Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jones, A. E.
Right arrow Articles by Kline, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jones, A. E.
Right arrow Articles by Kline, J. A.
(Chest. 2006;130:941-946.)
© 2006 American College of Chest Physicians

Emergency Department Hypotension Predicts Sudden Unexpected In-hospital Mortality*

A Prospective Cohort Study

Alan E. Jones, MD; Vasilios Yiannibas, BS; Charles Johnson, BS and Jeffrey A. Kline, MD

* From the Department of Emergency Medicine, Carolinas Medical Center, Charlotte, NC.

Correspondence to: Jeffrey A. Kline, MD, Director of Research, Department of Emergency Medicine, 1000 Blythe Blvd, MEB 304D, Carolinas Medical Center, Charlotte, NC 28203; e-mail: jkline{at}carolinas.org

Abstract

Objective: The prevalence and prognostic significance of nontraumatic hypotension measured in the emergency department (ED) have not been studied. We hypothesized that ED hypotension confers risk of in-hospital mortality.

Design: Prospective cohort study.

Setting: Large urban ED with 115,000 visits per year.

Participants: Nontrauma ED patients aged > 17 years admitted to the hospital were prospectively identified on a random sample of 24-h blocks during 2004 to 2005. "Exposures" had any systolic BP (SBP) < 100 mm Hg in the ED; "nonexposures" all had SBP ≥ 100 mm Hg in the ED. Deaths were classified as sudden and unexpected by independent observers using explicit criteria.

Outcome measures: Mortality rates were compared with confidence intervals (CIs), Kaplan-Meier survival curves, and multivariate logistic regression.

Results: A total of 4,790 patients were enrolled during the study period, with 887 patients (19%) in the exposure group. Exposures were more likely to die in the hospital compared with nonexposures (8% vs 3%; 95% CI for difference of 5%, 4 to 8%). Exposures were more likely to have sudden and unexpected death compared with nonexposures (2% vs 0.2%, 95% CI for difference of 1.8%, 1 to 3%). Kaplan-Meier estimates showed increased mortality in the exposure group at all time points (log-rank test, p < 0.001). Multivariate logistic regression revealed exposure to hypotension as an independent predictor of in-hospital mortality (odds ratio, 2.0; 95% CI, 1.3 to 2.8).

Conclusion: Nontraumatic hypotension was documented in 19% of a random sample of ED patients admitted to the hospital. Patients exposed to hypotension had a significantly increased risk of death during hospitalization.

Key Words: BP • emergency department • hypotension • mortality • risk stratification • shock

Symptomatic arterial hypotension evidences a profound derangement in hemodynamic homeostasis. A large multicenter study1 has identified systolic hypotension, defined as a systolic BP (SBP) < 100 mm Hg measured in the prehospital phase of patient care, as a significant risk factor for in-hospital mortality among patients with no history of trauma. We recently found a 16% in-hospital mortality rate among 200 symptomatic, hypotensive emergency department (ED) patients enrolled in a controlled diagnostic trial.2 No published studies have investigated the prevalence or prognostic significance of nontraumatic arterial hypotension measured in a large, prospectively collected, unbiased sample of ED patients.

In the present study, we prospectively quantify the prevalence and outcomes of hypotensive ED patients admitted to the hospital. The hypotheses were as follows: (1) in an ED population with no history of trauma, the presence of systolic arterial hypotension, defined as a SBP < 100 mm Hg, will be associated with a significant increase in the incidence of in-hospital death; (2) in-hospital mortality will increase with the severity (depth and duration) of hypotension; and (3) when compared against a predefined set of predictor variables in a multivariate analysis, systolic hypotension will have a significant bias-corrected odds ratio (OR) for the prediction of the outcome of in-hospital death.

Materials and Methods

We performed a prospective cohort analysis. The study design is summarized in Figure 1 . Patients were assigned to a study group according to their lowest recorded BP in the ED: (1) the "exposure" group required at least one SBP measurement of < 100 mm Hg; and (2) the "nonexposure" group required all SBP measurements to be ≥ 100 mm Hg. The main outcome measure of the study was in-hospital mortality. The study was approved by the Institutional Review Board and Privacy Board at Carolinas Healthcare System prior to conduct of the study.


Figure 1
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Flow diagram of the study design.

 
Study Setting and Subjects
All subjects were enrolled from the ED at Carolinas Medical Center, an 800-bed teaching and tertiary referral hospital with > 100,000 patient visits per year. Inclusion criteria were age > 17 years and admission to the hospital from the ED. Exclusion criteria were as follows: (1) history of blunt or penetrating trauma to the torso in the past 24 h; (2) direct admission to the hospital or transfer from another facility or lack of full evaluation in the ED (full evaluation required both complete nursing and physician assessment that generates a physician and facility charge); and (3) no vital signs measured or recorded in the ED.

