Association of Embolic Sources With Cause-Specific Functional Outcomes Among Adults With Cryptogenic Stroke

Abstract

Importance: It is unknown whether poststroke outcome varies between different potential causes in patients with cryptogenic stroke.

Objective: To investigate whether functional outcome differs according to potential embolic sources after cryptogenic stroke.

Design, setting, and participants: This multicenter, hospital-based, prospective stroke registry cohort study investigated potential embolic sources on admission and assessed 3-month outcome in patients with ischemic stroke hospitalized at 7 stroke centers in the Fukuoka Stroke Registry. This registry enlisted 9866 consecutive patients with acute ischemic stroke who were enrolled from June 11, 2007, to May 31, 2016, in Fukuoka, Japan. Patients with small vessel occlusion (n = 3130), extracranial and intracranial atherosclerosis causing at least 50% luminal stenosis in arteries supplying the area of ischemia (n = 2011), and other specific uncommon causes of stroke identified (n = 301) were excluded. Potential embolic sources were diagnosed in patients with embolic stroke of undetermined source (ESUS) based on the following criteria proposed by the Cryptogenic Stroke/ESUS International Working Group: minor-risk potential cardioembolic sources (MCS) (n = 209), covert paroxysmal atrial fibrillation (CPAF) (n = 43), cancer associated (CA) (n = 79), arteriogenic emboli (AE) (n = 522), paradoxical embolism (PE) (n = 190), and undetermined embolism (unidentified or ≥2 potential embolic sources) (UE) (n = 1120).

Main outcomes and measures: The association between potential causes and functional outcome was evaluated in reference to cardioembolic stroke (CE) caused by major-risk cardioembolic sources after adjusting for age, sex, National Institutes of Health Stroke Scale score on admission, and reperfusion therapy using logistic regression analysis. Functional dependency (modified Rankin Scale score, 3-5) was evaluated at 3 months after onset.

Results: The study enrolled 2261 patients with CE (mean [SD] age, 78.4 [10.7] years, 51.8% male) and 2163 patients with ESUS (mean [SD] age, 72.4 [12.6] years, 57.1% male). Compared with CE (median National Institutes of Health Stroke Scale score, 8 [interquartile range {IQR}, 3-17]), baseline neurological deficits did not differ in MCS (median, 7 [IQR, 2-18]), CPAF (median, 6 [IQR, 2-18]), and CA (median, 5 [IQR, 2-13]) but were less severe in AE (median, 2 [IQR, 1-4]), PE (median, 2 [IQR, 1-4]), and UE (median, 3 [IQR, 1-7]). Multivariable-adjusted odds ratios of functional dependency significantly increased in CA (3.61; 95% CI, 1.52-8.54 vs CE) but decreased in PE (0.33; 95% CI, 0.16-0.71 vs CE).

Conclusions and relevance: Potential causes are associated with poststroke outcome in patients with cryptogenic stroke. Embolic sources potentially underlying cryptogenic stroke should be considered significant variables associated with outcome.

Figure 1.. Neurological Severity

The National Institutes of Health Stroke Scale (NIHSS) scores on admission (A) and at discharge (B) are shown according to each potential cause compared with cardioembolic stroke (CE). The NIHSS score can range from 1 to 42 and measures neurological severity categorized into 3 clinically meaningful groups (mild if 0-4, moderate if 5-14, and severe if ≥15). In patients who died during hospitalization, the NIHSS score at discharge was assigned the maximum score of 42. The box indicates ranges between lower quartile score and upper quartile score, and the horizontal line in the box represents the median score. Lower and upper vertical bars indicate the 10th and 90th percentiles, respectively. AE indicates arteriogenic emboli; CA, cancer associated; CPAF, covert paroxysmal atrial fibrillation; MCS, minor-risk potential cardioembolic sources; PE, paradoxical embolism; and UE, undetermined embolism. ^aStatistically significant at P < .05 vs CE by multiple comparisons.

Figure 2.. Functional Outcomes

Functional dependency (A and B) and poor functional outcome (C and D) in embolic stroke of undetermined source are shown according to each potential cause compared with cardioembolic stroke (CE). Functional outcomes were evaluated at discharge (A and C) and at 3 months of stroke onset (B and D). Odds ratio (OR) (square) and 95% CI (bars) of functional outcomes are shown for each potential cause with reference to CE (diamond). The multivariable model included age, sex, National Institutes of Health Stroke Scale score (measuring neurological severity) on admission (mild if 0-4, moderate if 5-14, and severe if ≥15), and reperfusion therapy. The sizes of squares or diamonds are proportional to the sizes of the subgroups of each potential cause. Patients who died during hospitalization or within 3 months were excluded from the analysis for functional dependency. Patients who were lost to follow-up at 3 months were also excluded from the analysis for functional outcome at 3 months. AE indicates arteriogenic emboli; CA, cancer associated; CPAF, covert paroxysmal atrial fibrillation; MCS, minor-risk potential cardioembolic sources; PE, paradoxical embolism; and UE, undetermined embolism.

Figure 3.. Stroke Recurrence and Mortality

Rates of stroke recurrence (A) and mortality (B) are shown according to each potential cause compared with cardioembolic stroke (CE). Patients whose data regarding stroke recurrence or mortality at 3 months were missing were excluded from the analysis for the respective adverse events at 3 months. The number of patients in each group is shown below the graph. AE indicates arteriogenic emboli; CA, cancer associated; CPAF, covert paroxysmal atrial fibrillation; MCS, minor-risk potential cardioembolic sources; PE, paradoxical embolism; and UE, undetermined embolism. ^aStatistically significant at P < .05 vs CE by multiple comparisons.