Low-flow time and neurological outcomes in out-of-hospital cardiac arrest: implications for prehospital care quality and health outcomes

Abstract

Background

Out-of-hospital cardiac arrest (OHCA) remains a significant public health challenge with high mortality and morbidity. Low-flow time (LFT) is an important determinant of outcomes, yet the specific temporal thresholds for functional recovery across different cardiac rhythms remain poorly defined in metropolitan Asian settings. This study aimed to identify the optimal LFT cut point for predicting survival and favorable neurological outcomes using a large-scale emergency medical services (EMS) registry.

Methods

This retrospective cohort study analyzed 4,933 adult non-traumatic OHCA cases from the Taoyuan City EMS registry between January 2021 and October 2025. The primary exposure was LFT, defined as the duration from the initiation of CPR to hospital handover or return of spontaneous circulation. Multivariable logistic regression was utilized to evaluate the independent association between LFT and survival to discharge. Optimal LFT thresholds were identified using Receiver Operating Characteristic (ROC) curves and the Youden Index, stratified by initial cardiac rhythm (shockable vs. non-shockable).

Results

The overall survival rate to discharge was 22.8% (n = 1,123). A significant rhythm-specific divergence in the impact of LFT was observed. In the shockable rhythm group, LFT exceeding 27 minutes was independently associated with increased mortality (Adjusted Odds Ratio [AOR]: 2.25; 95% CI: 1.19–4.27; p = 0.013). The optimal LFT threshold for predicting survival to discharge was 27 minutes (AUC 0.58; Sensitivity 41%, Specificity 72%), while the threshold for favorable neurological outcome was 26 minutes. In contrast, LFT was not a significant independent predictor for the non-shockable rhythm group (AUC 0.52; p > 0.05), where outcomes were likely dominated by non-temporal factors.

Conclusion

Among OHCA patients presenting with shockable rhythms, a low-flow time of approximately 27 minutes was associated with survival and favorable neurological outcomes. However, given the modest predictive performance of this threshold, it should be interpreted as a supportive reference rather than a definitive clinical cutoff. These findings highlight the potential value of incorporating rhythm-specific temporal information into prehospital decision-making, EMS resource allocation, and future evaluations of termination-of-resuscitation protocols.

Supplementary information

The online version contains supplementary material available at 10.1186/s12245-026-01314-x.

Introduction

Out-of-hospital cardiac arrest (OHCA) remains one of the most formidable challenges to global public health. Its high mortality rate and the potential for severe neurological impairment impose a substantial burden on both healthcare systems and society at large. Although various links in the “Chain of Survival,” including the promotion of bystander CPR and advancements in Advanced Life Support (ALS), have significantly improved the rates of Return of Spontaneous Circulation (ROSC) over the past decades, the proportion of patients discharged with favorable neurological outcomes remains suboptimal [1]. In emergency medicine, “time is brain” is a widely accepted principle. Current clinical consensus emphasizes that cerebral injury worsens progressively with the duration of hypoxia from the moment of arrest.

A critical variable in the resuscitation process is Low-Flow Time (LFT), which we define in this study as the duration from the initiation of CPR by any rescuer (bystander or EMS) to the restoration of circulation or arrival at the hospital.

This definition intentionally excludes patients who did not receive immediate CPR after collapse, as LFT without active perfusion does not capture the clinically relevant time window for cerebral protection. Literature indicates a significant positive correlation between prolonged LFT and declining survival rates. According to large-scale registry studies (n = 9,433), survival decreases by approximately 3% for every 5-minute increment during the first 35 minutes of resuscitation [2]. In the context of Extracorporeal Cardiopulmonary Resuscitation (ECPR), survival rates decline more gradually than in Conventional CPR (CCPR) yet still drop from 37.2% at 15 minutes to 19.1% at 60 minutes [3]. Multiple studies have quantified this effect through regression analysis, finding that for every one-minute increase in LFT, the odds ratio (OR) for survival ranges between 0.88 and 0.97 [4, 5]. While specific populations, such as those with accidental hypothermia (<32°C), can maintain survival rates of around 44% despite LFT durations of 60 to 120 minutes [6–8], the pressure of time remains a major challenge for the majority of non-hypothermic OHCA patients.

