Abstract
Objective
Mechanical thrombectomy (MT) is the standard of care for acute large vessel occlusion (LVO); however, off-hours treatment may be delayed because of limited staffing. We compared MT outcomes between on-hours and off-hours at a single center and investigated factors associated with 90-day modified Rankin Scale (mRS) 0–3.
Methods
We retrospectively analyzed 101 consecutive patients with anterior circulation LVO and a pre-stroke mRS 0–2 who underwent MT at Tokorozawa Central Hospital between January 2020 and December 2025. Patients were classified into on-hours (9:00–17:00 on Mondays through Saturdays, excluding national holidays and the period from December 31 to January 3) and off-hours groups based on hospital arrival time. The primary endpoint was 90-day mRS 0–3. Secondary endpoints included Thrombolysis in Cerebral Infarction (TICI) grade 2b–3, symptomatic intracranial hemorrhage (sICH), and any intracranial hemorrhage. Time metrics, including onset-to-door time (ODT), door-to-image time, door-to-needle time, image-to-puncture time (IPT), door-to-puncture time (DPT), puncture-to-recanalization time (PRT), door-to-recanalization time (DRT), and onset-to-recanalization time (ORT), were compared between the 2 groups, and factors associated with 90-day mRS 0–3 were evaluated using univariable and multivariable logistic regression analyses.
Results
The on-hours group comprised 46 patients and the off-hours group 55 patients. Compared with the on-hours group, the off-hours group had significantly longer IPT (median 83 vs. 51 min, p <0.001), DPT (median 103 vs. 71 min, p <0.001), and DRT (median 151 vs. 124 min, p = 0.003). ODT and ORT did not differ significantly between the groups (median 105 vs. 89 min, p = 0.825, and median 285 vs. 234 min, p = 0.154, respectively). PRT did not differ significantly between groups (median 53 vs. 48 min, p = 0.626). The proportion of patients achieving 90-day mRS 0–3 did not differ significantly between groups (45.5% vs. 50.0%, p = 0.692). No significant differences were observed in TICI 2b–3 (80.0% vs. 82.6%, p = 0.802) or sICH (3.6% vs. 4.3%, p = 1.000). In the multivariable logistic regression, younger age (odds ratio [OR] 0.922, p = 0.002) was an independent predictor of 90-day mRS 0–3, whereas off-hours status was not (OR 0.917, p = 0.845).
Conclusion
Despite prolonged time from hospital arrival to treatment initiation in the off-hours group, the proportion of patients achieving 90-day mRS 0–3 did not differ significantly between groups. In multivariable logistic regression, younger age was an independent predictor of 90-day mRS 0–3. Further efforts are needed to identify the causes of off-hours delays and to improve our clinical care system.
Keywords: mechanical thrombectomy, large vessel occlusion, off-hours, acute ischemic stroke, endovascular treatment
Introduction
Mechanical thrombectomy (MT) is a standard treatment for acute large vessel occlusion (LVO) that can improve functional outcomes through early recanalization. A secondary analysis of the Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke Trials (HERMES) collaboration demonstrated that each 1-hour delay in successful recanalization was associated with a 19% reduction in the odds of achieving modified Rankin Scale (mRS) 0–2.1) Current guidelines recommend aiming for a door-to-puncture time (DPT) of ≤60 min and a door-to-recanalization time (DRT) of ≤90 min.2) Facilities performing MT are expected to organize their clinical care systems to achieve these benchmarks.2) However, during off-hours (defined as nights, weekends, and holidays), reduced in-hospital staffing may delay treatment initiation and lower recanalization rates compared with on-hours.3–12)
In Japan, the Work Style Reform for Physicians, which took effect in April 2024, in principle caps annual overtime at 960 hours per physician and obliges institutions to secure compensatory rest.13) Further, a system has been introduced that limits continuous duty to 28 hours and secures a rest interval of at least 9 hours between work shifts.14) Physicians whose annual overtime is within 960 hours are recommended to have at least 18 hours of rest after a night shift.15) However, for physicians permitted up to 1860 hours of annual overtime, securing this rest period is obligatory.15) Therefore, physicians who perform MT during off-hours may require compensatory rest thereafter, potentially reducing the number of physicians available to perform MT. A nationwide simulation has estimated that under these limitations, the physician workforce would be insufficient in several specialties, including neurosurgery, particularly in institutions that rely on physicians dispatched from other hospitals.13)
Our institution (Tokorozawa Central Hospital) has 2 Japanese Society for Neuroendovascular Therapy (JSNET)–certified specialists and 1 JSNET-certified instructor. However, these 3 physicians cover only approximately half of the overnight shifts, while the remaining shifts are covered by part-time neurosurgeons from other institutions or non-JSNET-certified physicians from our institution. Consequently, situations arise in which a full-time JSNET-certified specialist or instructor must be called from outside the hospital to perform MT during off-hours. In addition, catheterization laboratory staff (nurses and radiological technologists) are on standby at home during off-hours and must be called in before MT can be initiated. Although this staffing arrangement is reasonable from the perspective of preventing physicians from working overtime, it raises concerns that treatment delays during off-hours may adversely affect clinical outcomes.
