When you aspirate the medium before trypsinization don't discard it and use the same media of particular plate to neutralize trypsin. Now, mix and centrifuge the cell suspension in the same media, This way you can account for the pre-detached cells floating in the medium. I hope this helps. Good Luck.

Trypan blue dye exclusion assay is usually used to assess the "health" of cells harvested for cell culture purposes, i.e. to determine proportions of viable and dead cells among harvested cells, which are intended for further subcultures or cryopreservation. In this context, there is no need to include cells floating in the medium, if one intends to subculture adherent cells, because only harvested adherent cells are normally used for subculturing. Situation is different in case of cells that grow partially adherent , or fully in suspension, where floating cells have to be included in the assay. If cell viability/death are determined as a response to some experimental intervention (e.g. drug treatment), free floating cells are a part of the population of exposed cells and need to be included to the evaluation. In this case, cells floating in medium should be collected by centrifugation and combined with adherent cells (detached by appropriate cell dissociation reagent) and analyzed together. However, the use of trypan blue assay to measure cell viability/death in response to experimental interventions is obsolete, cumbersome and not robust for this purpose. It tends to overestimate viability of cultures, in which the actual cell viability is less than 70%. Another drawback is requirement for cell densities of 1-3 million cells per 1 mL.. There are numerous modern and available cell viability or cytotoxicity assays that can better serve the purpose.

Altman SA, Randers L, Rao G. Comparison of trypan blue dye exclusion and fluorometric assays for mammalian cell viability determinations. Biotechnol Prog. 1993; 9: 671-4.

Hello Peter, your silly question (just kidding, no question is silly as you put it) deserves a silly answer:  I would simply aspirate and count the cells in your spent media (direcly or centrifuge to concentrate it down to an acceptable range to count using a hemocytometer) then add the result to the trypsinized cell count  to end up with total number of cells in that volume of culture (of course you are applying the trypan blue assay along the process to account for death and viable cells in your total cell count.)  Best of luck

Dear all, thank you for your valuable and thoughtful replies. 

We ourselves actually use another method for the total cell viability assay (as Roman mentioned - for drug treatment studies). We add an equal amount of trypsin to the well with complete medium, and the cells detach nicely. This method is even faster than the one suggested by Kushal, and is surely cheaper and much faster then any method requiring a centrifuge. 

And in cases when we need to seed an exact amount of the cells - we surely do not include the detached cells. 

But what bothers me a lot is that I've never seen a protocol which would even consider the pre-detached cells. It's like a conspiracy, you know))

I asked the question to see how other labs overcome the issue. I also asked it because too much things are not good at all in publishing policies regarding the cell viability assays. As Roman said, 70% viability may even be a good viability, although for our routinely used cells cells, HeLa and K562, this is not a good viability at all. In our studies, the control viability ranges between some 80-90% (HeLa; DMEM; gentamycin; 10% FBS). And I've seen too much papers showing their viability to be 95-99-100%. Bad practice I say. Especially when treatment group viability reaches good 130%, and the control viability bar (100%) is present on the chart. Which means that the authors had 70% control viability at best.

Regarding the false-positives and false-negatives, I can't really say that I can agree that only trypan blue overestimates viable cells, since there are papers showing that PI is misleading, too (see PMIDs 21763280, 20078512). And automated methods based on PI are technicaly more complicated, so there are numerous protocol steps that can lead to biases. 

Thank you all again for your answers and your hints. I hope our little discussion will help someone or will even drag some attention to the problems in current cell viability protocols and publication practices. 

Hi Peter: you are right, many studies have not included the free-floating cells to analysis and examined only those cells that remained adherent after cell treatments. This problem affects many different types of cell analysis (e.g. cell cycle analysis)  and not only cell viability assays. Depending on circumstances, this may introduce considerable error, as an important sub-population of cells is essentially ignored. Inclusion of free-floating cells to analysis is often warranted, but it has already happened to me that a reviewer did not understand why I have performed this step in my experiment....Your last comment that not only trypan blue overestimates viability (again, under some circumstances) is correct. For instance, PI staining detects cells with damaged membrane integrity and it fails to stain apoptotic cells, which have not yet reached the late stage of apoptosis when cell membrane integrity is lost.