LOD vs LOQ: Limit of Detection and Quantitation Explained

The limit of detection (LOD) is the lowest amount of an analyte that a method can reliably distinguish from zero; the limit of quantitation (LOQ) is the lowest amount it can measure with acceptable accuracy and precision. LOD answers "is it there?", while LOQ answers "how much is there, with confidence?", and the LOQ is always equal to or higher than the LOD.

These limits matter whenever results sit near zero: pesticide residues against a maximum residue level, impurities in a drug substance, low-level clinical markers or a pathogen in food. Report a number below the LOQ and you may be reporting noise as fact.

LoB, LoD and LoQ: the CLSI EP17 terms

Clinical laboratories commonly use the three-level terminology of CLSI EP17, Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures. Check that you are using the current edition.

• Limit of blank (LoB): the highest apparent result expected when replicates of a blank sample, containing no analyte, are tested.
• Limit of detection (LoD): the lowest concentration that can be reliably distinguished from the LoB.
• Limit of quantitation (LoQ): the lowest concentration that meets defined goals for bias and imprecision, or total error.

In the parametric approach widely described for EP17, LoB = mean of blank + 1.645 × SD of blank, and LoD = LoB + 1.645 × SD of a low-concentration sample. The factor 1.645 corresponds to the 95th percentile of a one-sided normal distribution, so false positives and false negatives are each held to about 5%.

Terms vary by sector. Limit of quantitation, limit of quantification and quantitation limit mean the same thing, and ICH Q2 uses detection limit (DL) and quantitation limit (QL).

Quick reference table

Common estimation approaches

Standard deviation of the blank

Measure replicate blanks and low-level samples across several days and, where possible, several reagent lots and instruments. Simplified versions, such as LOD = mean blank + 3 SD, are widely used in testing laboratories. Whichever you choose, state it in the validation report.

Signal-to-noise ratio

In chromatography, a common approach described in ICH Q2 compares the analyte signal with baseline noise. A ratio of about 3:1 is generally accepted for the detection limit (2:1 is also mentioned) and about 10:1 for the quantitation limit. It only suits methods with a visible baseline, and the way noise is measured must be defined.

Calibration curve: standard deviation and slope

ICH Q2 also describes DL = 3.3σ/S and QL = 10σ/S, where σ is the standard deviation of the response, for example of blanks or the residual SD of a calibration curve near the limit, and S is the slope. ICH Q2 has been revised as Q2(R2), so check the current version.

Confirm by spiking

Whatever the calculation, confirm the LOQ by analysing samples spiked at that level and showing that recovery and precision meet your criteria. A calculated LOQ that has never been tested is an estimate, not a validated limit.

Qualitative vs quantitative methods

For quantitative methods, both limits matter, and the reportable range starts at the LOQ. For qualitative (positive or negative) methods there is no LOQ. The key figure is the LOD, often expressed as the concentration detected in a stated proportion of replicates, such as 95% in many molecular assays.

Near the LOD, detection becomes probabilistic. A sample at the LOD will sometimes test negative, so a negative result means "not detected at the method's detection capability", not "absent".

LOD in food microbiology

Detection methods for pathogens such as Salmonella or Listeria monocytogenes report results per test portion, for example detected or not detected in 25 g. Enrichment means, in principle, that very low numbers of viable cells can be found, but performance depends on the food matrix, injured cells and competing flora.

The ISO 16140 series on microbiological method validation uses LOD50, the level giving a 50% probability of detection, to characterise qualitative methods. For plate counts, the lower limit depends on dilution and volume plated: if 1 mL of a 1 in 10 dilution yields no colonies, the result is commonly reported as less than 10 CFU/g.

Reporting results below the LOQ

Write your reporting policy down and apply it consistently. Common practice:

• Below the LOD: report as not detected, stating the LOD, or as less than the LOD.
• Between LOD and LOQ: report as less than the LOQ, or detected below the LOQ, without a numerical value unless a client or regulation requires an estimate.
• At or above the LOQ: report the value with units and, where required, measurement uncertainty.
• Never report zero.

Some authorities and schemes specify how results near the LOQ must be reported, so confirm requirements with the relevant authority or client.

Worked example: pesticide residues in a Dubai food lab

A food testing laboratory in Dubai, accredited to ISO/IEC 17025:2017, validates an LC-MS/MS multi-residue method for pesticides in fresh vegetables using QuEChERS extraction. Some clients need results against EU limits, where a default maximum residue level of 0.01 mg/kg applies when no specific limit is set, so the lab targets an LOQ of 0.01 mg/kg for most compounds. Local limits are confirmed with the relevant authority, such as Dubai Municipality.

The team first estimates the LOD from signal-to-noise in matrix extracts: for imidacloprid, 3:1 is reached at about 0.003 mg/kg. It then spikes blank tomato samples at 0.01 mg/kg, five replicates on each of two days. Mean recovery is 94% and the relative standard deviation is 11%, within the 70 to 120% recovery and 20% or lower RSD criteria commonly applied under EU SANTE guidance for pesticide residues. The LOQ is set at 0.01 mg/kg.

For one compound, recovery at 0.01 mg/kg is only 58%. The lab sets its LOQ at the next validated level, 0.02 mg/kg, and tells clients that results for that compound cannot demonstrate compliance with a 0.01 mg/kg limit.

In routine work, a sample shows imidacloprid at a calculated 0.006 mg/kg, above the LOD but below the LOQ. The report states less than 0.01 mg/kg (LOQ). Each batch includes a matrix-matched standard at the LOQ to confirm the method still performs at that level.

Common mistakes

• Calculating LOD and LOQ without confirming them with spiked samples
• Using solvent standards when the matrix suppresses the signal
• Reporting numerical values below the LOQ as if they were reliable
• Reporting zero instead of less than the LOD or LOQ
• Setting an LOQ above the regulatory limit you need to assess
• Treating a negative qualitative result as proof of absence

Key takeaway: The LOD tells you whether an analyte is present, the LOQ tells you the lowest level you can trust as a number, and both must be confirmed experimentally in the matrix you actually test.

Frequently Asked Questions

What is the difference between LOD and LOQ?

The LOD is the lowest concentration a method can reliably distinguish from a blank, so it answers whether the analyte is present. The LOQ is the lowest concentration that can be measured with acceptable accuracy and precision. The LOQ is equal to or higher than the LOD, and quantitative results are reported only from the LOQ upwards.

How do you calculate LOD and LOQ from a calibration curve?

A common approach described in ICH Q2 uses LOD = 3.3σ/S and LOQ = 10σ/S, where σ is the standard deviation of the response, such as the residual SD of the regression or the SD of blanks, and S is the slope. Always confirm the calculated limits by analysing samples spiked at those levels.

What signal-to-noise ratio is used for LOD and LOQ?

In chromatography, a signal-to-noise ratio of about 3:1 is generally accepted for the detection limit and about 10:1 for the quantitation limit, as described in ICH Q2. It only works for methods with measurable baseline noise, and the noise measurement should be defined.

What is the limit of blank?

The limit of blank, or LoB, is a CLSI EP17 term for the highest apparent result expected from a sample containing no analyte. It is commonly estimated as the mean of blank replicates plus 1.645 times their SD. The LoD is then set above the LoB using replicates of a low-concentration sample.

Next steps

To validate detection limits and defend them at assessment, our ISO/IEC 17025 Advanced: Laboratory Competence & Accreditation course covers method validation and reporting, and Laboratory Quality Management, GLP & ISO Best Practices builds the wider system. For related reading, see method validation vs verification, measurement uncertainty in testing labs and OOS vs OOT results. For help choosing, speak to a counsellor.