Risk Factors, Incidence, and Effect of Cardiac Failure and Myocardial Infarction in Aneurysmal Subarachnoid Hemorrhage Patients
Kim, Young Woo MD*; Neal, Dan MS‡; Hoh, Brian L. MD‡
*Department of Neurosurgery, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Bucheon, Republic of Korea;
‡Department of Neurosurgery, University of Florida, Gainesville, Florida
Correspondence: Brian L. Hoh, MD, Department of Neurosurgery, University of Florida, PO Box 100265, Gainesville, FL 32610. E-mail: email@example.com
Received July 25, 2012
Accepted May 22, 2013
BACKGROUND: Cardiac dysfunction is a well-known complication of aneurysmal subarachnoid hemorrhage (aSAH). However, the clinical significance of cardiac complications is largely unknown.
OBJECTIVE: To determine whether cardiac complications are independently related to outcomes and to identify potential predictors associated with these complications.
METHODS: We extracted all hospitalizations for aSAH from the National Inpatient Sample database for years 2002 to 2009. We used generalized estimating equations to determine whether cardiac complications were associated with the patient outcomes and to evaluate potential predictors of cardiac complications.
RESULTS: Among 53 713 cases of aSAH, there were 3609 (6.72%) and 151 (0.28%) incidences of cardiac failure (CF) and myocardial infarction (MI), respectively. The overall in-hospital mortality rate was 24.8%, whereas the mortality rate for patients with CF was 34.4% and the mortality rate for patients with MI was 29.8%. Patients who experienced CF were significantly more likely than other patients to die in the hospital (odds ratio: 1.6, 95% confidence interval: 1.47-1.68; P < .001). The difference in mortality rates between MI patients and other patients, however, was not statistically significant. The generalized estimating equation model identified 7 factors that were predictive of CF: age, sex, race, primary payer, diabetes, smoker, and cardiac disease. For MI, the model identified age, race, and primary payer as significant predictors of MI.
CONCLUSION: Our results suggest that an important association exists between cardiac complications and mortality/morbidity in aSAH patients. aSAH patients with CF appear to have a higher mortality rate, longer hospital length of stay, and higher hospitalization costs compared with those without CF.
ABBREVIATIONS: aSAH, aneurysmal subarachnoid hemorrhage
AUC, area under the receiver-operating characteristic curve
CF, cardiac failure
CI, confidence interval
GEE, generalized estimating equation
LOS, length of stay
MI, myocardial infarction
NIS, Nationwide Inpatient Sample
SAH, subarachnoid hemorrhage
Medical (non-neurological) complications are common after aneurysmal subarachnoid hemorrhage (aSAH). Although the majority of the mortality and morbidity from complications in aSAH patients is attributed to neurological complications (primary hemorrhage, rebleeding, and cerebral vasospasm), medical complications have been thought to represent potentially life-threatening complications after aSAH.1,2 The most frequent medical complications include cardiac and pulmonary complications,3-5 electrolyte disturbance,6,7 and hematological change.8,9 Among these, cardiac dysfunction is a well-known complication of aSAH and has been reported regularly for many years. In aSAH patients, electrocardiogram (ECG) changes, such as T-wave inversion and QT interval prolongation, can be found in 50% to 100% of patients, troponin elevation is seen in 20% to 40%, and regional wall-motion abnormalities occur in 10% of patients.10,11 It is, however, not clear in which patients with aSAH cardiac complications will develop. Furthermore, the clinical significance of these cardiac dysfunctions is largely unknown. Complications may contribute to sudden death. aSAH patients who have signs of cardiac complications may benefit from early, appropriate, and intensive monitoring. We reviewed the incidence of cardiac complications in aSAH using a nationwide administrative database to evaluate the impact of these complications on patient outcome, as well as to identify potential predictors associated with these complications.
