Point-of-care monitoring of warfarin treatment in community dwelling elderly

Abstract

The objective of this study was to assess clinical effectiveness and costs of launching point-of-care monitoring of warfarin treatment in community dwelling frail elderly patients. A prospective multicentre controlled randomised study over 12 months comparing a point-of-care strategy with usual monitoring routines was carried out in primary healthcare centres and anticoagulation clinics in southeast Sweden. The subjects were community dwelling elderly across rural southeast Sweden on chronic warfarin treatment. Main outcome measures were time in therapeutic range (TTR), rate of treatment-related adverse events and costs. The study comprised 103 elderly people (61% women) mean age 86 yrs (range 75--98) treated with warfarin for median 9 yrs (range 1--18). Patients randomised to start point-of-care monitoring (n = 55) showed 75.9% in TTR before trial vs. 72.6% during trial (ns). The patients randomised to continue on usual monitoring routines (n = 48) showed 75.2% in TTR prior to trial vs. 72.9% during trial (ns). The point-of-care monitoring showed potential savings of SEK 624 per patient annually (based partly on effects that were not statistically significant). The study shows that point-of-care monitoring of warfarin treatment in community dwelling elderly in rural areas is as effective as usual monitoring routines and that it may offer savings to society.

Introduction

Warfarin is an anticoagulant drug affecting the synthesis of clotting factors II, VII, IX and X by interfering with the vitamin K metabolism. Treatment with warfarin is effective in preventing thromboembolism and in patients with atrial fibrillation. In Sweden approximately 180,000 patients are treated with warfarin corresponding to 2.0% of the population¹ and the numbers are still not decreasing despite the non-vitamin K dependent oral anticoagulants that have been introduced in recent years. For these patients, warfarin treatment reduces the risk for embolic stroke events by two-thirds.² However, bleeding events during warfarin treatment are frequent, appearing at the rate of 1.0–1.5 per 100 treatment years^3,4 with the most severe bleeding complications often explained by excessive anticoagulation due to drug-drug interactions.⁵ Accordingly; close monitoring of the warfarin treatment is essential to reduce the risk for either bleeding or thrombo-embolic events.

This may be particularly true for frail (physically weak) elderly patients as advanced age and deteriorating kidney function are known to be risk factors for serious warfarin-related bleeding complications.⁶ In Sweden patients once regarded suitable for chronic warfarin treatment are expected to continue on warfarin life-long unless the physician in charge finds the risks to outweigh any potential benefits.

Point-of-care monitoring, i.e. blood sampling and measurement outside hospital laboratories, has been shown to have at least the same quality as monitoring at anticoagulation clinics for patients who perform their own testing.^7,8 Point-of-care monitoring has also been found to be cost-effective in some studies,^9,10 even though a modelling study from the UK calculated that point-of-care monitoring failed to show cost-effectiveness.¹¹ However, point-of-care monitoring has not been studied for community dwelling frail elderly patients.

All of this led us to embark on a controlled randomised prospective trial to study the feasibility, quality and clinical effectiveness of a point-of-care monitoring strategy as compared with the usual monitoring routines.

Material and Methods

CoaguChek XS

CoaguChek XS measures prothrombin time on fresh whole blood, drawn by a finger-stick. The blood drop is applied on one end of a test strip, and the other end is inserted in the instrument before blood collection. The test strip contains human recombinant thromboplastin and a thrombin substrate, which upon cleavage increases conductivity. The change in conductivity is transformed to International Normalized Ratio (INR) and presented on the screen of the instrument by the soft-ware. Each lot of bar-coded test strips is calibrated by the manufacturer.

Study design and population

A randomised controlled prospective trial was carried out in 2008-2010 designed to comprise 100 patients aged 70 years or above on chronic warfarin treatment for minimum 2 years living in community settings across rural southeast Sweden. After written informed consent the patients were randomised to continue with the usual monitoring routines (see below) or start a nurse-led monitoring of the warfarin treatment over the next 12 months using CoaguChek®XS Plus (Roche Diagnostics Scandinavia, Stockholm Sweden) that enabled capillary blood samples to be drawn and analysed on site and dosing by an anticoagulation clinic on line (‘Point-of-care monitoring strategy’). The results read on Coaguchek were immediately reported by telephone to the anti-coagulation clinic (run by the local hospital), which typed the results on their computer. The INR-values are interpreted by a trained person, who has access to all previous readings and dosages of warfarin. All dosages, INR-values, side-effects etc, were registered in the data program “AVK-brevet” (Journalia, Kil, Sweden).