Given this background, we retrospectively compared MT outcomes between on- and off-hours at our institution and examined the factors associated with 90-day mRS 0–3.
Materials and Methods
Study population
We retrospectively reviewed consecutive patients who underwent MT at our institution between January 1, 2020, and December 31, 2025. The inclusion criteria were as follows: 1) age ≥18 years; 2) symptom onset or last known well time within 24 hours; 3) anterior circulation occlusion involving the internal carotid artery (ICA) or the 1st or 2nd segments of the middle cerebral artery (M1 or M2); and 4) pre-stroke mRS 0–2. The patients were classified into on-hours and off-hours groups based on hospital arrival time. At our institution, regular working days are defined as Mondays through Saturdays, excluding national holidays and the period from December 31 to January 3. On-hours were defined as 9:00–17:00 on regular working days, while all other times were classified as off-hours.
Stroke care system
Patients with suspected LVO referred to our institution undergo evaluation by a neurosurgeon upon arrival. However, the neurosurgeon responsible for the initial management is not necessarily a JSNET-certified specialist or instructor; the responding neurosurgeon is assigned based on clinical availability, in consideration of outpatient duties and scheduled surgeries. Two emergency department (ED) nurses are assigned during on-hours, and additional personnel are mobilized from the general outpatient department as necessary. MRI is performed according to standard imaging protocols on all days. When intravenous thrombolysis with tissue-type plasminogen activator (IV tPA) and MT are deemed indicated, catheterization laboratory staff (nurses and radiological technologists) and a JSNET-certified specialist are notified, and MT is subsequently performed.
During off-hours, 2 ED nurses are on duty, while an on-duty neurosurgeon manages the emergency cases. The workflow from hospital arrival to MRI acquisition is identical to that during on-hours; however, the catheterization laboratory staff must be called from standby at home. Furthermore, when the on-duty neurosurgeon is not a full-time JSNET-certified specialist or instructor, the MT operator must be called from outside the hospital. In accordance with national guidelines, the indications for IV tPA and MT are determined by a JSNET-certified specialist or instructor, and MT is performed exclusively by a JSNET-certified specialist or instructor. Each MT procedure is performed by 1 or 2 operators; whether a 2nd operator is available depends on neurosurgeon scheduling and can be affected by concurrent outpatient duties, scheduled surgeries, or absences. Device preparation, including unpacking and setup, is performed exclusively by neurosurgeons. The choice of specific devices is not standardized by an institutional protocol; however, all MT procedures employed a combined (stent retriever and aspiration) technique.
Data collection
Baseline characteristics included age, sex, hypertension, diabetes mellitus, dyslipidemia, pre-stroke mRS, occlusion site, National Institutes of Health Stroke Scale (NIHSS) score at admission, diffusion-weighted image–Alberta Stroke Program Early CT Score (DWI-ASPECTS), stroke subtype classified according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification, and the number of MT operators.
The following time metrics were evaluated: onset-to-door time (ODT), door-to-image time (DIT), door-to-needle time (DNT), image-to-puncture time (IPT), DPT, puncture-to-recanalization time (PRT), and DRT. The onset-to-recanalization time (ORT) was defined as the interval from the reference onset time to recanalization, because several of the intervals listed above overlap and therefore cannot be summed. When the exact time of symptom onset was unknown, the last known well time was used as the reference onset time for the calculation of ODT and ORT. In patients in whom recanalization was not achieved (Thrombolysis in Cerebral Infarction [TICI] 0 or 1), the time of the final angiographic run was used as a surrogate for the time of recanalization. All time intervals were recorded in minutes.
Given the single-center retrospective design of this study and the potential for model instability arising from a limited number of outcome events, the primary endpoint was defined as 90-day mRS 0–3. Secondary endpoints included TICI grade 2b–3, symptomatic intracranial hemorrhage (sICH), and any intracranial hemorrhage. sICH was defined as parenchymal hematoma type 2 (PH2) with a neurological deterioration of ≥4 points on the NIHSS, according to the Safe Implementation of Thrombolysis in Stroke-Monitoring Study (SITS-MOST) criteria.16) Intracranial hemorrhage was defined as hemorrhagic infarction type 1 (HI1), HI2, PH1, or PH2 according to the SITS-MOST classification, or as subarachnoid hemorrhage.
Statistical analysis
Continuous variables were compared between the 2 groups using the Mann–Whitney U-test, while categorical variables were compared using Fisher’s exact test. Continuous and categorical variables are expressed as the median (interquartile range [IQR]) and as the number (percentage), respectively.
Univariable and multivariable logistic regression analyses were conducted to identify the factors associated with 90-day mRS 0–3. To facilitate clinical interpretation, the odds ratios for time variables were calculated per 10-min increment, rather than per min. Because our institution routinely uses a combined technique for thrombectomy, the type of thrombectomy device used was not included as a variable in the analysis; a comparison of different thrombectomy techniques was beyond the scope of this study.