Patients were enrolled from July 2004 to June 2005, divided into four blocks with allocation to ensure equal sampling during each season: June to August, September to November, December to February, and March to May. Patients were enrolled during 24-h periods chosen from a standard random sample of 24-h periods (12 AM to 12 AM), comprising one half of the number of days in the block, which resulted in 182 total enrollment days.

Data Collection
On enrollment days, all patients admitted to the hospital from the ED were identified by query of the hospital computer tracking system (HBOC Star Navigator; McKesson; Alpharetta, GA). This query provided a text file containing patient names, personal and hospital identifiers, demographics, hospital location, admitting service, time of ED arrival and disposition, and all laboratory parameters ordered in the ED. An author manually compared each electronic query with the handwritten ED admission log to ensure completeness.

On the morning after each enrollment day, a research coordinator (V.Y.) downloaded this data into a custom programmed electronic data form (e-form, programmed in hypertext markup language with active server page overlays) that allowed hand input of additional data using a handheld wireless tablet personal computer (Toshiba Portege M200; Toshiba America; New York, NY). Additionally, the e-form was populated manually with mandatory data from the patient’s medical record in real time, including chief complaint, comorbidities, all vital signs with times of measurement, and administered therapies. The software prevented the e-form from being saved with empty data fields, nonsense characters, or illogical entries using predefined values (eg, pulse range required to be 20 to 200/min, all ages between 17 and 110 years). Using the tablet, the form was stored on the hard drive until docked synchronization at which time the polling program executed data upload to a Web-based server. This system and its method of protecting personal health identifiers has been described previously.3

We employed a rolling follow-up method to establish outcome of each patient. At the beginning of each month, a query of the study database was performed to identify new enrollees. A unique patient medical record number allowed access to each patient’s follow-up data, archived in a commercial database (HBOC Star Navigator).

Outcome Measures and Data Analysis
The main outcome measure was in-hospital mortality. Secondary outcomes included the following: (1) the incidence of sudden and unexpected deaths between the groups; (2) the relationship between the decile of lowest ED SBP and in-hospital mortality; (3) the relationship between duration of ED hypotension and in-hospital mortality rate in the exposure group; and (4) the sensitivity and specificity of ED hypotension for in-hospital mortality.

Sudden unexpected death was determined by modification of a published definition.4 Briefly, this definition required agreement of two physician observers, independent of the other’s decision and independent of knowledge of the patient’s group assignment, using the following published explicit criteria: (1) age < 85 years; (2) absence of a defined end-stage disease being specifically documented to be treated with comfort care only; (3) absence of hospice care; and (4) absence of do-not-resuscitate order. All four criteria were required to be absent for the death to be designated as sudden and unexpected. The duration of ED hypotension was determined by classifying patients into three groups: (1) transient, only one documented SBP < 100 mm Hg; (2) sustained, SBP < 100 mm Hg for ≥ 60 consecutive min; and (3) episodic, ED hypotension not categorized as transient or sustained.

Continuous data are presented as means ± SD, and when appropriate were compared for statistical differences using the unpaired t test. Categorical data are reported as proportions and 95% confidence intervals (CIs) where applicable (Clopper-Pearson method). Relative risks and their corresponding 95% CIs were calculated. The Kaplan-Meier survival estimates and log-rank test for comparison were used for time-to-primary outcome analysis. ORs were calculated to determine independent predictors of in-hospital mortality by using logistic regression with bootstrap correction for 95% CIs.5 Fourteen variables known to predict in-hospital mortality were entered into the regression analysis.16 For all statistical tests, p < 0.05 was considered significant. All statistical analyses were done with a commercially available software package (v 2.3.2; StatsDirect; Cheshire, UK).

The sample size was calculated based on preliminary data suggesting that 15% of hypotensive patients and 8% of nonhypotensive patients would suffer the primary outcome of death. Additionally, we estimated that 60 of 1,000 monthly admissions have the primary outcome. To find a significant difference with an {alpha} = 0.05 and a ß = 0.20, we planned enrollment of an exposure group of at least 325 subjects. With an approximate 12:1 ratio of nonexposures to exposures, we anticipated a final cohort of approximately 4,225 total patients.

Results

A total of 4,790 patients were enrolled during the study period with 887 patients (18.5%) in the exposure to hypotension group and 3,903 patients (81.5%) in the nonexposure to hypotension group. All patients had valid follow-up. Table 1 summarizes the clinical and demographic characteristics of all study subjects.


