Could prescribing mix help explain missed blood-pressure targets?

It counts a practice’s dispensed items for the top strengths of the common calcium-channel blockers, amlodipine 10 mg, felodipine 10 mg and lercanidipine 20 mg, and compares that with how many ACE inhibitor and ARB items the same practice dispensed over the year. Datim-QI shows it as upper-strength CCB items per 100 ACEi/ARB items, with both item counts on the card, so you can always see what the ratio is actually made of.

I linked 12-month OpenPrescribing windows with CVDPREVENT practice data at March 2024 and December 2025. The outcome was the percentage of people with recorded hypertension whose measured blood pressure met the age-appropriate threshold. CVDPREVENT is a national audit built from rolling extracts of routinely recorded general-practice data; NHS England describes the collection and its limitations on the CVDPREVENT dataset page.

Weighted models adjusted for age structure, neighbourhood-weighted ethnicity, deprivation, QOF hypertension, diabetes and CKD prevalence, 42 ICB fixed effects, and prescribing intensity within the CCB and ACEi/ARB pair. A one-standard-deviation higher log ratio was associated with 0.57 percentage points lower blood-pressure control in 2024, and 0.66 points lower in 2025. Substituting combined pair items per 1,000 patients for the primary intensity term produced very similar estimates.

The adjustment was built up in stages, which is the more informative way to read it. Starting from age and ethnicity alone the estimate was 0.50 percentage points, and adding deprivation, disease prevalence, ICB fixed effects and prescribing intensity in turn left it at 0.57. An association driven by the case mix a practice serves should fade as those controls go in. This one did not.

A second ratio was tested alongside it, and this is where the story gets more complicated than a single result. NICE NG136 prefers a thiazide-like diuretic such as indapamide over bendroflumethiazide, so a practice with a high bendroflumethiazide to indapamide ratio is guideline-discordant in the same broad sense. I expected it to track with poorer control too. It did not. The estimate was close to null after equivalent adjustment in both snapshots.

The ratio is a rear-view mirror. By the time a practice appears at the high end of it, a series of individual decisions to increase a CCB dose has already been taken, one appointment at a time, each of them entirely reasonable in isolation. The clinically useful moment sits earlier, and it is worth stating plainly because it is where the actual choice lives:

Evidence from 42 factorial trials involving 10,968 participants found that adding a medicine from a different class produced roughly five times the additional blood-pressure reduction achieved by doubling the dose of a single medicine. Dose-related adverse effects move the other way: CCB ankle oedema increases with dose, and adding renin-angiotensin-system blockade reduces it, while uptitrating the CCB does not.

The uncontrolled-on-low-dose consultation is therefore worth naming explicitly, because it is entirely invisible to this metric. The upper-strength ratio only detects the aggregate residue of that decision having already gone the other way, repeatedly, over a long period. A practice acting on this analysis should not confine itself to the patients the metric happens to point at.

That is an argument for widening the review, not for redirecting it. For someone already at the top dose, adding a complementary class carries a second and quite specific benefit: it addresses the dose-related ankle oedema that the higher strength may itself be causing, which the next section covers, and which is a common reason people quietly stop taking a CCB at all. Reviewing that group can improve blood pressure and tolerability in the same conversation. Both groups are worth the time, for different reasons.

Which complementary class to add is not a choice a dataset can make. It turns on kidney function, potassium, pregnancy potential, previous angioedema, postural symptoms, frailty, comorbidity, the indication for the original prescription, and the person’s own preference. See NICE NG136 recommendations.

The factorial meta-analysis above is the clearest statement of the dose-versus-class trade-off: read the meta-analysis on PubMed. It supports asking the review question. It does not identify which second class is right for any individual, and it was not designed to.

On tolerability, NHS Specialist Pharmacy Service notes that ACE inhibitors and ARBs can reduce the incidence of CCB-related ankle oedema in some patients, while diuretics have little effect on that particular mechanism. A meta-analysis of 25 randomised trials reported 38% less peripheral oedema with a CCB plus renin-angiotensin-system blockade than with CCB monotherapy. See the SPS review and the trial meta-analysis. This matters for the low-dose decision too: oedema is a common reason people quietly stop taking a CCB, and an uptitration that is never actually swallowed lowers nobody’s blood pressure.

The trial evidence on dose versus class is reasonably consistent, and it points one way: combining classes at lower doses achieves more blood-pressure reduction, with fewer dose-dependent adverse effects, than pushing a single medicine towards its ceiling.

The 2024 ESC guidelines for elevated blood pressure and hypertension build that directly into the structure of treatment. For most adults they recommend starting with a two-drug combination, ideally as a single pill, typically a renin-angiotensin-system blocker together with a CCB or a thiazide or thiazide-like diuretic, rather than monotherapy escalated by dose. The stated reasoning is the one this article keeps returning to: faster control, better tolerability than high-dose monotherapy, and better adherence from single-pill combinations. See the ESC guideline page.

NICE NG136 retains a stepped approach that begins with monotherapy and adds at step 2, so UK practice is organised differently. That difference is the reason a metric like this one has anything to detect at all. Under a combination-first structure, the choice between adding a class and increasing a dose largely does not arise. Under a stepped structure it arises constantly, in individual consultations, and the aggregate of those decisions is visible in dispensing data.

Datim-QI shows the ratio on a practice’s prescribing page alongside its ICB and England comparators, with both underlying item counts, in the same format as every other prescribing measure. That much is simply description, and it is available whether or not anything else is true.

It becomes a suggested review only when two conditions hold together. The first does the selecting: an earlier version compared the ratio with the England average, which splits the country roughly in half and therefore selected nothing, so it is now the highest quartile.

The second condition is deliberately generous. Almost every practice in England misses the 85% ambition for this outcome, so it works as a floor rather than a filter: it excludes practices that already meet the highest published ambition, and passes the rest through to the quartile test. That is the intended behaviour. A practice at 72% is already ahead of most of England, and there is still real headroom between that and the 85% ambition. Medicines optimisation is one of the most direct ways to work into it, which is why those practices are worth including rather than filtering out.

On the December 2025 data the two conditions together select 1,506 of 6,086 practices, about one in four. The review itself should check home or ambulatory confirmation, adherence, indication, tolerability and the whole regimen before considering any change, and should record why an ACEi/ARB or another complementary option is unsuitable where that is the case. The aggregate ratio is never used to issue, stop or switch a medicine, and the product sets no target to lower it.

Two groups are worth separating, because the useful question differs between them. The first is where the decision is still open. The second is where the metric points, and where tolerability and blood pressure can often be addressed together.