Despite existing research on the adverse effects of LFT, a significant gap remains in current scientific knowledge. Most studies treat LFT as a continuous variable or employ arbitrary cut-offs (e.g., 20 or 30 minutes), failing to precisely define clinical predictive thresholds tailored to different cardiac rhythm characteristics. Although it is known that the initial rhythm modulates the impact of LFT on prognosis [7], a critical question remains for frontline emergency physicians: is there a specific time window beyond which the probability of achieving favorable neurological outcomes declines substantially? Research focusing on precise cut-points for such clinical heterogeneity, especially within high-density Asian metropolitan EMS systems, remains scarce in evidence-based medicine [8].

This study aims to analyze clinical data from 4,933 OHCA patients to explore the association between LFT, survival to discharge, and favorable neurological prognosis. We hypothesize that the impact of LFT varies significantly across different rhythm groups and that a clinically relevant temporal threshold may exist for patients with shockable rhythms to support outcome prediction. The findings of this study will help optimize pre-hospital resuscitation decisions and provide a more robust evidence-based framework for the termination of resuscitation or the transition to advanced interventions.

Methods

Study population and setting

This retrospective cohort study was conducted using the EMS registry of Taoyuan City, Taiwan, a metropolitan area with a population of approximately 2.3 million. We analyzed all non-traumatic OHCA cases recorded between January 2021 and October 2025.

A total of 8,681 OHCA cases were initially identified. Patients were included if they received resuscitation efforts by EMS personnel and had documented LFT data. To ensure internal validity, we excluded patients with traumatic causes of arrest, pediatric cases (age < 18 years), and those with incomplete medical records regarding primary outcomes. The final analysis included 4,933 patients, categorized into survivors (n = 1,123) and non-survivors (n = 3,810) (Table 1). This study was approved by the Institutional Review Board (IRB No. TYGH115022), and the requirement for informed consent was waived due to the retrospective nature of the de-identified data.

Table 1.

Background and resuscitation data of enrolled patients

Characteristics		All patients	Survived to discharge	Deceased	p-value
Number of cases		4,933	1123	3810
Patient characteristics (age: mean ± SD; n (%))
Age, mean ± SD		68.2 ± 16.3	65.4 ± 16.0	69.0 ± 16.3	<0.001*
Sex-male	male	3140 (63.7%)	724(64.5%)	2416 (63.4%)	0.517
	female	1793 (36.3%)	399(35.5%)	1394 (36.6%)
Comorbidities
Cancer		473 (9.6%)	97(8.7%)	376(9.9%)	0.221
Hypertension		1840 (37.3%)	396(35.4%)	1444(38.0%)	0.112
Diabetes		1397 (28.3%)	295(26.3%)	1102(29.0%)	0.085
Heart disease		1209 (24.5%)	286(25.5%)	923(24.3%)	0.387
Stroke		385 (7.8%)	87(7.8%)	298(7.8%)	0.941
Number of Comorbidities
0		1773 (35.9%)	423(37.8%)	1350(35.5%)	0.375
1		1600 (32.4%)	356(31.8%)	1244(32.7%)	0.895
≥2		1550 (31.4%)	341(30.4%)	1209(31.8%)	0.941
Event Characteristics
Witnessed Arret Missing = 10(0.2%)	Yes	2029 (41.1%)	701(62.6%)	1328(34.9%)	<0.0001*
	No	2894 (58.7%)	419(37.4%)	2475(65.1%)
Public location Missing = 10 (0.2%)	Yes	468 (9.5%)	163(14.6%)	305(8.0%)	<0.0001*
	No	4455 (90.3%)	957(85.4%)	3498(92.0%)
PAD Missing = 32 (0.6%)	Yes	367 (7.4%)	111 (9.9%)	256 (6.8%)	<0.0001*
	No	4534 (91.9%)	1005 (89.5%)	3529 (93.2%)
Initial Rhythm Missing = 213(4.3%)	Yes	975 (19.8%)	351(33.2%)	624(17.0%)	<0.0001*
	No	3745 (75.9%)	706(66.8%)	3039(83.0%)
EMS variables (time: median (Q1–Q3); n (%))
Response time (mins)		6.7 (5,6,7,8,9)	6.1 (5 - 8)	7.0 (5 - 9)	<0.0001*
Airway Management Missing = 21(0.4%)	Advanced Airway (ETI + SGA)	3451 (70.0%)	678 (60.4%)	2773 (72.8%)	<0.0001*
	Basic Airway	1461 (29.6%)	437 (38.9%)	1024 (26.9%)
Epinephrine administration	Yes	1808 (36.7%)	416 (37.0%)	1392 (36.5%)	0.756
	No	3125 (63.3%)	707 (63.0%)	2418 (63.5%)
Primary Exposure
Low flow time (n (%);
≤20 mins		425(8.6%)	114 (10.2%)	311(8.2%)	0.007
20–30 mins		2513(51.0%)	592 (52.7%)	1921(50.4%)	0.223
31–40 mins		1585(32.1%)	331 (29.5%)	1254(32.9%)	0.115
>40 mins		410(8.3%)	86 (7.7%)	324(8.5%)	0.842
Prehospital ROSC Missing = 19(0.4%)	Yes	473 (9.6%)	407 (36.3%)	66 (1.7%)	<0.0001*
	No	4441 (90.0%)	712 (63.4%)	3729 (97.8%)
Cerebral performance category	CPC 1–2		262(23.3%)
	CPC 3–5		861(76.7%)

