Comparison of Point-of-Care PT-INR by Hand-Held Device with Conventional PT-INR Testing in Anti-phospholipid Antibody Syndrome Patients on Oral Anticoagulation

Abstract

Evidence on agreement of point-of-care (POC) INR testing with laboratory testing in APS patients on oral anticoagulation (OAC), is scarce. This study assessed agreement of paired PT INR testing by a POC device vs. conventional platform-based laboratory test, in APS patients on OAC using a pre-determined definition of agreement. Simultaneous paired PT INR estimation in 92 APS patients was carried out, during October 2020-September 2021. POC INR was performed on capillary blood (pin prick) using the qLabs® PT-INR hand-held device, while laboratory INR estimation was performed using citrated blood (venepuncture) on STA-R Max Analyzer® using STA-NeoPTimal thromboplastin reagent®. Concordance was defined no greater than ± 30% (as per international standards ISO 17593:2007 guidelines) for each paired INR estimation. Agreement between the two was defined as ≥ 90% of paired INR measurements being concordant. 211 paired estimations were performed, within which 190 (90%) were concordant. Good correlation was seen between the 2 methods of INR estimation on Bland Altman plot analysis with an Intra-class correlation coefficient (95% CI) of 0.91(0.882, 0.932). Lab INR range > 4 (P = 0.001) was a significant predictor of higher variability between both methods of INR estimation. Lupus anti-coagulant, other anti-phospholipid antibodies (APL) or triple APL positivity did not result in any statistically significant variation in paired measurements. This study demonstrated good correlation between POC INR measurement and Lab INR estimation and agreement was ascertained between the 2 methods in APS patients on OAC.

Introduction

Anti-phospholipid syndrome (APS), an autoimmune disorder, is characterized by frequent occurrence of vascular thromboses and/or pregnancy morbidity [1]. Oral anticoagulation (OAC) therapy is of utmost importance in preventing future recurrent thromboses [2]. While on OAC, APS patients need to maintain a narrow therapeutic range of International Normalized Ratio (INR), to ensure no future risk of thromboses or bleeding.

Despite adherence to OAC, a 29% cumulative incidence of subsequent thromboembolic events has been reported among APS patients [3]. Among many causes for the same, one important reason could be a sub-therapeutic INR. This can result from lack of adherence to monitoring and subsequent OAC dose adjustments. Frequent laboratory visits, repeated venipuncture, distance to laboratory, direct or indirect costs involved in the process etc. could be hindrances to conventional laboratory platform-based INR (hereafter called Lab INR) testing. Point-of-Care (POC) or home-based INR testing with handheld portable coagulometers, having rapid turn-around times (typically within seconds) and ease of frequent out-of-hospital testing, could simplify the life of APS patients; especially those who require indefinite anticoagulation. POC INR testing has been deemed effective and safe in everyday clinical practice, for monitoring patients on OAC in situations other than APS [4–9]. However the accuracy of POC INR testing in APS patients has thus far been assessed only in a few studies [10–13] with no present literature from Indian sub-continent/Asian region.

Therefore, primary objective of the current study was to check agreement of INR testing by POC qLABS® PT INR hand-held device (hereafter called device) and lab INR in APS patients on OAC.

Methods

Research Setting

The current study was conducted at Christian Medical College Hospital, Vellore; a tertiary level rheumatology service centre, in Southern India, following approval by institutional ethics committee (IRB 12,811, dated June 15, 2020). Patients and controls were recruited from October 2020 to September 2021, following written informed consent.

Study Population

We included APS patients satisfying the revised Sapporo criteria [1] with age ≥ 18 years, on OAC > 7 days (with Vitamin K antagonists [VKA] only). APS patients on (a) concomitant heparin therapy and (b) Novel anti-coagulants (NOACs) were excluded from the study, preceding the INR measurement. Post- test exclusions were done following error codes for the device pertaining to insufficient blood sample (EC1), internal quality check failure (EC2), defective strip (EC3), device temperature out of operating range (EC4) or, hematocrit outside the operational range (EC5).

Controls were patients ≥ 18 years on OAC with VKA > 7 days, for any indication with a diagnosis other than APS. Exclusion Criteria for controls was same as followed for patients.

