Predicting the risk of apparent treatment-resistant hypertension: a longitudinal, cohort study in an urban hypertension referral clinic.

Apparent treatment-resistant hypertension (aTRH) is associated with higher prevalence of secondary hypertension, greater risk for adverse pressure-related clinical outcomes, and influences diagnostic and therapeutic decision-making. 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. aTRH was assessed in an urban referral hypertension clinic using a 140/90 mm Hg goal blood pressure target in 745 patients with uncontrolled blood pressure, who were predominately African-American (86%) and female (65%). Analyses were stratified according to existing prescription of diuretic at initial visit. Risk for aTRH was estimated using logistic regression with patient characteristics at index visit as predictors. Among those prescribed diuretics, 84/363 developed aTRH; the risk score discriminated well (area under the receiver operating curve = 0.77, bootstrapped 95% CI [0.71, 0.81]). In patients not prescribed a diuretic, 44/382 developed aTRH, and the risk score showed a significantly better discriminative ability (area under the receiver operating curve = 0.82 [0.76, 0.87]; P < .001). In the diuretic and nondiuretic cohorts, 145/363 and 290/382 of patients had estimated risks for development of aTRH <15%. Of these low-risk patients, 139/145 and 278/290 did not develop aTRH (negative predictive value, diuretics - 0.94 [0.91, 0.98], no diuretics - 0.95 [0.93, 0.97]). We created a novel clinical score that discriminates well between those who will and will not develop aTRH, especially among those without existing diuretic prescriptions. Irrespective of baseline diuretic treatment status, a low-risk score had very high negative predictive value.