| Precision |
Within‐run: N = 10 replicates, assessing at ±40 ng/ml and ±200 ng/ml or lower and upper AMR Alternatively, for two levels of control material: two runs per day in triplicate for 5 days is another suitable precision assessment Between‐run: N = 10 days minimum, N = 20 days optimal
|
Within‐run: ≤10% CV Between‐run: ≤15% CV
|
| Limit of detection |
Drug‐naïve sample testing. Minimum N = 10, optimal N = 20 samples Not applicable for chronometric or clot‐based assays reporting in seconds
|
|
| Reference interval |
Not required for quantitative drug measurements Required for dTT and ECT testing if these tests are not dabigatran calibrated and report results in seconds or ratio At a minimum, N = 20, optimally N = 40, ostensibly healthy adults
|
If normally distributed, then mean ± 2 × SD would acceptably range If not normally distributed, then 10th–90th percentile should be used
|
|
LLOQ
linearity
|
Predicated on calibration. Commercial DOAC calibrator sets may not provide a 0 ng/ml concentration Consider 0 ng/ml calibration point to improve LLOQ if performance criteria are acceptable Linearity not required for chronometric or clot‐based methods reporting in seconds Assessing the level of dabigatran required to elevate the ECT or dTT beyond the upper limit of the RI would be desirable to indicate test sensitivity threshold
|
LLOQ: Desirable to be ±20 ng/ml or less Linearity within 10% of theoretical (recovery) values This element addresses analytical sensitivity and reportable range This element will address whether extended measurement interval can be applied
|
| Method comparison |
At least 20 samples, optimally 40 samples, from DOAC‐treated patients spanning the measurement range. Ideally, comparator method is mass spectrophotometer measurements considered the gold standard
47
Use of other commercial calibrators, controls or other assayed material for method comparison may be acceptable in lieu of patient samples
|
Correlation coefficient >0.90; slope 1.0 ± 0.15; Bias ≤15% between paired results. Recommend Bland–Altman bias plots This element addresses analytical sensitivity
|
| Carryover |
Reagent carryover required if new automated platform is used. Reagent carryover is not required if kit method approved for other indication (e.g., anti‐FXa kit used for heparin) Sample carryover—may be required unless published studies on local instrument has already been assessed
|
No carryover detected |
| Stability |
If reagent is to be maintained on‐board for extended periods, then on‐board stability must be assessed. Use of longitudinal material (e.g., controls) can be used and measured at defined frequency (0, 4, 8, 16 h intervals, etc.) Not required when modifying regulatory approved methods for intended use on an IFU listed instrument with stated stability limits
|
Recovery of longitudinal material should be within 15% CV (between‐run precision acceptability criteria) |
| Interfering substances |
For dabigatran testing—assess whether heparins or other direct thrombin inhibitors affect the assay (likely) For FXa DOACs—assess effect of UFH, LMWH and/or pentasaccharide will affect the assay (likely) Assess test method interferences such lipemia, icterus or hemolysis on result. Not required when modifying regulatory approved methods Determine if ultracentrifugation will alter reported results when used for clarifying lipemic samples
|
Concomitant drug effect is expected, and findings shared with clinical team Method interferences are expected, especially with chromogenic assays (ECA, anti‐FXa). Consider seeking other IFUs to aid in method interferences This element addresses some aspects of analytical specificity
|
| Other |
|
Within EQA acceptability limits as determined by either peer group or method |
