Creatinine Quality-Control Shift After a Calibrator Lot Change
A 24-hour laboratory uses a single chemistry analyser for creatinine. The two-level quality control material has a level 1 mean of 0.90 mg/dL (SD 0.03) and a level 2 mean of 4.20 mg/dL (SD 0.12). Today level 1 is 0.99 mg/dL and level 2 is 4.28 mg/dL. Reviewing the Levey–Jennings chart, level 1 has been above the mean for six consecutive days: +1.0, +1.2, +1.4, +1.6, +1.8, and +2.0 SD. The calibrator lot was changed eight days ago; no reagent lot has changed. A patient with known chronic kidney disease has a creatinine of 3.1 mg/dL yesterday and 4.9 mg/dL today, with no change in weight, blood pressure, or urine output. The clinician calls the laboratory asking whether the rise is real. The laboratory uses an enzymatic creatinine method and has an independent quality-control material and a fresh vial of calibrator available.
1. Evaluate the QC data using rules. Is the run acceptable, and what type of error is present?
Read the worked answer
Level 1 today is 0.99 mg/dL against a mean of 0.90 and an SD of 0.03, which is exactly +3 SD. This violates the 1:3s rule and is a rejection criterion under Westgard rules. Level 2 is 4.28 mg/dL against 4.20 with an SD of 0.12, which is only +0.67 SD, so the second level alone would not reject the run. However, the six consecutive points above the mean for level 1 form a trend or shift. Six consecutive points on one side is not one of the classic Westgard rejection rules in every scheme, but it is a recognised warning of systematic error and should trigger investigation. The run is therefore not acceptable because of the 1:3s failure, and the trend indicates a systematic bias rather than a random error. Patient results should not be released until the problem is resolved. The pattern is more consistent with a calibration or reagent-related bias than with random imprecision, because the level 1 values have moved steadily in one direction.
2. What is the most likely cause, and what investigation should be performed before patient results are released?
Read the worked answer
The most likely cause is a calibration bias introduced when the calibrator lot was changed eight days ago. A fresh calibrator lot can differ in assigned value, reconstitution, storage, or stability. The laboratory should first check the calibrator certificate, expiry, storage, reconstitution volume, and the calibration record for errors. It should then recalibrate with a fresh vial of calibrator, run both QC levels, and confirm with the independent quality-control material. If the bias persists, the laboratory should check the reagent lot, the photometer lamp, the reaction temperature, the sample and reagent probes, and the linearity of the method. A split-sample comparison with another method or a reference laboratory can help confirm the patient results. Because the creatinine method is enzymatic, interfering substances such as high bilirubin, haemolysis, or glucose should also be considered, but the QC trend points to a system problem rather than a single patient interference. Patient results should be held until QC is acceptable and the cause is documented.
3. How should the patient’s creatinine result and the clinician’s query be handled? Include documentation and corrective action.
Read the worked answer
The patient’s creatinine should not be released until the QC problem is resolved. The sample should be repeated in duplicate on the same analyser, and the delta check should be reviewed: a rise from 3.1 to 4.9 mg/dL in one day is a large change and exceeds most delta-check limits. If the repeat confirms the result after recalibration and acceptable QC, the result can be reported with a comment that the sample was repeated because of a QC shift. If the repeat does not confirm the result, a new sample should be requested. The laboratory should notify the clinician of the investigation and should not simply state that the result is “real” without verification. Documentation should include the QC rule violated, the calibrator lot change, the recalibration, the independent QC result, the patient repeat result, and the time the clinician was informed. Corrective action should include reviewing calibrator lot validation, updating the QC acceptance criteria for lot changes, and training staff to recognise trends. The final report should state whether the patient result was verified or replaced.