Looking at the paper again:
- “Earlier chronotypes (the proverbial “morning larks”) are individuals who rise earlier and show peak alertness in the mid-morning hours, whereas later chronotypes (“night owls”) are late risers who exhibit peak alertness later in the day, often late into the evening.”
- “TIME was a large, prospective, pragmatic, decentralised, parallel-group study which assessed, in more than 21,000 hypertensive adults in the UK who took their usual prescribed antihypertensive medications at a single time of day (excluding night shift workers), whether there are differences in major cardiovascular outcomes between those randomised to take their usual antihypertensive medications in the morning (06:00–10:00) or in the evening (20:00–00:00).”
- “Individual chronotype was quantitatively assessed using the ultra-short version of the Munich ChronoType Questionnaire (μMCTQ)33 (Figure S1, Appendix p 16).”
- “662 (23.8%) participants in the morning dosing group and 673 (26.1%) in the evening group had MSFsc >30 min before the median and were considered ‘earlier’ chronotype. 750 (27.0%) participants in the morning dosing group and 664 (25.7%) in the evening group had MSFsc >30 min after the median and were considered ‘later’ chronotype. 1272 (45.8%) participants of the morning dosing group and 1170 (45.3%) of the evening dosing group had MSFsc within 30 min of the median and were considered an ‘intermediate’ chronotype.”
- “In line with the TIME study results, we found no effect of dosing time on cardiovascular outcomes in intermediate chronotypes (∼50% of the study population).”
- “Conversely, the risk of non-fatal stroke was not influenced by the interaction of chronotype with the administration time of antihypertensives, despite our observation that later chronotypes had a higher risk of stroke compared to earlier chronotypes, both in males and females.”
I’m rather a morning person, but according to their questionnaire, I’m actually “intermediate.” So calculate your MSFsc (mid-sleep time on free days corrected for sleep debt on workdays; ChatGPT can help 
) to see whether you fall in the “earlier” or “later” chronotype category.
Then, if you don’t have an intermediate chronotype, do you need to change your dosing schedule based on that? There’s probably no risk in doing so, so why not?
Still, I’m skeptical about the potential benefits. I’ll quote another paper: Lowering Nighttime Blood Pressure With Bedtime Dosing of Antihypertensive Medications: Controversies in Hypertension - Con Side of the Argument 2021
Why would taking BP-lowering medications, regardless of class, regardless of pharmacodynamics or pharmacokinetics, work better when taken at night? For example, enalapril, an angiotensin-converting enzyme inhibitor, and amlodipine, a calcium channel blocker, are commonly used antihypertensive agents in Spain. Enalapril has a half-life of ≈11 hours and reaches steady state in 4 days.34 Amlodipine has a half-life of ≈50 hours and reaches steady state in 1 to 2 weeks. Once at steady state, how could the timing of administration matter? In Table 3, we show the half-life of the commonest drugs used in the Hygia trial, and the data on nocturnal BP lowering with nighttime dosing from previous studies. Indeed, a comprehensive analysis of all time-of-day trials in humans showed that medications with a half-life of <8 hours were likely to have a time-of-day dosage effect, while those with a half-life of >15 hours were not. Apart from pharmacokinetic aspects, there indeed might be pharmacodynamic effects based on time-of-day. Paradoxically, however, administration of perindopril 4 mg at 9 in the morning had a more complete 24-hour BP-lowering effect than the same dose administered at 9 in the evening in a small mechanistic randomized clinical trial. A longer discussion of pharmacological aspects of time-of-day dosing in hypertension as it relates to Hygia has been made elsewhere for the interested reader. The Hygia research team has not addressed the mechanism(s) of time-of-day dosage effects of antihypertensive medications or how the effect of a drug with a ≈50-hour half-life could be influenced by timing of administration.
It turns out that the antihypertensive drugs that many people on this forum seem to use have long half-lives:
- Telmisartan: ~24h, longest acting ARB (vs 2 to 9 hours for losartan, the most used ARB)
- Amlodipine: 35–50h, longest acting CCB (vs less than 12 hours for most other CCBs)
- Indapamide SR: 14–24h (vs 6–15h for hydrochlorothiazide and just a few hours for loop diuretics and potassium-sparing diuretics). [Chlorthalidone is the longest-acting diuretic with a half-life of 40 to 60 hours.]
So, if you use one of the above drugs, is the chronotype-based dosing still relevant? The 2024 actually looked at this question, but only in a sub-study in the appendix!
The dosing time seems potentially critical to managing disease, such as hypertension, considering the short half-life of several medications used in daily practice.
This sub-study will address the following research questions: […] Are any of these effects dependent on the half-life of the antihypertensive drug?
This study is exploratory and is intended to identify hypothesis that can be tested by further research. The analyses will use an as-treated approach and the emphasis will be on parameter estimation rather than hypothesis testing.
Effects of drug half life will be explored using the entire TIME cohort.
The models will include: evening vs morning dosing (last known dosing time), adherence (fully adherent, switched dosing time but last know time same as the allocated time, switched dosing time and last known time different from the allocated time), mid-sleep time, social jetlag, self-reported chronotype and half-life of antihypertensive.
And yet… they didn’t publish the adjusted model’s results (not even in this appendix)! So, we don’t know if the findings still hold after adjusting for sex, age, smoking status, history of heart attack or stroke, number of antihypertensive drugs, and their mean half-life. That’s really intellectually dishonest, and it’s a disgrace that a paper can be published in The Lancet with only unadjusted event rates. My guess is that the adjusted model would not be statistically significant and that if you use a long-acting drug (and you should), this paper is irrelevant.