Study design and measurement

The primary exposure of interest was LFT, defined as the duration (in minutes) from the initiation of CPR by either bystanders or EMS providers to the achievement of sustained ROSC or the handover of the patient to emergency department staff at the hospital. LFT was analyzed both as a continuous variable and categorized into time intervals (≤20 minutes, 21–30 minutes, 31–40 minutes, >40 minutes) to evaluate potential threshold effects.

The primary outcome was survival to hospital discharge. The secondary outcome was favorable neurological status, defined as CPC 1–2 at discharge. Data were extracted from the high-reliability Taoyuan EMS database, which includes electronic patient care records and hospital-linked outcome registries. Covariates retrieved included patient demographics (age, sex), comorbidities (hypertension, diabetes, heart disease, malignancy, and stroke), and event characteristics (witnessed status, bystander CPR, and initial cardiac rhythm). Initial rhythm was dichotomized into shockable (ventricular fibrillation or pulseless ventricular tachycardia) and non-shockable (asystole or pulseless electrical activity) (Fig. 1).

Fig. 1.

Flowchart of enrolled patients with non-traumatic out-of-hospital cardiac arrest (OHCA). Notes: EMS = emergency medical services; OHCA = out-of-hospital cardiac arrest; CPR = cardiopulmonary resuscitation

Statistical analysis

Continuous variables were presented as means with standard deviations (±SD) and compared using t-tests, while categorical variables were expressed as frequencies and percentages, compared using the Chi-square test (Table 1).

To identify the independent association between LFT and clinical outcomes, we performed multivariable logistic regression analysis to calculate adjusted odds ratios (AOR) with 95% confidence intervals (CI), adjusting for age, sex, comorbidities, and event characteristics (Table 2). Given the physiological differences in resuscitation potential, we further performed stratified analysis based on the initial rhythm (Shockable vs. non-shockable) (Tables 3, 4).

Table 2.

Associations of low-flow time and covariates with mortality and functional impairment

