Background. Stroke is a leading cause of death and disability in developed countries. The major factor affecting long-term survival other than age is the disability severity caused by stroke. The modified Rankin Scale (mRS) is a global functional endpoint measurement used in acute stroke to evaluate the degree of disability or dependence in daily life activities. The objective of this study was to assess the effects of sociodemographic factors, concomitant disease states, and some measures performed in the emergency department (ED) on patients’ disability. Methods. We conducted a retrospective study on ischemic stroke patients admitted to Intensive Care Unit of three Lebanese university hospitals between June and December 2016. Patients were excluded if they had been discharged from ED without hospital admission or if mRS was not performed. The mRS was further subdivided into two categories considered as “good prognosis” (0-2 or 0-3) and “poor prognosis” (>2 or > 3). Results. 204 patients were included in the study with mean age of years, hypertension was the most previous concomitant past medical disease (77.1%), and 27.1% of these patients had previous history of stroke. No significant differences were found in both mRS categories for all sociodemographic factors, and past medical history except that arrhythmia was significantly more common in the higher mRS and > 3. Based on multivariable analysis, there was a trend for previous intake of calcium channel blocker to be associated with lower mRS at admission (beta -0.586). However, intracranial arterial stenosis, ED mg/dL, and performing brain imaging above 20 minutes after patient presentation to ED were significantly associated with higher mRS scores at discharge with an ORa and (confidence interval) of 2.986 (0.814, 10.962), 3.301 (1.072, 1.261), and 1.138 (1.071, 9.080), respectively. Conclusion. mRS is affected by previous disease states, prescribed medications, and acute measures performed in ED. It is also influenced by intracranial arterial stenosis etiology, which is associated with worse outcome.
1. Introduction
Stroke is the third leading cause of mortality after heart diseases and cancer and a principal cause of severe long-term disability in adults [1]. The majority of strokes (80–85%) are of ischemic origin [2, 3].
In the last decade, several outcome measures were developed to assess the degree of poststroke disability, such as modified Rankin Scale (mRS), Barthel Index (BI), health-related quality of life (EQ5D-3L) [4]. However, mRS is a global functional endpoint measurement and the most frequently used index in acute stroke to evaluate the degree of disability or dependence in daily life activities following stroke and thereby easily communicate effects of treatments to physicians and patients [4, 5]. This scale is also a standard element in clinical practice and an outcome measure that must be used in all stroke survivors according to the Get with the Guidelines [6]. The mRS is an ordinal disability score which categorizes patients among 7 levels ranging from 0 “no disability” to 6 “death” [7].
In clinical practice, interpretation of mRS is challenging since it is a nonlinear scale; thus, dichotomization of outcomes has been performed in research studies to ensure consistent scoring, minimization of both subjective judgment, and variability in score assignment [8–11]. Indeed, in clinical trials, mRS is often further subdivided into two categories according to a cut-off value of either 2 or 3 defining good prognosis as or < 4 and poor prognosis otherwise [12]. However, in current practice, even if patients with mRS of 3 are grouped with patients with mRS of 4-6 with the basic assumption that score of 3 is more similar to 4 than to 2 in terms of clinical outcomes, it raises concerns regarding the validity of this dichotomization given that mRS of 3 is typically considered of good clinical outcome and that patients with mRS 2 or 3 share a similar 7-year survival [13, 14]. The optimal and most objective cut-off value for mRS dichotomization is debatable as it depends on stroke severity, lacks the incorporation of the entire possible range of outcomes across the mRS, and raises difficulty as regards the interpretation of borderline score values of 2 and 3.
Several investigators performed poststroke outcome measurement following the entire ordinal distribution values of mRS, as this strategy is more powerful than the dichotomization approach, especially because it takes into consideration treatment effect that might occur over the entire range of mRS [15]. Result performance using the entire distribution of mRS values has greater statistical power as well than the dichotomized approach, mainly when treatment benefit is a continuous process [16].
Although mRS is widely applied for evaluating stroke patient outcomes, it has several pitfalls when used to measure poststroke disability. An extensive literature documents the negative effect of patient comorbidities including cardiovascular disease, diabetes, arthritis, surgery, and socioeconomic factors on physical functioning, cognitive abilities, and overall health status so that these factors may have detrimental effect on the mRS [17–22]. This is extremely important as comorbidities are common in stroke patients, and stroke incidence in low socioeconomic populations is especially high [23]. It is essential for clinicians to take into account these various attributes and avoid misapplication and misinterpretation of mRS. Hence, the objective of this study was to compare the effects of sociodemographic factors, concomitant disease states, and some measures performed in the ED on mRS.
2. Methods
2.1. Study Design
This was a retrospective observational study in which hospital records of all patients admitted to the Intensive Care Unit (ICU) of three Lebanese university hospitals between June and December 2016 were reviewed for inclusion. The list of hospitals, provided by the Lebanese Ministry of Public Health, was used to randomly select the centers. All adult patients presented to the emergency department with the diagnosis of ischemic stroke during the study period and who were subsequently admitted to the ICU were included in the study. No attempt was made to verify accuracy of the physician’s diagnosis, because the aim of this study was to assess the impact of sociodemographic factors and medical history on mRS prognosis. Patients were excluded if the patient had been diagnosed with transient ischemic attack, had been discharged from the ED without hospital admission, or if the mRS was not available. Eligible participants were randomly selected from each center using the list generated from the hospital administrator for all patients diagnosed with ischemic stroke from the medical record department for possible enrollment in the study.
2.2. Data Collection
A medical record review was performed on site by the principal investigator, through a standardized data collection sheet. We collected baseline information including patient demographics, vascular risk factors, stroke severity (mRS at admission and upon discharge), and acute stroke management. mRS was documented by the attending physician on all patients’ medical charts both upon presentation to the emergency department and upon discharge.
