Device-dependent activated clotting time measurements during transseptal ablation under continuous anticoagulation: Comparison of mechanical-optical and microfluidic electrochemical systems

Abstract

Background

Monitoring activated clotting time (ACT) is essential during transseptal ablation. Despite having distinct analytical principles and potentially non-equivalent values, different point-of-care ACT systems are used interchangeably.

Objective

This study aimed to compare ACT values from a mechanical-optical and a microfluidic electrochemical system in patients treated with direct oral anticoagulant (DOAC) undergoing transseptal ablation and to quantify the impact of device-related bias on guideline-recommended ACT targets.

Methods

30 consecutive patients undergoing left-sided transseptal catheter ablation under uninterrupted DOAC therapy were studied. After transseptal puncture, unfractionated heparin was administered, and ACT was measured simultaneously with both systems at predefined time points. The primary analysis included 42 paired post-heparin measurements; a complementary analysis included all 55 paired measurements (pre- and post-heparin). Agreement was assessed using paired t-tests, linear regression, and Bland–Altman analysis. Thresholds of ≥250 seconds and ≥300 seconds were evaluated.

Results

In the primary post-heparin analysis, mean ACT was 316.3 ± 51.5 seconds with the mechanical-optical system and 259.3 ± 29.8 seconds with the microfluidic system, yielding a mean bias of 57.0 seconds (95% confidence interval, 41.7–72.3; P < .001). Including all 55 measurements, the mean bias remained large at 48.4 seconds with strong per-measurement correlation. Using a threshold of ≥250 seconds, 90.0% vs 46.6% of measurements met the target; at ≥300 seconds, 60.0% vs 10.0%.

Conclusion

Mechanical-optical and microfluidic electrochemical ACT systems show a consistent positive offset of roughly 50–60 seconds. Uniform ACT cutoffs across devices may misclassify anticoagulation adequacy, supporting device- and cartridge-specific ACT targets during DOAC-treated transseptal ablation.

Graphical abstract

Key Findings.

• ▪The analytical values of point-of-care activated clotting time (ACT) during transseptal ablation are dependent upon the measurement system employed, given that mechanical-optical and microfluidic electrochemical platforms adhere to divergent analytical principles and are not interchangeable.
• ▪Higher ACT values are consistently reported by the mechanical-optical system than the microfluidic electrochemical system across the clinically relevant intraprocedural range, indicating a systematic device-related bias rather than random variation.
• ▪Due to this inherent bias, the implementation of a uniform, device-agnostic ACT target can result in markedly divergent assessments of whether anticoagulation is “sufficient,” contingent exclusively on the employed ACT system.
• ▪These observations highlight the importance of having device- and cartridge-specific ACT target ranges, as well as the need for consistent use of the same ACT system within and across procedures when managing anticoagulation during transseptal ablation procedures.

Introduction

Catheter ablation via transseptal access has become a cornerstone in the management of supraventricular (SVT)¹ and ventricular arrhythmias (VT),² with pulmonary vein isolation (PVI) established as the standard approach for atrial fibrillation (AF) ablation.³ The growing number of electrophysiological procedures (EP), especially PVIs, coupled with the emphasis on same-day discharge and low complication rates, has intensified the focus on intraprocedural anticoagulation. Although advances in technique and periprocedural management have decreased the incidence of thromboembolic and bleeding events, these remain clinically relevant.4, 5, 6

Oral anticoagulation strategies have shifted from vitamin K antagonists (VKAs) to direct oral anticoagulants (DOACs). Large randomized and observational studies have demonstrated that uninterrupted or minimally interrupted DOAC therapy is at least as safe, and in some settings even safer, as uninterrupted VKA therapy with respect to stroke and major bleeding during AF ablation.7, 8, 9, 10, 11, 12

To further reduce thromboembolic complications during transseptal puncture (TSP), guidelines recommend the administration of intravenous unfractionated heparin (UFH) before or following TSP, alongside the maintenance of an activated clotting time (ACT) of at least 300 seconds, or even 350 seconds.³ The main source of this ACT cutoff has been established in VKA-treated cohorts and may not fully account for the pharmacodynamic differences between VKAs and DOACs, or among different DOACs, particularly factor Xa inhibitors, which are now used in the majority of patients undergoing AF ablation.^6,8,11, 12, 13, 14