A multivariable logistic regression model was subsequently constructed, with 90-day mRS 0–3 as the dependent variable. To avoid the instability of parameter estimates and overfitting that can arise from automated variable selection procedures, forced entry was employed instead of stepwise methods. Covariates were selected from factors with an established association with clinical outcome after MT in prior studies, with a target of events per variable (EPV) ≥10. The following 4 variables were entered into the model: off-hours status, age, NIHSS score at admission, and DWI-ASPECTS score. Standard logistic regression was applied for all analyses. All statistical analyses were conducted using EZR on R Commander version 1.70 (programmed by Y. Kanda), and statistical significance was set at p <0.05. This study was approved by the institutional review board of Tokorozawa Central Hospital (approval no. 2025-22-001-49).
Results
Baseline characteristics
During the study period, 132 patients underwent MT. Seven patients with basilar artery occlusion and 24 patients with a pre-stroke mRS ≥3 were excluded, yielding 101 patients with anterior circulation LVO, of whom 46 and 55 were classified into the on-hours and off-hours groups, respectively (Fig. 1). The baseline characteristics of the participants are summarized in Table 1. Figure 2 shows the 24-hour distribution of hospital arrival times. The median age was 80 years in both groups, with no significant difference (p = 0.899). The proportion of male patients was significantly higher in the on-hours group (73.9% vs. 43.6%, p = 0.003). No significant differences were observed in the prevalence of hypertension, diabetes mellitus, or dyslipidemia between groups. Left-sided occlusion was present in 24 patients (52.2%) in the on-hours group and 24 patients (43.6%) in the off-hours group (p = 0.429). The admission NIHSS score did not differ significantly between the groups (median 16 vs. 19, p = 0.133). The DWI-ASPECTS score did not differ significantly between the groups (median 8 vs. 7, p = 0.079).
Fig. 1. Flow chart: exclusion and inclusion criteria. Of 132 patients who underwent mechanical thrombectomy between January 2020 and December 2025, 31 were excluded (7 with basilar artery occlusion and 24 with a pre-stroke mRS ≥3). The remaining 101 patients were classified into the on-hours group (n = 46; 9:00–17:00 on regular working days) and the off-hours group (n = 55; all other times) according to hospital arrival time.
M1, 1st segment of the middle cerebral artery; M2, 2nd segment of the middle cerebral artery; mRS, modified Rankin Scale
Table 1. Baseline characteristics.
| On-hours (n = 46) | Off-hours (n = 55) | p Value | |
|---|---|---|---|
| Age, median (IQR) | 80 (72–85) | 80 (76–84) | 0.899 |
| Male | 34 (73.9%) | 24 (43.6%) | 0.003 |
| Hypertension | 22 (47.8%) | 32 (58.2%) | 0.323 |
| Dyslipidemia | 12 (26.1%) | 15 (27.3%) | 1.000 |
| Diabetes mellitus | 8 (17.4%) | 11 (20.0%) | 0.802 |
| Left-sided occlusion | 24 (52.2%) | 24 (43.6%) | 0.429 |
| NIHSS, median (IQR) | 16 (10–24) | 19 (13–26) | 0.133 |
| DWI-ASPECTS, median (IQR) | 8 (6–10) | 7 (6–9) | 0.079 |
| Stroke subtype | 0.736 | ||
| Cardioembolism | 26 (56.5%) | 37 (67.3%) | |
| Large artery atherosclerosis | 6 (13.0%) | 6 (10.9%) | |
| Undetermined source | 13 (28.3%) | 11 (20.0%) | |
| Others | 1 (2.2%) | 1 (1.8%) | |
| Occlusion site | 0.205 | ||
| ICA | 8 (17.4%) | 18 (32.7%) | |
| M1 | 27 (58.7%) | 25 (45.5%) | |
| M2 | 11 (23.9%) | 12 (21.8%) |
Total number is 101. Continuous variables are presented as median (IQR). Categorical variables are presented as n (%). Continuous variables were compared using the Mann–Whitney U-test and categorical variables using Fisher’s exact test.
DWI-ASPECTS, diffusion-weighted image–Alberta Stroke Program Early CT Score; ICA, internal carotid artery; IQR, interquartile range; M1, 1st segment of the middle cerebral artery; M2, 2nd segment of the middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale
Fig. 2. Time of arrival at the ED (n = 101). (A) Histogram of arrival times grouped into 2-hour intervals. Time is displayed using a 24-hour clock. (B) The proportion of patient arrivals during on-hours (46%) vs. off-hours (54%).
ED, emergency department
Stroke subtypes were classified as cardioembolism, large artery atherosclerosis, embolic stroke of undetermined source, and others (including Trousseau syndrome), with no significant differences in distribution between the 2 groups. The occluded vessels were ICA in 8 (17.4%), M1 in 27 (58.7%), and M2 in 11 (23.9%) patients in the on-hours group, and ICA in 18 (32.7%), M1 in 25 (45.5%), and M2 in 12 (21.8%) patients in the off-hours group. The proportion of ICA occlusion tended to be higher in the off-hours group, but the difference was not statistically significant (p = 0.205).