Performance of Paired INR Testing

Paired INR estimation in APS patients was done either simultaneously, or within 30-min difference from either method.

POC INR testing was performed on capillary blood obtained by finger prick with a lancing device on a qLabs® PT-INR electrometer device (Micropoint Biotechnologies, Inc. China). The INR value was displayed on the screen less than 2 min from the time of application of blood. On insertion of a proprietary INR test strip, the device automatically detected and heated it to a pre-set operating temperature. The drop of blood added to the sample well of the strip, flowed through the test channels to two reaction zones: Test Zone and a Control Zones, where the blood reacted with pre-printed reagents and began to coagulate. Each reaction zone contained a pair of electrodes to which a constant voltage was applied by the electrometer. As the coagulation of the blood proceeded, the current monitored across the two electrodes changed. The device detected the change of the current and determined PT/INR. The operational specification ranges of the device were as follows: haematocrit (30–55%), temperature (10–35 °C), humidity (10–90%) and maximum altitude (4300 m/14,000ft). The device could detect INR values 1–7. It also measured the clotting time in the Quality Control (QC) zone. If the QC result failed to fall within a predetermined range, the device would display an error code. The INR test strips were all within their shelf life period at the time of usage and stored as per manufacturer’s instructions of storage environment temperature: − 10 to + 40 °C and relative humidity 10 to ~ 90%.

Lab INR estimation was performed on citrated blood, obtained by venepuncture by using STA-R Max Analyzer® with STA-NeoPTimal thromboplastin reagent®. A mixture of thromboplastin was added to citrated plasma and the time of clot formation was determined. The STA–R Max Analyzer, a fully automated coagulation platform is based on electromagnetic viscosity detection measuring the oscillation amplitude of a steel ball in the cuvette with thromboplastin and plasma, within an electromagnetic field. An algorithm determined the clotting time based on the changes in oscillation amplitude following changes in viscosity due to coagulation.

Study Definitions

Concordance was defined as no greater than ± 30% (as per international standards ISO 17593:2007 guidelines) for each paired INR estimation [14]. Any INR value displayed by the device as Error code corresponding to INR value being too high and corresponding lab INR being ≥ 7, was also taken as a concordant measurement. Agreement between the 2 methods was defined as ≥ 90% of paired INR measurements being concordant.

Statistical Methods

Data for continuous variables were as summarized using mean ± SD/Median (with inter quartile range) and percentage/frequency for categorical variables. Comparison of continuous variables was performed using independent t-test, while chi-square test was used for categorical variables. The consistency between the two methods was checked with Bland–Altman plot. Significance was considered at p < 0.05.Data analysis was undertaken using SPSS version 25.

Results

Two hundred eleven and 51 paired INR estimations performed in 92 APS patients and 27 controls respectively, were found valid as per the inclusion criteria and thereby, considered for statistical analysis. Thirty-five APS patients and 4 controls were excluded on screening, since they utilized concomitant heparin/NOAC. Post-test exclusions were done for error codes pertaining to POC device; EC1 (n = 4 in APS patients & n = 2 controls), EC2 (n = 1 APS patient), EC3 (n = 1 APS patient), EC4 (n = 1 APS patient) and EC5 (n = 1 APS patient & n = 2 control). Patient and control characteristics are described in Table 1.

Table 1.

Demographic and clinical characteristics of APS patients and controls in the study

