Higher systolic BP on ICU admission may predict favorable prognosis

Swedish researchers suggested incorporating systolic BP readings at ICU admission into risk-scoring models after finding that patients with chest pain and high supine systolic BP had better prognosis and lower mortality risk at one-year follow-up compared with those who had lower systolic BP.

Researchers from Linköping University in Linköping, Sweden studied the relationship between long-term mortality and supine systolic BP among 119,151 patients who were treated at an ICU for chest pain from 1997 to 2007. They analyzed data from the Registry of Information and Knowledge About Swedish Heart Intensive Care Admissions, a registry which contains information on all patients admitted to medical ICUs in all Swedish hospitals, the Swedish National Patient Registry and the National Death Registry.

Participating hospitals consecutively enrolled patients and assigned them to one of four groups based on systolic BP readings: first quartile (Q1; BP<128 mm Hg; n=29,802), second quartile (Q2; BP=128 mm Hg to 144 mm Hg; n=32,165), third quartile (Q3; BP=145 mm Hg to 162 mm Hg; n=27,522) or fourth quartile (Q4; BP≥163 mm Hg; n=29,662). Only patients with a pulse pressure >9 mm Hg were included in the analyses.

At a mean follow-up of 2.47 years, Cox proportional hazard models revealed that Q1 patients had the highest risk for death (HR=1.46; 95% CI, 1.39-1.52) and that Q4 patients had the best prognosis (HR=0.76; 95% CI, 0.72-0.80) even after adjusting for potential confounding variables including age, sex, smoking, diastolic BP and medication use at admission and discharge.

Compared with Q2 patients, Q4 patients had a 21.7% lower absolute risk for death and Q3 patients had a 15.2% lower risk, whereas those in Q1 had 40.3% higher risk for mortality.

The association between high-admission BP and lower mortality was unaffected when the researchers adjusted for the presence of known diabetes, dementia, malignancy, previous MI and stroke, and the lower risk identified in Q3 and Q4 patients compared with Q2 patients remained when data for participants with diabetes and congestive HF were excluded.

The association also remained among patients who were discharged with a diagnosis of MI (n=43,987; one-year mortality HR for Q4 compared with Q2, 0.74; 95% CI, 0.68-0.79), those discharged with a diagnosis of ischemic heart disease (n=56,585; HR for Q4 compared with Q2, 0.75; 95% CI, 0.71-0.80) and when CVD mortality was set as the outcome (HR for Q4 compared with Q2, 0.66; 95% CI, 0.58-0.74).

“The results presented here show that the prognostic value of the recorded systolic admission BP is still valid when adjusted for a large set of potentially confounding variables including diastolic BP,” the researchers wrote.

“The fact that the measurement of BP and other variables was performed in an ordinary clinic setting speaks in favor of the usefulness, for general prognostic purposes, in the corresponding clinic setting as compared with results from randomized controlled trials, which ordinarily use specific inclusion and exclusion criteria in screening participants.”

Stenestrand U. JAMA. 2010;303:1167-1172.

Follow CardiologyToday.com on Twitter.