Procedural data and time metrics
The procedural data and time metrics are presented in Table 2. Overall, IV tPA was administered to 18 (39.1%) and 21 (38.2%) patients in the on- and off-hours groups, respectively (p = 1.000). The proportion of MT performed with 2 operators did not differ significantly between the groups (37.0% vs. 36.4%, p = 1.000). The exact time of onset was unknown in 38 patients; in these patients, the last known well time was used as the reference onset time for the calculation of ODT and ORT. In the 13 patients in whom recanalization was not achieved (TICI 0 or 1), the time of the final angiographic run was used as a surrogate for the time of recanalization. ODT did not differ significantly between the groups (median 105 [IQR 51–352] vs. 89 [IQR 49–276] min, p = 0.825). DIT did not differ significantly between the groups (median 20 [IQR 17–28] vs. 21 [IQR 16–27] min, p = 0.865). DNT did not differ significantly between the groups (median 54 [IQR 45–69] vs. 48 [IQR 39–59] min, p = 0.229). IPT was significantly longer in the off-hours group (median 83 [IQR 60–99] vs. 51 [IQR 45–64] min, p <0.001), as was DPT (median 103 [IQR 82–130] vs. 71 [IQR 66–92] min, p <0.001). PRT did not differ significantly between the groups (median 53 [IQR 34–72] vs. 48 [IQR 33–64] min, p = 0.626). DRT was significantly longer in the off-hours group (median 151 [IQR 125–202] vs. 124 [IQR 100–156] min, p = 0.003), whereas ORT did not differ significantly (median 285 [IQR 222–658] vs. 234 [IQR 177–450] min, p = 0.154). The median number of device passes was 2 (IQR 1–2) in the on-hours group and 1 (IQR 1–2) in the off-hours group (p = 0.307).
Table 2. Procedural data and clinical outcomes.
| On-hours (n = 46) | Off-hours (n = 55) | p Value | |
|---|---|---|---|
| IV tPA | 18 (39.1%) | 21 (38.2%) | 1.000 |
| Unknown onset | 17 (37.0%) | 21 (38.2%) | 1.000 |
| ODT, median (IQR) | 89 (49–276) | 105 (51–352) | 0.825 |
| DNT, median (IQR) | 48 (39–59) | 54 (45–69) | 0.229 |
| DIT, median (IQR) | 21 (16–27) | 20 (17–28) | 0.865 |
| IPT, median (IQR) | 51 (45–64) | 83 (60–99) | <0.001 |
| DPT, median (IQR) | 71 (66–92) | 103 (82–130) | <0.001 |
| PRT, median (IQR) | 48 (33–64) | 53 (34–72) | 0.626 |
| DRT, median (IQR) | 124 (100–156) | 151 (125–202) | 0.003 |
| ORT, median (IQR) | 234 (177–450) | 285 (222–658) | 0.154 |
| Number of operators | 1.000 | ||
| 1 | 29 (63.0%) | 35 (63.6%) | |
| 2 | 17 (37.0%) | 20 (36.4%) | |
| TICI 2b–3 | 38 (82.6%) | 44 (80.0%) | 0.802 |
| Number of passes, median (IQR) | 2 (1–2) | 1 (1–2) | 0.307 |
| Any ICH | 15 (32.6%) | 22 (40.0%) | 0.535 |
| Symptomatic ICH | 2 (4.3%) | 2 (3.6%) | 1.000 |
| 90-day mRS 0–3 | 23 (50.0%) | 25 (45.5%) | 0.692 |
Time intervals are presented as median (IQR) in minutes. Categorical variables are presented as n (%). Continuous variables were compared using the Mann–Whitney U-test and categorical variables using Fisher’s exact test. In the 38 patients whose exact onset time was unknown, the last known well time was used as the reference onset time for ODT and ORT. In patients with TICI 0 or 1, the time of the final angiographic run was defined as the time of recanalization.
DIT, door-to-image time; DNT, door-to-needle time; DPT, door-to-puncture time; DRT, door-to-recanalization time; ICH, intracranial hemorrhage; IPT, image-to-puncture time; IQR, interquartile range; IV tPA, intravenous thrombolysis with tissue-type plasminogen activator; mRS, modified Rankin Scale; ODT, onset-to-door time; ORT, onset-to-recanalization time; PRT, puncture-to-recanalization time; TICI, Thrombolysis in Cerebral Infarction
Clinical outcomes
The patients’ clinical outcomes are presented in Table 2. Overall, 90-day mRS 0–3 was achieved in 23 (50.0%) and 25 (45.5%) patients in the on- and off-hours groups, respectively, with no significant difference (p = 0.692). Successful recanalization (TICI 2b–3) was achieved in 38 patients (82.6%) in the on-hours group and 44 patients (80.0%) in the off-hours group (p = 0.802). sICH occurred in 2 patients (4.3%) in the on-hours group and 2 patients (3.6%) in the off-hours group (p = 1.000). Any intracranial hemorrhage was observed in 15 patients (32.6%) in the on-hours group and 22 patients (40.0%) in the off-hours group (p = 0.535).
