LDH Cytotoxicity Assay: How to Measure Cell Damage

The LDH cytotoxicity assay measures cell damage by detecting a household enzyme that spills out of cells when their membranes break. Lactate dehydrogenase (LDH) sits inside virtually every cell in the body, and when a cell is injured or dies, LDH leaks into the surrounding fluid. The more LDH you find in the culture medium, the more cells have been damaged. This makes the assay one of the simplest and most widely used tools in toxicology and drug development for gauging how harmful a compound is to living cells, though its apparent simplicity hides real pitfalls that can skew your results.

Why LDH Leaks When Cells Are Damaged

LDH is a soluble enzyme that lives in the cytoplasm of nearly all mammalian cells. Under normal conditions the cell membrane keeps it inside. When something disrupts that membrane, whether a toxic compound, physical stress, or an immune attack, LDH escapes into the extracellular space.¹ The enzyme is remarkably stable once it’s outside the cell: activity drops by no more than about 5% per day in standard culture medium, which gives you a comfortable window to collect and measure your samples.²

This stability is a big part of why LDH became a go-to marker. Unlike some intracellular molecules that degrade quickly once exposed to culture conditions, LDH hangs around long enough that you don’t have to race to the plate reader the moment you pull samples. It also doesn’t require you to destroy the cells to take a measurement. You simply collect a small volume of the medium sitting above the cells, run the detection reaction, and read the result. The cells themselves stay in the well, which opens the door to running additional tests on the same plate afterward.

The Two-Step Detection Reaction

You can’t see LDH directly, so the assay uses a coupled enzymatic reaction to convert LDH activity into a color change you can measure with a standard plate reader. In the first step, LDH in your sample catalyzes the conversion of lactate to pyruvate, generating a molecule called NADH in the process. In the second step, that freshly produced NADH drives the reduction of a tetrazolium salt into a colored formazan product, with an electron-coupling agent acting as the go-between.¹ The more LDH that leaked out, the more NADH gets made, the more formazan accumulates, and the darker the color. You read absorbance at a specific wavelength (commonly around 490 nm, depending on the kit), and the intensity maps directly to the amount of cell damage in the well.

Commercial kits bundle the substrates, coupling reagents, and stop solutions together so the hands-on work is minimal. You typically add the reaction mixture to your collected supernatant, incubate for 10 to 30 minutes at room temperature, add the stop solution, and read. The whole detection phase from pipetting to plate reading takes well under an hour.

Essential Controls for Meaningful Data

An LDH reading by itself doesn’t tell you much. The number only becomes useful when you compare it to the right controls. Every experiment needs at least three reference points.

Spontaneous release: Untreated cells in medium, showing how much LDH leaks out under normal culture conditions with no added insult. This is your background level.
Maximum release: Cells lysed with a detergent (commonly Triton X-100 at around 1%) to blow open every membrane. This tells you the total LDH available in that well if every cell were dead.
Medium-only blank: Culture medium with no cells at all, which accounts for any background absorbance from the medium itself or from serum components.

Cytotoxicity is then calculated as a percentage: you subtract the spontaneous release from your experimental reading, divide by the difference between maximum release and spontaneous release, and multiply by 100. If the resulting number is 50%, roughly half the cells in that well were damaged. Getting the maximum-release control right matters more than people realize. If your lysis is incomplete, you’ll underestimate the total LDH pool and inflate your cytotoxicity percentages.

Where the Assay Falls Short

The LDH assay’s biggest blind spot is sensitivity. Head-to-head comparisons across multiple cell lines consistently show that LDH leakage detects toxicity at later timepoints or higher concentrations than some competing methods. In hepatoma cell lines exposed to cadmium chloride, for example, the neutral red uptake assay and the MTT assay flagged toxicity hours before the LDH assay registered a signal.³ A separate automated comparison using four mammalian cell lines and four standard toxicants confirmed a significant lack of sensitivity in LDH relative to WST-1 reduction and intracellular ATP assays.⁴ In some tissue models, intracellular LDH measurement was the least sensitive method tested, and extracellular LDH couldn’t be used at all because the test compound interfered directly with the enzyme.⁵

The reason is straightforward: LDH only escapes once the membrane is already compromised. Many toxic insults damage mitochondria, deplete energy stores, or trigger early apoptotic signaling well before the membrane actually breaks. Assays that measure metabolic activity (like MTT or ATP) can pick up those earlier events. The LDH assay, by contrast, is essentially asking “has the cell’s outer wall fallen apart yet?” and by that point the damage is often severe. This doesn’t make the assay bad. It means it’s measuring a specific stage of injury, late-stage membrane failure, rather than all forms of cellular harm.

