Background:
Apparent treatment resistant hypertension (aTRH) is associated with increased prevalence of secondary hypertension and adverse pressure-related clinical outcomes. We previously showed that cross-sectional prevalence estimates of aTRH are lower than its true prevalence as patients with uncontrolled hypertension undergoing intensification/optimization of therapy will, over time, increasingly satisfy diagnostic criteria for aTRH.
Methods:
aTRH (SBP and/or DBP at or above a clinically defined goal BP [140/90, 130/85, 130/80, or 125/75 mmHg] over two consecutive office visits when on ≥ 3 antihypertensive drug classes, including a diuretic; or SBP and DBP below goal when on ≥ 4 drug classes, including a diuretic) was assessed in an urban referral hypertension clinic in 924 patients ≥ 30 years old (57.7 ± 12.6) with at least two follow-up visits over 240 days. Patients were mostly African-American (86%; 795/924) and female (65%; 601/924). A minority (28.7%; 265/924) were taking diuretics at their index visit, and analyses were stratified according to this use. Risk for aTRH was estimated using logistic regression with patient characteristics at index visit as predictors. Performance of this risk score at discriminating aTRH status over follow-up was assessed using AUC and was internally validated using bootstrapping.
Results:
Amongst those on diuretics, 80/265 (30.2%) developed aTRH; the risk score discriminated well (AUC = 0.79, bootstrapped 95% CI [0.73, 0.84]). In patients not on a diuretic, 151/659 (22.9%) developed aTRH, and the risk score showed moderate, but significantly lower, discriminative ability (AUC = 0.71 [0.66, 0.74]; p < 0.001). In the diuretic and non-diuretic cohorts, 43/265 (16.2%) and 101/265 (38.1%) of patients, respectively, had estimated risks for development of aTRH < 10%. Of these low-risk patients, 42/43 (97.7%) and 97/101 (96.0%) did not develop aTRH (negative predictive value, diuretics – 0.95 [0.93, 1.00], no diuretics – 0.96 [0.91, 1.00]).
Conclusions:
We created a novel clinical score that discriminates well between those who will and will not develop aTRH, especially amongst those taking diuretics initially. Irrespective of diuretic treatment status, a low risk score had very high negative predictive value.