Univariable and multivariable logistic regression (Tables 3 and 4)
Table 3. Univariable logistic regression analysis: factors associated with 90-day mRS 0–3.
| mRS 0–3 (n = 48) | mRS 4–6 (n = 53) | Univariable OR (95% CI) | p Value | |
|---|---|---|---|---|
| Age, median (IQR) | 77 (69–81) | 81 (78–85) | 0.925 (0.880–0.972) | 0.002 |
| Male | 25 (52.1%) | 33 (62.3%) | 0.659 (0.298–1.456) | 0.302 |
| Hypertension | 21 (43.8%) | 33 (62.3%) | 0.471 (0.213–1.045) | 0.064 |
| Dyslipidemia | 14 (29.2%) | 13 (24.5%) | 1.267 (0.524–3.062) | 0.599 |
| Diabetes mellitus | 10 (20.8%) | 9 (17.0%) | 1.287 (0.473–3.496) | 0.621 |
| NIHSS, median (IQR) | 15 (11–22) | 21 (14–26) | 0.956 (0.909–1.004) | 0.073 |
| DWI-ASPECTS, median (IQR) | 8 (6–9) | 7 (6–9) | 1.099 (0.929–1.300) | 0.271 |
| Left-sided occlusion | 28 (58.3%) | 20 (37.7%) | 2.310 (1.039–5.134) | 0.040 |
| ODT, median (IQR) | 75 (48–295) | 129 (54–349) | 0.996 (0.982–1.010) | 0.588 |
| DNT, median (IQR) | 53 (43–63) | 48 (42–64) | 0.928 (0.639–1.347) | 0.693 |
| DIT, median (IQR) | 21 (17–27) | 22 (16–28) | 0.956 (0.689–1.325) | 0.785 |
| IPT, median (IQR) | 66 (49–94) | 63 (51–95) | 0.977 (0.932–1.026) | 0.353 |
| DPT, median (IQR) | 86 (69–119) | 88 (71–119) | 0.978 (0.932–1.025) | 0.355 |
| PRT, median (IQR) | 48 (38–67) | 46 (32–71) | 1.010 (0.890–1.147) | 0.874 |
| DRT, median (IQR) | 136 (114–193) | 147 (109–185) | 0.982 (0.945–1.021) | 0.366 |
| ORT, median (IQR) | 241 (192–484) | 268 (195–679) | 0.994 (0.981–1.008) | 0.412 |
| TICI 2b–3 | 45 (93.8%) | 37 (69.8%) | 6.486 (1.754–23.982) | 0.005 |
| Any ICH | 16 (33.3%) | 21 (39.6%) | 0.762 (0.337–1.720) | 0.513 |
| Symptomatic ICH | 1 (2.1%) | 3 (5.7%) | 0.355 (0.036–3.530) | 0.377 |
| Off-hours | 25 (52.1%) | 30 (56.6%) | 0.833 (0.380–1.826) | 0.649 |
Continuous variables are presented as median (IQR), with time intervals in minutes, and categorical variables as n (%). Odds ratios for time variables are calculated per 10-min increment, for age per year, and for NIHSS and DWI-ASPECTS per point. Standard logistic regression was applied for all analyses. The 95% CIs and p Values were derived from the same models. DNT was calculated only in the 39 patients who received intravenous thrombolysis. The odds ratio for symptomatic ICH should be interpreted with caution because of the small number of events.
CI, confidence interval; DIT, door-to-image time; DNT, door-to-needle time; DPT, door-to-puncture time; DRT, door-to-recanalization time; DWI-ASPECTS, diffusion-weighted image–Alberta Stroke Program Early CT Score; ICH, intracranial hemorrhage; IPT, image-to-puncture time; IQR, interquartile range; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; ODT, onset-to-door time; OR, odds ratio; ORT, onset-to-recanalization time; PRT, puncture-to-recanalization time; TICI, Thrombolysis in Cerebral Infarction
Table 4. Multivariable logistic regression analysis: 90-day mRS 0–3.
| Variable | Adjusted OR (95% CI) | p Value |
|---|---|---|
| Age (per year) | 0.922 (0.876–0.970) | 0.002 |
| NIHSS (per point) | 0.974 (0.922–1.030) | 0.335 |
| DWI-ASPECTS (per point) | 1.130 (0.931–1.380) | 0.213 |
| Off-hours | 0.917 (0.386–2.179) | 0.845 |
Variables were entered using forced entry method.
