Building the comparison: specimens, criteria, and limits

EP26 runs in two stages. Stage 1 is a one-time setup per assay: the laboratory sets the target concentration intervals, the critical difference, the expected imprecision, the statistical power, the rejection limit, and the sample size. Stage 2 is the evaluation itself, repeated for every new lot using that established design. The guided example below works through Stage 2 reasoning; it does not reproduce EP26's proprietary Stage 1 statistical worksheet, and no single formula here should be read as CLSI's mandated calculation. EP26 is written for quantitative measurement procedures interpreted against a cutoff. It explicitly excludes methods where a lot-related shift is already expected and clinically accommodated, such as coagulation testing (PT and aPTT), where lot-specific reference intervals are the normal handling.

Patient samples for the comparison should span the clinically important concentration range and cluster around medical decision points, not just the population average, because a small bias near a cutoff matters more than the same bias far from any cutoff. Specimens need adequate volume, documented stability, and a minimal time gap between the current-lot run and the candidate-lot run, the guided case below uses under two hours, so that specimen aging is not mistaken for lot bias. Pooled or processed patient material can be useful when fresh individual specimens are not practical, but it should never be assumed automatically commutable; whenever a nonpatient material substitutes for fresh plasma or serum, commutability has to be actively considered, not assumed.

CAP requires new reagent lots and shipments to be checked against the previous lot or a suitable reference material before or concurrently with being placed into service, for any reagent that produces a chemical or biological reaction to detect or measure an analyte. For CAP-accredited quantitative nonwaived testing, patient specimens are the preferred comparison material; acceptable alternatives when patient specimens are unavailable include QC products with lot-specific target values, proficiency-testing material with a peer-group mean, or third-party reference material shown commutable for that method. For qualitative nonwaived testing, the checklist minimum is at least one known positive and one known negative specimen against the new lot, with a weak positive recommended when weak-positive results are clinically reported. CAP also requires that components of a multi-component reagent kit be used only within their own kit lot unless the manufacturer states otherwise, because mixing kit lots introduces an uncontrolled variable into the comparison.

A related but separate question is specimen-type equivalence: whether plasma can substitute for serum, or one anticoagulant tube for another, on a given method. CLSI EP35 (2nd ed., 2025, corrected November 2025) is the current guideline for that question. EP35 judges equivalence or suitability against predefined, clinically justified acceptance criteria, bias limits, agreement near decision points, allowable analytical error, rather than a bare correlation coefficient, and it separates similar-matrix comparisons (equivalence) from dissimilar-matrix comparisons (suitability). EP35 answers a different question than EP26: EP35 asks whether a specimen type is acceptable for a method, EP26 asks whether a reagent lot change moved that method's results.

Under 42 CFR 493.1255(b)(3)(i), a complete reagent change triggers calibration verification unless the laboratory documents that the lot change does not affect the reportable range and that control values are not adversely affected. When calibration verification is required, 42 CFR 493.1255(b)(2)(ii) sets the minimum materials at a minimal or zero value, a mid-point value, and a value near the upper limit of the reportable range. An Individualized Quality Control Plan (IQCP) under CLIA can let a laboratory use an alternative number, concentration, or frequency of controls, but only when a documented risk assessment and quality control and assessment plans demonstrate equivalent quality, and an IQCP can never be less stringent than the manufacturer's own instructions.

Delta checks add a second, ongoing layer after cutover: CLSI EP33 describes comparing a patient's current result to a recent prior result on the same analyte to flag an unexpected change, which may reflect a specimen problem, an analytical problem such as a lot shift, or a genuine clinical change. That is exactly how the opening scenario's lot problem first surfaced. Parallel operation, running the outgoing and incoming lot side by side on the same patient samples before cutover, is the practical mechanism the Stage 2 comparison depends on, and it requires enough overlapping inventory and expiration dating for both lots to be tested together before the old lot runs out.

Before the first specimen is run, write down what concentration range you are testing, which decision points you are protecting, what material you will use if patient specimens are limited, and what difference would be too large to accept, because deciding that after seeing the data invites bias in the decision, not just the assay.