Furthermore, there is a growing awareness that ACT itself is not a standardized measurement. Several point-of-care ACT systems are used interchangeably in electrophysiology, interventional cardiology, and cardiac surgery, but they differ in terms of their activators, detection principles, and cartridge configurations.15, 16, 17, 18, 19, 20, 21, 22, 23 Comparative studies show that substantially different ACT values can be obtained with different systems, especially during TSP where strict ACT management is essential.^20,22,23

It has been argued that ACT thresholds should be adapted to the different measurement devices and should not be applied uniformly across platforms.^20,23,24 Nevertheless, current AF ablation guidelines recommend numerical ACT cutoffs without specifying device-specific adjustments, and centers may even switch between ACT devices based on availability or institutional preference without adapting their specific ACT cutoffs.³ Still, many practitioners are not aware of these device-specific differences and the possible clinical implications for patients treated with DOAC. Against this backdrop, we undertook a comparative analysis of ACT values obtained from 2 systems: a microfluidic electrochemical cartridge-based ACT system (i-STAT Alinity; Abbott) and a mechanical-optical clot detection system (GEM Hemochron 100; Werfen). We aimed to quantify the systematic bias between these 2 commonly used systems across the clinically relevant ACT range and assess how this bias affects the proportion of patients achieving the recommended ACT thresholds during transseptal ablation procedures performed under uninterrupted DOAC therapy.

Methods

Study population and procedural setting

We performed a retrospective analysis using prospectively collected data from 30 consecutive patients undergoing left-sided transseptal catheter ablation at a single tertiary electrophysiology center in June 2025. Procedures included PVI using cryoablation or radiofrequency energy, VT-ablation, and SVT-ablation, all requiring TSP. ACT was measured simultaneously with both systems at all predefined time points, and all patients continued their DOAC uninterrupted on the day of the procedure. Patients were included if they had an indication for transseptal ablation and excluded in the presence of a bleeding diathesis, hypercoagulable state, creatinine clearance of <15 mL/minute, or severe liver disease. The study was approved by the Brandenburg Ethics Committee (approval number 2026-15-BO-ff) and conducted in accordance with the Declaration of Helsinki; written informed consent was obtained from all, and the ethics committee waived the requirement for additional informed consent for the retrospective analysis.

Ablation procedure

All procedures were performed using a single transseptal technique under fluoroscopic guidance. PVI was performed with either an open-irrigated contact force catheter (TactiFlex or TactiCath, St. Paul, MN) guided by electroanatomical mapping (EnSite NavX, St. Paul, MN) via a steerable sheath (Agilis, St. Paul, MN), or a 28 mm cryoballoon (Medtronic, Minneapolis, MN, USA) with an inner-lumen mapping catheter (AchieveVR, MedtronicVC), which was advanced to each pulmonary vein (PV) ostium through a steerable 15-Fr sheath (FlexCath AdvanceVR, MedtronicVC, Minneapolis, MN). The endpoint was entrance and exit block of all PVs. Additional ablation lines were performed at the operatorś discretion. VT ablation was substrate-based, targeting late potentials and local abnormal ventricular activities (LAVAs), guided by electroanatomical mapping (EnSite NavX, St. Paul, MN) via a steerable sheath (Agilis, St. Paul, MN) with non-inducibility of clinical VT as the procedural endpoint. The single SVT case involved ablation of a left lateral accessory pathway in a patient with concealed Wolff–Parkinson–White syndrome.