PATIENTS AND METHODS
We extracted all hospitalizations from the National Inpatient Sample database, years 2002 to 2009, in which the patient was 18 years of age or older and the hospital record included ICD-9 diagnosis code 430 (subarachnoid hemorrhage [SAH]). We did not include traumatic SAH (diagnosis codes 852.0 and 852.1) in this study. We also used ICD-9 diagnosis codes to identify patients who experienced cardiac failure (CF) (diagnosis codes 428.0-428.4, or 428.9) or myocardial infarction (MI) (diagnosis codes 410.0-410.9, 411.0, 411.1, or 411.8) and to identify current or former smokers (code 305.1 and V15.82) (Table 1). We obtained information on patient age, sex, race, median income in patient's ZIP code, expected primary payer, diabetes with/without complications, cardiac disease, hypertension, length of stay (LOS), total hospital charges, as well as outcome (discharge disposition classification including in-hospital mortality). We defined patient outcome as favorable if the patient was sent home under his or her own care or under the care of a home health service, if he or she was sent to a short-term facility, sent to inpatient rehabilitation, sent to another institution for outpatient care, or left against medical advice. We defined patient outcome as unfavorable if the patient was sent to a skilled nursing facility, to an intermediate care facility, to hospice, or to a Medicare approved swing bed (a hospital bed used for nursing home–type services), or if the patient died in the hospital. We excluded from the analysis any hospital records in which the patient's outcome did not fall into one of these categories. We increased total hospital charges by 3% per year for each year before 2009 to adjust for inflation.
We used the SAS statistical software package version 9.3 (SAS Institute, Cary, North Carolina) to calculate all descriptive statistics and to perform 2 separate analyses: (1) investigate potential predictors of CF and MI in patients with aSAH and (2) investigate the effects of CF and MI on mortality, unfavorable discharge, LOS, and total hospital charges for these patients.
To evaluate potential predictors of CF and MI (treated as dichotomous variables), we used generalized estimating equations ([GEEs] SAS PROC GENMOD) to develop predictive models for each outcome. To account for multiple observations on hospitals, we treated hospital as a repeated factor and assumed an exchangeable working correlation. We assumed a binary distribution for the outcome variables, using a logit link function, and we retained the default convergence criteria set by the SAS system. We included sex, age, race (white, black, Hispanic, or other), median income in the patient's ZIP code (low, low-medium, medium-high, or high), diabetes with or without complications (yes or no), hypertension (yes or no), current or former smoker (yes or no), cardiac disease (yes or no), and primary payer as independent variables. The primary payer covariate included categories for Medicare (government-sponsored care for the elderly), Medicaid (government-sponsored care for the poor), private insurance (care paid for by a for-profit insurance company), self-pay, and other. We considered any variable whose model coefficient had a P value <.05 as a significant predictor. We used area under the receiver-operating characteristic curve (AUC) to evaluate the model's predictive ability.
To evaluate the effects of MI and CF on mortality, unfavorable discharge, LOS, and total charges, we also used GEE models with a repeated hospital factor. In these analyses, we took CF or MI as the single predictor. For mortality and discharge, we assumed a binary distribution for the outcome and used a logit link function. For LOS and charges, we used the natural log of the response as the outcome, and we assumed a normal distribution and an identity link.
Between 2002 and 2009, there were a total of 53 713 cases of aSAH with 3609 (6.72%) and 151 (0.28%) incidences of CF and MI, respectively. The overall incidence rate of cardiac complications (CF and MI) was 7%. The clinical characteristics and outcomes of patients are shown in Tables 2 and 3. The mean age of the study patients was 58.8 ± 16.6 years, and 61.1% were women, a typical sex distribution for aSAH. Hypertension was the most common of the comorbidities that we studied, with 28 455 cases (53.6%). Only 1828 (3.4%) cases of cardiac disease appeared in this database.