Vascular risk factors included hypertension, dyslipidemia, atrial fibrillation, previous stroke, smoking history, alcohol consumption, marital status, and body mass index (BMI) at admission which was categorized into underweight, normal, overweight, and obese [24]. Hypertension was defined as systolic mmHg or diastolic mmHg, any use of antihypertensive drug, or self-reported history of hypertension. We also included detailed assessment of antihypertensive medications including dose, frequency of administration, and therapeutic class [25]. Diabetes mellitus was defined as fasting glucose mmol/L, nonfasting glucose mmol/L, use of any glucose-lowering drugs, or self-reported history of diabetes [26]. Dyslipidemia was defined as serum mmol/L, low-density lipoprotein mmol/L, high-density lipoprotein mmol/L, use of any lipid-lowering drugs, or self-reported history of dyslipidemia [27]. Atrial fibrillation was defined as history of atrial fibrillation confirmed by at least one electrocardiogram or presence of arrhythmia during hospitalization. The treatment of atrial fibrillation defined by the use of anticoagulation agents or antiplatelet drugs during hospitalization and after discharge was also considered [28]. We followed the diagnosis of ischemic stroke set by the physician on the medical chart which is “an episode of neurological dysfunction caused by focal cerebral, spinal, or retinal infarction based on pathological, imaging, or other objective evidence of ischemic injury in a defined vascular distribution.” Etiologic subtypes of ischemic stroke were classified into cardioembolic, intracranial arterial stenosis, or other origins [29].
In addition, neurological assessment was documented through mRS. We then subdivided the mRS into two categories as “good prognosis” (either 0-2 or 0-3) and poor prognosis (either >2 or> 3) [12].
Data collection also gathered information about all measures done in the ED including brain imaging performance delay (within 20 minutes versus >20 minutes from hospital presentation) and administration of antiplatelet and antihypertensive medications.
2.3. Outcomes
The primary outcome was to assess the impact of different factors including prestroke sociodemographic factors and medication prescriptions along with etiologic subtypes of stroke on mRS scores both at hospital admission and discharge. The secondary outcome was to assess the impact of different measures performed in the hospital like brain imaging delay, glucose level, and antithrombotic therapy in the ED on mRS at discharge.
2.4. Statistical Analysis
The questionnaires were coded, and the collected data were introduced using the Statistical Package for Social Sciences (SPSS) software, version 23.0 by an independent person who was unaware of the objectives of the study. All continuous variables were presented as mean and standard deviation (SD), and categorical variables were presented as percentages. Correlations between sociodemographic factors, medical history, and mRS categories were determined by the Pearson chi-square test or Fisher’s exact test when Pearson chi-square test could not be applied. Paired sample -test was used to assess mean difference between mRS values at hospital admission and discharge.
Multiple logistic regression models were used to assess association between sociodemographic factors, medical history that showed a in the bivariate analysis and mRS both at hospital admission and discharge (taken as the dependent variable). Potential confounders may be eliminated only if , in order to protect against residual confounding. We also performed a linear regression taking both mRS at hospital admission and discharge as the dependent variable and all other variables as independent. We ensured model adequacy by the use of the Hosmer Lemeshow test which is done for the logistic variables (mRS categories) to calculate if the observed event rates match the expected event rates. A was considered as statistically significant.
3. Results
3.1. Sociodemographic Characteristics of Participants
From a total number of 400 patients screened for possible inclusion in the study, 204 subjects had documented mRS values at hospital admission and discharge and were enrolled in the study. Among the 204 participants, 132 (64.7%) were males. Mean age was years. Concerning BMI categories, 12 (29.3%) patients were normal, 17 (41.4%) overweight, and 12 (29.3%) obese. Almost half of the participants were smokers (48.5%), while the majority was nonalcoholic (64.7%), 75.1% were married, and 52.9% lived in Beirut district. Only 0.5% of the participants were physically active. As regards past medical history, hypertension was present in 77.1%; previous stroke or TIA prevalence was 27.1%. Atrial fibrillation, hyperlipidemia, and coronary artery bypass graft (CABG) were, respectively, found in 12.5%, 19.8%, and 4.2% of the participants. Regarding past medication history, 5.7% of the patients were on oral vitamin K antagonists, 0.5% on Low Molecular weight Heparin (LMWH), 38.8% on aspirin, 19.2% on clopidogrel, and 36.1% on lipid lowering therapy. Among patients with hypertension, 55.4% were on beta blockers, 34% were on calcium channel blockers (CCBs), and 25.3% were on dual antihypertensive drugs. The mean mRS scores at baseline and upon discharge were, respectively, and (see Table 1).
Variable
(%)
Mean age
65.40 years
Gender
Male
132 (64.7)
Female
72 (35.3)
BMI
Normal
12 (29.3)
Overweight
17 (41.4)
Obese
12 (29.3)
Marital status
Single
39 (21.1)
Married
139 (75.1)
Divorced
7 (3.8)
Residence area
Beirut
108 (60.3)
Mount Lebanon
33 (18.4)
Bekaa
5 (2.8)
North Lebanon
2 (1.1)
South Lebanon
31 (17.3)
Cigarette smoker
Yes
99 (48.5)
Alcohol consumption
Yes
21 (11.4)
Past medical history
Hypertension
74 (77.1)
Stroke or TIA
26 (27.1)
Atrial fibrillation
12 (12.5)
Hyperlipidemia
19 (19.8)
Baseline mRS upon admission mean
1 interquartile range (2, 6)
Baseline mRS upon hospital discharge mean
2 interquartile range (0, 6)