From outside, a chemistry analyser is a big box that swallows tubes and spits out numbers. Inside, it’s a choreographed sequence of tiny, precise steps. Knowing them helps you understand why results sometimes go wrong.
The sample path
A tube’s journey usually looks something like this:
- Identification. A barcode links the tube to the patient and the tests requested.
- Preparation. Most chemistry tests use serum or plasma, so the tube is centrifuged first. In bigger labs, a dedicated, automated pre-analytical line may also decap, aliquot and sort tubes.
- Loading. The tube is presented to the analyser, which reads the barcode and works out which tests to run.
- Sampling. A probe dips into the tube and aspirates a small volume.
Analysers sense where the liquid surface is so the probe dips just far enough. Roche’s ISE method sheet, for example, warns that gel from a separator tube can coat the sample probe and interfere with “proper sample level detection”, and that fibrin microclots from poorly mixed plasma tubes can clog the system. Many analysers also have clot or aspiration-error detection, but how it works varies by platform.
Pipetting samples and reagents
Each test needs a tiny, exact volume of sample, often just a few microlitres, mixed with an exact volume of reagent in a reaction cuvette. Abbott’s Alinity c Creatinine (Enzymatic) assay, for instance, needs just 3.6 µL of serum or plasma for a single test, and uses two separate reagents.
Since results are calculated from the ratio of sample to reagent, any pipetting error goes straight into the answer. That’s one reason Westgard lists probe and syringe variation and bubbles among the causes of random error.
Carryover: the ghost of the last sample
Most analysers reuse the same probes for thousands of samples, washing them between each one. If the washing isn’t perfect, a little of the previous sample or reagent can be “carried over” into the next test.
- Sample carryover: a very high sample (say, a huge troponin or hCG) contaminates the next one and makes it read falsely high.
- Reagent carryover: a component of one test’s reagent interferes with the next test’s chemistry. In one published case, carryover of creatine kinase (CK) reagent intermittently caused falsely low cholesterol results. An extra reagent probe wash stopped the problem, and replacing the probes eliminated it (Kyle, 2010).
Manufacturers handle this with wash routines. Some method sheets make “special wash” steps mandatory when certain test combinations run together, and point to a carry-over evasion list. Labs check carryover when they evaluate a new method. CLSI EP10 includes a protocol that estimates carryover alongside other performance characteristics.
Incubation
Once sample and reagent are mixed, the reaction needs time and a stable temperature. The cuvettes sit in a temperature-controlled reaction area, usually at 37 °C. Temperature matters most for enzyme activity tests (like ALT or ALP), because reaction rate changes with temperature. Incorrect temperature settings are also a recognised cause of systematic error.
Some assays are read once the reaction has finished (an endpoint method). Others are read repeatedly while the reaction is happening, using the rate of change (a kinetic method). Many Jaffe creatinine methods are kinetic. In Abbott’s Jaffe-based Creatinine2 assay, creatinine reacts with picric acid at alkaline pH, and “the rate of increase in absorbance” at 500 nm is proportional to the creatinine concentration.
Photometry and Beer-Lambert, in plain words
Most chemistry tests are read by photometry. The analyser shines light of a chosen wavelength through the cuvette and measures how much gets through. The more of the coloured substance there is, the more light it absorbs.
The Beer-Lambert law puts numbers on that idea:
A = ε × l × c
- A is absorbance (how much light is soaked up)
- ε is how strongly that substance absorbs at that wavelength
- l is the path length through the cuvette
- c is the concentration
Keep the substance, wavelength and cuvette the same, and absorbance tracks concentration. Calibrators with known concentrations let the analyser turn absorbance into a result.
The caveats
That tidy equation comes with fine print. IUPAC’s definition assumes collimated, monochromatic light in a homogeneous medium, and real samples don’t always cooperate.
-
It’s only linear over a range. At high concentrations the relationship bends. That’s why every assay has a stated measuring range, and why high samples get diluted and rerun. The enzymatic creatinine assay, for example, has a serum and plasma measuring interval of 8.8–3536 µmol/L. Higher results are flagged and can be diluted automatically or by hand, and urine samples are automatically diluted 1:10.
-
Colour and cloudiness get in the way. Haemolysis (red), icterus (yellow-brown bilirubin) and lipaemia (turbid fat) all add their own absorbance or scatter light, so the analyser “sees” more or less than the reaction produced. Lipaemia interferes mainly through light scattering. Bilirubin interferes both spectrally and chemically.
-
We usually measure a product, not the analyte. Most analytes aren’t coloured, so the reagent turns them into something that is. The enzymatic creatinine method does it in a chain of steps:
- creatininase converts creatinine to creatine
- creatinase splits creatine into sarcosine and urea
- sarcosine oxidase converts sarcosine to glycine and formaldehyde, releasing hydrogen peroxide
- peroxidase uses that hydrogen peroxide to join 4-aminoantipyrine and ESPMT into a coloured quinoneimine dye
The change in absorbance at 548 nm is proportional to the creatinine concentration. Any chemistry that relies on a coloured product can be fooled by other substances that make the same colour. The Jaffe reaction is the best-known example, because proteins, ketoacids and some drugs also react with alkaline picrate. These are the “non-creatinine chromogens”.
-
Choosing a more specific chemistry helps, but it isn’t magic. Enzymatic creatinine methods are increasingly widely used, largely because their different reaction chemistry is less affected by interferents such as bilirubin and non-creatinine chromogens. In Sweden, for example, 90% of labs used non-compensated Jaffe methods in 1996–2001, but by 2014, 87% of hospital labs used enzymatic methods (Helmersson-Karlqvist et al). Uptake still varies by country: about 35% of UK labs were still using a compensated Jaffe assay in 2022 (Marrington and MacKenzie). In Australia, a 2012 position statement endorsed by the RCPA, AACB and Kidney Health Australia recommends enzymatic assays for children and young people under 18. It also notes that an enzymatic method may be preferable wherever interference is likely (Johnson et al). The evidence is nuanced, though. When seven commercial methods were compared against an isotope dilution mass spectrometry reference, interference was less frequent with the enzymatic methods, but no method was unaffected, and the details of each manufacturer’s version mattered (Greenberg et al). Enzymatic methods have their own interferents. For the enzymatic assay above, N-acetylcysteine and alpha-methyldopa can cause falsely low results, and IgM or IgG monoclonal proteins have very rarely caused falsely high ones. Other reported problems with enzymatic methods include high lithium heparin concentrations in under-filled tubes and haemolysis.
Many analysers also read at two wavelengths, a main wavelength and a secondary one, to help correct for background absorbance.
Key takeaways
- An analyser is a chain of steps: identify, prepare, sample, pipette, incubate, read. Each step can introduce error.
- Carryover is real, and it can come from reagents as well as samples.
- Photometry relies on Beer-Lambert, which only holds within a linear range and in a clear sample.
- Most methods measure a coloured reaction product, so lookalike substances and sample colour can bias results. More specific chemistries, like enzymatic creatinine, reduce that risk but don’t remove it.