The usual hospital based monitoring routines involved district nurses taking venous samples at the patient’s home and bringing them to the nearest hospital-affiliated laboratory facility for analysis and dosing before returning within 24 hours to the patients for administrating warfarin tablets according to any change in the dosing.

All patients were included between December 2008 and November 2009 and followed over 12 months.

The reasons for the warfarin treatment are listed in Table 1.

Table 1.

List of stated indications for the warfarin treatment.

Indication for the warfarin treatment	No. patients
Atrial fibrillation	67
Venous thrombo-embolism	22
Stroke	7
Miscellaneous	7

Outcome measures

Outcomes were determined by measuring percentage of days in therapeutic range (TTR) as outlined by Rosendaal,¹² rate of treatment-related adverse events, overall processing times expressed as costs, and health-related quality of life among patients for either strategy.

Data collection

Baseline demographic details are shown in Table 2. Health status at entry was evaluated using the Barthel Index¹³ and the EQ-5D instrument.¹⁴ The EQ-5D instrument was also used to create quality-adjusted life-years (QALY) weights which represent health-related quality of life. The baseline values of the Barthel Index and the EQ-5D indicates that the studied group belong to the “frail elderly”, as the results are far below average. Repeat evaluations were carried out at end of trial. Details of the International normalised ratio (INR) measures prior to trial were obtained from the medical records carefully taking into account any individually set deviation from the INR 2.0 - 3.0 common target range. The INR was measured within the Point-of-care monitoring strategy by means of CoaguChek®XS Plus and within the usual hospital based monitoring routines by means of analysis of citrated blood using a combined rabbit thromboplastin (i.e. an Owren method) and a spectrophotometric end-point utilising the instrument Simple Simon™ (Zafena, Borensberg, Sweden).

Table 2.

Patient demographics.

Patients	Point-of-care monitoring (n = 55)	Usual monitoring (n = 48)	Total (n = 103)
Women	33 (60%)	29 (60%)	62 (60%)
Men	22 (40%)	19 (40%)	41 (40%)
Age (years)	85.7	84.8	85.3
Living alone	40 (73%)	31 (65%)	71 (68%)
Home care services	35 (64%)	35 (73%)	70 (68%)
Nursing home	17 (31%)	13 (27%)	30 (29%)
Years on warfarin (median)	9	9	9
Years on warfarin (range)	1–18	1–16	1–18

Costs were calculated in Swedish Krona (SEK), in 2014 year values (exchange rate SEK1=Euro 0,105 as of December 15^th 2014). Only costs that were expected to differ between point-of-care monitoring and usual monitoring routines were estimated, which consisted of the equipment, the time used by nurses and the cost of transportations. Time used by nurses consisted of time for travel and time for others, which includes the time to wait to leave blood tests at the lab, waiting for an answer, and checking faxes etc. Time used by nurses was calculated using an average monthly salary of SEK 28,000 plus 50 percent due to labour fees and other costs. This lead to a cost per hour of SEK 262.50. Cost per kilometre was set at SEK 1.85 according to the recommendation by the Swedish Tax Agency.

Statistics

The t-test (double-sided within groups) was used to compare days spent within the therapeutic range 12 months prior to trial and corresponding data during the 12 month trial. The t-test (double-sided within groups) was also applied to the QALY weights and Barthel’s Index. Number of kilometres travelled by nurses and time used were compared using the independent sample Mann-Whitney U test. All tests were done using the Statistica Version10 (StatSoft, Tulsa, OK USA).

Ethical considerations

The study was approved by the ethics committee of Linköping University, Sweden.

Results

The study comprised a total of 103 elderly people (61% women) mean age 86 yrs (range 75-98) treated with warfarin for median 9 yrs (range 1-18). No differences could be found between the two groups (see Table 2).

Patients randomised to start point-of-care monitoring (n = 55) showed 75.9% in TTR before trial vs. 72.6% during trial. The patients randomised to continue on usual monitoring routines (n = 48) showed 75.2% in TTR prior to trial vs. 72.9% during trial. No significant differences were found for either before vs during trial or between the randomised groups.

