Cardiovascular outcomes in adults with hypertension with evening versus morning dosing of usual antihypertensives in the UK (TIME study): a prospective, randomised, open-label, blinded-endpoint clinical trial

The researchers tested whether the *time of day* you take your blood pressure medication changes your risk of major cardiovascular events. The intervention was simply switching the timing of all existing antihypertensive medications to the evening (between 8 PM and midnight). The comparator was taking the exact same medications in the morning (between 6 AM and 10 AM). The primary outcome was a composite of: death from cardiovascular causes, hospitalisation for a non-fatal heart attack (myocardial infarction), or hospitalisation for a non-fatal stroke. Secondary outcomes included each component separately, plus all-cause mortality, hospitalisation for heart failure, and hospitalisation for unstable angina. The study also tracked blood pressure control (though this was not the primary focus) and safety events like falls, fractures, and fainting.

**Sample size:** 21,104 adults with hypertension (10,503 assigned to evening dosing, 10,601 to morning dosing)

**Population:** Adults aged 18 years or older living in the UK, with a diagnosis of hypertension, currently taking at least one antihypertensive medication. They were recruited through primary care practices and via media campaigns.

**Setting:** A "decentralised" trial — participants were recruited, enrolled, and followed up remotely (by post, phone, or online) without requiring in-person clinic visits. This was a pragmatic, real-world study conducted across the UK.

**Demographics:** Mean age 65.1 years (SD 9.3). 57.5% men, 42.5% women. 90.5% White, 0.5% Black (ethnicity not reported for 7.8%). 13.0% had a history of prior cardiovascular disease. Participants were taking a wide range of antihypertensive drugs (ACE inhibitors, angiotensin receptor blockers, calcium channel blockers, beta-blockers, diuretics, and combinations).

**Exclusion criteria:** People who worked night shifts (because their "morning" and "evening" would be different), those with a life expectancy less than 6 months, and those who were already taking their medications at a specific time for a clinical reason (e.g., to avoid side effects).

**Primary endpoint:** A composite of vascular death (death from cardiovascular causes, confirmed by death certificate or medical records), hospitalisation for non-fatal myocardial infarction (heart attack), or hospitalisation for non-fatal stroke. These were identified through two methods: (1) participant self-report via annual postal or online questionnaires, and (2) electronic record linkage to NHS England, NHS Scotland, and NHS Wales hospital admission and death registry datasets.

**Endpoint adjudication:** All potential primary endpoint events were reviewed by a clinical endpoints committee who were *masked* (blinded) to the participant's treatment allocation (morning vs. evening). They applied standardised diagnostic criteria (e.g., universal definition of myocardial infarction, WHO criteria for stroke) to confirm or reject each event.

**Blood pressure measurement:** Participants were asked to provide self-measured home blood pressure readings (using their own monitors) at baseline and at follow-up timepoints. However, this was not a primary outcome and the data were incomplete (only ~60% of participants provided usable BP data at follow-up).

**Safety monitoring:** Participants were asked about adverse events (falls, fractures, fainting, dizziness) on annual questionnaires. Serious adverse events were also captured through record linkage.

**Adherence monitoring:** Participants were asked to self-report whether they were taking their medications at the assigned time on annual questionnaires. This was a pragmatic trial, so adherence was not strictly enforced or monitored with pill counts or electronic monitors.

**Study design:** Prospective, randomised, open-label, blinded-endpoint (PROBE design) parallel-group trial. This is a pragmatic, real-world trial design that sacrifices some internal validity (no placebo, no blinding of participants or clinicians) to gain external validity (generaliseability to real-world practice).

**Randomisation:** Participants were randomly assigned 1:1 to morning or evening dosing. The randomisation was simple — no stratification, minimisation, or restriction was used. This means that by chance, the two groups might differ slightly in baseline characteristics (e.g., age, sex, medication types), but with over 10,000 participants per group, the groups were well-balanced. The randomisation sequence was computer-generated and concealed from the study team until allocation.

**Blinding:**

*Participants and clinicians:* NOT blinded. Participants knew whether they were taking their pills in the morning or evening. This is a major limitation because it could influence behaviour (e.g., people who forget evening doses might be different from those who forget morning doses).

*Outcome assessors:* BLINDED. The clinical endpoints committee that adjudicated heart attacks, strokes, and deaths did not know which group each participant was in. This is the "blinded-endpoint" part of the design — it prevents bias in deciding whether an event "counts" as a primary outcome.

