A recent article found that certain antihypertensives were better than others regarding cognitive impairment (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8800076/)



— a secondary analysis of the SPRINT trial assessed cognitive impairment in those on antihypertensive regimens that stimulated angiotensin II receptors 2 and 4 (ARBs, dihydropyridine calcium channel blockers such as amlodipine and nifedipine, and thiazide diuretics) versus those that inhibited these receptors (ACE inhibitors, beta blockers, and non-dihydropyridine calcium channel blockers such as diltiazem or verapamil)

— 8685 participants: 2644 (30%) were exclusively on stimulating meds, 1536 (18%) exclusively on inhibiting meds, and 4505 (52%) on mixed antihypertensive regimens

     — for details of the SPRINT trial, see http://gmodestmedblogs.blogspot.com/2015/11/tighter-blood-pressure-control-sprint.html, but basically  this was a pivotal trial finding that lower blood pressures (attained SBP 121 mmHg vs 136mm Hg) was associated with approx 30% decrease in many cardiovascular outcomes and all-cause mortality (though their methodology was eclectic/not really reproducible in a real clinical setting: their stated blood pressures probably translated to about 5mmHg higher values in normal clinical settings); and for tthe SPRINT substudy in their elderly subgroup: https://bucommunitymed.wpengine.com/sprint-trial-elderly-subgroup-study-of/

    — but, of note, the SPRINT study did not include diabetics or those with history of stroke

    — there was discretion by clinicians on which meds to use to treat the hypertension

— visit frequency was every 3 months for the duration of the trial

— mean age 68, 64% male, 11% Hispanic/30% non-Hispanic black/57% of non-Hispanic white

— follow-up 4.8 years (the SPRINT trial itself stopped early because of evident early benefit of many outcomes in those with tighter BP control, but there was an extended followup visit 2-3 years later

— confounding was assessed at baseline and was statistically minimized by propensity scoring

    — they matched for age, sex, race/ethnicity, history of CVD or CKD, BMI, number of antihypertensive meds classes, SBP tertile, and treatment group in the study

— main outcome of this study: a composite of adjudicated amnestic mild cognitive impairment (MCI) or probable dementia

    — mild cognitive impairment (MCI): noticeable symptoms but not interfering with everyday activities

    — amnestic MCI: memory that is significantly impaired but other cognitive functions are spared: eg language, visual-spatial skills, executive functioning (ie capacity to plan ahead, display self-control, follow multiple-step directions, stay focused despite distractions). Amnesic MCI seems to have a higher risk of progressing to Alzheimers and a higher mortality risk



— amnestic MCI or probable dementia incidence, comparing stimulating versus inhibiting meds:

    — 428 events (45 events per 1000 person-years) versus 355 events (59 events per 1000 patient years): risk reduction of 24%, HR 0.76 (0.66-0.87)

        — overall a 25% to 35% decreased risk for the array of unadjusted and several different adjusted models

— amnestic MCI alone, comparing stimulating only versus inhibiting only users:

    —  373 events (40 events per 1000 person-years) versus 312 events (54 events per 1000 person-years): risk reduction 26%, HR 0.74 (0.64-0.87)

— probable dementia alone: 8 versus 10 cases per 1000 person-years: HR 0.80 (0.57 1.14), not statistically significant [but small numbers of people]


Their figure showing that after about 1.5 years there was an impressive splaying of the curves, showing benefit for those on the stimulating hypertension meds:




— subgroup analyses for the combination of probable dementia or amnestic mild cognitive impairment, comparing stimulating to inhibiting antihypertensive users:

    — by age groups:

        — <75yo: 3026 patients, 232 versus 168 events, 20% decrease, HR 0.80 (0.65-0.99)

        — >75yo: 1123 patients, 196 versus 187 events, 23% decrease, HR 0.33 (0.63-0.93)

    — females had a 33% decrease, and males a 20% decrease


— negative control outcome analysis suggested the presence of residual confounding (by assessing some outcomes felt to be very unlikely to be causally related to antihypertensive use):

    –comparing patients only on stimulating vs inhibiting antihypertension meds:

        — infectious negative control outcomes: 90 patients (12%) versus 77 patients (16%), strong trend favoring stimulating only with HR 0.78 (0.55-1.11)

        — orthopedic negative control outcome: 102 patients (13%) versus 77 patients (18%), HR 0.72 (0.52-1.00), statistically significant 28% decrease

        — hematologic negative control outcome: 48 patients (7%) versus 40 patients (9%), HR 0.83 (0.50-1.37), a trend favoring those on stimulating medications only

    — there were similar findings for these 3 negative controls, skewed towards the benefit of those on stimulating medications comparing those on mixed regimens (those on combinations of inhibiting and stimulating) versus those on inhibiting medications only, though this time the hematologic negative control outcome was the only one that was statistically significant, the other two had very strong trends to being significant

        — this all raises the possibility that the 2 groups of patients (on stimulating vs inhibiting meds) were different for some reason, since there appeared to be possible benefit from the stimulating meds though this is highly unlikely, especially from 3 different negative control outcomes



— several studies have found that more intensive BP control overall is associated with less MCI

— and several past observational studies have suggested that blocking types 2 and 4 angiotensin II receptors lowers dementia risk; and, for what it’s worth, a slew of animal studies have found similar results

— this study, a secondary analysis of the well–done SPRINT hypertension study with lots of granular data, did find consistent decreases in amnestic MCI or probable dementia across subgroups by age, sex, race/ethnicity, the presence of cardiovascular or kidney disease, BMI, number of antihypertensive medication classes, systolic blood pressure, or SPRINT treatment group. And these results were consistent across their various sensitivity analyses with different statistical adjustment strategies


