ELISA results are reported as optical density (OD) values, which represent how much colored signal developed in each well of the test plate. A higher OD value means more light was absorbed, which in most ELISA formats means more of your target substance is present. But turning that raw number into a meaningful answer depends on the type of ELISA, the controls included, and whether the result is qualitative or quantitative.
What the Raw Numbers Actually Measure
At the end of an ELISA, a chemical reaction produces a color change in each well of the plate. A plate reader (spectrophotometer) shines light through each well and measures how much light gets absorbed. This measurement is the optical density, or absorbance value, typically read at a wavelength of 450 nm. Wells with more target substance produce a stronger color and a higher OD reading. Wells with little or none stay pale and read close to zero.
By itself, a single OD number doesn’t tell you much. It only becomes useful when you compare it to the controls and standards run on the same plate. Every plate includes blank wells (no sample), negative controls (samples known to lack the target), and in quantitative assays, a series of standards at known concentrations. These reference points are what transform a raw absorbance reading into an interpretable result.
Qualitative Results: Positive or Negative
Many diagnostic ELISAs give a simple yes-or-no answer. HIV screening tests, for example, report whether a sample is positive or negative for antibodies. To make that call, the lab calculates a cutoff value from the negative controls on the plate. The most common approach is to take the average OD of the negative controls and add a multiple of the standard deviation. A typical formula looks like this: cutoff = mean of negatives + 3 × standard deviation. Any sample with an OD above that threshold is called positive; anything below it is negative.
Different labs and kit manufacturers use slightly different formulas. Some set the cutoff at two or three times the mean absorbance of the negative controls, skipping the standard deviation entirely. Others use statistical methods that account for the number of controls run. The specific formula matters because it determines how conservative the test is. A higher cutoff reduces false positives but may miss weakly positive samples. Kit instructions always specify which formula to use, and deviating from it changes the test’s validated sensitivity and specificity.
Some results fall in a gray zone, just above or just below the cutoff. These are typically reported as “equivocal” or “borderline,” and the standard practice is to retest the sample or follow up with a confirmatory test using a different method.
Quantitative Results: Using the Standard Curve
When you need to know exactly how much of a substance is in your sample, the ELISA includes a set of standards at known concentrations. These might range from, say, 15 pg/mL to 1,000 pg/mL. After the plate is read, you plot the OD values of these standards against their known concentrations to create a standard curve.
The standard curve is rarely a straight line. Most quantitative ELISAs use a four-parameter logistic (4PL) regression, which produces an S-shaped curve that accounts for the natural flattening at very low and very high concentrations. Software built into most plate readers handles this calculation automatically. Once the curve is generated, you find where each sample’s OD falls on the curve and read across to get the corresponding concentration. This process is called interpolation.
A good standard curve has an r² value of 0.98 or higher, meaning the data points fit the curve tightly. If the curve is poor, none of the sample concentrations derived from it are reliable. Common units on the final report include pg/mL, ng/mL, μg/mL, IU/mL (international units), or EU/mL (arbitrary ELISA units) for assays calibrated against international reference standards.
One critical rule: you can only trust concentrations that fall within the range of your standards. If a sample’s OD is higher than your highest standard or lower than your lowest, the result is outside the measurable range. Samples reading too high need to be diluted and rerun. If you diluted a sample before testing, remember to multiply the result by the dilution factor to get the true concentration.
Sandwich vs. Competitive: The Signal Goes Opposite Ways
Not all ELISAs work the same way, and the format determines how you interpret the color intensity.
- Sandwich ELISA: The most common format for measuring proteins like cytokines and hormones. The target substance gets “sandwiched” between two antibodies. More target means more color, so the signal increases with concentration. Higher OD = more analyte.
- Competitive ELISA: Used when the target molecule is small or when only one antibody is available. The sample competes with a labeled version of the target for binding sites. More target in your sample means less labeled target binds, producing less color. The signal decreases with increasing concentration. Higher OD = less analyte.
Misreading a competitive ELISA as if it were a sandwich ELISA (or vice versa) will invert your conclusions entirely. Always check the kit documentation for the assay format before interpreting results.
Common Problems That Distort Results
The Hook Effect
When a sample contains an extremely high concentration of the target, something counterintuitive can happen: the reading drops instead of climbing. This is called the hook effect (or prozone effect). It occurs because the overwhelming amount of target saturates both the capture and detection antibodies independently, preventing them from forming the proper “sandwich.” The result is a falsely low or even false-negative reading for a sample that actually has very high levels of the analyte. If you suspect a sample might have unusually high concentrations, running it at multiple dilutions will reveal the hook effect: the diluted sample will paradoxically give a higher reading than the undiluted one.
High Background and False Positives
A common source of error is elevated background signal across the plate. Proteins in serum and other biological fluids can stick nonspecifically to the plastic surface of the wells, producing color even when no target is present. This drives up the OD of all wells, including negative controls, and can push borderline samples over the positive threshold. Proper blocking steps during the assay and thorough washing between steps minimize this problem. If your blank wells are reading above 0.1 OD, something in the procedure likely needs optimization.
Edge Effects and Plate Variability
Wells along the edges of the plate tend to evaporate faster during incubation, which can produce slightly different readings than wells in the center. Running standards and critical samples away from the outermost rows and columns helps reduce this variability. Duplicate or triplicate wells for each sample also let you spot outliers and calculate a more reliable average.
Reading Results Step by Step
Putting it all together, here’s the practical sequence for interpreting an ELISA plate:
- Subtract the blank: Take the average OD of your blank wells (no sample, just reagents) and subtract it from every other well’s reading. This removes background noise from the plate and reagents themselves.
- Check your controls: Negative controls should read low (close to zero after blank subtraction). Positive controls should read at the expected level stated in the kit documentation. If either is off, the plate may not be reliable.
- For qualitative assays: Calculate the cutoff from your negative controls using the formula specified in the kit insert. Compare each sample’s OD to the cutoff.
- For quantitative assays: Plot your standard curve, confirm the fit is good (r² ≥ 0.98), and interpolate sample concentrations. Multiply by any dilution factor. Flag any samples outside the standard range for retesting.
- Average your replicates: If you ran duplicates or triplicates, average them. If one replicate is wildly different from the others (more than 20% coefficient of variation is a common threshold), investigate before including it.
The wavelength at which the plate is read also matters. Most kits use 450 nm as the primary wavelength, but some recommend a secondary reference wavelength (often 570 or 630 nm) to correct for optical imperfections in the plate. When a reference wavelength is used, the plate reader automatically subtracts it from the primary reading, giving a cleaner result.