**Duration:**

Recruitment: December 17, 2011 to June 5, 2018 (over 6.5 years)

Follow-up: Until March 31, 2021

Median follow-up: 5.2 years (IQR 4.9–5.7 years). This is a long follow-up for a hypertension trial, which is a strength.

Total follow-up: Over 100,000 patient-years of data.

**Statistical approach:** The primary analysis was intention-to-treat (ITT) — all participants were analysed in the group they were assigned to, regardless of whether they actually took their medications at that time. This is the gold standard for RCTs because it preserves the benefits of randomisation. The primary endpoint was analysed using Cox proportional hazards models, with time to first event as the outcome. The hazard ratio (HR) with 95% confidence intervals (CI) was reported. A pre-specified non-inferiority margin was not set; the study was designed as a superiority trial (to see if evening dosing was *better* than morning dosing).

**What this design can prove:**

This design can prove that, in a real-world setting, advising patients to take their medications in the evening does not lead to a clinically meaningful reduction in cardiovascular events compared to morning dosing over ~5 years.

The large sample size and long follow-up mean the study had high statistical power to detect even small differences (e.g., a 10–15% relative risk reduction).

The blinded endpoint adjudication ensures that the classification of heart attacks, strokes, and deaths was not biased by knowledge of the dosing time.

**What this design cannot prove:**

It cannot prove that timing *never* matters for specific subgroups (e.g., people with nocturnal hypertension, people on specific drug classes like ACE inhibitors vs. calcium channel blockers). Subgroup analyses were exploratory and underpowered.

It cannot prove that timing does not matter for other outcomes not measured, such as 24-hour blood pressure profiles, sleep quality, or kidney function.

Because it was open-label, it cannot rule out that differences in adherence between groups (e.g., people in the evening group forgetting doses more often) masked a true benefit of evening dosing. The withdrawal rate was slightly higher in the evening group (5.0% vs. 3.0%), which could indicate lower acceptability.

It cannot prove that timing does not matter for people with specific chronotypes (e.g., "night owls" vs. "morning larks") because chronotype was not assessed.

**Major methodological weaknesses:**

1. **Open-label design:** Participants and their doctors knew the dosing time. This could affect adherence (people might be more or less likely to remember evening doses) and co-interventions (doctors might adjust other medications differently based on dosing time).

2. **Self-reported adherence:** Adherence was assessed only by annual questionnaire, which is prone to recall bias and social desirability bias. No objective adherence measures (pill counts, electronic monitors, drug levels) were used.

3. **Incomplete blood pressure data:** Only ~60% of participants provided follow-up BP readings, and these were self-measured with home monitors of unknown calibration. This means we cannot confirm that the two groups actually had different BP profiles (e.g., evening dosing might lower nocturnal BP but not clinic BP).

4. **Low event rate:** The primary endpoint occurred in only 3.4–3.7% of participants over 5 years. This is lower than expected, which reduces statistical power for subgroup analyses and means the study was actually underpowered to detect very small differences (e.g., <10% relative risk reduction).

5. **Lack of washout or run-in period:** Participants were already on antihypertensives before randomisation. They were simply asked to switch their dosing time. This means that any acute effects of changing dosing time (e.g., temporary BP fluctuations) were not captured.

**Primary endpoint (composite of vascular death, non-fatal MI, non-fatal stroke):**

- Evening group: 362 events (3.4% of participants), incidence rate 0.69 per 100 patient-years (95% CI 0.62–0.76)

- Morning group: 390 events (3.7% of participants), incidence rate 0.72 per 100 patient-years (95% CI 0.65–0.79)

- Unadjusted hazard ratio: 0.95 (95% CI 0.83–1.10), p = 0.53

- **Conclusion:** No statistically significant difference. The 95% CI includes 1.0 (no effect) and ranges from a 17% reduction to a 10% increase in risk with evening dosing.

**Secondary outcomes (all non-significant):**

- Vascular death alone: HR 0.97 (95% CI 0.75–1.25)

- Non-fatal MI alone: HR 0.93 (95% CI 0.74–1.17)

- Non-fatal stroke alone: HR 0.96 (95% CI 0.76–1.21)

- All-cause mortality: HR 0.94 (95% CI 0.82–1.08)

- Hospitalisation for heart failure: HR 0.93 (95% CI 0.68–1.26)

- Hospitalisation for unstable angina: HR 1.10 (95% CI 0.80–1.52)

**Subgroup analyses (all exploratory, no significant interactions):**

- No difference by age, sex, BMI, smoking status, diabetes status, prior CVD, number of antihypertensives, or specific drug class (ACE inhibitors, ARBs, calcium channel blockers, beta-blockers, diuretics).

