Hi Ka,

Could you explain a little bit more how the different concentrations relate to each other? Which one is higher? How many measurements and sites do you have? "Arterial" is peripheral or coronary artery? When are you taking your samples (how long after induction)? Are the patients undergoing extra-corporeal circulation? Is the heart beating or stopped?

Thanks,

Matthieu

Hi Matthieu, 

Thank you very much for your comment. 

The radial artery and coronary sinus blood samples were taken before the application of aortic cross-clamping during cardiopulmonary bypass. 

The arterial isoflurane concentration is higher than the venous one. 

i have total 9 patients... 6 of them were mean 82 min after isoflurane started, 3 of them were mean 4 min after isoflurane started while sevoflurane stopped.

- my judgements are:

Both arterial and venous isoflurane had not equilibrated due to ?

1. isoflurane solubility difference

2. circulatory factor as underlying cardiovascular disease/ coronary steal phenomenon of isoflurane

3. skewed by those 3 samples where uptake of isoflurane into tissue is high during induction time (mean 4 min after isoflurane started)

Kindly please advise further.

Do you observe the same arterio-venous gap in all your subjects? Is there a difference between your 2 groups (first 6 subjects and last 3 subjects)? If yes, this could simply mean that equilibrium was not reached for the second group.

In any case I think the most obvious explanation is tissular uptake of isoflurane. Isoflurane will diffuse in all tissues following the concentration (or partial pressure) gradient, so it makes sense that your arterial concentrations are higher. This is confirmed in this book, p190: http://bit.ly/1mGjr02

Hope this helps.

Matthieu

Matthieu, 

I run them in spss separately. 

Suprisingly, their differences were not significant in 6 and 3 patients. 

But, combined them, their difference were significant. 

What do you think about that?

KT

I guess the size effect. But the main problem is that using a t-test with only 9 subjects is a risky business and should be avoided. You may want to switch to non-parametric testing. For paired samples, the usual one to use would be the Wilcoxon signed-rank test (available in SPSS): https://en.wikipedia.org/wiki/Wilcoxon_signed-rank_test. Let me know if your difference is still significant.

The reasons of me using paired student T test because both arterial and CS isoflurane concentrations are normally distribution (Shapiro test,p>0.05).

I have tried the Wilcoxon signed rank as you suggested. Their differences were still significant.

.

Ok, this is good news. I would suggest putting both tests in your report. Do you have any other question?

No at the moment. 

thank you so much Matthieu. 

You're welcome!