Research concepts
Quantification limits in analytical reports
Limit of quantification or LOQ: What can be stated when a result falls below the LOQ?
Source editorial review:
Before interpreting the result
What can be stated when a result falls below the LOQ?
- Record the limit with units and distinguish the complete method from the instrument.
- Review the matrix, preparation and precision or accuracy criteria used to establish it.
- Keep 'detected', 'not detected' and 'not quantifiable' as distinct result categories.
Limit of quantification or LOQ
LOQ defines quantitative capability under stated performance criteria and conditions. It is not a universal boundary between presence and absence. A signal may be detected without its quantity being estimable at the required performance. Instrument limits can also differ from whole-procedure limits because preparation, matrix and recovery are part of measurement.
A reading case: what to check
If a table reports 'below LOQ,' do not replace the cell with zero to simplify an average. Preserve the limit, its units and the rule used for censored values. Ask whether it applies to the final sample or requires a preparation factor. This avoids turning unavailable quantification into an absence claim or an artificial comparison between laboratories.
What to preserve in the record
A result below the quantification limit is not zero. It also does not support exact quantity comparisons when the method cannot quantify the samples within a valid range.
Questions and answers
What can be stated when a result falls below the LOQ?
LOQ defines quantitative capability under stated performance criteria and conditions. It is not a universal boundary between presence and absence. A signal may be detected without its quantity being estimable at the required performance. Instrument limits can also differ from whole-procedure limits because preparation, matrix and recovery are part of measurement.
What should the review record preserve?
A result below the quantification limit is not zero. It also does not support exact quantity comparisons when the method cannot quantify the samples within a valid range.