- The p-value for interaction was >0.05 for all subgroups tested, meaning there was no statistical evidence that any subgroup benefited more from evening dosing.

**Adherence:**

- At 1 year, 81.5% of the evening group and 82.7% of the morning group reported taking their medications at the assigned time.

- At 5 years, adherence was ~70% in both groups. This means ~30% of participants were not taking their medications at the assigned time by the end of the study, which dilutes any true effect of timing.

**Safety:**

- No significant differences in falls, fractures, fainting, or dizziness between groups.

- No safety concerns identified.

The difference between groups was tiny and statistically indistinguishable from zero. The absolute risk reduction for the primary endpoint was 0.3 percentage points (3.7% in the morning group vs. 3.4% in the evening group). This means that if 1,000 people took their medications in the evening instead of the morning for 5 years, you would expect about 3 fewer cardiovascular events (heart attacks, strokes, or cardiovascular deaths). However, the confidence interval is wide enough that the true effect could range from 17 fewer events to 10 more events. In plain terms: the data are consistent with evening dosing having no meaningful effect on cardiovascular outcomes. The number needed to treat (NNT) for 5 years to prevent one event is approximately 333 — but this is not statistically significant, so the NNT could be infinite (no benefit) or even negative (harm).

**Acknowledged by authors:**

Open-label design (participants and clinicians knew the dosing time)

Self-reported adherence (no objective measures)

Incomplete blood pressure data (only ~60% provided follow-up BP readings)

Low event rate (fewer events than anticipated, reducing power for subgroups)

Predominantly White population (90.5%), limiting generaliseability to other ethnic groups

Decentralised design meant no standardised BP measurement protocol

**Critical reader observations:**

**No 24-hour ambulatory blood pressure monitoring (ABPM):** The study did not measure whether evening dosing actually lowered nocturnal blood pressure more than morning dosing. This is a critical mechanistic gap — if evening dosing does not actually change the 24-hour BP profile in this real-world setting, then the null result is expected.

**Adherence decay:** By 5 years, ~30% of participants were not taking their medications at the assigned time. This is a massive dilution of the intervention. If the true effect of evening dosing is small (e.g., 10–15% risk reduction), this level of non-adherence could completely mask it.

**No chronotype assessment:** The study did not measure whether participants were "morning types" or "evening types." Chronotype could modify the effect of dosing time on BP and cardiovascular risk. For example, evening dosing might be more effective in "morning types" who have a different circadian BP rhythm.

**Drug class heterogeneity:** The study included all classes of antihypertensives. Some drugs (e.g., ACE inhibitors, ARBs) have a longer half-life and might be less affected by dosing time, while others (e.g., short-acting calcium channel blockers) might be more sensitive. The subgroup analysis by drug class was underpowered.

**No placebo control:** This was a comparison of two active dosing schedules, not a placebo-controlled trial. The study cannot answer whether *any* dosing schedule is better than no treatment (which is already well-established).

**Industry funding:** Funded by the British Heart Foundation (a charity), so no pharmaceutical industry influence. This is a strength.

For someone running their own n=1 experiment:

**What to test:**

Test whether taking your blood pressure medication in the evening (8 PM–midnight) vs. morning (6 AM–10 AM) changes your blood pressure readings, how you feel, or your side effect profile. Do NOT test this as a way to reduce heart attack or stroke risk — the TIME study shows that for cardiovascular outcomes, timing does not matter at the population level. However, you might find individual benefits in terms of side effects (e.g., some people get dizzy from morning doses, others get insomnia from evening doses).

**Minimum meaningful duration:**

For blood pressure effects: 2–4 weeks. Blood pressure adapts to a new dosing schedule within 1–2 weeks, but a 4-week trial gives you enough data to see a stable pattern.

For side effects: 1–2 weeks. Most side effects (dizziness, fatigue, nocturia) appear quickly after a dosing change.

For cardiovascular outcomes: Do not attempt to test this at home. The TIME study followed 21,000 people for 5 years and found no difference. You cannot detect a difference in heart attack or stroke risk in an n=1 experiment.

**What to measure (specific metrics):**

**Morning blood pressure:** Take 3 readings in the morning (before medication, after sitting quietly for 5 minutes) using a validated home BP monitor. Record the average.

**Evening blood pressure:** Take 3 readings in the evening (before bed,