CI, confidence interval; DWI-ASPECTS, diffusion-weighted image–Alberta Stroke Program Early CT Score; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio
In the univariable logistic regression, younger age, left-sided occlusion, and successful recanalization were found to be significantly associated with 90-day mRS 0–3. Sex, hypertension, diabetes mellitus, dyslipidemia, NIHSS, DWI-ASPECTS, the individual time metrics (ODT, DIT, DNT, IPT, DPT, PRT, DRT, and ORT), intracranial hemorrhage, and off-hours status were not significantly associated with 90-day mRS 0–3. The results of the multivariable logistic regression are presented in Table 4. Younger age (OR 0.922, p = 0.002) was identified as an independent predictor of 90-day mRS 0–3. The NIHSS, DWI-ASPECTS, and off-hours status were not significant predictors.
Discussion
In this single-center retrospective study, we compared MT outcomes in 101 patients with anterior circulation LVO who presented during on- and off-hours and investigated factors associated with 90-day mRS 0–3. Although the IPT, DPT, and DRT were significantly prolonged in the off-hours group, the proportion of patients achieving 90-day mRS 0–3, successful recanalization, and complication rates did not differ significantly between the groups. In the multivariable logistic regression, younger age was identified as an independent predictor of 90-day mRS 0–3, whereas off-hours status was not.
In terms of baseline characteristics (Table 1), the DWI-ASPECTS tended to be higher in the on-hours group. The off-hours group had a higher, albeit nonsignificant, proportion of ICA occlusions (32.7% vs. 17.4%), which may involve larger ischemic territories; this may have contributed to the numerical difference in DWI-ASPECTS between the groups.
The ODT did not differ significantly between the groups, indicating that the off-hours delay occurred predominantly after hospital arrival. However, the ORT did not differ significantly between the groups, as it was strongly influenced by the substantial variability in prehospital time. The DPT was significantly prolonged in the off-hours group. Because DIT did not differ significantly between the groups, and IPT was significantly prolonged, the workflow from hospital arrival to MRI acquisition was likely performed comparably in both groups, indicating that a delay arose between imaging completion and procedural initiation. PRT was not prolonged, and the successful recanalization rate did not appreciably differ between the groups.
At our institution, the principal tasks within the IPT include patient transfer to the catheterization laboratory, calling on-call staff from home, preparing the catheterization laboratory, obtaining informed consent for MT, performing nursing handovers for catheterization laboratory admission, and unpacking of thrombectomy devices. These tasks involve substantial contributions from non-physician healthcare professionals, in addition to physicians. Hinsenveld et al. attributed part of the in-hospital delay to the on-call, rather than in-house, availability of the interventionist,8) while Benali et al. attributed the longer admission-to-puncture time during nighttime hours to the time required for the on-call interventionist to reach the hospital.9) Kim et al. further reported that hospitals with 3 neurointerventionalists had a shorter DPT than those with 1 or 2 neurointerventionalists, both during on- and off-hours, and that these hospitals also had substantially more neurointerventional technologists and dedicated nurses.5) Accordingly, the prolonged DPT at our institution could be explained by the time required for on-call staff to arrive from home and the possibility that ED nurses simultaneously manage patients from other departments. However, detailed records that would allow these factors to be retrospectively analyzed are not available, and future initiatives should, therefore, include documentation and evaluation of each process following patient arrival, including the time required for on-call staff to arrive. Reducing DPT at our institution may require an increase in the number of full-time JSNET-certified specialists and dedicated nurses, as suggested by Kim et al.,5) although such staffing is not readily achievable. The American Heart Association proposed that DPT could be shortened by optimizing in-hospital protocols from arrival to treatment initiation, cross-training staff, and improving workflow through the collection and analysis of process data.17) Implementing such measures may help to shorten the DPT delay observed during off-hours.
Numerous studies have evaluated MT performed off-hours; however, there is currently no consensus on the association between off-hours treatment and clinical outcomes.3–12,18–19) Zaeske et al. previously reported that procedural time metrics and the proportion of patients achieving a 90-day mRS 0–2 did not significantly differ between regular working hours and on-call duty, although in-hospital mortality was higher during on-call duty.19) In contrast, a systematic review and meta-analysis by Ghozy et al. reported a significantly lower rate of 90-day functional independence in the off-hours group.3) In the present study, despite the prolonged DRT in the off-hours group, the proportion of patients who achieved 90-day mRS 0–3 did not differ significantly. Hinsenveld et al. similarly reported that DPT was mildly prolonged during off-hours, without any significant difference in outcomes.8) Citing the MR CLEAN Registry finding that each 1-hour delay in onset-to-groin-puncture time was associated with a 5.3% absolute reduction in the probability of functional independence,20) the authors discussed the possibility that the observed time difference was not sufficiently large to reach statistical significance. In the present study, the magnitude of the off-hours delay may not have been sufficient to produce a detectable difference in the 90-day mRS 0–3.
In both univariable and multivariable logistic regression, younger age was identified as a strong predictor of 90-day mRS 0–3, suggesting that patient-related factors may have influenced the outcome more strongly than time-related factors. Age has also been reported as a prognostic factor after MT. In the study by Zaeske et al., age was independently associated with in-hospital mortality.19) In the present study, both groups had a median age exceeding 80 years, which may have attenuated the relative impact of time delay on the outcomes. These observations suggest that, in this older population, advanced age may have had a greater influence on outcomes than the off-hours time delay and that the likelihood of achieving a 90-day mRS of 0–3 decreases with increasing age. Taken together, the limited magnitude of the delay and the advanced age of the cohort may explain why the off-hours delay observed in the present study did not translate into a significant difference in the 90-day mRS 0–3.