Mortality (All OHCA Patients, n = 4654)					Functional Impairment (Survivors Only, n = 1041)
	Adjusted		95% CI	p-value	Adjusted OR		* 95% CI			p-value
Patient Characteristics
Age (per year)	1.01	1.01–1.02		<0.0001*		1.03		1.02–1.05			<0.0001*
Male Gender	1.20	1.01–1.42		0.04		0.90		0.59–1.37			0.633
Number of Comorbidities
0	Reference					Reference
1	1.04	0.85–1.26		0.734		1.28		0.80–2.04			0.302
≥2	1	0.82–1.23		0.994		1.04		0.64–1.70			0.865
Event Characteristics
Witnessed Arrest	0.43	0.36–0.50		<0.0001*		0.88		0.58–1.33			0.547
Public Location	0.90	0.68–1.1.9		0.461		0.37		0.22–0.62			<0.0001*
Bystander PAD	1.02	0.74–1.39		0.925		0.51		0.28–0.93			0.027
EMS & Resuscitation Factors
Response Time	1.04	1.01–1.08		0.008		1.06			0.98–1.15		0.13
Shockable Rhythm	0.66	0.54–0.80		<0.0001*		0.18			0.12–0.28		<0.0001*
Pre-hospital ROSC	0.04	0.03–0.05		<0.0001*		0.18			0.12–0.28		<0.0001*
Advanced Airway	1.07	0.88–1.29		0.496		2.03			1.32–3.13		<0.0001*
Epinephrine	1.05	0.88–1.25		0.62		4.39			2.82–6.84		<0.0001*
Primary Exposure
Low flow time
≤20 mins	Reference					Reference
20–30 mins	1.03	0.77–1.37		0.867		0.79			0.42–1.47		0.453
31–40 mins	1.25	0.91–1.72		0.178		1.19			0.59–2.40		0.635
>40 mins	1.30	0.83–2.02		0.252		1.22			0.43–3.42		0.709

Table 3.

Associations of low-flow time and covariates with mortality in OHCA patients stratified by initial rhythm

Mortality (All OHCA Patients, n = 4654)
Non-Shockable Rhythm (n = 3694)				Shockable Rhythm (n = 960)
	Adjusted OR	95% CI	p-value	Adjusted OR	95% CI	p-value
Patient Characteristics
Age (per year)	1.01	1.00–1.02	0.005	1.03	1.01–1.04	<0.0001
Male Gender	1.20	1.01–1.42	0.04	0.90	0.59–1.37	0.633
Event Characteristics
Witnessed Arrest	0.39	0.32–0.48	<0.0001	0.57	0.41–0.80	<0.0001
Public Location	1.13	0.76–1.69	0.557	0.75	0.50–1.13	0.165
EMS & Resuscitation Factors
Response Time	1.06	1.02–1.10	0.003	1.01	0.95–1.07	0.782
Pre-hospital ROSC	0.03	0.02–0.05	<0.0001	0.05	0.03–0.08	<0.0001
Advanced Airway	1.10	0.89–1.37	0.385	1.01	0.67–1.52	0.967
Epinephrine	1.03	0.84–1.27	0.788	1.15	0.81–1.64	0.436
Primary Exposure
Low flow time
≤ 20 min	Reference			Reference
20–30 min	0.95	0.67–1.34	0.758	1.35	0.75–2.43	0.322
31–40 min	1.01	0.69–1.48	0.948	2.25	1.19–4.27	0.013
>40 min	1.01	0.59–1.70	0.985	2.50	1.07–5.83	0.034

Table 4.

Associations of low-flow time and covariates with mortality in OHCA patients stratified by initial rhythm

Survivors Only, n = 1041
Non-Shockable Rhythm (n = 698)				Shockable Rhythm (n = 343)
	Adjusted OR	95% CI	p-value	Adjusted OR	95% CI	p-value
Patient Characteristics
Age (per year)	1.04	1.02–1.06	<0.0001	1.02	1.00–1.05	0.025
Event Characteristics
Witnessed Arrest	0.90	0.49–1.67	0.739	0.78	0.43–1.41	0.412
Public Location	0.34	0.13–0.87	0.025	0.39	0.21–0.73	0.003
EMS & Resuscitation Factors
Response Time	1.15	1.00–1.31	0.046	1.03	0.93–1.14	0.578
Pre-hospital ROSC	0.14	0.07–0.26	<0.0001	0.20	0.12–0.35	<0.0001
Advanced Airway	2.74	1.44–5.19	0.002	1.46	0.80–2.65	0.216
Epinephrine	13.05	5.74–29.7	<0.0001	2.34	1.34–4.09	0.003
Primary Exposure
Low flow time
≤ 20 min	Reference			Reference
20–30 min	0.71	0.26–1.92	0.495	0.81	0.35–1.87	0.618
31–40 min	0.72	0.24–2.20	0.565	1.55	0.61–3.99	0.36
>40 min	0.64	0.12–3.45	0.604	1.57	0.42–5.93	0.503

Abbreviations: OR, odds ratio; CI, confidence interval; ROSC, return of spontaneous circulation

