Ambulatory blood pressure monitoring and mortality

​​A large Spanish study confirmed the importance of 24-hour ambulatory blood pressure monitoring (ABPM) as a superior predictor of all-cause and cardiovascular mortality (see htn ambulat higher mortality nejm2018 in dropbox, or DOI: 10.1056/NEJMoa1712231).

Details:

--63,920 Spanish adults in the ongoing Spanish Ambulatory Blood Pressure Registry, from 223 primary care centers within the Spanish National Health System; all enrolled in the registry from 2004-2010

--58% male, BMI 29, current smoker 16%, diabetes 20%, dyslipidemia 42%; prior cardiovasc disease 11%

--clinic BP 147.9/86.7, 24-hr ABPM 129.2/76.5, daytime ABPM 132.3/79.4, nighttime ABPM 120.2/68.4

--clinic and 24h ABPM were examined, in the following categories (normal clinic BP is defined as <140/90, ABPM is <130/80):

    --sustained hypertension (both clinic and ABPM were high)

    --"white-coat" hypertension (elevated clinic but normal ABPM)

    --masked hypertension (normal clinic but elevated ABPM)

    --normotension (both normal)

--clinic blood pressure was measured after the patient had been resting in a seated position for 5 minutes, using the mean of 2 recordings by a validated oscillometric or mercury sphygmomanometer.

--median followup of 4.7 years

Results:

--3808 patients died overall, 1295 from cardiovascular causes (440 from ischemic heart disease, 291 from stroke, 123 from heart failure)

--all-cause mortality:

    --24-hr systolic ABPM was strongly associated with all-cause mortality: HR 1.58 (1.56-1.60), after adjusting for clinic blood pressure

    --clinic blood pressure was barely associated with all-cause mortality: HR 1.02 (1.00-1.04), after adjusting for 24-hr ABPM

    --HR for 1-SD increase in BP was:

        --1.55 (1.53-1.57) for nighttime ambulatory systolic blood pressure, after adjustment for clinic and daytime ambulatory blood pressures

        --1.54 (1.52-1.56) for daytime ambulatory systolic pressure, after adjusting for clinic and nighttime ambulatory blood pressures

    ​--these relationships were consistent across subgroups of age (<60 vs >60yo), BMI (<30 vs >30), sex, diabetes, cardiovascular disease, antihypertensive treatment

    ​--overall, systolic blood pressure was a better predictor of cardiovasc disease, particularly systolic 24-hr ABPM [systolic blood pressure has been found to be a stronger predictor than diastolic in prior observational studies of clinic-based BP]

--by types of hypertension:

    --masked hypertension was most strongly associated with all-cause mortality, HR 2.83 (2.12-3.79)

    --sustained hypertension, HR 1.80 (1.41-2.31)

    --white-coat hypertension, HR 1.79 (1.38-2.32)

--the "rate advancement period" (approx additional years of chronologic age that would be required to yield the equivalent mortality per 1-SD increase in BP as compared with normotension): 1.4 years by clinic systolic blood pressure and 8.5-10.2 years by systolic ABPM

--for the population attributable fractions (the fraction of mortality in the population attributable to each hypertension phenotype): sustained hypertension (observed in 15.6% of patients who died) accounted for 7.0% of deaths in the whole cohort, and masked hypertension (observed in 3.0% of deaths) accounted for 1.9% [the population attributable risk of sustained hypertension was higher than masked just because it was more common, but masked hypertension had higher relative mortality]

--cardiovascular mortality (similar results for pretty much all of the above for all-cause mortality)

Commentary:

--this study adds to the increasing numbers of studies finding ABPM is far superior to clinic-based blood pressure measurements in predicting clinical outcomes (hard cardiovascular events, mortality), presenting data from a large cohort followed closely in an ongoing clinical registry. Many of the prior studies showing benefit were much smaller and had some variations from the current one (eg, the actual clinical effect of white-coat hypertension, or the relative prevalence of masked hypertension).