In the present study, left-sided occlusion was significantly associated with 90-day mRS 0–3 in the univariable logistic regression; however, the relationship between lesion laterality and clinical outcomes remains unclear. Jehkonen et al. previously reported that hemispatial neglect due to right hemispheric damage can impede independence in activities of daily living.21) Conversely, in the HERMES meta-analysis, Almekhlafi et al. reported that stroke laterality was not an independent prognostic factor.22) To clarify the relationship between lesion laterality and outcome, detailed evaluation of higher cortical dysfunctions, including hemispatial neglect, and their impact on 90-day mRS would be desirable; however, this was not performed in the present study.
In recent years, the relationship between post-MT blood pressure variability and prognosis has attracted considerable attention; however, the association between preexisting hypertension and clinical outcomes after MT remains poorly established. In the present study, preexisting hypertension was not significantly associated with 90-day mRS 0–3 in univariable logistic regression. Although hypertension and post-MT blood pressure variability have been associated with outcomes after MT in previous reports,23,24) the present retrospective dataset did not include postoperative blood pressure measurements; a history of hypertension alone is insufficient to evaluate this relationship, which therefore remains a subject for future prospective study.
Limitations
This study has several limitations. First, this was a single-center retrospective study with a limited sample size, which constrained the statistical power. As such, the precision of the estimates was correspondingly reduced, and clinically relevant differences in the outcomes between the on- and off-hours groups could not be excluded. Second, the symptom onset time was unknown in 38 patients (37.6%), meaning the last known well-time had to be used instead. As this metric may substantially precede the actual onset, ODT and ORT may have been overestimated and widely dispersed in these patients, which could have limited our ability to detect between-group differences in these metrics. Third, in 13 patients in whom recanalization was not achieved, the time of the final angiographic run was used as a surrogate for the time of recanalization. Because this timepoint reflects the operator’s decision to terminate the procedure, rather than achieve reperfusion, the PRT, DRT, and ORT in these patients are not strictly equivalent to those in patients with successful recanalization. Fourth, because of EPV constraints, covariate selection in the multivariable logistic regression was limited to 4 variables. Baseline characteristics that were imbalanced between the groups, such as sex (73.9% vs. 43.6% male), as well as factors found to be significant in the univariable logistic regression, such as left-sided occlusion and successful recanalization, were not adjusted for. Consequently, residual confounding factors may have influenced our results. Fifth, because the retrospective records did not capture whether the JSNET-certified specialist was called from outside the hospital during off-hours, the time from call to arrival, or the time from when the catheterization laboratory staff was called to their arrival, we were unable to assess the impact of these factors separately.
Conclusion
Despite the prolonged time from hospital arrival to treatment initiation in the off-hours group, the proportion of patients achieving 90-day mRS 0–3 did not differ significantly between the groups. In the multivariable logistic regression, younger age was an independent predictor of 90-day mRS 0–3, while off-hours status was not. Because time to treatment initiation has been identified as an important prognostic factor in previous studies, the present findings should not be interpreted as negating the importance of early recanalization. Further efforts are needed to identify the causes of off-hours delays and to improve our clinical care system.
Acknowledgments
We would like to thank Eleanor Hawkins, MSc, PhD, from Editage (www.editage.jp) for English editing.
During the manuscript preparation, the authors used Claude (Claude Opus 4.6 and 5.0, Anthropic) for proofreading the Japanese manuscript and assisting with English writing. The authors take full responsibility for the content of this manuscript.
Disclosure Statement
The authors declare that they have no conflicts of interest.
