Controls for Molecular Assays

What controls prove

Every control answers one question about the run:

Internal controls come in two kinds. An endogenous IC is a human gene present in the specimen itself, which also shows the specimen contained cells. An exogenous IC is a synthetic sequence spiked into each sample before extraction, which tests extraction and inhibition but says nothing about specimen adequacy ¹.

What CLIA requires

At least once each day patient specimens are tested, CLIA requires ²:

• Quantitative procedures: two control materials at different concentrations.
• Qualitative procedures: a negative and a positive control.
• Procedures with an extraction phase: two controls, one able to detect extraction errors.
• Molecular amplification procedures: two controls, plus a control capable of detecting inhibition if inhibition is a significant source of false negatives.

Acceptability criteria must be set before patient results are reported, and control results documented ².

Good practice goes further ¹:

• Rotate positive controls through different alleles in genotyping assays, since running every genotype in every run often isn’t feasible.
• For quantitative assays, use controls at more than one level (low and high) across the measuring range.
• Run controls after maintenance, a new operator, or a new reagent lot.
• CAP requires cut-off verification for qualitative tests with every new lot or at least every six months; an external low-positive near the threshold satisfies it (MOL.34516) ¹.

Reading a run

Controls first, then samples.

Run-level failures invalidate the whole run:

• Positive control negative or out of range: reagent failure, instrument fault, or wrong setup. Nothing in the run can be reported negative.
• Negative control or NTC positive: contamination. Positives from the run can’t be trusted; investigate before repeating.

Sample-level interpretation with an internal control:

The last row is the one that matters most. A sample with neither signal is not a negative.

Worked example

A quantitative HIV-1 viral load run includes a high positive control at 5.2 log copies/mL (expected 5.0 ± 0.3), a low positive at 2.9 (expected 3.0 ± 0.3), and a negative control. All controls pass. Sample 14 reads target not detected, and its internal control Cq is 38 against a usual 30 ± 1.5.

The run is valid, but sample 14’s IC is delayed by 8 cycles, about a 250-fold drop: strong partial inhibition. A “not detected” here could be a false negative. Report it as invalid or repeat after re-extraction or dilution, per the lab’s procedure.

What to take from this

Each control tests something specific: positive (detection works), negative and NTC (no contamination), extraction (recovery), internal control (that sample worked), low-positive (sensitivity at the cutoff). CLIA requires two controls daily for molecular amplification and an inhibition control when inhibition causes false negatives. Read controls before samples, and never call a sample negative when its internal control also failed.