid hemorrhage, intraventricular hemorrhage, and ventriculoperitoneal shunts were excluded. Forty-four of these isolated SDHs had surgical lesions, which were defined as SDHs ≥1 cm with mass effect
and midline shift. We compared outcomes across patients treated surgically and conservatively across age groups. Patient gender, race, comorbidities, use of anticoagulation, and type of SDH (acute v. chronic) were analyzed. The primary outcome was in-hospital all-cause death. We examined one year survival rates through the social security database. We obtained IRB approval for this study. 2.1.6. Statistical analysis All data was analyzed using Student’s t-tests/Chi Square were performed with P < 0.05 as statistically significant. Patients were analyzed based on age, gender, anticoagulant use, type of subdural, Glasgow coma score and baseline cognitive status. 3. Results 3.1. NIS database Between 1998 and 2007, 52,166 people over 60 years old in the NIS database were discharged with a primary diagnosis of SDH and had valid mortality information. Patients between 80 and 95 made up 42.7% (22,321) of this group. Of those, 3612 died during their hospital stay (16.1% of patients) which was the highest rate of mortality of all age groups. Of the patients who underwent SDH evacuation (n = 2596), 927 patients (35.7%) were ≥80. Patient gender, race, and hospital teaching status did not affect mortality rates. A significant difference in mortality rate did exist between different age groups (Table 1). Mortality was significantly increased across all groups by congestive heart failure (P < 0.0001), liver disease (P < 0.01), and renal failure (P < 0.0001). Octogenarians with ≥3 comorbidities had an inhospital mortality rate of 36.6% as compared to 11.8% in the same age group without any comorbidities. 3.2. Multivariate analysis Using multivariate logistic regression, we assessed the impact of age v. comorbidities on in-hospital mortality and complications on SDH surgical patients in NIS (Table 2). Age appeared to increase the risk of in-hospital mortality, but not complications, while comorbidities increased both rates. Table 2 Multivariate analysis of NIS data showing effect of patient age and comorbidities on mortality and complications after subdural hematoma evacuation. Variable
Mortality Odds ratio (95% C.I.)
Complications Odds ratio (95% C.I.)
0.606 (0.421–0.873) 0.755 (0.566–1.007)
0.959 (0.754–1.219) 0.940 (0.766–1.153)
0.859 (0.485–1.523) 1.832 (1.095–3.065) 2.550 (1.564–4.159)
1.858 (1.245–2.771) 3.249 (2.212–4.770) 4.460 (3.076–6.466)
Age 60–69 vs ≥ 80 70–79 vs ≥ 80 Comorbidities 1 vs 0 2 vs 0 ≥3 vs 0
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
1431
Table 3 Institutional mortality rates for patients with subdural hematoma by age. Factor
Total No. Patients age 60–79 ≥80
All cases
Patients undergoing SDH evacuation In-hospital mortality (%)
One year mortality (%)
60
15
12
31 29
5 (16.1) 10 (34.5)
6 (19.4) 6 (20.7)
In-hospital mortality (%)
One year mortality (%)
29
2
6
21 8
2 (9.5) 0
5 (23.8) 1 (12.5)
Table 4 Impact of age on patients with surgical lesions on progression to surgery.
Number of patients with surgical lesions proceeding to surgery Number of patients with surgical acute SDHs proceeding to surgery Number of patients with surgical chronic SDHs proceeding to surgery Number of patients with surgical lesions who did not undergo surgery and died in-hospital Number of patients with surgical lesions who did undergo surgery and died in-hospital Number of patients with surgical lesions who did undergo surgery and died by 1 year
60–79 year
≥80
21 of 23 10 of 12 11 of 11 1 of 2 2 of 21 5 of 19
8 of 21 4 of 14 4 of 6 8 of 13 0 of 8 1 of 8
3.3. Retrospective study
4. Discussion
Our institutional data confirmed higher in-hospital mortality rates in patients ≥ 80 with SDHs as a group. (Table 3). In-hospital complication rates were equivalent across age groups (35% in patients < 80 and 38% in patients ≥ 80). Patients ≥ 80 with surgical lesions that did not have surgery had an in-hospital mortality rate of 61.5%, while none of the 8 patients ≥ 80 that underwent surgery died in the hospital; 7 were discharged to rehab and 1 went home. According to the social security database, only one patient in this cohort had died at one year follow-up. Patients with another neurological condition including previous stroke, Parkinson’s, Alzheimers, and epilepsy had a higher one year mortality (P < 0.05). Other medical comorbidities did not correlate with one year outcome. We were able to quantitate factors that related to which patients were offered surgery (Table 4) and found that 87% of the patients that had surgical lesions that did not have operations were ≥80. In patients over 80, type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on decision making (Table 5).