Variability is another concern. In a zebrafish liver cell line study comparing colorimetric and fluorometric assays, the LDH assay showed the highest intra- and inter-assay variability and the lowest signal-to-noise ratio of the methods tested.⁶ And in multi-lab validation studies, different laboratories running the same protocol on the same cell type arrived at divergent cytotoxic concentrations and selected different top-dose limits, underscoring that lab-to-lab reproducibility remains a genuine challenge.

Serum in Culture Medium Can Quietly Inflate Your Readings

One interference that catches people off guard is fetal bovine serum (FBS), the standard supplement in most cell culture media. FBS itself contains LDH. Adding it at 5%, 10%, or 15% increased culture medium LDH content in a dose-dependent manner, meaning your baseline readings climb before your cells have done anything at all.⁷ If your spontaneous-release control uses the same serum-supplemented medium (and it should), the math mostly cancels out. But if you switch serum lots between experiments, or run one arm in serum-free conditions and another in full serum, you can introduce systematic bias that’s invisible in your final numbers.

Other common interferences include test compounds that absorb at the same wavelength as the formazan product (deeply colored drugs or nanoparticles), compounds that directly inhibit or activate LDH enzymatic activity, and high concentrations of reducing agents like ascorbic acid that can drive the tetrazolium reaction independently of LDH. Before trusting an LDH result, it’s worth running a cell-free control with the test compound and the detection reagents to check whether the compound alone generates a signal.

How LDH Compares to MTT and Other Viability Assays

The most common alternative to LDH is the MTT assay (and its relatives like MTS and WST-1), which measures mitochondrial metabolic activity rather than membrane integrity. Because these assays track different events, they often agree on whether a compound is toxic but can disagree sharply on the potency or on whether a protective agent is working. In neuronal cultures, for example, LDH release and MTT reduction gave similar readings for the overall level of injury from several apoptosis-inducing insults. But when researchers added anti-apoptotic drugs, the two assays told different stories: compounds that appeared protective by LDH showed no benefit or even worsened the signal by MTT.⁸ Trypan blue exclusion, a simple membrane-integrity stain counted under a microscope, gave results that tracked with LDH rather than MTT, supporting the idea that the discrepancy came from MTT measuring something downstream of mitochondrial metabolism that the protective agents didn’t fully rescue.

Metabolism-based assays also struggle to separate true cell killing from growth inhibition. A drug that simply stops cells from dividing without actually killing them will reduce the MTT or ATP signal in a growing culture, mimicking cytotoxicity. Researchers have noted that MTS- and ATP-based assays can overestimate a drug’s effectiveness when it alters cell metabolism but not cell viability, and that these assays fail to distinguish between cell death and growth inhibition.⁹ The LDH assay sidesteps this problem because LDH only appears in the medium when cells actually lyse. A growth-inhibited but intact cell releases no extra LDH. For drug screening, this distinction can matter enormously: you want to know if a compound is a cytostatic agent (stops growth) or a cytotoxic one (kills cells), and LDH gives you a cleaner read on the killing part.

In practice, the strongest approach is not to pick one assay and rely on it exclusively but to pair LDH with a metabolic assay. That way you catch both early metabolic disruption and late membrane damage, and disagreements between the two readouts actually tell you something useful about the mechanism of injury.

Multiplexing With Other Endpoints on the Same Plate

Because the LDH assay only requires a small sample of the culture supernatant, you can combine it with cell-based measurements in the same well without sacrificing either readout. Several groups have built multiplexed panels that run LDH alongside caspase-3/7 activation (an apoptosis marker), resazurin reduction (a viability metric), and DNA-binding dyes in a single 96-well plate.¹⁰ The workflow typically starts by removing an aliquot of medium for the LDH reaction, then adding the fluorescent viability reagent to the remaining cells, followed by the caspase substrate. Each endpoint hits a different emission wavelength, so the plate reader can separate them.

More elaborate panels add further enzyme leakage markers like aspartate aminotransferase (AST) and glutamate dehydrogenase (GLDH) alongside LDH and ATP from the same set of wells.¹¹ The rationale is that different cell types express these enzymes at different levels, and a compound might cause selective organ-level toxicity that one marker catches better than another. A compound that damages liver-like cells, for instance, might cause a disproportionate spike in AST relative to LDH. Running multiple markers in parallel gives you a richer picture of what the compound is doing and to which cellular compartments.

Multiplexing also saves time and reagents. Instead of plating four identical sets of cells for four separate single-endpoint assays, you plate once and read everything from the same wells. For large compound libraries where plate count is a bottleneck, this is more than a convenience; it can be the difference between a screening campaign being feasible or not.

