Haemolysis, Icterus & Lipaemia Interpreter
Haemolysis, Icterus & Lipaemia Interpreter
A haemolysis, icterus or lipaemia index is not a number you can look up a threshold for. The units are manufacturer-specific and do not transfer between platforms, and the concentration at which interference becomes clinically significant is different for every analyte — potassium and LDH fail at a degree of haemolysis that troponin tolerates. So this page asks for your laboratory’s own alert and hold indices for the analyte in question, and then tells you which way the result is wrong and what the comment should say.
Report, comment or reject?
Index + analyte + your limits → actionA sample with a haemolysis index of 60, as reported by the laboratory’s own analyser. The request is for potassium. The laboratory has verified an alert index of 50 and a hold index of 200 for potassium on this platform.
The three-threshold structure, and why the numbers are yours
Between the alert and hold indices → report with a comment naming the interferent and the direction of the error
At or above the hold index → suppress the analyte and request a fresh specimen
No universal value can be printed for either threshold, because the index units are manufacturer-specific and the threshold is analyte-specific.
- alert index
- the lowest tested concentration of haemoglobin, bilirubin or lipid at which the analyte’s result is altered to a clinically significant degree. It is the point at which a comment becomes necessary, and it is different for every analyte on the same platform — LDH and potassium reach it at a fraction of the haemolysis that troponin tolerates
- hold index
- the index above which the result carries no useful clinical information and should not be reported at all. Reporting above this level with a caveat is worse than not reporting, because a number gets acted on and a comment often does not
- why the units do not transfer
- index results are produced by different sample dilutions, different dilution factors, different measured wavelength pairs and different correction equations on different platforms, and are variously reported semi-quantitatively or quantitatively, in mg/dL, in micromol/L or as an arbitrary scale. An index of 50 on one analyser is not the same sample as an index of 50 on another, and neither an index nor a threshold can be carried across
- the three mechanisms
- RELEASE — cell contents enter the plasma, raising analytes concentrated inside red cells (potassium, LDH, AST, iron) and lowering those that are not (sodium, chloride). SPECTRAL — haemoglobin absorbs roughly 320 to 580 nm, bilirubin roughly 400 to 540 nm with a peak near 460, and lipid scatters across the spectrum with more effect at short wavelengths. VOLUME AND CHEMISTRY — lipid displaces aqueous volume, biasing indirect ISE electrolytes down; bilirubin consumes hydrogen peroxide in peroxidase-linked assays; haemoglobin’s pseudoperoxidase activity degrades bilirubin, and released proteases degrade troponin T and insulin
- how to set your own thresholds
- spike a pooled patient sample with mechanically lysed red cells, with bilirubin, or with a lipid emulsion; measure the analyte across a range of index values; and find the index at which the change exceeds what you are prepared to tolerate. Set that tolerance from the analyte’s allowable total error or from the size of change that alters a clinical decision, not from a generic ten per cent
Worked example
A sample with a haemolysis index of 60, as reported by the laboratory's own analyser. The request is for potassium. The laboratory has verified an alert index of 50 and a hold index of 200 for potassium on this platform.
Both thresholds exist for this analyte on this platform, so a decision can be made
The index of 60 is at or above the alert index of 50 → the result is altered to a clinically significant degree
It is below the hold index of 200 → the result still carries usable information
So the action is report with a comment, and the comment has to say which way the value is wrong
Mechanism: intracellular potassium is roughly thirty times the plasma concentration, so lysis adds potassium and the reported value is higher than the patient's
Change only the analyte, to cardiac troponin, and the same index of 60 is almost certainly below troponin's alert index on the same platform — the sample would be reportable without comment. Change only the platform, and the index of 60 has no meaning at all until it is re-expressed on that platform's scale
Direction of interference by index and mechanism
| Index | Raises | Lowers | Direction depends on the assay |
|---|---|---|---|
| Haemolysis | Potassium, AST, LDH, iron, lipase, GGT | Sodium, chloride, alkaline phosphatase, bilirubin; insulin and troponin T by proteolysis | Creatinine, and any assay reading 320-580 nm |
| Icterus | — | Phosphate, Jaffe and enzymatic creatinine, cholesterol, triglycerides, urate | Any assay reading 400-540 nm; conjugated and unconjugated bilirubin can act oppositely on the same assay |
| Lipaemia | — | Sodium, potassium and chloride by INDIRECT ISE only | Any assay reading near 340 nm; steroid hormones and hydrophobic drugs by partitioning into the lipid phase |
What European laboratories actually do
| Practice | Proportion |
|---|---|
| Use automated HIL indices only | 43% (493 of 1,160) |
| Visual inspection only | 30% (348 of 1,160) |
| Both | 28% (319 of 1,160) |
| Adopt the manufacturer’s cut-offs | 54% (624) |
| Verify those cut-offs | 39% |
| Derive their own cut-offs | 7.2% (83) |
| Suppress only the affected analytes, with a comment | 77% |
| Release everything with a general HIL note | 14% |
| Reject the whole haemolysed sample | 7.4% |
| Release everything with no comment | 1.2% |
Why this page will not print a cut-off
The most useful thing a page about serum indices can do is refuse to give you a number. Haemolysis, icterus and lipaemia indices are measured by reading the sample’s absorbance at selected wavelengths, but the dilution used, the dilution factor, the wavelength pairs and the correction equations all differ between manufacturers, and the results are variously reported as a semi-quantitative grade, as an estimated haemoglobin concentration in mg/dL, as micromoles per litre, or on an arbitrary scale. An index of 50 on one analyser and an index of 50 on another are not the same sample. A threshold carried from one platform’s documentation to another’s results is not a cautious approximation; it is a number with no defined meaning.