Footnote: Adjusted ORs were controlled for all variables listed in the table using multivariable logistic regression

The optimal clinical cut-point for LFT to predict survival and CPC 1–2 at discharge was determined using Receiver Operating Characteristic (ROC) curves and the Youden Index. The Cochran-Armitage trend test was employed to evaluate the linear trend between LFT categories and outcomes (Fig. 2). To assess the robustness of our findings to the potential measurement uncertainty of low-flow time in unwitnessed cases, we performed a sensitivity analysis restricted to patients with witnessed OHCA. Using the same multivariable logistic regression models as in the primary analysis, we re-estimated the associations between low-flow time categories and the outcomes of mortality and CPC 1–2 at discharge, stratified by initial rhythm (shockable vs. non-shockable). Results are presented in Supplementary Table S1. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), and a two-sided p-value < 0.05 was considered statistically significant.

Fig. 2.

Combined analysis of clinical outcomes by low-flow time and initial heart rhythms. Consistency of time thresholds: for patients with a shockable rhythm, the optimal cut-off for predicting both mortality (Fig a) and poor functional outcome (Fig b) consistently falls between 26 and 27 minutes (AUC range: 0.574–0.577). This suggests a critical “golden window” for intervention in this subgroup, providing a robust clinical indicator for prognosis assessment. Prognostic heterogeneity: in contrast, for patients with non-shockable rhythms, the time thresholds (ranging from 22 to 29 minutes, Fig c & d) failed to reach statistical significance (AUC ≈ 0.52). These findings underscore the poor overall prognosis of this cohort, suggesting that shortening the low-flow duration alone may be insufficient to improve neurological outcome. Clinical implications: these results highlight that while resuscitation efforts should be intensified within the first 26 minutes for shockable cases, alternative interventions or early discussions regarding the termination of resuscitation (TOR) may be more appropriate for non-shockable cases where time-dependent recovery is less evident

Results

Baseline characteristics

Between January 2021 and October 2025, a total of 8,681 non-traumatic OHCA cases were assessed by Taoyuan City EMS. After excluding patients who did not receive bystander CPR (n = 3,510), those under 18 years or with missing age data (n = 89), and those with missing or physiologically implausible LFT data (n = 149), a final cohort of 4,933 adult OHCA cases was analyzed, comprising 1,123 survivors (22.8%) and 3,810 non-survivors (77.2%). Survivors were characterized by younger age (65.4±16.0 vs. 69.0 ± 16.3 years, p < 0.0001) and higher rates of witnessed arrests (62.6% vs. 34.9%, p < 0.0001) compared to non-survivors. Additionally, arrests occurring in public locations (14.6% vs. 8.0%, p < 0.0001) and the use of PAD (9.9% vs. 6.8%, p < 0.0001) were significantly more frequent among survivors. Notably, baseline comorbidities (such as cancer, hypertension, diabetes, heart disease, and stroke) showed no significant statistical difference between the two groups.

Association of low-flow time with clinical outcomes

Initial shockable rhythms were present in 33.2% of survivors versus 17.0% of non-survivors (p < 0.0001). Survivors also had a significantly shorter EMS response time (median 6.1 vs. 7.0 minutes, p < 0.0001) and a substantially higher rate of prehospital ROSC (36.3% vs. 1.7%, p < 0.0001). Among the 1,123 survivors, 262 patients (23.3%) achieved favorable neurological outcomes (CPC 1–2). The distribution of LFT significantly differed between groups, particularly within the 20 minutes interval (10.2% in survivors vs. 8.2% in non-survivors, p < 0.0001), as detailed in (Table 1).

Survival and Neurological Trends by Rhythm (Fig. 3). The impact of LFT on clinical outcomes exhibited a stark divergence when stratified by initial cardiac rhythm. As demonstrated in (Fig. 3a), patients with a shockable rhythm showed a significant time-dependent decline in survival to hospital discharge (p < 0.004). A marked decline in survival rates was observed after LFT exceeded 20 minutes. In contrast, patients in the non-shockable group maintained consistently low survival rates across all LFT intervals, with no statistically significant temporal trend (p = 0.105).

Fig. 3.

