There are two groups of patients with equal number of patients on betablockers. Will this effect the difference of heart rate variability in the two groups if both are equally on betablockers? I believe that the difference will not be effected although the absolute HRV will be effected.
Frequency domain parameters are not being tested and only time domain is being analysed.
If you can comment on this with some evidence linked please it will be very helpful.
Hi,
There must not be a considerable difference attributed to beta-blockers if the two groups are the same regarding their drug history (type of drug, dosage, duration of treatment):
https://www.ncbi.nlm.nih.gov/pubmed/8176095
also consider other patients' characteristics like age, gender, BMI, physical activity, smoking, etc. (if such data are available) which might be different between the two groups and could explain the difference in HRV.
I believe the difference may be altered, in addition to absolute HRV. Let's assume your groups have baseline HRV x and y (x>y). Let's assume that beta-blockers increase HRV by, say, 50%. If 0% patients in both groups are taking them, the difference in HRV between groups is is x-y. If all patients are taking them, their HRV difference would be 1.5*x - 1.5*y, = 1.5*(x-y) making the difference 50% larger than if nobody was taking beta-blockers. For intermediate values of beta-blocker usage, you'll get respectively intermediate increase in the total difference.
Also, it might be important to know what the patients on beta-blockers are; it would be good to have a similar composition of diagnoses/degrees of sympathetic dysfunction. E.g., in congestive heart failure, where strong sympathoexcitation translates into low response to beta-adrenergic stimulation via downregulation of beta-adrenoreceptors; beta-blockers were shown to >>increase
Dear Jakub,
thank you for the explanation. I agree that the HRV might increase. The structure of the patients is normal LV, AF history (but in sinus rhythm when being analysed), and hypertension (the reason of them taking betablockers).
However, looking at your explanation - if I was to use real values that that significance between two groups and then if I was to increase the factor by 50% and retest using T.Test (for example), the statistical result remains the same (p value). Although P value is not the only thing to look at, the mean, SD would be altered but not the significance between the two groups. I hope this make sense. I have attached an example.
Please do correct me if I am understanding this incorrectly.
Dear Habib,
agreed, there should be no issue with a t-test (precisely because both mean an SD increase at the same time). If the effect of beta-blockers was very strong, it might make the distribution of HRV values non-normal, making it not that suited for a t-test, but it might not be too bad in practise, I suppose. And it shouldn't affect nonparametric tests anyway, I think - it's just the absolute difference that might get affected.
Thank you Ali for your answer. All the patients are randomised and hence everything should ideally be balanced.
Habib, for randomization to ideally balance the study groups the sample size should also be ideal. If it is, and no other difference between the two groups exists, then the difference in HRV cannot be attributed to the measured factors and can be attributed to the actual intervention/condition which is different between the two groups.
logarithmic transformation can be helpful for frequency-domain indexes.
Habib, it's missing here what time-domain indexes are different between the two groups? is it only one index or most indexes are different?
the HRV is being derived from SDARR surrogate only. That is the only way it is being documented and is different
So I guess it's from 24-hour holter monitoring?
to my knowledge SDARR (or SDANN) represents variation in HR in cycles longer than 5 minutes (if predefined as 5-min samples), and its meaning is very different from RMSSD or HF-HRV. This must be considered in interpretation of the results !
standard deviation of the 5 minute medians for the day is the measurement that I am being told. What other way can I measure or interpret HRV from this dataset?
if it is 24-h holter monitoring and you also have access to the RR-intervals, then you can extract all the usual time-domain and frequency domain indexes. Usually, the 24-h is splited to 5-min intervals. The reason is that, first you can exclude those intervals with more than 5-15% artifacts since artifacts can considerably affects the HRV results. second you can have better temporal resolution since HRV has diurnal variation which may be important with regards to the interventions or comparing patients with controls.
but if you don't have access to the RR-intervals but still have medians of the 5-min intervals, you may be able to increase your temporal resolution, for example by looking at the SD of the medians in each hour, or during the day and night (sleep). I don't know if this is a standard way to do that but who knows, maybe you can get some interesting findings ! but I would suggest also to not do it blindly, only if it's relevant to your study aims and hypotheses.
For the fetus, atenolol blocks accelerations but labetalol does not.
Michelle L Murray, PhD, RNC-OB
Author of 2 books related to fetal monitoring