ACT monitoring protocol

ACT was measured using a microfluidic electrochemical cartridge-based device (i-STAT Alinity; Abbott Point of Care Inc., Princeton, NJ, USA) and a mechanical-optical clot-detection system (GEM Hemochron 100; Werfen, Bedford, MA). The i-STAT Alinity system uses the principle of microfluidic electrochemistry, whereby blood is introduced into a single-use cartridge containing an activator and embedded electrodes. Coagulation-induced changes in the electrical properties are then detected over time, and an ACT value is calculated based on the clotting behavior within the microfluidic channel (measurement range 50–1000 seconds).^16,18 The GEM Hemochron 100 uses a mechanical-optical clot detection principle. Activated whole blood is contained in a cuvette that rotates or oscillates, and clot formation is inferred from changes in blood movement and optical density. Low-dose cartridges designed for monitoring heparinized whole blood in electrophysiology and interventional settings were used in accordance with the manufacturer’s recommendations (practical range ≤400 seconds).^16,22 At each time point, a single blood sample was drawn from the venous sheath and transferred simultaneously to both systems. All ACT measurements were performed by trained nursing staff in the electrophysiology laboratory. Measurements were conducted under standardized ambient conditions (20°C–23°C) and blood samples were processed immediately after collection, in accordance with the manufacturer’s instructions, to minimize pre-analytical variability.

Anticoagulation regimen

UFH was administered as a weight-adjusted intravenous bolus after TSP (120–130 U/kg for pre-ACT ≥150 seconds; 140–150 U/kg for pre-ACT <150 seconds), targeting an ACT of 300–400 seconds.²⁵ ACT was measured 15–30 minutes after the bolus and at 30-minute intervals thereafter; additional boluses were given for ACT of <300 seconds (3000 U for ACT 250–274 seconds; 2000 U for ACT 275–299 seconds). All decisions regarding heparin dosing were based exclusively on ACT values obtained from the mechanical-optical system (GEM Hemochron 100). The microfluidic electrochemical system (i-STAT Alinity) was used in parallel for research purposes only and did not affect the management of procedural anticoagulation. If ACT exceeded 350 seconds at procedure end, protamine was administered (3000 IU for ACT >350 seconds; 2000 IU for ACT >320 seconds). Venous hemostasis was achieved with a figure-of-eight suture, and DOACs were resumed per product characteristics and contemporary peri-ablation protocols.

Data collection and statistical analysis

Statistical analyses were performed in R version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are reported as mean ± standard deviation, categorical variables as counts and percentages. The primary analysis involved comparing the mean ACT difference between systems across all post-heparin ACT measurements obtained after TSP. This was achieved using a paired t-test with 95% confidence intervals (CIs). Complementary analyses were then performed using all paired measurements, including pre-heparin values. Agreement was assessed by linear regression (microfluidic electrochemical s mechanical-optical values, compared with the line of identity), Bland–Altman analysis, and Pearson correlation coefficient. To characterize the heparin–ACT dose-response relationship, separate linear regressions were performed for each system with ACT as the dependent variable and cumulative UFH dose per kilogram as the independent variable; R² quantified the proportion of explained variance. As multiple ACT measurements were obtained from the same patients, an additional linear mixed-effects model was fitted (using the R package lme4), with each patient acting as a random intercept. The dependent variable was the paired difference between the mechanical-optical and microfluidic electrochemical ACT values. This model was applied to all post-heparin paired measurements to obtain a bias estimate and a 95% CI, while accounting for within-patient correlation. The proportion of measurements achieving ACT thresholds of ≥250 seconds and ≥300 seconds was calculated separately for each system to assess the clinical implications of the observed bias. A 2-sided P-value of <.05 was considered statistically significant.

Results

Patient and procedural characteristics

A total of 30 patients were included, yielding 55 paired ACT measurements (mean 1.87 ± 0.97 pairs per patient); 42 paired ACT measurements after TSP and 13 paired measurements before TSP. The baseline characteristics reflected those of a typical contemporary transseptal ablation cohort (see Table 1). The mean age was 67.4 ± 10.6 years, with a male predominance (73.3%), and a mean body mass index of 29.8 ± 5.7 kg/m². All patients were treated with a DOAC due to a history of AF or planned ventricular tachycardia (VT) ablation. Apixaban was used in 60% of cases, followed by rivaroxaban (16.6%) and edoxaban (10%), reflecting current treatment patterns for AF in Germany. Procedures included 25 PVIs for AF, treated with either the 28 mm cryoballoon (20%) or the radiofrequency technique (63.3%), as well as 4 patients with VT ablation (13.3%) and 1 patient with SVT ablation (3.3%). All ablations required transseptal access. There were no serious adverse events, specifically no clinically relevant groin bleeding, significant pericardial effusion or documented cerebral ischemic events, during the peri-procedural period.

Table 1.