The term “elderly” has been too vaguely and inconsistently defined to be helpful in decision making for patients with TBI. Many studies report conflicting age-related mortality data based on various types of TBI, such as a study of acute SDHs in 28 patients > 65 with a mortality rate of 82%, and only 5% having a functional survival [8]. Amacher and Bybee showed 12% mortality in patients > 80 after head injury, even if minor [9], whereas a case-study reports a 102 year old woman surviving two acute SDHs with craniotomies in 12 months [10]. While these studies offer important information, it is difficult to interpret the data because the studies have different parameters and exclude details on decision making, which is no doubt important in the extremes of age. NIS data shows that patients ≥ 80 both who undergo surgery and do not undergo surgery for any reason (i.e. medically unstable, SDH does not warrant surgery, advanced directives) have significantly increased mortality rates. In sicker patients, i.e. those who have 3 or more comorbidities, mortality rates are nearly equivalent between groups. The impact of other variables is difficult to assess in NIS or any administrative database [11,12]. In order to gain a better
Table 5 Decision making in patients over 80 with SDHs.
Total No. patients Gender Male Female Comorbidities 0 1 2 Anticoagulation Yes No Mechanism of injury Fall Motor vehicle crash Other high energy injury GCS at presentation GCS 3–8 GCS 9–11 GCS 12–15 Baseline cognitive status Good Marginal Poor a
Total Number of patients with SDH diagnosis (%)
Surgery (%)
Patients with surgical lesions who did not undergo surgery (%)
n = 29 n = 29 13 (44.8) 16 (55.2) n = 25 13 (52.0) 8 (32.0) 4 (16.0) n = 27 20 (74.1) 7 (25.9) n = 27 25 (92.6) 1 (3.7) 1 (3.7) n = 26 5 (19.2) 3 (11.5) 18 (69.2) n = 27 15 (55.6) 5 (18.5) 7 (25.9)
n=8 n=8 6 (75.0) 2 (25.0) n=8 5 (62.5) 2 (25.0) 1 (12.5) n=8 6 (75.0) 2 (25.0) n=8 6 (75) 1 (12.5) 1 (12.5) n=8 1 (12.5) 0 7 (87.5) n=8 8 (100.0)a 0 0
n = 13 n = 13 6 (46.2) 7 (53.8) n = 10 4 (40.0) 5 (50.0) 1 (10.0) n = 10 8 (80.0) 2 (20.0) n = 11 11 (100.0) 0 0 n = 10 2 (20.0) 3 (30.0) 5 (50.0) n = 11 4 (36.4) 2 (18.2) 5 (45.5)
Patients > 80 who underwent surgery presented with significantly better baseline cognitive statuses than patients who had surgical lesions but did not have surgery (P < 0.05).
1432
J. Pilitsis et al. / Clinical Neurology and Neurosurgery 115 (2013) 1429–1432
understanding of these factors including mechanism of injury, baseline cognitive status, type of subdural, admission GCS and presence of anticoagulation, we utilized institutional data. Acute SDH cases accounted for 90% of deaths in our institutional patients ≥ 80 and 66% of complications. High rates of mortality and complications were also found in patients ≥ 80 on anticoagulant treatment (62.5%). Many studies have shown an increase in complications associated with anticoagulant treatments [8,9,13–15]. One study showed that 1 in 10 older patients (≥75) on anticoagulant treatment had bleeding complications and the rates of complication rose with increasing age [14]. In our study, anticoagulation use did not significantly influence decision making, while admission GCS, age, and baseline cognitive status significantly did. Our data found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. The mortality rate for this group was 61.5%, while none of the surgical patients ≥ 80 died in the hospital. These rates were no doubt influenced by medical decision making taking into account the surgeon’s assessment of survivability and disability from the lesion as well as the patient’s comorbid status. Interestingly, the presence of an underlying neurological disorder but no other medical comorbidity (renal, cardiac, pulmonary, diabetes, ophthalmological) led to a significantly higher mortality at one year. We hypothesize that this is due to a continued high fall risk in this group. In the elderly, falls have been shown to be the most common cause of traumatic brain injury, especially SDHs [2,4,8,9]. Falls can also lead to serious injuries like hip fractures. Prevention of falls in the home can be as simple as installing grab bars in bathrooms and anti-slip appliques in bath tubs, in addition to repairs of any environmental hazards and maintenance of walking aids. Another useful intervention is exercise, which can be focused to improve gait, balance and strength. Lastly, providers should closely monitor medications which may cause dizziness, unsteadiness and hypotension as these also may lead to a rise in falls. One study documented a significant decrease in fall risk after a 3 month intervention period in both high and low risk patients as well as a greater time until first fall [16]. As neurosurgeons caring for these patients, it is essential that we provide education on fall prevention in conjunction with the patient’s primary doctor. 5. Conclusion Our review of NIS data shows that patients ≥ 80 both who undergo and do not undergo surgery have an increased in-hospital mortality rate as compared to patients 60–79. In patients with 3 or more comorbidities, mortality rates are nearly equivalent between groups. Our institutional data showed an increased mortality rate in those ≥80 who did not undergo surgery, but not in those who did.