Effects of CF and MI on Outcomes
The overall in-hospital mortality rate was 24.8%, while the rate for patients with CF was 34.4% and the rate for patients with MI was 29.8%. Patients who experienced CF were significantly more likely than patients without CF to die in the hospital (odds ratio [OR]: 1.6, 95% confidence interval [CI]: 1.47-1.68, P < .001). The difference in mortality rates between MI patients and other patients, however, was not statistically significant (P = .154). Patients who experienced CF also were significantly more likely to have unfavorable discharge (OR: 2.3, 95% CI: 2.09-2.47, P < .001), but we found no significant association between MI and discharge disposition (P = .144). Similarly, aSAH patients with CF experienced significantly longer hospital stays than other patients (P < .001), and their hospital charges were significantly higher (P < .001). Patients who experienced MI also had significantly higher charges than patients without MI (P = .047), but the LOS was not significantly different (P = .153).
Potential Predictors of CF and MI
There were 53 713 observations involving aSAH in the Nationwide Inpatient Sample (NIS) database. Of these, 16 027 were missing information on 1 or more of the predictor variables (14 302 were missing information on race), leaving 37 686 observations available for analysis. Of these, 2574 patients experienced CF and 35 112 did not. The multivariate GEE model identified 7 factors as predictive of CF (P < .05): age, sex, race, primary payer, diabetes, smoker, and cardiac disease (Figure 1 and Table 4). The model predicted the outcome correctly for 72.5% of the observations in the dataset (AUC = 0.725).
We followed a similar strategy to determine predictors of MI. After observations with missing information were excluded, there were 37 686 observations available for analysis. Of these, 116 patients experienced MI and 37 570 did not. The model identified age, race, and primary payer as significant predictors (P < .05) of MI (Figure 2 and Table 5). The model predicted the outcome correctly for 68.6% of the observations in the dataset (AUC = 0.686).
Although many attempts have been made to clarify whether cardiac abnormalities after an intracerebral event are predictive of mortality in patients with SAH, the relationship of cardiac abnormalities (abnormal ECG, elevated cardiac enzyme, and wall-motion abnormality) to morbidity and mortality has been a matter of debate, and the significance of cardiac abnormalities remains uncertain. The degree of damage may be minor, evident only by asymptomatic elevation of cardiac enzymes, or major, causing life-threatening severe CF with left ventricular dysfunction.12,13 This study presents the first nationwide report of the association of cardiac complications with outcomes in SAH in the United States. In our study, cardiac complications, such as CF and MI, were associated with an increased risk of in-patient mortality and severe disability at discharge in patients with aSAH. In addition, our results also show that aSAH patients with CF had longer in-hospital LOS compared with those without CF. Patients with CF may experience other medical complications, such as pulmonary complications, renal failure, and hepatic dysfunction, which may require observation in intensive care units and thus longer periods of hospitalization and higher charges. Although patients who experience MI had significantly higher charges, they did not have significantly longer LOS. This result, however, might be due to the low frequency of MI compared with CF in the database, which yielded relatively little statistical power to detect such an association.
These findings are consistent with those of previous studies that described associations between cardiac dysfunction and adverse outcomes after aSAH, possibly by a reduction of cerebral perfusion pressure during periods of cerebral vasospasm.14,15 In a recent study of 588 patients with SAH from the Intraoperative Hypothermia Aneurysm Surgery Trial, Coghlan et al16 reported that specific cardiac dysfunctions (bradycardia, relative tachycardia, and nonspecific ST- and T-wave abnormalities) are strongly and independently associated with mortality after SAH. In contrast, other studies have suggested that the myocardial irregularities occur from temporary repolarization that has little prognostic implication for cardiac alterations with SAH.17-19 These studies showed that patients with evidence of cardiac dysfunction may not have a permanent cardiac complication. The several types of cardiac manifestation (myonecrosis, left ventricular wall-motion abnormalities, electrocardiographic changes) generally resolve within days to weeks.20,21 Additionally, the most recent prospective observational study of 108 patients with nontraumatic intracranial hemorrhage reported that each repolarization abnormality has characteristic predisposing factors and that specific ECG changes were associated with a poorer functional outcome, which indicated that these changes may be linked to neurological compromise.22
It is not clear in which patients with SAH important cardiac complications will develop. This study identified potential predictors of cardiac complications, which may worsen neurological outcomes, as shown in Figures 1 and 2. We observed a higher average age for aSAH patients with CF or MI (vs those without CF or MI) admitted to US hospitals. The odds of CF and MI are estimated to multiply by 1.03 (increase by 3%) and by 1.02 (increase by 2%) for each additional year of age, respectively. These results might be due to the positive correlation between prevalence of cardiac abnormality and age; thus, older patients have a higher incidence of cardiac complications in the setting of acute aSAH.23 We also observed that the incidence of CF was higher among men than among women, higher among blacks than among whites, but lower among races other than blacks or Hispanics than among whites. For MI, the incidence was higher among blacks than among whites with aSAH. Unfortunately, the high percentage of unrecorded race information in the NIS prevents a definitive conclusion about whether race is a risk factor for cardiac complications in aSAH patients.
