Distinguish detection from reliable quantitation
A detection limit addresses the lowest amount or concentration that can be detected under defined conditions, not necessarily quantified with suitable accuracy and precision. A quantitation limit addresses the lowest level that can be quantified with suitable performance.
LLOQ is often used for the lower validated or reportable calibration level in a specific bioanalytical context. Define the term in the procedure rather than assuming all LOQ and LLOQ uses are interchangeable.
Choose an approach appropriate to the procedure
Approaches may use signal-to-noise, blank and low-level variability, calibration slope and response standard deviation, or empirical performance across low levels. The assumptions and constants differ by framework and method.
For nonlinear curves, local sensitivity and variance can change across the range. Applying one linear slope formula mechanically may not represent inverse-prediction performance near the lower asymptote.
One conventional linear approach: DL = 3.3σ ÷ S; QL = 10σ ÷ SWorked linear example: calculate an estimate, then verify it
If the response standard deviation σ is 0.006 response units and the calibration slope S is 0.020 response units per ng/mL, the conventional estimates are DL = 3.3 × 0.006 ÷ 0.020 = 0.99 ng/mL and QL = 10 × 0.006 ÷ 0.020 = 3.0 ng/mL.
Those values are estimates under a particular linear convention. They do not prove that 0.99 ng/mL is detected reliably or that 3.0 ng/mL meets the required accuracy and precision. Independent low-level samples must test those claims.
| Input or estimate | Value |
|---|---|
| Response SD, σ | 0.006 response units |
| Calibration slope, S | 0.020 response units per ng/mL |
| Estimated DL | 0.99 ng/mL |
| Estimated QL | 3.0 ng/mL |
Verify the proposed limit with samples
A calculated estimate is a starting point. Prepare independent samples at or around the proposed level and evaluate the relevant accuracy, precision, detection, classification, or total-error behavior under routine variation.
Include matrix, instruments, analysts, days, reagent lots, preparation, and other sources needed for intended use. Define handling of values below the limit and avoid reporting unqualified numbers beyond supported range.
Report the definition, method, and evidence
State the term, analyte, matrix, unit, calculation or study design, raw/derived data, model, software version, estimate, verification results, acceptance criteria, and final approved limit.
Keep display qualifiers such as <LLOQ separate from the stored numerical estimate and status. Downstream exports should transmit both machine-readable qualifier and unit.
Frequently asked questions
Is LOQ always 10 times standard deviation divided by slope?
No. That is one conventional relationship under specific assumptions. Use the approach appropriate to the procedure and applicable framework, then verify performance.
Can the lowest calibration standard automatically be called LLOQ?
Only if its role and performance are established under the method. A nominal level is not made reportable by its position in the plate layout.
Are LOD and LOQ required for every potency assay?
Not necessarily. Select performance characteristics based on intended use and applicable guidance, documenting the rationale.