One warfarin-related adverse non-fatal bleeding event was reported in each strategy.

There were no differences between patients’ QALY weights or Barthel’s Index between the groups (see Table 3). Distance travelled differed between the two groups, where the nurses involved in point-of-care monitoring had fewer travelled kilometres than the other group. The time used by nurses did not differ between the groups, even though a tendency of less time was found for the Point-of-care monitoring.

Table 3.

Mean values of quality of life and cost parameters. Independent sample t-test (QALY weights and Barthel’s Index) and Mann-Whitney U test (kilometres and time used) of differences between the two groups.

	Usual care	Point-of-care
	Mean (St.d.)	Mean (St. d.)	P
QALY weight (EQ-5D)	0.42 (0.36)	0.44 (0.32)	0.821
QALY weight (VAS)	0.59 (0.18)	0.57 (0.16)	0.505
Barthel’s Index	68.2 (29.9)	60.5 (34.1)	0.231
Kilometres	406 (409)	292 (427)	0.015
Time for travel, minutes	659 (461)	602 (575)	0.172
Time for other, minutes	325 (617)	168 (181)	0.553

Cost of the Point-of-care monitoring equipment is SEK 15,900 plus SEK 30.33 for every test. If the equipment can be used for 5 years and for 1,000 tests per year, the cost per test would be SEK 33.51 (SEK 15,900 / 1,000 tests / 5 years + SEK 30.33). If a patient on average controls his or her PK 16 times a year, the annual cost per patient with Point-of-care monitoring becomes SEK 536.

Point-of-care monitoring leads to less kilometres travelled by the nurses, and by using the mean difference this accrue to a mean cost difference of SEK 210. However, this saving due to fewer kilometres is less than the extra cost of SEK 536 by using the equipment. By only using significantly verified data, point-of-care monitoring would therefore not be assumed cost-effective as no improvements are shown. However, if the tendencies of a faster procedure with point-of-care would be true, this may reduce the total costs with point-of-care monitoring. Costs related to the time used by nurses and distance traveled for point-of-care per patient and year is estimated at SEK 3930 and for usual care SEK 5090. The savings in time used by nurses and distance traveled of point-of-care (SEK 1160) is therefore larger than the extra cost of the point-of-care equipment (SEK 536), and point-of-care monitoring could therefore be assumed cost-effective as it shows a potential of saving SEK 624 per patient and year and with equal clinical outcomes.

Discussion

The main new finding of the paper is that employing a point-of-care strategy in monitoring elderly patients on warfarin outside hospital is feasible and of the same quality as the usual hospital-affiliated monitoring routines as assessed by the TTRs.

The study was conducted in rural areas of southeast Sweden and where district nurses visit warfarin treated elderly in community settings to draw venous samples before administrating warfarin according to the current INR value.

The results showed that the TTR during the trial of this aged population (72.6%) was close to the TTR (76.2%) observed in a national survey of anticoagulation quality in Sweden in 18,391 patients at a mean age of 70 years.¹⁵

Moreover the results are in good keeping with the conclusion in a study that treatment outcomes do not differ between patients monitored in the primary care or dedicated hospital anticoagulation departments.⁶ The results furthermore relates to the findings by some systematic reviews comparing point-of-care monitoring with monitoring at anticoagulation clinics.^7,8 Finally, a recent study has shown that anticoagulation treatment in the south of Sweden generally has a high quality, with a mean TTR of 75.6%.¹⁶

The cost analysis indicates potential savings by introducing a point of-care management into the caring of old and frail patients needing lifelong warfarin therapy such as stroke victims and those otherwise disabled. Similar results were found in a study by Boman et al.¹⁷, where the use of telemedicine for anticoagulant treatment in primary health care centres improved the monitoring process and indicated a tendency to reduce overall costs for primary health care centres. Indeed the results of our study would alert any health care provider to consider options to facilitate patient access to the technology to promote self-care and self-monitoring within warfarin treatment knowing that vast experience already exists (I assume) across Europe.¹⁸

The strengths include the originality of conducting a randomized controlled study in frail elderly with results that can be generalized across Scandinavia.

The weaknesses include the limited size of the study population and the short follow-up time although the advanced age of the patients may be an inherent obstacle in this regard.

