Discrepant Potassium in a Haemolysed, Delayed Sample from a Patient with CKD
A 62-year-old man with type 2 diabetes and known chronic kidney disease (baseline creatinine 1.4 mg/dL, eGFR about 57 mL/min/1.73 m²) attends a district hospital with vomiting and reduced oral intake. At 08:00 a phlebotomist draws blood into a plain tube and a fluoride-oxalate tube from a difficult vein, then leaves the samples on the OPD counter. They reach the laboratory at 12:40, having been transported in a warm box without ice. The serum is visibly pink (haemolysis index 3+). Results: sodium 136 mmol/L, potassium 6.4 mmol/L, creatinine 2.7 mg/dL, glucose 3.1 mmol/L, bicarbonate 18 mmol/L. The analyser's potassium QC that morning was within 2 SD. The clinician phones asking whether to give intravenous calcium and insulin-dextrose immediately. A repeat sample collected at 13:30 from the other arm, separated within 30 minutes, gives potassium 5.2 mmol/L, creatinine 2.6 mg/dL and glucose 7.8 mmol/L.
1. Which of the first results are likely to be pre-analytically false, and how would you handle the clinician's request? State what you would report and what you would telephone.
Read the worked answer
Interpretation. The first potassium (6.4 mmol/L), glucose (3.1 mmol/L) and bicarbonate (18 mmol/L) are all affected by the 4.5-hour delay and visible haemolysis. Potassium rises as intracellular potassium leaks from red cells; glucose falls through glycolysis; bicarbonate falls as lactate is generated. Creatinine is much less affected by haemolysis, so 2.7 mg/dL is probably real, but it still needs confirmation. The morning QC was within 2 SD, which supports a pre-analytical rather than an analytical cause.
Action. Do not simply ignore the 6.4, but do not authorise calcium and insulin-dextrose on the basis of a haemolysed sample alone. Telephone the clinician and explain that the specimen was haemolysed and delayed, that the potassium is likely falsely high, and that a repeat is already available. Ask about ECG changes, urine output, and drugs such as ACE inhibitors or potassium-sparing diuretics. Report the repeat potassium 5.2, creatinine 2.6 and glucose 7.8 as the validated results. Report the first set only with a comment: 'Haemolysed/delayed specimen; potassium and glucose not reliable.' Document the critical-value call and the advice given. If the repeat had been ≥6.0 mmol/L with ECG changes, treatment should start immediately while the sample is repeated.
2. Calculate the CKD-EPI 2021 eGFR for the repeat creatinine and classify the CKD stage. Explain why the first creatinine should not be used for staging, and what the change from baseline suggests.
Read the worked answer
For a male, CKD-EPI 2021 (race-free) is: eGFR = 142 × min(Scr/0.9, 1)^−0.302 × max(Scr/0.9, 1)^−1.200 × 0.9938^age. With Scr 2.6 mg/dL: 2.6/0.9 = 2.89; 2.89^−1.2 = 0.280; 142 × 0.280 = 39.8; 0.9938^62 = 0.680; eGFR ≈ 27 mL/min/1.73 m², which is CKD G4. At the baseline creatinine of 1.4 mg/dL: 1.4/0.9 = 1.56; 1.56^−1.2 = 0.589; 142 × 0.589 × 0.680 ≈ 57 mL/min/1.73 m², CKD G3a.
The first creatinine should not be used for staging because the specimen was delayed and haemolysed and was not validated by the laboratory. The repeat is the defensible value. The change from about 57 to about 27 mL/min/1.73 m² over a short illness suggests acute kidney injury on top of CKD, not true progression of CKD, and requires clinical correlation: volume status, obstruction, sepsis, and nephrotoxins. A single acute value should not be used to label permanent stage change.
3. Design corrective actions for the pre-analytical failure in this district hospital, covering sample type, transport, separation and monitoring.
Read the worked answer
- Phlebotomy SOP: label at the bedside with two identifiers and the draw time; avoid prolonged tourniquets and vigorous fist clenching, which raise potassium; draw distal to any infusion site.
- Sample type and order of draw: serum or lithium heparin for chemistry; fluoride-oxalate for glucose; avoid EDTA contamination by following the correct order of draw. EDTA contamination causes a high potassium and low calcium picture.
- Transport and separation: separate serum or plasma within 2 hours at room temperature. Do not refrigerate whole blood for potassium, because cold storage promotes potassium leakage; do not leave tubes in sunlight or on a hot counter. If transport is unavoidable, centrifuge and separate at the peripheral site using a validated bench centrifuge.
- Receipt and rejection: record arrival time and haemolysis index; reject grossly haemolysed samples for potassium with a comment and request a repeat.
- Monitoring: audit monthly rejection rates, repeat rates and turnaround times; review critical-value calls; retrain phlebotomists and ward nurses; display a simple flow chart for sample handling. Review the incident as a quality-improvement activity, not as an individual blame exercise.