Baseline patient characteristics

Baseline characteristics	n = 30
Age (years)	67.4 ± 10.6
Men, n (%)	22 (73.3)
BMI (kg/m2)	29.8 ± 5.7
Left ventricular ejection fraction (%)	49.67 ± 13.7
Direct oral anticoagulation (DOAC)
Apixaban	18 (60.0)
Rivaroxaban	5 (16.6)
Edoxaban	3 (10.0)
Dabigatran	1 (3.3)
Laboratory parameters
Hemoglobin (mmol/L)	8.9 ± 1.0
Platelets (×109/L)	215.2 ± 60.7
Quick (%)	68.5 ± 16.6
INR	1.25 ± 0.29
aPTT (s)	35.8 ± 8.7
GFR (mL/min per 1.73 m2)	71.9 ± 23.0
Comorbidities
Arterial hypertension, n (%)	22 (73.3)
Diabetes mellitus, n (%)	4 (13.3)
Coronary artery disease, n (%)	8 (25,6)
Chronic pulmonary disease, n (%)	2 (6.6)
Renal insufficiency, n (%)	3 (10)
Heart failure, n (%)	11 (36.6)
Procedural data
Pulmonary vein isolation (Cryo), n (%)	6 (20.0)
Pulmonary vein isolation (Radiofrequency), n (%)	19 (63.3)
Ventricular tachycardia ablation, n (%)	4 (13.3)
Supraventricular tachycardia ablation, n (%)	1 (3.3)
Procedure duration (min)	82.3 ± 27.2
Ultrasound-guided puncture, n (%)	13 (43.3)

Values are n (%), n, mean ± standard deviation.

BMI = body mass index; GFR = glomerular filtration rate; INR = International Normalized Ratio; aPTT (s) = activated partial thromboplastin time (seconds).

Comparison of ACT values between measurement systems

For the primary analysis, our focus was on intraprocedural anticoagulation following TSP and UFH administration. Of the 42 paired post-heparin ACT measurements, the mean ACT was 316.3 ± 51.5 seconds when measured with the mechanical-optical system, compared with 259.3 ± 29.8 seconds with the microfluidic electrochemical system (Figure 1). This corresponds to a mean paired difference (bias) of 57.0 seconds (95% CI, 41.7–72.3 seconds; P = 2.21 × 10^-8). Linear regression analysis showed that, across the measured range, the regression line for mechanical-optical vs microfluidic electrochemical ACT values consistently lay above the line of identity (Figure 2). The Pearson correlation coefficient between the 2 systems was moderate (r = 0.60; 95% CI, 0.31–0.79; P = .0004) and Bland–Altman analysis revealed a consistent positive bias of around 57 seconds across clinically relevant ACT values (see Figure 3). In a linear mixed-effects model, with each patient included as a random intercept and all 42 post-heparin paired measurements as observations, the estimated mean bias (mechanical-optical minus microfluidic electrochemical system) was 62.0 seconds (95% CI, 47.6–76.4 seconds; P = 6.1 × 10^-9). This confirmed the presence of a highly significant systematic offset, even when accounting for repeated measurements within patients.

Figure 1.

Comparison of mean activated clotting time (ACT) values, as measured using a microfluidic electrochemical system and a mechanical-optical system, was conducted in 30 patients. Boxplots illustrate the higher ACT values obtained using the mechanical-optical system. Each dot represents the mean ACT value for an individual patient, with lines connecting corresponding values.

Figure 2.

Scatter plot comparing mean ACT values obtained with a microfluidic electrochemical system and a mechanical-optical system. The solid regression line lies above the identity line (dashed), indicating that the mechanical-optical system systematically overestimates ACT across the observed range. ACT = activated clotting time.

Figure 3.

Bland–Altman plot for mean ACT values measured with a microfluidic electrochemical system and a mechanical-optical system. The differences between the 2 methods are plotted against the mean ACT, showing a positive bias of 57 seconds that is consistent with a systematic, device-related offset rather than random variation. ACT = activated clotting time.