If there are signs of cardiac complications, management of these complications should be based on the patient's clinical condition, especially in the setting of vasospasm. Intensive cardiac monitoring with serial enzymes, ECGs, and echocardiograms is recommended. Additionally, continuous monitoring of hemodynamic variables, such as pulmonary capillary wedge pressure and cardiac index, will allow aggressive but careful fluid management while monitoring cardiac status.24 Even though the manifestations may be transient, management should focus on supportive care that balances cardiac needs with the underlying neurological condition.25
Unlike many previous studies, this study included patient's cardiac history and coronary risk factors (hypertension, DM, smoking status) in the analysis. Although cardiac complications, of course, occur in patients without a history of cardiac disease,26 in this study, as expected, patients with cardiac history and DM were more likely to experience CF. In addition, there was a greater incidence of CF in men with aSAH than in women, which is not different from the incidence of CF in the general population.27
Although we believe that our study has significant value, there are limitations to this study. The primary limitations are directly related to the use of the NIS data—a large database with a predefined data structure. The predefined data structure of the NIS does not include the nature, timing, and degree of cardiac dysfunction in patients and prevents firm conclusions about the causal nature of the observed associations between cardiac abnormalities and mortality/morbidity. Second, the type of cardiac abnormalities, such as abnormal ECG, elevated cardiac enzymes, and wall-motion abnormality, could not be classified. The NIS data do not include more detailed information, such as electrocardiography, echocardiography coronary angiography findings and the level of cardiac enzyme, which would let us better define CF or MI. Third, the database is vulnerable to coding errors, which are a well-established limitation of such data sources,28 although it seems likely that such errors are random rather than systematic and therefore have little effect on the statistical results. The use of the NIS specifically has been validated previously in the literature, including several articles discussing the treatment of intracerebral aneurysms.29,30 Finally, we did not include SAH severity (World Federation of Neurological Surgeons, Hunt-Hess, or Fisher grade), as these variables are not available in the NIS. Poor clinical grade is associated with the occurrence of at least 1 severe medical complication.31,32 Furthermore, the NIS offered no way to identify patients with the presence of a diffuse, thick deposition of clot demonstrated by computed tomography at the time of admission, which is more likely to occur in those who later experience severe complications.
This study presents the first nationwide report of the association of cardiac complications with outcomes in aSAH, as well as the first attempt to identify potential predictors associated with cardiac complications in such a large sample of patients. The results of our study suggest that an important association exists between cardiac complications and mortality/morbidity in aSAH patients. aSAH patients with cardiac complications appear to have a higher mortality rate, longer LOS, and higher hospitalization costs compared with those without cardiac complications. Therefore, any type of cardiac dysfunction should be managed aggressively in aSAH patients who present with clinical signs or symptoms of potentially devastating cardiovascular complications. The mechanism of the higher mortality and morbidity in patients with CF requires further research to identify modifiable factors clearly and to determine better treatments.
The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
1. Solenski NJ, Haley EC Jr, Kassell NF, et al.. Medical complications of aneurysmal subarachnoid hemorrhage: a report of the multicenter, cooperative aneurysm study. Participants of the Multicenter Cooperative Aneurysm Study. Crit Care Med. 1995;23(6):1007–1017.
2. McDonald RJ, Cloft HJ, Kallmes DF. Impact of admission month and hospital teaching status on outcomes in subarachnoid hemorrhage: evidence against the July effect. J Neurosurg. 2012;116(1):157–163.