Scaling Up for High-Throughput Screening

Drug discovery campaigns routinely test thousands of compounds, and the LDH assay adapts reasonably well to automated platforms. Miniaturization efforts have successfully moved the assay from a 150 µL reaction volume in 96-well plates down to 30 µL in 384-well format, cutting reagent costs by roughly 80% while maintaining comparable data quality.¹² Robotic liquid handlers manage the supernatant transfer and reagent addition steps that would otherwise be tedious at scale.

That said, the LDH assay’s inherent variability becomes more noticeable at high throughput. When your signal-to-noise ratio is lower than competing assays, as the zebrafish cell line comparison showed, small pipetting errors or edge effects on a 384-well plate can push borderline wells across a hit-call threshold. Quality-control metrics like the Z’-factor (a statistical measure of assay separation between positive and negative controls) need to be monitored closely. If Z’ values sag below acceptable levels, it’s usually a sign that the seeding density, lysis control, or incubation timing needs adjustment for that particular plate format.

Challenges With 3D Cultures and Organoids

The LDH assay was designed for flat, monolayer cell cultures where the medium sits in direct contact with every cell. In three-dimensional models like spheroids, organoids, and scaffold-based cultures, the geometry changes the game. Cells in the interior of a spheroid may lyse without their LDH efficiently diffusing to the surrounding medium. And because 3D cultures are often maintained for days or weeks, the LDH that does reach the medium has more time to degrade, even at a loss rate of only a few percent per day.

Researchers adapting the assay for 3D work have flagged two specific obstacles: inappropriate normalization (since cell number in a spheroid is harder to control than in a monolayer) and low LDH stability over the extended culture periods that 3D models require.¹³ Normalization is particularly tricky because 3D structures grow unevenly, and the standard maximum-release control (lyse everything and call it 100%) may not fully solubilize cells buried deep in a dense aggregate. Some groups have moved to enzymatic or mechanical dissociation of the 3D structure followed by a total-protein or DNA-content measurement to get a reliable denominator, but each extra step adds hands-on time and introduces its own variability.

Despite the challenges, the demand for 3D-compatible cytotoxicity readouts is growing as the field moves away from flat-culture models. Optimization strategies include more frequent medium sampling (rather than a single endpoint measurement), inclusion of a parallel spheroid set that gets fully disrupted at each timepoint for normalization, and switching from colorimetric to more sensitive fluorometric detection to improve the signal-to-noise ratio in these low-volume samples.

Reproducibility Across Laboratories

Even in well-resourced labs running validated protocols, LDH-based cytotoxicity numbers can vary strikingly from site to site. Interlaboratory datasets have documented cases where different groups using the same cell line, the same endpoint, and the same positive-control compound identified divergent cytotoxic concentrations and chose different maximum-dose limits for downstream testing. When slight differences in seeding density, passage number, serum lot, incubation humidity, or plate-reader calibration all nudge the signal in different directions, the cumulative effect is enough to shift a compound from “cytotoxic at this concentration” to “non-cytotoxic” depending on which lab ran the test.

This isn’t unique to LDH. Inter-lab variability plagues virtually all in vitro cytotoxicity assays, including the neutral red uptake assay used in OECD-validated phototoxicity testing, where coefficients of variation around 47% have been reported across participating laboratories. But LDH’s lower inherent sensitivity amplifies the problem, because modest absolute shifts in signal can push results across decision thresholds more easily than in a higher-dynamic-range assay. For anyone using LDH data to support regulatory submissions or go/no-go decisions in drug development, pairing it with at least one orthogonal endpoint is not just good science but practical risk management.

When LDH Is the Right Choice and When It Is Not

The assay earns its place when you specifically need to measure membrane damage. If you’re studying necrotic cell death (where the membrane is the primary casualty), evaluating physical disruption from nanoparticles or mechanical forces, or screening for compounds that cause outright cell lysis, LDH gives you a direct, quantitative readout of the event you actually care about. It’s also the right pick when you want to run further assays on the same cells afterward, since it’s non-destructive to the culture.

It’s less suited as a standalone assay for detecting early or subtle toxicity. Compounds that induce apoptosis, mitochondrial dysfunction, or oxidative stress without immediately compromising the membrane can register as non-toxic by LDH while other assays already show damage. In the cadmium-chloride study mentioned earlier, one cell line showed zero toxicity by LDH at all timepoints tested, while the neutral red assay flagged damage within three hours. If you rely on LDH alone in that scenario, you’d conclude the compound is safe when it isn’t.

For most practical purposes, the assay works best as one member of a small panel rather than a solo performer. Pairing LDH (membrane integrity) with an ATP or MTT assay (metabolic activity) and a caspase readout (apoptosis) covers the three major dimensions of cell injury with minimal extra effort, especially if multiplexed on a single plate. That combination gives a timeline of the damage: metabolic disruption first, then apoptotic signaling, then membrane collapse. Each assay fills the others’ blind spots, and the pattern across all three tells a more complete story than any single number could.