To assess robustness across all measurement time points, we performed a complementary analysis that included all 55 paired ACT measurements taken before and after TSP and heparin administration. Each pair was treated individually. Even in this expanded dataset, the mechanical-optical system yielded higher ACT values than the microfluidic electrochemical system on average, with a mean bias of 48.4 seconds and a standard deviation of the differences of 44.9 seconds. The correlation between the 2 systems was strong on a per-measurement basis (r = 0.85; 95% CI, 0.75–0.91; P < 2.2 × 10^-16) (see Figure 4), yet the Bland–Altman plot demonstrated a persistent positive offset of the mechanical-optical system across the entire ACT range (see Figure 5), indicating close tracking but poor interchangeability.

Figure 4.

Scatter plot comparing all individual ACT measurements obtained with a microfluidic electrochemical system and a mechanical-optical system. The regression line again lies above the line of identity, demonstrating that the mechanical-optical system yields higher ACT values than the microfluidic system at all time points. ACT = activated clotting time.

Figure 5.

Bland–Altman plot for all individual ACT measurements obtained with a microfluidic electrochemical system and a mechanical-optical system. The plot shows that the mechanical-optical system produces higher ACT values across the full range, indicating a consistent positive bias of 48 seconds for all measurements. ACT = activated clotting time.

Impact on achievement of ACT thresholds

When applying a threshold of ≥300 seconds, adequate anticoagulation was achieved in 60% of cases using the mechanical-optical system and in only 10% of cases using the microfluidic electrochemical system. Using the less stringent threshold of ≥250 seconds, 90% of measurements were classified as therapeutic using the mechanical-optical system, compared with 46.6% using the microfluidic electrochemical system (Table 2). Therefore, applying identical numeric ACT targets to both systems resulted in a difference of more than 2-fold in the apparent proportion of patients with adequate anticoagulation at ≥250 seconds and an even more pronounced discrepancy at ≥300 seconds.

Table 2.

Proportion of measurements achieving guideline-recommended ACT thresholds

ACT threshold	Mechanical-optical system (%)	Microfluidic electrochemical system (%)
≥300 s	90.0	46.6
≥250 s	60.0	10.0

ACT = activated clotting time.

Discussion

The results of our study demonstrate a substantial and systematic bias between 2 widely used point-of-care ACT measurement systems in patients treated with DOAC undergoing transseptal catheter ablation. Our key findings are as follows. (1) In the primary analysis of mean ACT measurements obtained after TSP and UFH administration, the mechanical-optical system consistently yielded ACT values approximately 57 seconds higher than the microfluidic electrochemical system. In a complementary analysis including all 55 paired measurements before and after TSP and heparin administration, the mean bias remained large at 48 seconds, and the per-measurement correlation was strong (r ≈ 0.85), indicating that the 2 systems correlate with a constant positive bias of the mechanical-optical device across the ACT range. (2) This pattern, confirmed by both regression and Bland–Altman analyses, supports the presence of a systematic, device-related difference rather than random variation. (3) When identical ACT thresholds were applied, this bias resulted in significant discrepancies in the proportion of patients considered adequately anticoagulated, with substantially higher rates of “therapeutic” ACT documented by the mechanical-optical system than by the microfluidic electrochemical system at both ≥250 seconds and ≥300 seconds.

Comparison with previous literature and device-specific findings

Our findings are consistent with prior comparative studies of ACT systems in different clinical settings. Dirkmann et al¹⁶ reported systematically higher ACT values with Hemochron Signature Elite vs i-STAT in cardiac surgery (mean differences 40–50 seconds), and Kemna et al¹⁷ confirmed significant discordance between the 2 platforms in a multicenter comparison of over 1000 paired samples. In percutaneous coronary intervention (PCI), Schussler et al²¹ observed a mean 44-second excess with Hemochron, and Fülkell et al²² demonstrated consistently higher Hemochron Signature readings with poor agreement at therapeutic ACT levels during AF ablation. Most recently, Kovacs et al published a systematic review and international survey on ACT platform variability in the EP setting, reporting mean paired differences of 50–63 seconds with weak-to-moderate concordance.²³ Our Pearson correlation of r = 0.604 (95% CI, 0.31–0.79; P = .0004) is consistent with the upper end of this range. Notably, Sakanoue et al reported r = 0.90 yet still demonstrated a 50-second systematic bias, confirming that even high correlation does not preclude clinically meaningful disagreement, and that the ∼57-second Bland–Altman bias, rather than the correlation coefficient, carries clinical significance.²⁰ Our additional analyses demonstrate a bias of 48 seconds when both pre- and post-heparin measurements are included, and a bias of 57–62 seconds for the mean therapeutic ACT after heparin administration. These results are consistent with the 40–60 second excess reported for Hemochron-type systems in previous studies and highlight that the magnitude of the offset may be greater after heparin administration. Previous publications have shown that variability increases at higher ACT levels across point-of-care platforms.¹⁸ Our Bland–Altman plots (Figures 4 and 5) indicate a similar trend: the discrepancy between the ACT systems broadens above approximately 200–250 seconds. This suggests a possible proportional bias, with greater discordance at therapeutic and supratherapeutic ACT values. When interpreting ACT in this range, this should be taken into account.