This difference is likely secondary to the impact on medical decision making on which patients to take to surgery. We found that patients ≥ 80 made up 87% of all patients with “surgical lesions” that were not operated on. Type of subdural, admission GCS, and baseline cognitive status appeared to have a significant impact on surgical decision making. Conflict of interest The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. References [1] Tallon JM, Acokroyd-Stolarz S, Karim SA, Clarke DB. The epidemiology of surgically treated acute subdural and epidural hematomas in patients with head injuries: a population-based study. Canadian Journal of Surgery 2008;51(5):339–45. [2] Jamjoom A, Nelson R, Stranjalis G, Wood S, Chissell H, Kane N, et al. Outcome following surgical evacuation of traumatic intracranial haematomas in the elderly. Journal of Neurosurgery 1992;6(1):27–32. [3] Miranda LB, Braxton E, Hobbs J, Quigley MR. Chronic subdural hematoma in the elderly: not a benign disease. Journal of Neurosurgery 2011;114(January (1)):72–6. [4] Jones S, Kafetz K. A prospective study of chronic subdural haematomas in elderly patients. Age Ageing 1999;28(October (6)):519–21. [5] HCUP Nationwide Inpatient Sample (NIS). Healthcare cost and utilization project (HCUP). Rockville, MD: Agency for healthcare research and quality; 1997–2007 www.hcup-us.ahrq.gov/nisoverview.jsp [6] International classifications of diseases, ninth revision, clinical modification. Washington, DC: Public health service, US dept of health and human services; 1998. [7] Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Medical Care 1998;36:8–27. [8] Wilberger JE, Harris M, Diamond D. Acute subdural hematoma: morbidity, mortality, and operative timing. Journal of Neurosurgery 1991;74:212–8. [9] Amacher AL, Bybee DE. Toleration of head injury by the elderly. Neurosurgery 1987;20:954–8. [10] Vyas N, Chicoine M. Extended survival after evacuation of subdural hematoma in a 102-year-old patient: case report and review of the literature. Surgical Neurology 2007;67:314–7. [11] Iezzoni L, Foley SM, Daley J, Hughes J, Fisher ES, Heeren T. Comorbidities, complications, and coding bias: does the number of diagnosis codes matter in predicting in-hospital mortality? JAMA 1992;267:2197–203. [12] Jencks SF, Williams DK, Kay TL. Assessing hospital-associated deaths from discharge data: the role of length of stay and comorbidities. JAMA 1988;260:2240–6. [13] Bershad EM, Farhadi S, Suri FK, Feen ES, Hernandez OH, Selman WR, et al. Coagulopathy and inhospital deaths in patients with acute subdural hematoma. Journal of Neurosurgery 2008;109:664–9. [14] Palareti G, Neali N, Coccheri S, Poggi M, Manotti C, D’Angelo A, et al. Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaboration study (ISCOAT). Lancet 1996;348:423–8. [15] Flaherty ML, Tao H, Haverbusch M, Sekar P, Kleindorfer D, Kissela B, et al. Warfarin use leads to larger intracerebral hematomas. Neurology 2008;71:1084–9. [16] Jensen J, Nyberg L, Gustafson Y, Lundin-Olsson L. Fall and injury prevention in residential care – effects in residents with higher and lower levels of cognition. Journal of the American Geriatrics Society 2003 May;51(5):627–35.