View this table:
[in this window]
[in a new window]

  		Table 1. Patient Demographics and Initial Vital Signs*

 
Figure 2 shows the Kaplan-Meier survival curves for both groups. The exposure group had significantly higher in-hospital mortality at all time points (log-rank test, p < 0.001) with a hazard ratio of 2.6 (95% CI, 1.8 to 3.7). Overall, 184 of 4,790 patients (3.8%; 95% CI, 3 to 4%) had the primary outcome of in-hospital mortality. Patients in the exposure group had a threefold increased incidence of in-hospital mortality: 8.3% vs 2.8% (95% CI for difference, 4 to 8%).


Figure 2
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Kaplan-Meier estimates of survival among 887 patients in the exposure group and 3,903 patients in the nonexposure group. There was a statistically significant higher in-hospital mortality rate in the exposure group vs the nonexposure group (log-rank test, p < 0.0001).

 
Twenty-four patients (0.5%; 95% CI, 0.3 to 0.7%) had sudden unexpected death (Table 2 ). The majority of patients (17 of 24 patients; 71%) with sudden unexpected death were exposed to hypotension in the ED. Patients exposed to hypotension had a tenfold higher risk of sudden, unexpected death, compared with nonexposed patients (2% vs 0.2%; 95% CI for difference, 1 to 3%).


View this table:
[in this window]
[in a new window]

  		Table 2. Sudden Unexpected In-hospital Deaths*

 
In exposure patients, the lower their nadir of ED SBP, the higher the probability of in-hospital death (Fig 3 ). Patients with a lowest SBP < 80 mm Hg had a sixfold increase in mortality rate compared with patients with a lowest SBP ≥ 80 mm Hg (18 vs 3%; 95% CI for difference, 9 to 25%).


Figure 3
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Relationship between the lowest ED SBP and in-hospital mortality (n = 4,790).

 
Likewise, longer exposure to hypotension in the ED corresponded to increased probability of in-hospital death (Fig 4 ). Patients exposed to hypotension (< 100 mm Hg) for ≥ 60 min had an in-hospital mortality rate of 14% (95% CI, 9 to 19%), compared with a 5% (95% CI, 3.3 to 8.1%) mortality rate observed in ED patients who were exposed to only one measured SBP < 100 mm Hg. Compared to nonexposures, the risk (OR) of in-hospital death in patients with transient hypotension was 2.0 (95% CI, 1.2 to 3.2), with episodic hypotension was 3.2 (95% CI, 1.9 to 5.1), and with sustained hypotension was 5.4 (95% CI, 3.4 to 8.5). Taken together, Figures 3, 4 suggest a dose-response relationship between severity of hypotension and in-hospital mortality.


Figure 4
View larger version (15K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. In-hospital mortality rate among patients categorized by duration of hypotension: sustained hypotension (continuous SBP < 100 mm Hg for ≥ 60 min), transient hypotension (only one SBP < 100 mm Hg followed by no subsequent readings < 100 mm Hg), and episodic hypotension (any patient exposed to ED hypotension and not categorized as transient or sustained).

 
To determine the diagnostic accuracy of hypotension for in-hospital death, two-by-two tables were constructed at various deciles of lowest SBP. The summary results of these analyses are shown on Table 3 . A lowest SBP < 80 mm Hg demonstrated the highest specificity, 97% (95% CI, 96 to 97%), and the highest likelihood ratio positive, 4.8 (95% CI, 3.3 to 6.8). Multivariate logistic regression analysis indicated that the presence of hypotension (SBP < 100 mm Hg) was an independent predictor of adverse mortality (OR, 2.0; 95% CI, 1.3 to 2.8) [Table 4 ].


View this table:
[in this window]
[in a new window]

  		Table 3. Performance of Various Deciles of SBP for the Prediction of In-hospital Mortality

 

View this table:
[in this window]
[in a new window]

  		Table 4. Results of Logistic Regression Analysis

 
Discussion

This single-center prospective study documents the frequency and prognostic significance of arterial hypotension measured in the ED. We found that 19% of a random sample of adults with no history of trauma had hypotension while in the ED. Patients exposed to hypotension had a threefold increased risk of in-hospital death and a 10-fold increased risk of sudden, unexpected in-hospital death. Patients with any one SBP < 80 mm Hg had a sixfold-increased incidence of in-hospital death, and patients with a SBP < 100 mm Hg for > 60 min had almost a threefold-increased incidence of in-hospital death. These data are the first to quantify the relationship between severity of hypotension and outcome.

We actively employ the term exposure to hypotension to draw attention to its significance as an independent predictor of death. Previously, several large studies789 in patients with specific diagnoses such as acute myocardial infarction, pulmonary embolism, and sepsis have documented the presence of hypotension to be a predictor of poor outcome. We previously found that arterial hypotension measured in the prehospital setting was a significant predictor of in-hospital mortality.1 Arterial BP measurements are performed in virtually every patient who visits an ED in the United States. Nonetheless, no prior published data have demonstrated the prognostic significance of ED hypotension in nontrauma hospitalized patients. Moreover, no severity of illness scoring systems have been systematically derived and validated for use in the ED in nontrauma patients being admitted to the hospital.10 The present findings strongly support the role of arterial hypotension to estimate acuity of illness in adult medical patients admitted to the hospital from the ED.

