Every dose starts with a count
In cell therapy, the dose is a number of cells: often a number of viable cells, sometimes per kilogram of patient body weight. The same is true in less regulated settings. Seeding a bioreactor, plating an assay or loading a single-cell sequencing run all depend on delivering a planned number of cells.
No one counts those cells directly. The dose is calculated from a measured concentration and viability, then delivered as a volume. Any error in the count therefore moves straight into the dose. Understanding how large that error is, and where it comes from, is the first step toward controlling it.
How a count becomes a dose
The calculation is simple. If the target is a fixed number of viable cells, the volume to deliver is:
Volume = target viable cells ÷ (total cell concentration × viability)
Because the measured concentration and viability sit in the denominator, a relative error in either one becomes roughly the same relative error in the delivered dose, in the opposite direction.
Consider a target of 1.0 × 10⁸ viable cells. A count reports 2.0 × 10⁶ viable cells/mL, so 50 mL is formulated. If the true concentration is 2.2 × 10⁶ viable cells/mL, the count was 10% low, and the patient or culture receives 1.1 × 10⁸ cells, 10% more than intended. An overestimate does the opposite and delivers too few cells.
Where the uncertainty comes from
Counting error has several independent sources. Some are random and shrink with replication; others are systematic and do not.
Random error
- Counting statistics. Cells in a counting chamber follow approximately a Poisson distribution, so the relative standard deviation from counting alone is about 1/√N, where N is the number of cells counted. Counting 100 cells gives about 10% uncertainty, 400 cells about 5%, and 1,000 cells about 3%.
- Sampling. Cells settle within minutes. A sample drawn from a poorly mixed tube may not represent the bulk suspension.
- Pipetting and dilution. Small volumes, viscous samples and serial dilutions each add their own error.
Systematic error (bias)
- Method differences. Trypan blue and acridine orange/propidium iodide (AO/PI) can report different viabilities for the same sample, especially for cryopreserved or apoptotic cells.
- Misclassification. Debris, red blood cells or platelets counted as cells, or clumps counted as single cells, shift the result consistently in one direction.
- Analysis settings. Size gates, focus and fluorescence thresholds change what the counter accepts as a cell.
Replicate counts reduce random error but leave bias untouched. A method can be very precise and still consistently wrong, which is why both precision and accuracy need attention.
Estimating the combined uncertainty
When error sources are independent, their relative uncertainties combine as the square root of the sum of squares. For example:
| Source | Relative uncertainty |
| Concentration (one count) | 8% |
| Viability (one count) | 3% |
| Formulation volume | 1% |
| Combined, single count | about 8.6% |
| Combined, mean of three independent counts | about 5.0% |
Two lessons follow. First, the largest single source dominates the total, so effort is best spent on the biggest contributor, which is usually concentration. Second, averaging independent replicate counts reduces the random part of each term by roughly the square root of the number of replicates. Three counts cut it by about 40%.
This calculation covers random error only. Any bias in the method adds directly on top of it, which is why method qualification matters as much as replication.
Don’t forget what happens after the count
The count is usually taken before final formulation. Washing, centrifugation, filtration and transfer through tubing all lose cells, and viability can fall while a product waits at room temperature or after thaw. A dose calculated from an early count can therefore overstate what is delivered. Counting the final formulated product, or at least characterizing typical losses between the count and administration, closes this gap.
Practical ways to reduce dose uncertainty
- Count more cells. Aim for several hundred to a thousand cells per measurement. Analyzing a larger sample volume, or adding chambers, reduces counting statistics error directly.
- Run independent replicates. Take each replicate as a separate sample from the mixed suspension, not repeated readings of the same loaded chamber.
- Mix thoroughly and sample immediately. Resuspend gently and draw the sample from the middle of the tube.
- Dilute into the linear range. Too dense a sample causes overlapping cells; too dilute a sample gives too few events.
- Use fluorescence for complex samples. AO/PI staining counts only nucleated cells and separates live from dead, removing much of the bias caused by debris and red blood cells.
- Qualify the method. ISO 20391-2 describes a dilution-series approach for checking that a counting method responds proportionally to concentration and for estimating its precision.
- Fix and record settings. Lock analysis protocols, record stain, dilution and instrument settings with each result, and use the same method across sites.
An automated cell counter supports most of these steps by removing operator judgment from focus and classification and by applying the same analysis settings every time. Fluorescence instruments such as the LUNA-FX7™ also analyze a larger sample volume per count than a standard hemocytometer, which increases the number of cells counted and lowers the statistical error of each measurement.
Conclusion
A cell dose is only as reliable as the count behind it. Counting statistics, sampling, pipetting and method bias all feed into the delivered number of cells, and losses after the count add more. Estimating each contribution, counting enough cells, replicating independently, qualifying the method and recounting the final product turn a single number into a dose with known uncertainty.
To learn more about getting reliable concentration and viability results, explore this automated cell counter overview from Logos Biosystems.