Survival and neurological trends by rhythm. (a). Survival to hospital discharge by low-flow time stratified by initial rhythm. The mortality/survival analysis shows a significant time-dependent decline for patients with a shockable rhythm (p < 0.004), with a notable drop in survival after 20 minutes of low-flow time. Conversely, patients with a nonshockable rhythm exhibited consistently low survival rates regardless of duration (p < 0.105). (b). Outcome (CPC 1–2) by low-flow time stratified by initial rhythm. Consistent with the survival trends, favorable neurological outcomes (CPC 1–2) for shockable patients demonstrated a clear “golden window” before 30 minutes, although the overall trend narrowly missed the standard significance threshold (p < 0.0537). For the non-shockable group, functional recovery remained extremely limited across all time intervals (p < 0.102)

A parallel trend was identified for favorable neurological outcomes (CPC 1–2). (Fig. 3b) illustrates a distinct time-dependent pattern for shockable cases, where the probability of functional recovery remained highest within the first 30 minutes but declined as LFT increased (p < 0.0537). For the non-shockable cohort, however, the rate of favorable neurological outcomes remained minimal and relatively stable regardless of the resuscitation duration (p = 0.102), suggesting that shortening the low-flow duration alone may be insufficient to improve neurological outcome in this subgroup. These findings align with prior large-scale observations suggesting that while LFT is an important factor, its independent impact can be masked by strong predictors such as age and bystander intervention [5, 9].

Stratified analysis by initial cardiac rhythm

Threshold Analysis via ROC Curves (Fig. 2) To identify the optimal clinical cut-points for LFT, ROC analysis was performed. For patients presenting with a shockable rhythm, LFT demonstrated modest predictive value for both mortality (Fig. 2a, AUC = 0.574) and neurological outcome defined as CPC 1–2 at discharge (Fig. 2b, AUC = 0.577). The optimal threshold for predicting these outcomes was observed between 26 and 27 minutes. Specifically, using an LFT cut-point of 27 minutes may serve as an exploratory indicator for prognosis assessment in shockable cases, but it should not be considered a definitive or decisive clinical threshold.

Conversely, LFT showed negligible predictive utility for the non-shockable rhythm group. The ROC curves for both mortality (Fig. 2c) and functional status (Fig. 2d) demonstrated poor discriminatory performance. Although mathematical modeling suggested thresholds between 22 and 29 minutes, these failed to reach statistical significance, underscoring the poor overall prognosis of this cohort. These findings may have clinical implications: while favorable outcomes were more commonly observed within the first 26 minutes for shockable cases, these thresholds remain exploratory and should not directly guide termination-of-resuscitation (TOR) decisions without further validation. These findings suggest that the temporal relationship between LFT and outcomes may differ according to the initial cardiac rhythm based on the metabolic state of the myocardium at the time of arrest [10, 11].

Optimal threshold and trend analysis

Multivariable Analysis of LFT and Covariates Multivariable logistic regression confirmed that for the shockable cohort, LFT categories of 31–40 min (AOR 2.25, 95% CI 1.19–4.27, p = 0.013) and >40 min (AOR 2.50, 95% CI 1.07–5.83, p = 0.034) were independent predictors of mortality (Tables 3 and 4). However, in the general population and the non-shockable group, LFT did not emerge as a significant independent factor. Beyond LFT, advanced age and the absence of pre-hospital ROSC were consistently strong independent predictors of mortality and functional impairment across models. Furthermore, an initial shockable rhythm proved to be a highly protective independent factor against both overall mortality (Adjusted OR 0.66, p < 0.0001) and functional impairment (Adjusted OR 0.18, p < 0.0001). In a sensitivity analysis restricted to witnessed OHCA cases (Supplementary Table S1), the association between prolonged low-flow time and mortality in patients with shockable rhythms remained significant (AOR 2.67, 95% CI 1.26–5.67 for 31–40 min; AOR 3.19, 95% CI 1.16–8.77 for >40 min), supporting the robustness of our primary findings. These findings further support the observed association between prolonged low-flow time and adverse outcomes in patients with shockable rhythms, while indicating limited predictive value of LFT in non-shockable cases [3, 12].