3. Mutoh T, Kazumata K, Ajiki M, Ushikoshi S, Terasaka S. Goal-directed fluid management by bedside transpulmonary hemodynamic monitoring after subarachnoid hemorrhage. Stroke. 2007;38(12):3218–3224.
4. McLaughlin N, Bojanowski MW, Girard F, Denault A. Pulmonary edema and cardiac dysfunction following subarachnoid hemorrhage. Can J Neurol Sci. 2005;32(2):178–185.
5. Kim DH, Haney CL, Van Ginhoven G. Reduction of pulmonary edema after SAH with a pulmonary artery catheter-guided hemodynamic management protocol. Neurocrit Care 2005;3(1):11–15.
6. Qureshi AI, Suri MF, Sung GY, et al.. Prognostic significance of hypernatremia and hyponatremia among patients with aneurysmal subarachnoid hemorrhage. Neurosurgery. 2002;50(4):749–755; discussion 755-756.
7. Nakagawa I, Kurokawa S, Nakase H. Hyponatremia is predictable in patients with aneurysmal subarachnoid hemorrhage—clinical significance of serum atrial natriuretic peptide. Acta Neurochir. 2010;152(12):2147–2152.
8. McGirt MJ, Mavropoulos JC, McGirt LY, et al.. Leukocytosis as an independent risk factor for cerebral vasospasm following aneurysmal subarachnoid hemorrhage. J Neurosurg. 2003;98(6):1222–1226.
9. Parkinson D, Stephensen S. Leukocytosis and subarachnoid hemorrhage. Surg Neurol. 1984;21(2):132–134.
10. Franco C, Khaled B, Afonso L, Raufi M. Acute subarachnoid hemorrhage and cardiac abnormalities: takotsubo cardiomyopathy or neurogenic stunned myocardium? A case report. Cases J. 2010;3:81.
11. Khechinashvili G, Asplund K. Electrocardiographic changes in patients with acute stroke: a systematic review. Cerebrovasc Dis. 2002;14(2):67–76.
12. Brouwers PJ, Westenberg HG, Van Gijn J. Noradrenaline concentrations and electrocardiographic abnormalities after aneurysmal subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry. 1995;58(5):614–617.
13. Busani S, Rinaldi L, Severino C, Cobelli M, Pasetto A, Girardis M. Levosimendan in cardiac failure after subarachnoid hemorrhage. J Trauma. 2010;68(5):E108–E110.
14. Yarlagadda S, Rajendran P, Miss JC, et al.. Cardiovascular predictors of in-patient mortality after subarachnoid hemorrhage. Neurocrit Care. 2006;5(2):102–107.
15. Crago EA, Kerr ME, Kong Y, et al.. The impact of cardiac complications on outcome in the SAH population. Acta Neurol Scand. 2004;110(4):248–253.
16. Coghlan LA, Hindman BJ, Bayman EO, et al.. Independent associations between electrocardiographic abnormalities and outcomes in patients with aneurysmal subarachnoid hemorrhage: findings from the intraoperative hypothermia aneurysm surgery trial. Stroke. 2009;40(2):412–418.
17. Mayer SA, Fink ME, Homma S, et al.. Cardiac injury associated with neurogenic pulmonary edema following subarachnoid hemorrhage. Neurology. 1994;44(5):815–820.
18. Doshi R, Neil Dwyer G. A clinicopathological study of patients following a subarachnoid hemorrhage. J Neurosurg. 1980;52(3):295–301.
19. Zaroff JG, Rordorf GA, Ogilvy CS, Picard MH. Regional patterns of left ventricular systolic dysfunction after subarachnoid hemorrhage: evidence for neurally mediated cardiac injury. J Am Soc Echocardiogr. 2000;13(8):774–779.
20. Banki N, Kopelnik A, Tung P, et al.. Prospective analysis of prevalence, distribution, and rate of recovery of left ventricular systolic dysfunction in patients with subarachnoid hemorrhage. J Neurosurg. 2006;105(1):15–20.