The second reason is analyte specificity. Even on one platform, the degree of interference that matters is entirely different from one test to the next. Lactate dehydrogenase and potassium are shifted appreciably by lysis that leaves most of the chemistry panel untouched, because both are concentrated inside the red cell — potassium at roughly thirty times its plasma concentration. Some high-sensitivity troponin assays tolerate a haemolysis index at which a potassium would have to be suppressed. A single sample-level threshold therefore either suppresses results that were perfectly reportable or releases results that were not, and usually both at once.
What a page can usefully give you is the mechanism, because the mechanism predicts the direction and the direction is what the comment has to say. There are three. Release: lysis puts cell contents into the plasma, raising analytes that are concentrated inside the cell and lowering those that are not, so potassium, AST, LDH and iron go up while sodium, chloride, alkaline phosphatase and bilirubin go down. Spectral: free haemoglobin absorbs across roughly 320 to 580 nm, bilirubin from about 400 to 540 with a peak near 460, and lipid scatters light across the spectrum with more effect at shorter wavelengths, so any assay reading in those regions is susceptible in a direction set by its wavelength pair. Chemical and volumetric: bilirubin consumes hydrogen peroxide in peroxidase-linked assays, haemoglobin’s pseudoperoxidase activity degrades bilirubin, released proteases degrade troponin T and insulin, and lipid displaces aqueous volume so that indirect ion-selective electrodes under-read electrolytes.
Setting your own thresholds is not difficult and it is the only defensible option. Spike a pooled sample with mechanically lysed cells, with bilirubin or with a lipid emulsion, measure the analyte across a range of index values, and find where the change exceeds what you are prepared to tolerate — with the tolerance taken from the analyte’s allowable total error or from the size of change that alters a decision, rather than from a generic ten per cent. Record two thresholds: an alert index above which the result goes out with a comment, and a hold index above which it does not go out at all. A European survey of laboratory practice found that most laboratories adopt the manufacturer’s figures and well under half verify them, which is the gap this page is trying to name.
Frequently asked questions
What haemolysis index should I use as a cut-off?
There is no answer to that question that does not name your platform and your analyte. Index units are manufacturer-specific — different dilutions, wavelength pairs and correction equations — so a value from one analyser cannot be read on another’s scale, and the degree of interference that matters differs by an order of magnitude between analytes on the same platform. Take the manufacturer’s stated limit for each analyte as a starting point, verify it by spiking, and record an alert and a hold index for each test.
Should I reject the whole sample or just suppress the affected tests?
Suppress the affected analytes and report the rest. Rejecting the whole sample discards results that were never affected and delays care for no benefit. A European survey of laboratory practice found 77% of laboratories suppress only the affected parameters with an explanatory comment, while 7.4% reject the entire haemolysed sample. Where the index is above your hold threshold for an analyte, suppress that analyte rather than releasing it with a caveat — a number on a report gets acted on, and the caveat frequently does not.
How do I tell in vivo haemolysis from in vitro haemolysis, and does it matter?
It matters a great deal, because recollection fixes one and not the other. In vitro haemolysis arises from the collection or the transport — difficult venepuncture, a small-bore needle, vigorous aspiration, collection from a line, rough or delayed transport — and a properly taken repeat sample will be clean. In vivo haemolysis reproduces in every sample, and repeated failed recollections in such a patient are a clinical finding being mistaken for a preanalytical one. Look at the clinical picture, the reticulocyte count, the haptoglobin and the blood film, and note that LDH cannot arbitrate because it is both the marker and the artefact.
Why is my sodium low in a lipaemic sample, and how do I get the right answer?
Lipoprotein particles occupy volume in the non-aqueous phase. An indirect ion-selective electrode dilutes the sample and reports per litre of total sample, so the electrolyte appears diluted by lipid it never occupied — pseudohyponatraemia. A direct ISE, as used on blood gas analysers and point-of-care devices, measures activity in the aqueous phase without dilution and is unaffected, so running the sample on one gives the right answer. Ultracentrifugation or a lipid-clearing reagent is the alternative.
My analyser applies an automatic correction for the index. Can I just report the corrected result?
Only if you know the correction is enabled for that assay, know what it does, and have verified it. A corrected and an uncorrected result are different numbers, and only one should be leaving the laboratory. Corrections are also assay-specific and are usually validated over a limited index range, so applying one beyond that range is an extrapolation. Where a correction exists and is validated it is preferable to suppression; where its provenance is unclear, treat the result as uncorrected.
Related calculators
References
- Farrell CL, Carter AC. Serum indices: managing assay interference. Ann Clin Biochem. 2016;53(Pt 5):527-538.
- Cadamuro J, Simundic AM, von Meyer A, et al. European survey on preanalytical sample handling — Part 2: Practices of European laboratories on monitoring and processing haemolytic, icteric and lipemic samples. On behalf of the EFLM Working Group for the Preanalytical Phase (WG-PRE). Biochem Med (Zagreb). 2019;29(2):020705.
- Lippi G, Cadamuro J, von Meyer A, Simundic AM. Practical recommendations for managing hemolyzed samples in clinical chemistry testing. Clin Chem Lab Med. 2018;56(5):718-727.
- Lippi G, Cadamuro J. To report or not to report: a proposal on how to deal with altered test results in hemolytic samples. Clin Chem Lab Med. 2017;55(8):1112-1114.
- Monneret D, Mestari F, Atlan G, et al. Hemolysis indexes for biochemical tests and immunoassays on Roche analyzers: determination of allowable interference limits according to different calculation methods. Scand J Clin Lab Invest. 2015;75(2):162-169.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.