	APS patients (n=211)	Controls (n=51)
Female: Male	166:45	44:7
Mean age (SD)	34 ± 9.7	38.5 ± 10.7
Primary: Secondary APS	74: 137	–
Indication for anticoagulation n (%)	DVT 97 (46)	Cardiac valve replacement 14 (27.4)
	Pulmonary embolism 50 (23.7)	Atrial fibrillation 4 (7.8)
	Budd chiari syndrome 6 (2.8)	DVT 6 (11.8)
	Arterial stroke 30 (14.2)	Pulmonary embolism 5 (9.8)
	CVT 11 (5.2)	CVT 9 (17.6)
	Arterial ulcer 7 (3.3)	Budd chiari syndrome 4 (7.8)
	Myocardial infarction 1 (0.5)	Arterial stroke 2 (3.9)
	Gangrene 19 (9.0)	Arterial ulcer 1 (1.9)
	Splenic infarct 5 (2.4)	Myocardial infarction 1(1.9)
	Arterial vessel thrombosis 7 (3.3)	Arterial vessel Thrombosis 5(9.8)
	Bone infarcts 4 (1.9)	Vasculitis 3 (5.9)
	Hepatic infarct 4 (1.9)
	CRAO 3 (1.4)
Nicoumalone: Warfarin	165:46	31:20
LA positivity	194/210 (92.4%)	–
ACL positivity (IgM/IgG)	99/205 (48.3%)	–
Beta glycoprotein I IgG positivity	73/175 (41.7%)	–
Beta glycoprotein I IgM positivity	60/151 (39.7%)	–
Triple APL positivity	70/161 (43.5%)	–

APS Antiphospholipid antibody syndrome, SD Standard deviation, DVT Deep venous thrombosis, CVT Cortical venous thrombosis, CRAO Central retinal artery occlusion, LA Lupus anticoagulant, ACL Anticardiolipin, APL Antiphospholipid antibody

Range of Laboratory INR in Patients and Controls

Median (IQR) Lab INR value noted was 2.11(1.49, 2.74) with highest and least lab INR values being 10.0 and 0.94 respectively, in APS patients. Median (IQR) Lab INR value noted was 2.3(1.63, 3.38) with highest and least lab INR values being 10.0 and 1.15 respectively, among controls.

Concordance Rate of Paired INR Estimations in Controls

Forty-nine (96.1%) paired INR estimations were concordant. Median percentage (IQR) variability seen between paired INR estimations was 7.1(− 3.1, 14.8).

Concordance Rate of Paired INR Estimations and Agreement in APS Patients

One hundred ninety (90%) paired INR estimations were concordant. The predefined target for agreement was met at 90% concordance rate of paired INR estimations between the 2 methods. Median percentage (IQR) variability seen between paired INR estimations was 8.9(− 1.7, 17.6). The median percentage variability in paired INR, seen in APS patients was not statistically significant, in comparison to that seen among controls (p = 0.45).

Bland Altman Analysis and Predictors of Variability Between the Two Methods

Good correlation was seen between the 2 methods on Bland Altman plot analysis (Fig. 1a) with an Intra-class correlation co-efficient (ICC) (95% CI) of 0.91(0.882, 0.932), however poor correlation was seen when lab INR > 4 (Fig. 1b). Univariate analysis also showed that Lab INR range > 4 being significant predictor of variability in paired INR estimations (p < 0.001).Lupus anti-coagulant (LA) or other anti-phospholipid antibodies (APL) or triple APL positivity did not result in any statistically significant variation in paired INR measurements.

Fig. 1.

a Bland Altman plot analysis between the paired INR estimations between the INR device and conventional laboratory INR method b Bland Altman plot analysis between the paired INR estimations between the INR device and conventional laboratory INR method (for lab INR ranges of < 2, 2–4 and > 4)

Discussion

In this comparative observational study, assessing agreement between paired POC INR with Lab INR testing in 211 paired INR estimations in APS patients on OAC, good correlation and agreement was established between both methods.

Studies performed thus far, evaluating agreement between POC INR and lab INR in APS patients on OAC, are summarized in Table 2. Each study used a different POC device and different thromboplastin reagents, with varying sample sizes and differing definitions of agreements. The most robust study among these is by Isert et al. [12], who examined Coagucheck XS device and lab INR platform using 5 different thromboplastins. Further, 211 paired INR comparisons were carried out for each lab thromboplastin agent and POC device. Although, the present study evaluated a single device POC INR versus single thromboplastin lab INR platform, it had the same number of paired INR measurements as the study by Isert et al. [12], which can be counted as its strength.

Table 2.