During this study preferences of the treatments were collected during interviews with patients and nurses. Only positive preferences for the use of point-of-care monitoring were mentioned. The patients especially mentioned the quick procedure and that it was almost pain free. For many of these frail elderly the new oral anticoagulants are not an alternative due to poor kidney function and/or high risk of deteriorating kidney function.

Footnotes

Acknowledgements

We wish to thank all those participating in this study working with collecting and analyzing samples. We also wish to thank Roche Diagnostics Scandinavia for providing CoaguChek®XS Plus to each study site. The study was enabled by research grants from FORSS and Östergötland County Council.

References

Ericson

Bergfeldt

Bjorholt

. Atrial fibrillation: the cost of illness in Sweden. Eur J Health Econ 2011; 12: 479–87.

Hart

Pearce

Aguilar

. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007; 146: 857–67.

Hollowell

Ruigomez

Johansson

Wallander

Garcia-Rodriguez

. The incidence of bleeding complications associated with warfarin treatment in general practice in the United Kingdom. Br J Gen Pract 2003; 53: 312–4.

Kucher

Castellanos

Quiroz

Koo

Fanikos

Goldhaber

. Time trends in warfarin-associated hemorrhage. Am J Cardiol 2004; 94: 403–6.

Jonsson

Spigset

Jacobsson

Hagg

. Cerebral haemorrhage induced by warfarin - the influence of drug-drug interactions. Pharmacoepidemiol Drug Saf 2007; 16: 309–15.

Wallvik

Sjalander

Johansson

Bjuhr

Jansson

. Bleeding complications during warfarin treatment in primary healthcare centres compared with anticoagulation clinics. Scand J Prim Health Care 2007; 25: 123–8.

Garcia-Alamino JM, Ward AM, Alonso-Coello P, Perera R, Bankhead C, Fitzmaurice D, et al. Self-monitoring and self-management of oral anticoagulation. Cochrane Database Syst Rev 2010:CD003839.

SBU. [Self-testing and self-management of anticoagulant drugs], SBU Alert 2007:5. 2007 Swedish Council on Health Technology Assessment (SBU).

Craig

Chaplin

Jenks

. Warfarin monitoring economic evaluation of point of care self-monitoring compared to clinic settings. J Med Econ 2014; 17: 184–90.

10.

Davidson T, Levin L, Bergström A. [A pilot study of self testing of treatment with oral anticoagulantia. Health economic aspects]. 2013 Center for medical technology assessment (CMT), Linköping University.

11.

Connock

Stevens

Fry-Smith

Jowett

Fitzmaurice

. Clinical effectiveness and cost-effectiveness of different models of managing long-term oral anticoagulation therapy: a systematic review and economic modelling. Health Technology Assessment 2007; 11: 86–86.

12.

Rosendaal

Cannegieter

van der Meer

Briet

. A method to determine the optimal intensity of oral anticoagulant therapy. Thromb Haemost 1993; 69: 236–9.

13.

Collin

Wade

Davies

Horne

. The Barthel ADL Index: a reliability study. Int Disabil Stud 1988; 10: 61–3.

14.

Brooks

. EuroQol: the current state of play. Health Policy 1996; 37: 53–72.

15.

Wieloch

Sjalander

Frykman

Rosenqvist

Eriksson

Svensson

. Anticoagulation control in Sweden: reports of time in therapeutic range, major bleeding, and thrombo-embolic complications from the national quality registry AuriculA. Eur Heart J 2011; 32: 2282–9.

16.

Timberg

Schedvin

Svensson

. [Anticoagulant treatment audited in the south of Sweden. A well functioning practice--which can improve even more]. Lakartidningen 2012; 109: 1411–3.

17.

Boman

Davidson

Gustavsson

Olofsson

Renstrom

Johansson

. Telemedicine improves the monitoring process in anticoagulant treatment. J Telemed Telecare 2012; 18: 312–6.

18.

Heneghan

Alonso-Coello

Garcia-Alamino

Perera

Meats

Glasziou

. Self-monitoring of oral anticoagulation: a systematic review and meta-analysis. Lancet 2006; 367: 404–11.

Point-of-care monitoring of warfarin treatment in community dwelling elderly – A randomised controlled study

Abstract

Introduction

Material and Methods

CoaguChek XS

Study design and population

Outcome measures

Data collection

Statistics

Ethical considerations

Results

Discussion

Footnotes

Acknowledgements

References