--part of the reason that white-coat hypertension was associated with increased mortality in this study is likely because of the higher mean ABPM in this group: 119.9/71.9 vs 116.6/70.7 in normotensives, though both were in the "normal range"; observational studies do find fewer clinical cardiovascular events as the systolic continues to decrease to the 100-110 mmHg range

--one advantage of this study is that it was a pragmatic one, more likely to reflect real clinical practice

--some limitations of the study: clinic BP was the average of only 2 measurements; ABPM was performed only once (though some of the patients had two tests done, and similar results). also, perhaps selection bias for which patients had ABPM in the registry. and as with any study in a single population, there is risk in generalizing the results to others

--nonetheless, there is adequate support, accepted by many major recommending agencies (eg NICE in the UK, USPSTF and Am Heart Assn in the US) strongly recommending the use of ABPM in the diagnosis of hypertension, which leads to a more appropriate therapy.  that being said, i would stress a couple of things:

    ​--most studies do not find that clinic-based blood pressure diagnosis (including this one) adds anything (in this case minimally) to hypertension-related clinical outcomes

    --studies have pretty consistently found that for patients with stage 1 hypertension in a clinic visit (the vast majority of hypertensive patients overall), even using the strict definition of how to measure clinic blood pressure accurately (which is largely not done in routine clinical care), more than 30% do not have hypertension by ABPM.

    --though, in all patients it is really important to reinforce non-pharmacological therapy, whether they need meds or not. as is pretty clear in a prior blog, relaxation itself seems to lower blood pressure even more than expected by a single medication (see http://gmodestmedblogs.blogspot.com/2018/04/relaxation-lowers-blood-pressure.html ). And, alerting a patient to a potential diagnosis of hypertension by clinic BP, even if that is not backed up by ABPM, may help motivate and support patients in trying diet, weight loss, exercise, and relaxation to lower their blood pressure.  Older studies may be mixed on the clinical effects of white-coat hypertension, for example, though some have found that its presence does presage the development of sustained hypertension in the future. So, these patients, i think, would best be served by stressing nonpharmacologic therapy and the real possibility of their developing hypertension without this.

--masked hypertension does remain a conundrum for us.  several studies (including this one) have suggested that it is really dangerous/leads to lots of cardiovascular events. some (less so with this one) have found it to be remarkably common, as well. but does this mean everyone should get ABPM at some point? every 3-5 years? if risk factors change (like, they gain weight)??? perhaps health centers should at least lend validated BP cuffs to patients to check their home-based BP, and perhaps then check ABPM if high???

so,

this study reinforces the current clear recommendation consensus to perform ABPM pretty regularly in diagnosing hypertension. it seems abundantly clear (and it seems pretty universally accepted) that ABPM really is the best (by far) in predicting actual important clinical events, and that not using it is very likely to lead to overtreatment of many patients unnecessarily (leading to potential adverse effects, medicalization of the patients, perhaps dysphoria and perhaps even patients' feeling that they are "sick" in their self-conception and have more sick behavior to the world around them). and also lead to undertreatment of masked hypertension, which seems to be rather nefarious

but the real issue is how to make the culture change within health care. how do we transition from the old (and not-so-clinically-relevant model) of clinic-based BP measurements to one which largely disregards this in the move to the more evidence-based ABPM?? and this is harder/more work/takes more time than just a BP check. But we do need to figure out how to institutionalize this transition. it turns out, on my brief review, that these ABPM's are not so expensive (in the $1500 range, a pittance in our over-priced health care market), and are reimbursable to boot....

see:

-- http://gmodestmedblogs.blogspot.com/2018/02/ambulatory-blood-pressure-monitoring.html , highlighting a study showing the benefits of ABPM in high-risk women as well as reinforcing the high prevalence of masked hypertension, and further references the many prior blogs on ABPM

-- http://gmodestmedblogs.blogspot.com/2017/11/new-aha-hypertension-guidelines.html , which reviews the new AHA hypertension guidelines, along with references to USPSTF, and the lack of correlation between ABPM and clinical outcomes in the SPRINT study

-- http://gmodestmedblogs.blogspot.com/2018/03/home-bp-monitoring.html for a blog on home-based BP monitoring, with references to prior blogs including one showing that home-based monitoring is probably not quite as good as ABPM (very different approaches in different studies; no real standardization), but seems much better than clinic-based BP measurement.