Portal de Periódicos da CAPES

Comparison of ‘time within therapeutic INR range’ with ‘percentage INR within therapeutic range’ for assessing long‐term anticoagulation control in children

2011; Elsevier BV; Volume: 9; Issue: 5 Linguagem: Inglês

10.1111/j.1538-7836.2011.04248.x

1538-7933

T.T. BISS, P. J. Avery, P.M. WALSH, Farhad Kamali,

Blood Pressure and Hypertension Studies

Maintenance of anticoagulation response to warfarin within target therapeutic range (TTR) is essential to minimize the risk of either thromboembolism or hemorrhage caused by under- or over-dosing, respectively. Anticoagulation control in children receiving chronic warfarin therapy has been reported by many studies to be poor [1-3]. The increasing use of point-of-care testing devices and parent/patient education programmes for anticoagulated children has somewhat improved control but it remains suboptimal when compared with anticoagulated adults [4-7]. Anticoagulant control in children is usually reported as the percentage of International Normalised Ratio (INR) values within TTR (%ITTR) [1-4, 6]. However, the advent of point-of-care testing devices has allowed INR testing to occur more frequently during periods of instability, which might render the use of %ITTR an inappropriate form of assessing anticoagulation control. An alternative method described by Rosendaal et al. [8] measures percentage time in therapeutic range (%TIR) using linear interpolation. This method allocates an INR value to each day, including days between INR tests, and is therefore likely to minimize the disproportionate effect of temporary frequent INR testing on the assessment of long-term anticoagulant control. This study compared the %ITTR method with the %TIR method for assessing anticoagulant control in a cohort of children treated with warfarin. Anticoagulant records and hospital notes for children who were anticoagulated with warfarin under the care of paediatric cardiology services at The Freeman Hospital, Newcastle upon Tyne, were examined retrospectively. Data were collected from 3 months after the start of anticoagulant therapy and included demographic details, indication for warfarin therapy, target INR range, INR values, frequency of INR tests and occurrence of thromboembolic and hemorrhagic events. The cohort was divided into subgroups according to target INR range (2.0–3.0 or 2.5–3.5) and indication for anticoagulant therapy (Fontan circulation, other cardiac indication or non-cardiac indication). The patient cohort was divided into four age groups (≤ 1, 2–5, 6–12 or 13–17 years) according to age at the time of each INR measurement. Data were inputted into 4S Dawn Clinical Software (4S Information Systems Ltd, Milnthorpe, UK) in order to calculate %TIR using the linear interpolation methodology described previously by Rosendaal et al. [8]. The average %ITTR and %TIR were determined for each patient for the duration of their anticoagulant monitoring. %ITTR was compared to %TIR using a paired t-test and the differences between %ITTR and %TIR were compared between subgroups. Mean number of dose changes per month was compared between subgroups, following logarithmic transformation to achieve approximate normality within groups, using the t-test or one-way analysis of variance (anova). Mean number of INR measurements per month was compared between subgroups using the Mann–Whitney test or Kruskall–Wallis test. Statistical analysis was performed using Minitab v15.0 (Coventry, UK) and P-values < 0.05 were considered significant. Anticoagulation records for 38 consecutive children (21 boys) with a median age of 8.3 years (range, 1.1–17.2 years), anticoagulated with warfarin between January 1996 and April 2009, were reviewed. All children were monitored using a point-of-care testing device (Coaguchek®S or Coaguchek®XS; Roche Diagnostics Ltd, Burgess Hill, UK) and received advice on warfarin dose by telephone contact with the same two pediatric cardiology nurse specialists. Indications for anticoagulant therapy were: Fontan procedure, 16 patients; prosthetic mitral valve replacement, eight; primary pulmonary hypertension, four; cardiomyopathy, three; coronary aneurysm, one; thrombotic stroke, one; and truncus arteriosus repair, one. Patients were monitored for a median of 29 months (range, 2–115 months) and data were collected for a total of 112 years of warfarin therapy. Treatment outcome variables are shown in Table 1. P-values correspond to the differences between %ITTR and %TIR for the entire cohort and for each of the subgroups. For the entire cohort, mean %TIR was significantly higher than the mean %ITTR (63.8% vs. 55.0%; P = 0.002). %TIR was higher than %ITTR for each of the two target INR ranges, indication and age cohorts, reaching statistical significance for the following subgroups: target INR 2.0–3.0 and 2.5–3.5; other cardiac indication; and 6–12 years of age and 13–17 years of age. When the differences between %ITTR and %TIR were compared between the subgroups for target INR range, indication and age, the indication for anticoagulant therapy was the only factor that had a significant effect (Table 1). Children with target INR 2.5–3.5 and cardiac indication other than Fontan circulation had a greater number