21. Sommargren CE. Electrocardiographic abnormalities in patients with subarachnoid hemorrhage. American journal of critical care: an official publication, Am J Crit Care. 2002;11(1):48–56.
22. Junttila E, Vaara M, Koskenkari J, et al.. Repolarization abnormalities in patients with subarachnoid and intracerebral hemorrhage: predisposing factors and association with outcome. Anesth Analg. 2013;116(1):190–197.
23. Pullicino PM, Halperin JL, Thompson JL. Stroke in patients with heart failure and reduced left ventricular ejection fraction. Neurology. 2000;54(2):288–294.
24. Jain R, Deveikis J, Thompson BG. Management of patients with stunned myocardium associated with subarachnoid hemorrhage. AJNR Am J Neuroradiol. 2004;25(1):126–129.
25. Lee VH, Oh JK, Mulvagh SL, Wijdicks EF. Mechanisms in neurogenic stress cardiomyopathy after aneurysmal subarachnoid hemorrhage. Neurocrit Care. 2006;5(3):243–249.
26. Kono T, Morita H, Kuroiwa T, Onaka H, Takatsuka H, Fujiwara A. Left ventricular wall motion abnormalities in patients with subarachnoid hemorrhage: neurogenic stunned myocardium. J Am Coll Cardiol. 1994;24(3):636–640.
27. Mercuro G, Deidda M, Bina A, Manconi E, Rosano GM. Gender-specific aspects in primary and secondary prevention of cardiovascular disease. Curr Pharm Des. 2011;17(11):1082–1089.
28. Berthelsen CL. Evaluation of coding data quality of the HCUP National Inpatient Sample. Top Health Inf Manag. 2000;21(2):10–23.
29. Andaluz N, Zuccarello M. Recent trends in the treatment of cerebral aneurysms: analysis of a nationwide inpatient database. J Neurosurg. 2008;108(6):1163–1169.
30. Hoh BL, Rabinov JD, Pryor JC, Carter BS, Barker FG II. In-hospital morbidity and mortality after endovascular treatment of unruptured intracranial aneurysms in the United States, 1996-2000: effect of hospital and physician volume. AJNR Am J Neuroradiol. 2003;24(7):1409–1420.
31. Tung P, Kopelnik A, Banki N, et al.. Predictors of neurocardiogenic injury after subarachnoid hemorrhage. Stroke. 2004;35(2):548–551.
32. Naidech AM, Kreiter KT, Janjua N, et al.. Cardiac troponin elevation, cardiovascular morbidity, and outcome after subarachnoid hemorrhage. Circulation. 2005;112(18):2851–2856.
It is already well-known that SAH patients have cardiac complications that should be treated based on the condition of the patient, which are the main results of the current study. Despite the high number of SAH patients in the database analyzed, there are major limitations due to the lack of essential information like nature, timing, and degree of cardiac dysfunction of the patients. Also, the severity of SAH itself cannot be classified so no reliable causality with classification can be made nor precise recommendations and guidelines. It is cumbersome to establish a database not to talk about 1 with tens of thousands of patients, and naturally many important variables may be missing. Notably, some of the variables measured in this study (eg, primary payer) are applicable only in the United States with limited value elsewhere. It would be important to identify early those SAH patients in whom clinically relevant cardiac complications will actually develop, and we still need further studies to elucidate that.
1. What percentage of patients may have serum troponin elevations following aneurysmal subarachnoid hemorrhage?
2. Which of the following is a risk factor for myocardial infarction following aneurysmal subarachnoid hemorrhage?
a. History of smoking
b. History of diabetes mellitus
c. Older age
d. Male gender
e. White race
3. Following aneurysmal subarachnoid hemorrhage, which cardiac complication is the most significant independent predictor of mortality?
a. Myocardial infarction
b. Cardiac failure
c. Heart block
e. Myocardial stunning
Cardiac failure; Myocardial infarction; Subarachnoid hemorrhage
Copyright © by the Congress of Neurological Surgeons
Highlight selected keywords in the article text.