Our study adds 2 novel aspects. First, we used the GEM Hemochron 100 with low-dose cartridges, which is increasingly employed in contemporary EP and may exhibit different calibration characteristics compared with the high-dose cartridges studied previously.^16,19 Still, we observed a mean positive bias within the upper end of the 40–60 second range previously reported for mechanical-optical systems.^16,17,20,22 Second, our cohort consisted entirely of patients on uninterrupted DOAC therapy, predominantly apixaban (60%), whereas most prior comparisons involved VKA-treated or cardiac surgery populations, limiting their direct applicability to contemporary EP practice.^16,17,19,21 The consistency of findings across cardiac surgery,^16,17,19 PCI,^15,21,26 and AF ablation^20,22,23 supports the view that the bias is an intrinsic property of the measurement technologies rather than a procedural or pharmacological artefact. The consistent positive bias of the mechanical-optical system may be explained by fundamental differences in the detection endpoint and the design of the assay. While the microfluidic electrochemical system detects early electrochemical changes associated with fibrin polymerization within a microfluidic channel, the mechanical-optical system detects macroscopic clot formation via changes in blood movement and optical density, which is a later endpoint in the coagulation cascade. Furthermore, differences in contact pathway activator composition and independent calibration frameworks mean that the 2 systems do not share a common biological reference point. Therefore, the consistently lower values obtained using the microfluidic electrochemical system may reflect an earlier detection threshold rather than underestimation of anticoagulation per se.

Anticoagulation management in patients treated with DOAC

All patients continued DOAC therapy uninterrupted (apixaban 60%, rivaroxaban 16.6%, edoxaban 10%), reflecting contemporary evidence supporting uninterrupted DOAC strategies for EP procedures.7, 8, 9, 10, 11, 12 Despite the shift from VKA to DOAC, guideline ACT targets remain extrapolated from VKA data,^6,13,14 and the pharmacodynamic interaction between DOACs and UFH differs substantially from that between VKA and UFH.⁶ Factor Xa inhibitors, particularly apixaban, are associated with lower ACT values for a given heparin dose. Martin et al found that patients on apixaban or rivaroxaban required higher cumulative heparin doses to reach target ACT,¹⁴ and Nagao et al confirmed that patients treated with apixaban achieved the lowest intraprocedural ACT values despite equivalent or higher heparin doses.²⁷ These observations raise the possibility that rigid adherence to VKA-era ACT targets may result in over-heparinization in patients treated with DOAC without a proven reduction in thromboembolic risk.⁶

In our cohort, mean ACT was 316.3 seconds (mechanical-optical) vs 254.8 seconds (microfluidic electrochemical), with only 10% of microfluidic electrochemical measurements exceeding 300 seconds compared with 60% of mechanical-optical measurements. Sole reliance on the microfluidic electrochemical system would likely have prompted additional heparin in the majority of cases, with potential bleeding implications. Winkle et al. reported that with interrupted anticoagulation and open-irrigated catheters, a target ACT of 225 seconds (average achieved: 202 ± 7.5 seconds) was associated with no strokes and a 1.9% overall complication rate, suggesting that very high ACT values may not be essential for thromboembolic prevention in all settings.²⁸ But these findings cannot be extrapolated to uninterrupted DOAC practice, and Winkle et al do not mention the ACT point-of-care system they were using. Nevertheless, Boyle et al have similarly called for a re-evaluation of ACT targets in the DOAC era, arguing that the 300–350 seconds range may be unnecessarily high.²⁴ Our data support the view that intraprocedural ACT is shaped by the interplay of DOAC pharmacology, heparin dosing, and device performance, and that current practice remains poorly standardized.^6,13,14

