Ten percent of patients admitted to the hospital for reasons other than hypertension develop severe hypertension (sHTN) following admission. Standard treatment for patients with sHTN is administration of intravenous (IV) antihypertensives. However, IV antihypertensive treatment has been associated with end-stage organ injury, including acute kidney injury (AKI). Results of a previous study suggested that compared to no treatment, treatment for patients with sHTN with IV antihypertensives was associated with an 18% greater risk of AKI at the population level.
Hospitalized patients who develop AKI are at increased risk for adverse outcomes, including prolonged hospital stay, persistent loss of kidney function, and death. Lama Ghazi, MD, PhD, and colleagues conducted a study to examine whether there is heterogeneity in the effect of the use of IV antihypertensives on the incidence of AKI in patients who develop in-hospital sHTN. Results were reported online in the American Journal of Kidney Diseases.
The study population included patients without kidney failure who developed sHTN. The study exposure was treatment with IV antihypertensives within three hours of blood pressure elevation. The primary outcome of interest was time to development of AKI.
The researchers captured the association between the time to develop AKI and predictors utilizing an accelerated failure time Bayesian Additive Regression Trees (AFT-BART) model. A counterfactual outcome framework was used to estimate individual treatment effects for each participant. Those estimates were used to identify patient characteristics associated with treatment effect heterogeneity in response to IV hypertensives and to explore potential subgroups of patients who may or may not benefit from the treatment.
Eligible participants were adult patients who were admitted to one of five large hospitals in the New England Health System between 2016 and 2020. Excluded criteria were patients admitted for a hypertensive emergency; those with sHTN on admission or in the emergency department (ED); those with no measurement of blood pressure; those admitted to an intensive care unit, maternity ward, or research unit; those with length of stay less than two days or more than 30 days; those who received vasopressors zero to six hours prior to developing sHTN; those with missing covariate data; and those with end-stage kidney disease, renal failure, or on dialysis.
In-hospital sHTN was defined as the first documentation of systolic blood pressure >180 mmHg or diastolic blood pressure >110 mmHg reported at least one hour following hospital admission. Blood pressure measurements recorded in the ED were excluded.
Patients in the group receiving IV antihypertensive treatment within three hours of development of sHTN were treated with hydralazine, labetalol, metoprolol, or nicardipine (treatment group). Those in the untreated group did not receive IV or oral antihypertensives (control group).
The analyses included 28 preselected covariates. The preselected covariates were age, sex, race, surgical ward, vital signs, and Elixhauser comorbidity index score. The 2021 Chronic Kidney Disease Epidemiology Collaboration creatinine equation that measures estimated glomerular filtration rate (eGFR) without race was used to determine eGFR, considered as a continuous variable.
AKI was ascertained at three hours following development of sHTN, continuing until the end of the follow-up period (30 days after hospital admission). The Kidney Disease: Improving Global Outcomes criteria were used to define AKI. AKI was diagnosed if a patient’s serum creatinine increased by ≥0.3 mg/dL within 48 hours or if serum creatinine level was 1.5 times the lowest measured serum creatinine within the previous seven days.
The cohort comprised 11,951 patients who developed sHTN during hospitalization. Of them, 741 were treated with IV antihypertensives within three hours of severe elevation in blood pressure (treatment group); 18.2% (n=135) developed AKI. There were 11,210 patients in the control group. Of them, 12.7% (n=1422) developed AKI. The median time from development of sHTN to time of development of AKI was 57.7 hours; median time to AKI development from IV antihypertensive treatment was 59.1 hours.
Median age of the overall cohort was 72.5 years, 19.9% identified as Black, 42.9% were male, median Elixhauser score was 4, and 70% had a history of hypertension. Median SBP/DBP at admission was 154/82 mmHg and at the time of sHTN was 185/91 mmHg. Those in the treatment group were older, more likely to be admitted to the surgical ward, and had higher systolic and diastolic blood pressure at the time of admission and at development of sHTN.
The first stage of the statistical analysis used the AFT-BART models to estimate each patient’s characteristics and treatment status. In the second stage, the researchers implemented an exploratory fit-the-fit approach using classification and regression trees to clarify the key moderators for conditional average treatment effects.
Analysis results demonstrated that most patients would have been harmed by treatment with IV antihypertensives with the exception of a small subgroup of patients (n=317). Those patients were White, had systolic blood pressure ≥156 mmHg on admission, eGFR ≥70.7 mL/min/1.73 m2, and a serum bicarbonate level of <21.7 mmol/L.
The authors cited some limitations to the study findings, including the use of observational data and causal inference methods, using a single data source limited to one health care system, and the inability to determine how blood pressure was measured.
In summary, the researchers said, “While there are several machine learning methods that allow for the estimation of individual treatment effects, we have considered one state-of-the art approach—AFT-BART—due to its flexibility and robust empirical performance in multiple previous simulation studies. Overall, this innovative machine learning approach within the causal modeling framework allowed us to assess whether there are subpopulations that might benefit from IV antihypertensive treatment following sHTN development. We mainly observed that IV antihypertensive treatment is not beneficial following sHTN development, Our findings are exploratory, thus hypothesis-generating; we encourage more future studies with different data sources to assess for effect heterogeneity when identifying treatment options, if any are needed, for sHTN.”
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.