of INR tests and more warfarin dose changes per month than children with target INR range 2.0–3.0 and those who were anticoagulated for Fontan procedure, respectively, although these differences were not statistically significant (Table 1). There were significantly larger differences between %TIR and %ITTR in children with target INR 2.5–3.5 and in those who were anticoagulated for cardiac reasons other than Fontan procedure, the majority of whom had artificial heart valves. These groups of children also had a greater frequency of INR tests and warfarin dose changes per month but there was no significant correlation between the degree of difference between %ITTR and %TIR and the frequency of testing (r = −0.1, P > 0.5, Pearson correlation coefficient). A statistically significant difference between %ITTR and %TIR was seen in the older age groups (≥ 6 years) but not in the younger age groups (< 6 years), which may have been due to the reduction in power seen as a result of smaller numbers of patients. Hemorrhagic and thromboembolic events occurred in seven patients and one patient, respectively, giving rates of 0.063% and 0.009% per year. When reported as %ITTR, anticoagulant control in children is poor, ranging from 40% to 63% [1-3]. The use of a point-of-care testing device would be expected to improve anticoagulation control in children through a greater frequency of INR testing during periods of instability, which allows for early warfarin dose adjustments in order to maintain therapeutic efficacy. The increasing use of point-of-care testing in the home setting, often with support for dosing decisions from a trained specialist nursing team, has apparently had little impact on anticoagulation control, with reported %ITTR ranging from 60% to 70% [4, 6]. In contrast, anticoagulation control has been reported to be significantly greater in children in more recent studies of the impact of a parent/patient education programme, with a child-focused approach, when anticoagulation control is assessed using the linear interpolation method, with %TIR ranging up to 87% [9, 10]. Our study results highlight the existent discrepancy between the two methods currently used for assessing chronic anticoagulation control in children. We are of the opinion that the quality of anticoagulation control in children should not be reported as %ITTR. The latter is a poor measure of control unless INR measurements are taken at regular, predetermined intervals, which is generally not the case for anticoagulated children monitored using a point-of-care testing device. Use of a linear interpolation method reduces the impact of multiple INR values over a short period of time, which are ‘out of range’, and places more emphasis on the longer periods of stability during which the INR is tested less frequently [11]. Measurement of TIR, and therefore the actual number of days during which TTR is achieved, may correlate better with adverse clinical events, as it would be expected to provide a more accurate indication of the proportion of time that a patient has supratherapeutic or subtherapeutic INR and is therefore at risk of hemorrhagic or thromboembolic events. Due to the very low frequency of adverse events in this cohort of children we were unable to evaluate this. Linear interpolation methodology has some limitations. It can be biased by individual INR values that are far outside of TTR and it assumes that the change in INR over time is linear between each time-point, which may not be true. Small departures from target range are considered identical to large departures, which may not be correct as a larger deviation from target range is more likely to result in an adverse clinical event. In addition, a more complex calculation is required to determine TIR whereas the %ITTR is a simple measure of the proportion of the total number of INR values that are within TTR [12]. This study has not made it possible for the authors to recommend one measure of quality of anticoagulant control over another. This would require knowledge of how each measure performs as a surrogate for the efficacy and safety of warfarin anticoagulation in children. The rates of thromboembolic and hemorrhagic events were too low for a valid comparison to be made, although it is noted that a previous study published by Barbui et al. [13] found no difference in quality of anticoagulant control as measured by %ITTR or %TIR in a cohort of adult patients who had experienced adverse events. In conclusion, the %ITTR method may underestimate the quality of anticoagulant control in children and the %TIR method may be more appropriate for this patient group. Methods of assessing anticoagulant control in children should be compared in larger numbers of children and should be correlated with adverse clinical outcomes in terms of hemorrhagic and thromboembolic events. F. Kamali and T. T. Biss designed the study. T. T. Biss and P. Walsh were responsible for data collection. P. Avery, F. Kamali and T. T. Biss performed data analysis. All authors were involved in the preparation of the final manuscript. T. T. Biss was supported by a Baxter Royal College of Pathologists research training fellowship. The authors state that they have no conflict of interest.