The Impact of Myocardial Infarction vs. Pneumonia on Outcome in Acute Ischemic Stroke

Abstract

The aim of this study was to examine the association between MI and PNA in the setting of acute ischemic stroke and patient outcome. Eligible patients were identified from a prospectively collected stroke registry and included if transthoracic echocardiography (TTE) was performed during their inpatient stay. 426 patients met inclusion criteria (mean age 64, 73% Black, 48% female). Twenty-one patients (4.9%) experienced an MI. Patients who later suffered a MI initially presented with more severe strokes (median NIHSS 7 vs. 5, p=0.014). More patients in the MI group experienced pneumonia (26% vs. 9%, p=0.004). After adjusting for age, baseline glucose and NIHSS, the odds of in-hospital mortality for patients with MI was 3 times that of those without MI (OR 3.2 95% CI 1.1–9.7, p=0.036). When adjustment was made for pneumonia, age, baseline glucose and NIHSS, MI was no longer significantly related to in-hospital mortality (OR 2.5 95% CI 0.8–8.2, p=0.131). In our sample, while MI was significantly associated with in-hospital mortality, this association was attenuated after adjusting for presence of pneumonia. Our findings raise the question as to whether the prevention of pneumonia could improve in-hospital mortality among patients who experience MI in the setting of ischemic stroke.

I. Introduction

A cute ischemic stroke (AIS) and myocardial infarction (MI) share similar atherosclerotic pathophysiologic mechanisms and risk factors, leading to an intricate and codependent relationship between the two.^1–3 Brain tissue affected by AIS loses auto-regulatory mechanisms, forcing cerebral blood flow to become more dependent on cardiac function for appropriate perfusion. Further, AIS patients may experience elevated afterload and systolic dysfunction, resulting in higher in-hospital mortality, and high blood pressure being associated with poor clinical outcomes.^{4, 5} Despite this, previous research has shown that low left ventricular ejection fraction (LVEF) is not an independent, significant predictor of short term functional outcome in ischemic stroke patients after adjusting for stroke severity and admission glucose.⁶

Cardiotoxic catecholamines induce a cyclic AMP-mediated increase in cellular Ca²⁺ overload as well as a vasospasm of epicardial coronary arteries, further exacerbating cardiac injury during AIS.⁷ This same overactivity of the sympathetic nervous system can cause immunodeficiency, rendering patients more vulnerable to infection, including pneumonia (PNA).⁸ PNA has been shown to influence outcomes in AIS patients with an increased odds of poor functional outcome even after adjustment.⁹ The aim of this study was to examine the association between MI in the setting of acute ischemic stroke, patient outcomes and other clinical factors.

II. Methods

We conducted a single-site cross-sectional analysis of patients with acute ischemic stroke admitted to our stroke center between July 1, 2008 and December 31, 2010. Eligible patients were identified retrospectively from a prospectively collected stroke registry and included if transthoracic echocardiography (TTE) was performed during their inpatient stay. MI was defined as troponin >1.0 ng/mL in the setting of clinical symptoms or electrocardiogram changes.¹⁰ Baseline demographic information, imaging studies, laboratory values, and early outcomes were collected.¹¹ Stroke severity was measured using the National Institutes of Health Stroke Scale (NIHSS). Stroke subtype was defined according to Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification.¹² As previously described, pneumonia was defined as a new infiltrate on chest radiography with appropriate clinical signs and symptoms.¹³ Good functional outcome was defined as a modified Rankin Scale score (mRS) of 0–2. MI and non-MI groups were compared using Pearson Chi-square and Wilcoxon Rank Sum. Logistic regression was used to assess the odds of in-hospital mortality.

III. Results

Four hundred and twenty-six patients met inclusion criteria (mean age 64, 73% Black, 48% female). Twenty-one patients (4.9%) had an MI during hospitalization for AIS. Table 1 provides the baseline comparison between the MI and non-MI groups. There was no significant difference in age, race, or gender between the MI and non-MI groups. A significantly higher proportion of patients in the MI group had a past medical history that included coronary artery disease (p<0.001) and diabetes (p=0.003). Further, a higher proportion of patients in the MI group were on a statin at the time of admission, as compared to the non-MI group (p=0.010). Patients who later suffered a MI initially presented with more severe strokes (median NIHSS 7 vs. 5, p=0.014) and higher median glucose levels (130 vs. 114, p=0.048).

Table 1.

Demographic Factors Influencing MI in AIS

	MI (N=21)	No MI (N=405)	p-value
Age, years, median (range)	69 (41–88)	64 (22–95)	0.141
Sex, No. Female (%)	10 (35.7)	194 (42.8)	0.460
Race, No. (%)			0.946
Black	19 (67.9)	291 (65.0)
White	8 (28.6)	143 (31.9)
Other	1 (3.5)	14 (3.1)
Past Medical History, No. (%)
Coronary Artery	14 (50.0)	89 (19.7)	<0.001
Disease
Diabetes	17 (60.7)	149 (33.2)	0.003
Hyperlipidemia	16 (61.5)	190 (42.5)	0.057
Stroke	12 (42.9)	170 (37.6)	0.579
Active Smoker	3 (11.5)	164 (36.8)	0.009
Hypertension	22 (84.6)	356 (79.3)	0.512
Admission NIHSS, median (range)	7 (1–31)	5 (0–29)	0.014
Home Medications, No. (%)
Statin	17 (63.0)	170 (38.1)	0.010
Laboratory Values
Admission glucose, mg/dL, median (range)	130 (81–477)	114 (46–663)	0.048
Fasting HDL, mean mg/dL (SD)	37 (11–89)	42 (12–140)	0.241
Hematocrit, mg/dL median (range)	35.0 (14.5–49.1)	39.5 (16.3–55.5)	0.004
TOAST			0.148
Cardioembolic	13 (46.4)	109 (24.2)
Large vessel disease	5 (17.9)	119 (26.4)
Small vessel disease	3 (10.7)	94 (20.8)
Cryptogenic no more than 1 cause	0 (0.0)	13 (2.9)
Cryptogenic no cause	5 (17.9)	75 (16.6)
Other	2 (7.1)	41 (9.1)