References
- 1).Saver JL, Goyal M, van der Lugt A, et al. Time to treatment with endovascular thrombectomy and outcomes from ischemic stroke: a meta-analysis. JAMA 2016; 316: 1279–1288. [DOI] [PubMed] [Google Scholar]
- 2).McTaggart RA, Ansari SA, Goyal M, et al. Initial hospital management of patients with emergent large vessel occlusion (ELVO): report of the Standards and Guidelines Committee of the Society of NeuroInterventional Surgery. J Neurointerv Surg 2017; 9: 316–323. [DOI] [PubMed] [Google Scholar]
- 3).Ghozy S, Hasanzadeh A, Kobeissi H, et al. The impact of off-hour mechanical thrombectomy therapy on outcomes for acute ischemic stroke: a systematic review and meta-analysis. J Neurol Sci 2023; 453: 120802. [DOI] [PubMed] [Google Scholar]
- 4).Bao J, Shen Y, Shi Z, et al. Endovascular treatment for basilar artery occlusion: whether the “weekend effect” affects time metrics and clinical outcomes at a comprehensive stroke center. Front Neurol 2024; 15: 1413557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5).Kim JG, Choi JC, Kim DJ, et al. Effect of the number of neurointerventionalists on off-hour endovascular therapy for acute ischemic stroke within 12 hours of symptom onset. J Am Heart Assoc 2019; 8: e011933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6).Ding Y, Gao F, Ji Y, et al. Workflow intervals and outcomes of endovascular treatment for acute large-vessel occlusion during on-vs. off-hours in China: the ANGEL-ACT registry. Front Neurol 2021; 12: 771803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7).Mpotsaris A, Kowoll A, Weber W, et al. Endovascular stroke therapy at nighttime and on weekends-as fast and effective as during normal business hours? J Vasc Interv Neurol 2015; 8: 39–45. [PMC free article] [PubMed] [Google Scholar]
- 8).Hinsenveld WH, de Ridder IR, van Oostenbrugge RJ, et al. Workflow intervals of endovascular acute stroke therapy during on- versus off-hours: the MR CLEAN registry. Stroke 2019; 50: 2842–2850. [DOI] [PubMed] [Google Scholar]
- 9).Benali A, Moynier M, Dargazanli C, et al. Mechanical thrombectomy in nighttime hours: is there a difference in 90-day clinical outcome for patients with ischemic stroke? AJNR Am J Neuroradiol 2021; 42: 530–537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10).Ishikawa S, Miyake S, Akimoto T, et al. Increased door-to-puncture time during off-duty hours results in poor treatment outcomes for acute ischemic stroke: a subanalysis of the K-NET registry. Interv Neuroradiol 2026; 32: 374–380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11).Almekhlafi MA, Hockley A, Desai JA, et al. Overcoming the evening/weekend effects on time delays and outcomes of endovascular stroke therapy: the Calgary Stroke Program experience. J Neurointerv Surg 2014; 6: 729–732. [DOI] [PubMed] [Google Scholar]
- 12).Raymond SB, Akbik F, Stapleton CJ, et al. Protocols for endovascular stroke treatment diminish the weekend effect through improvements in off-hours care. Front Neurol 2018; 9: 1106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13).Numata Y, Matsumoto M. Labor shortage of physicians in rural areas and surgical specialties caused by Work Style Reform Policies of the Japanese government: a quantitative simulation analysis. J Rural Med 2024; 19: 166–173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14).Okawara M, Ishimaru T, Yoshikawa T, et al. Working hours, side work, and depressive symptoms in physicians: a nationwide cross-sectional study in Japan. J Occup Health 2022; 64: e12377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15).Takahashi M. Work-interval system in Japan. Ind Health 2023; 61: 1–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16).Wahlgren N, Ahmed N, Dávalos A, et al. Thrombolysis with alteplase for acute ischaemic stroke in the Safe Implementation of Thrombolysis in Stroke-Monitoring Study (SITS-MOST): an observational study. Lancet 2007; 369: 275–282. [DOI] [PubMed] [Google Scholar]
- 17).Ashcraft S, Wilson SE, Nyström KV, et al. Care of the patient with acute ischemic stroke (prehospital and acute phase of care): update to the 2009 comprehensive nursing care scientific statement: a scientific statement from the American Heart Association. Stroke 2021; 52: e164–e178. [DOI] [PubMed] [Google Scholar]
- 18).Grandhi R, Ravindra VM, Ney JP, et al. Investigating the “weekend effect” on outcomes of patients undergoing endovascular mechanical thrombectomy for ischemic stroke. J Stroke Cerebrovasc Dis 2021; 30: 106013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19).Zaeske C, Goertz L, Kottlors J, et al. Mortality after mechanical thrombectomy in anterior circulation stroke may be higher at nighttime and on weekends. Eur Radiol 2021; 31: 4148–4155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20).Mulder MJHL, Jansen IGH, Goldhoorn RB, et al. Time to endovascular treatment and outcome in acute ischemic stroke: MR CLEAN registry results. Circulation 2018; 138: 232–240. [DOI] [PubMed] [Google Scholar]
- 21).Jehkonen M, Laihosalo M, Kettunen JE. Impact of neglect on functional outcome after stroke – a review of methodological issues and recent research findings. Restor Neurol Neurosci 2006; 24: 209–215. [PubMed] [Google Scholar]
- 22).Almekhlafi MA, Hill MD, Roos YB, et al. Stroke laterality did not modify outcomes in the HERMES meta-analysis of individual patient data of 7 trials. Stroke 2019; 50: 2118–2124. [DOI] [PubMed] [Google Scholar]
- 23).Yuan Z, Chen N, Zhou M, et al. Effects of hypertension in patients receiving mechanical thrombectomy: a meta-analysis. Medicine (Baltimore) 2020; 99: e19803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24).Mistry EA, Mehta T, Mistry AM, et al. Blood pressure variability and neurologic outcome after endovascular thrombectomy: a secondary analysis of the BEST study. Stroke 2020; 51: 511–518. [DOI] [PMC free article] [PubMed] [Google Scholar]