Discussion

To our best knowledge, this study is one of the largest metropolitan-based registry analyses in Asia to specifically quantify the time-dependent threshold of low-flow time on both survival and functional outcomes for non-traumatic OHCA patients. Our primary finding reveals an exploratory time-sensitive threshold for patients with initial shockable rhythms: an LFT exceeding 27 minutes is associated with reduced survival and lower probability of favorable neurological recovery. This finding provides a potentially useful temporal marker for emergency medical providers, suggesting that the physiological resilience of the myocardium in shockable rhythms may reflect a physiological transition point near the 30-minute mark in a conventional CPR setting.

The significance of our findings lies in the rhythm-specific impact of LFT. While previous literature has broadly suggested that “time is brain,” our data indicate that LFT is not a universal predictor for all OHCA victims. For patients with non-shockable rhythms, LFT did not emerge as a significant independent predictor (AUC 0.52), likely because these patients often suffer from prolonged “no-flow” states or irreversible underlying etiologies that mask the incremental effect of resuscitation duration. Conversely, for the shockable group, every minute beyond the 27-minute threshold may represent a loss of metabolic reserve. This interpretation aligns with the “three-phase model” of cardiac arrest, where the circulatory phase transitions into the metabolic phase, making successful defibrillation and functional recovery increasingly improbable as LFT extends [2, 13].

When comparing our results with global literature, several interesting similarities and differences emerge. Previous studies reported a survival decline of 3% every 5 minutes in an ECPR-heavy context [14, 15]. In contrast, our study focuses on conventional CPR in a high-density Asian city (Taoyuan). Interestingly, our identified 27-minute threshold is shorter than the 35–40 minutes often cited in ECPR studies [14–16]. This discrepancy suggests that without the physiological support of extracorporeal circulation, the window for “functional survival” in CCPR is considerably narrower. Furthermore, unlike some Western cohorts where age is the dominant predictor [17], our multivariable analysis confirmed that LFT remained an independent factor for shockable patients even after adjusting for age and bystander intervention, highlighting the unique operational characteristics of the Taoyuan EMS system. Restricting the analysis to witnessed cases in which low-flow time is most accurately measured did not materially alter our principal findings regarding the approximately 27-minute threshold observed in shockable rhythms, suggesting that the association was not substantially distorted by the inclusion of unwitnessed cases.

Despite its strengths, this study has several limitations. First, as a retrospective registry-based study, we relied on EMS-documented LFT, which may be subject to recall bias if bystander CPR start times were estimated. Second, while we adjusted for major comorbidities, we lacked data on the “quality of CPR” (e.g., compression depth or fraction), which could potentially moderate the effect of LFT. These limitations highlight the need for future research. Specifically, our findings pave the way for a prospective study to validate whether the 27-minute threshold can be integrated into termination of resuscitation protocols to optimize EMS resource allocation without compromising patient outcomes. However, given the modest predictive performance reflected by the AUC values, this threshold should be interpreted as a supportive reference rather than a definitive clinical cutoff.

In conclusion, our study suggests that among OHCA patients with shockable rhythms, prolonged low-flow time may be associated with reduced survival and poorer neurological outcomes, with an exploratory threshold observed around 27 minutes. In non-shockable rhythms, the prognostic value of low-flow time appeared limited. These findings support the concept that rhythm-specific resuscitation strategies may be clinically relevant; however, further prospective validation is required before these thresholds can be applied to EMS decision-making or termination-of-resuscitation protocols.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

S.-C. Yang drafted the initial manuscript. Y.-C. Chou was responsible for the production and refinement of Figs. 1–3. Y.-H. Chen contributed to the development of the manuscript and performed a critical revision. Y.-J. Hu provided comprehensive review and final approval of the manuscript. All authors read and approved the final manuscript.

Funding

No funding was received for this study.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to legal restrictions regarding patient privacy within the Taoyuan Fire Department but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of Taoyuan General Hospital (IRB No. TYGH115022). The requirement for informed consent was waived by the IRB due to the retrospective nature of the study using de-identified data. All procedures performed in this study were in accordance with the ethical standards of the institutional research committee and with the Declaration of Helsinki and its later amendments.

Competing interests

The authors declare no competing interests.