Review of literature of studies done thus far in Point of Care INR testing in APS patients

Study	Number of APS patients/Paired INR estimations	Number of controls/Paired INR estimations	POC INR device	Lab INR methods	Definition of agreement	Results
Barcellona et al. [10]	36/123	41/141	Coagucheck XS	RecombiPlasTin	Difference of ± 0.5 INR unit between the 2 systems	Agreement rates 97%, 84%, 58% & 40% for lab INR ranges < 2, 2–3,3–4 and > 4 respectively
Braham et al. [11]	29/29	31/31	ProTime InRhythmTM System	CFX assay	Absolute difference less than 0.4 units	97% of the device INR measurements were similar to laboratory INR values
Mecki Isert et al. [12]	40/928	100/433	Coagucheck XS	1) Neoplastin Plus 2) RecombiPlasTin 3) Innovin 4) Thromborel S 5) CFX	Deviation ≤ 0.5 INR units between 2 methods	Agreement of INR test results obtained from different PT assays ranged from 81.5% to 87.7%. Agreement between Coagucheck XS & CFX was just 55.6%
Perry et al. [13]	59/59	49/49	1) ProTime INR 2) Hemochron Signature	1) Simplastin-HTF 2) Chromogenic Factor X levels	Within 0.4 INR units	ProTime & Lab INR agreement rates 89%, 66%, 80% and 0% for lab INR ranges < 2, 2–3,3–4 and > 4 respectively Hemochron & Lab INR agreement rates 78%, 72%, 78% and 20% for lab INR ranges < 2, 2–3, 3–4 and > 4 respectively
Current study	92/211	27/51	qLABs®PT INR device	STA-NeoPTimal	Concordance defined no greater than ± 30% (ISO 17593:2007 guidelines) for each paired INR estimation. Agreement between the 2 methods defined as ≥ 90% of paired INR measurements being concordant	90% concordance rate among APS patients

APS Antiphospholipid antibody syndrome, INR International normalized ratio, POC Point of care, PT Prothrombin, CFX Chromogenic factor X assay

The device utilized in the current study, has been validated in non APS setting with proven clinical agreement (upto 93%) and reproduced in another similar experimental non-APS setting [15, 16]. Ours is the first study in which this device has been assessed and validated in the setting of APS, and the study results need to be validated and reproduced in a similar experimental setting in APS. All research work concerned with this study was conceptualized and completed during the ongoing COVID19 pandemic; with the intention of advising home based POC INR monitoring in APS patients, who observed difficulty in INR monitoring, given the travel restrictions during the pandemic. These two points can be considered novel aspects of the study.

With regard to other factors, it was concluded in a previous study by Perry et al. [13], that APLs do interfere in POC INR testing. Additionally Barcellona et al. [10] demonstrated high discordance rates in triple APL positive patients when compared to those with lesser APL positivity. Isert et al. [12] also noted the influence of LA on variability between device INR and lab INR with regard to three thromboplastin agents, namely Neoplastin Plus, RecombiPlasTin and Innovin, but not with Thromborel S. However, we did not observe the influence of any APL/Triple positivity on the variability in INR measurement by both methods; although it should be mentioned that LA negative population in the current study was less (7.6%), however keeping in mind that sample size of the current study was primarily not powered to comment on the influence of APLs on INR variability. The current study was limited due to non-performance of simultaneous CFX assay for comparison but as these assays are phospholipid-independent and not influenced by APLs, they would be best suited in APS patients on OAC for monitoring. Conventional PT INR method is known to be non-concordant with inhibition of acute thrombin generation particularly at VKA induction [17]. This POC device being based on the same principle as conventional PT INR, caution should be exercised in the context of home use of this device during initiation of VKA therapy. POC INR measurements in this study were done by trained investigators and should not be immediately extrapolated for self-testing INR by APS patients, as they will need specialized training to operate the device.

Conclusion

This study demonstrated good correlation between the POC and Lab INR estimations and ascertained agreement between both when assessed in APS patients on OAC. Home based self-monitoring of INR can therefore be explored and advised in APS patients on OAC, so as to maintain target INR and hold great relevance especially in the prevalent and future pandemic scenario with restricted mobility and access to laboratory evaluations.

Funding

The study was funded by the Intramural Fluid Research Grant of Christian Medical College, Vellore, India (IRB No 12811, dated June 15 2020).

Declarations

Conflict of interest

The authors have no competing interests to disclose.

Ethical Approval

Study was performed following due approval from Ethics committee, Christian Medical College, Vellore, India.