Abbreviations: MI: myocardial infarction, AIS: acute ischemic stroke, SD: standard deviation, NIHSS: National Institutes of Health Stroke Scale, LDL: low density lipoprotein, HDL: high density lipoprotein, TOAST: Trial of Org 10172 in Acute Stroke Treatment.

Table 2 compares in-hospital complications and short-term outcomes in patients who experienced an MI and those who did not experience an MI during their inpatient stay for ischemic stroke. More patients in the MI group experienced pneumonia (26% vs. 9%, p=0.004), as compared to the non-MI group. Further, patients in the MI group appeared to have worse short-term outcomes as demonstrated by discharge NIHSS, discharge mRS, the proportion with unfavorable discharge disposition, and in-hospital mortality (Table 2). In addition, patients in the MI group had longer LOS (9 vs. 5 days, p<0.001). After adjusting for age, baseline glucose and NIHSS, the odds of in-hospital mortality for patients with MI was 3 times that of those without MI (OR 3.2 95% CI 1.1–9.7, p=0.036). When adjustment was made for pneumonia, in addition to age, baseline glucose and NIHSS, MI was no longer significantly related to in-hospital mortality (OR 2.5 95% CI 0.8–8.2, p=0.131).

Table 2.

Outcomes for MI during hospitalization for AIS

	MI (N=21)	No MI (n=405)	P-value
Infection, No. (%)
Pneumonia, No. (%)	7 (25.9)	40 (8.9)	0.004
Discharge NIHSS, median (range)	10 (0–40)	2 (0–42)	0.001
Discharge mRS, median (range)	4 (6)	3 (6)	0.001
Good functional outcome (mRS 0–2), No. (%)	5 (15.8)	193 (43.7)	0.010
Length of stay, median (range)	9 (37)	5 (51)	<0.001
Discharge Disposition, No. (%)			0.075
Home	7 (31.8)	239 (56.6)
Inpatient Rehab	9 (40.9)	125 (29.6)
Unfavorable Disposition (all others)	6 (27.2)	58 (13.7)
In-hospital mortality, No. (%)	6 (21.4)	26 (5.8)	0.001

Abbreviations: MI: myocardial infarction, mRS: modified Rankin Scale, NIHSS: National Institutes of Health Stroke Scale.

IV. Discussion

Our study found that patients who experienced an MI in the setting of acute ischemic stroke more frequently developed pneumonia. This is not surprising given that sympathetic nervous system over-activity is thought to lead to immunodeficiency, leaving patients more susceptible to infection.⁸ Further, we found higher proportions of in-hospital mortality in MI patients, when compared to non-MI patients. The higher odds of in-hospital mortality remained after adjustment for age, stroke severity, and admission glucose, indicating that MI was a significant predictor of in-hospital mortality, independent of covariates previously known to affect short-term outcomes in ischemic stroke. This is in keeping with findings from the Framingham cohort where stroke patients with coronary heart disease or congestive heart failure had worse survival rates when compared to stroke without these cardiac comorbidities.¹⁴

Interestingly, we found that after adjustment for pneumonia, the association between MI and in-hospital mortality was no longer significant. This finding is in keeping with the well-described association between pneumonia and poor outcome in stroke patients.¹⁵ While aspiration is a common cause of pneumonia in stroke patients with difficulty controlling their secretions, the proportion of MI patients in our sample that developed pneumonia (26%) was higher than previous reports (9%).^{13, 16} While strict definitions were used to classify both MI and pneumonia, the possibility of misclassification bias cannot be excluded.

Unfortunately the cross-sectional nature of our study prevented us from being able to assess if the relationships we observed between MI, pneumonia, and in-hospital mortality were causal. Additionally, our study is limited by its small sample size and a relatively small number of events. Only 21 patients out of 426 patients evaluated experienced an MI, and a total of 47 out of 426 patients acquired pneumonia during their hospital stay. Our limited sample size may have prevented us from detecting existing differences in groups. Additionally, while efforts were made to maximize internal validity, our sample is likely representative of the catchment area of our urban academic medical center in the southeastern US. It remains to be determined whether our findings are generalizable to nonurban centers in other regions of the US.

Although not without limitations, our findings reinforce the need for identifying patients at high-risk for pneumonia and implementing preventative measures that help to reduce this risk. Clinical scores such as A2DS2 can successfully predict incidence of pneumonia in acute ischemic stroke by evaluating demographics, comorbidities and clinical characteristics.¹⁷ Further, dysphagia and laryngeal cough reflex evaluations in post-stroke patients should be implemented,^{18, 19} particularly in patients with lesions known to increase incidence of pneumonia.²⁰