Device-specific ACT targets: The danger of universal cutoffs

Current guidelines recommend ACT ≥300 seconds during transseptal ablation without specifying device-dependent thresholds.³ Our results demonstrate the clinical consequences of this omission: at ≥250 seconds, 90% of mechanical-optical measurements were classified as therapeutic vs only 46.6% of microfluidic electrochemical measurements; at ≥300 seconds, 60% vs 10%. Despite identical underlying anticoagulation, clinical management would differ in the majority of cases based solely on device selection.^20,22,24

Sakanoue et al showed that device-agnostic thresholds can misclassify anticoagulation adequacy in up to 40% of cases.²⁰ Fülkell et al highlighted that cartridge type within the same platform introduces further variability,²² and Korpi–Steiner et al¹⁸ documented substantial ACT differences across 4 point-of-care systems in routine clinical use. Operators relying on a microfluidic electrochemical system and targeting ≥300 seconds may administer excess heparin, while mechanical-optical system users may achieve guideline thresholds with doses that would appear inadequate on the microfluidic electrochemical system.

Practical solutions may include: (1) establishing device- and cartridge-specific ACT target ranges through local validation; (2) ensuring device consistency within and across procedures; (3) incorporating DOAC type into anticoagulation algorithms; and (4) supporting randomized trials or registries to define outcome-based, device-stratified ACT targets for DOAC-treated cohorts. Until such data are available, clinicians should interpret ACT in the context of the specific device and cartridge used, acknowledge the uncertainty of extrapolating VKA-era targets, and balance ACT-guided dosing with individualized thromboembolic and bleeding risk assessment.^6,20,23,24

Limitations

Several limitations of this study warrant consideration. First, the sample size was modest (30 patients, 55 paired ACT measurements) and reflects the predefined enrolment window of an ethics-approved pilot study. While this sample size was sufficient to demonstrate a highly significant and consistent bias between systems, it limited the possibility of conducting subgroup analyses and generalizing the findings. Further, larger, prospectively designed follow-up studies are required to confirm and refine these observations. Second, our cohort was heterogeneous in procedural indication, catheter technology, and operator technique, precluding procedure-specific analyses. Furthermore, ACT measurements were not repeated within the same analyzer system at each time point. Therefore, intra-system reproducibility and cartridge-to-cartridge variability could not be formally assessed. Future studies should incorporate duplicate measurements to enable the quantification of random within-device error. Third, DOAC regimens were clinically determined, precluding firm DOAC-specific conclusions. Fourth, heparin dosing was operator-directed, and findings are specific to low-dose mechanical-optical cartridges and the microfluidic electrochemical platform, limiting generalizability to other devices or dosing strategies. Finally, our study was not designed or powered to evaluate whether the ACT discrepancy between platforms translates into differences in thromboembolic or bleeding outcomes. Prospective registries or randomized studies involving larger patient groups and systematic endpoint collection that correlate platform-specific ACT values with clinical outcomes are urgently needed to establish evidence-based, device-stratified anticoagulation targets for patients treated with DOACs undergoing transseptal ablation. Despite these limitations, this study provides robust, clinically relevant evidence of a systematic bias between 2 widely used point-of-care ACT systems during transseptal ablation under uninterrupted DOAC therapy, underscoring the need for device-specific ACT interpretation in contemporary EP practice.

Conclusion

Our results demonstrate substantial and systematic differences between mechanical-optical (Hemochron) and microfluidic electrochemical (i-STAT) point-of-care ACT systems during transseptal catheter ablation under uninterrupted DOAC therapy. Applying uniform ACT thresholds irrespective of device may lead to markedly divergent classifications of anticoagulation adequacy and risks unintended under- or over-anticoagulation, supporting the need for device- and cartridge-specific ACT targets during EP procedures in patients treated with DOAC.

Declaration of AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used Perplexity and DeepL Write to improve the English language writing. After using these tools/services, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Disclosures

The authors state that they have no conflicts of interest.