— there is an argument in the medical literature that we clinicians should be prescribing ARBs over ACE inhibitors for hypertension:

    — a real-world effectiveness study found that ARB’s (vs ACE-I’s) should perhaps be the best choice in the treatment of hypertension, given their better safety profile: https://bucommunitymed.wpengine.com/hypertension-use-arbs-over-ace-i/

    — there are real concerns about the half-life of these medications. Other studies have found an increased risk of stroke in those on ACE inhibitors, which may be attributed to the fact that their effectiveness wears out after 12 to 18 hours and tends not to be effective in the early morning at the time of the cortisol (and hypertension) surge. The increased stroke risk was found in 4-5 studies I have seen,  including the ALLHAT study (ALLHAT was a community-based study comparing amlodipine to chlorthalidone and lisinopril to chlorthalidone).

        — one drawback of losartan is that it achieves its full 24–hour effectiveness only at its highest dose of 100 mg per day. Other ARBs do have longer half-lives, for example valsartan has a 24–hour effectiveness independent of its dose

        — but losartan (and not other ARBs) have the benefit of lowering serum uric acid levels (it is uricosuric) and is even mentioned in the 2012 American Rheumatology Assn guidelines for primary treatment of hyperuricemia, and is included in the 2020 guidelines on hyperuricemia as a preferred antihypertensive (https://wchh.onlinelibrary.wiley.com/doi/epdf/10.1002/psb.1887 )

        — hyperuricemia itself is associated with increased cardiovascular disease and overall mortality

            — for an interesting evolutionary perspective: http://gmodestmedblogs.blogspot.com/2019/04/uric-acid-lowering-cardiovasc-benefit.html  

            — allopurinol seems to decrease cardiac events in several studies: http://gmodestmedblogs.blogspot.com/2016/03/hyperuricemia-allopurinol-decreases.html

            — and another one documenting an increased incidence of cardiac mortality with hyperuricemia, with uric acid threshold over around 5 mg/dL and increasing mortality relatively linearly after that: http://gmodestmedblogs.blogspot.com/2020/01/uric-acid-threshold-for-increased-cad.html

            — And, perhaps another small benefit: if a patient has heart failure and would likely benefit from sacubitril/valsartan (entresto), there is a higher risk of angioedema in those on ACE inhibitors (the ACE should be stopped for 3 days, but that type of patient advice creates some increased risk of a potentially very serious error, and that possibility would not need to happen if they were on an ARB). See http://gmodestmedblogs.blogspot.com/2022/05/heart-failure-outpatient-initiation-of.html

— it would also be useful to have a breakdown of the other stimulating antihypertensives: amlodipine is the drug-of-choice for almost all hypertensives per the NICE (UK) guidelines (it has several great features, including a full 24 hour protection, decreasing blood pressure variability,etc: eg see https://gmodestmedblogs.blogspot.com/2015/07/blood-pressure-variability-and-heart.html) or chlorthalidone https://gmodestmedblogs.blogspot.com/2023/01/hypertension-hctz-vs-chlorthalidone.html 



— the SPRINT trial excluded patients who were diabetic (a large and growing group…) and those with strokes, limiting broad generalizability of their results (and these subsequent analyses)

— this extension of the SPRINT trial was after the original trial stopped 2-3 years previously, with a “final extended follow-up visit”. Not clear how extensive this visit was and if there were subsequent changes in antihypertensive meds (I suspect this was included) or the array of other medical/psychosocial variables that might have affected cognition (stress, sleep, depression…), which i suspect were not included

— the SPRINT study by design was not one of new med users (eg initiating an ARB vs ACE inhibitor), but one of people mostly  already on these meds and following them. So, not quite as easy to interpret the results: what meds were they on before? were they on an ARB because of problem with ACE? Or other issues? Or other covariates? How long were they already on meds that either were stimulating or inhibiting or combos? This would bias the results.

— the issue of residual confounding is a bit complex. The baseline groups did have different outcomes in the assessment of the 3 very different negative controls of infectious, hematologic and orthopedic diseases. and, their clinicians may have chosen different antihypertension meds for the  different patients for an unknown but important reason, which would inflict a bias on the results  The researchers did model their analyses by propensity matching, a mathematical approach to equalize the 2 groups. But this may not reflect the real world: see http://gmodestmedblogs.blogspot.com/2020/03/tramadol-fo-oa-inc-mortalityprobs-with.html .


So, given the reasonable interchangeability of ARBs and ACE inhibitors in the treatment of hypertension, but the real-world issue of cough as a common adverse effect of ACE inhibitors (with the associated patient anxiety, increased medical visits, perhaps increased chest xrays with their attendant radiation exposure, and now a reasonable likelihood that dementia may be increased), it does seem reasonable that an ARB should be prescribed vs ACE inhibitor. Of course, the absolute number of people in this SPRINT substudy getting MCI and amnestic MCI is pretty small, but:

    — cognitive impairment is a pretty terrible thing to get, and clearly has awful consequences for the patient, their self-esteem, their family and community, etc (though it would be interesting to know longer-term followup: is there a higher risk of Alzheimers??? Following the SPRINT cohort for another several years might be helpful to assess the more profound effect of developing moderate to severe dementia

    — a small increased absolute risk of MCI translates into very large numbers of people when these meds are prescribed to a huge numbers of people treated for hypertension (about 50% of the US adults have BP >130/80, 116 million people)

— and, by the way, still about 20% of older adults are still prescribe b-blockers as their initial monotherapy in the absence of  heart failure or coronary artery disease