We believe that the practical utility of this report lies in its simplicity and integration with current practice. Noninvasive BP measurements are obtained in all ED patients. In the United States, national hospital accreditation guidelines have long mandated that hospitals establish and follow standard practices regarding BP measurements in ambulatory care settings. Thus, our findings do not require a clinician to memorize, archive, or employ a scoring system or clinical prediction rule. For our results to be applied to daily practice, a clinician need only pay attention to all recorded BP measurements, and remember that medical patients with any measured SBP < 100 mm Hg are significantly more likely to die during hospital admission. The next logical question regarding what intervention is necessary when hypotension is present (eg, admit to higher level of care such as an ICUs) is the topic of a planned future investigation.

Several limitations of this study warrant discussion. First, this report is from a single urban tertiary center, and its results might not be generalizable to nonacademic centers or hospitals with a different acuity of illness. Second, in the hypotensive patients we made no physiologic measurements to establish the presence or absence of circulatory insufficiency. Indeed, the overarching aim of this project was to measure numeric relationship between hypotension and outcome in all admitted patients, not just those patients with hypotension and symptoms and signs of shock.6

Hypotension remains a hallmark of critical illness.1611 Although BP is routinely measured in clinical practice, this is the first large study to systematically document the value in considering hypotension for the prediction of in-hospital death in ED patients with no history of trauma. We found that a single episode of ED hypotension confers significantly increased risk of death during hospitalization, and the more severe and prolonged the hypotension, the higher the risk of death.

Footnotes

Abbreviations: CI = confidence interval; ED = emergency department; OR = odds ratio; SBP = systolic BP

None of the authors have any conflicts of interest to report.

Dr. Jones and Dr. Kline had full access to all of the data in the study, and take responsibility for the integrity of the data and the accuracy of the data analysis.

This work was supported by a Career Development Award from the Emergency Medicine Foundation and Carolinas Health Services Foundation to Dr. Jones. The study sponsors had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.

Received for publication March 25, 2006. Accepted for publication May 15, 2006.

References

  1. Jones, AE, Stiell, IG, Nesbitt, LP, et al (2004) Non-traumatic out-of-hospital hypotension predicts in-hospital mortality. Ann Emerg Med 43,106-113[CrossRef][ISI][Medline]
  2. Jones, AE, Tayal, VS, Sullivan, DM, et al Randomized controlled trial of immediate versus delayed goal-directed ultrasound to identify the cause of nontraumatic hypotension in emergency department patients. Crit Care Med 2004;32,1703-1708[CrossRef][ISI][Medline]
  3. Kline, JA, Johnson, CL, Webb, WB, et al Prospective study of clinician-entered research data in the emergency department using an internet-based system after the HIPAA privacy rule. BMC Med Inform Decis Mak 2004;4,17[CrossRef][Medline]
  4. Kline, JA, Webb, WB, Jones, AE, et al Impact of a rapid rule-out protocol for pulmonary embolism on the rate of screening, missed cases, and pulmonary vascular imaging in an urban US emergency department. Ann Emerg Med 2004;44,490-502[CrossRef][ISI][Medline]
  5. Efron, B, Tibshirani, R Improvements on cross-validation: the .632+ bootstrap method. Am Stat Assoc 1997;92,548-560[CrossRef]
  6. Jones, AE, Aborn, LS, Kline, JA Severity of emergency department hypotension predicts adverse hospital outcome. Shock 2004;22,410-414[ISI][Medline]
  7. Lee, KL, Woodlief, LH, Topol, EJ, et al Predictors of 30-day mortality in the era of reperfusion for acute myocardial infarction: results from an international trial of 41,021 patients: GUSTO-I Investigators. Circulation 1995;91,1659-1668[Abstract/Free Full Text]
  8. Goldhaber, SZ, Visani, L, De Rosa, M Acute pulmonary embolism: clinical outcomes in the International Cooperative Pulmonary Embolism Registry (ICOPER). Lancet 1999;353,1386-1389[CrossRef][ISI][Medline]
  9. Bernard, GR, Vincent, JL, Laterre, PF, et al Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344,699-709[Abstract/Free Full Text]
  10. Jones, AE, Fitch, MT, Kline, JA Operational performance of validated physiologic scoring systems for predicting in-hospital mortality among critically ill emergency department patients. Crit Care Med 2005;33,974-978[CrossRef][ISI][Medline]
  11. Moore, CL, Rose, GA, Tayal, VS, et al Determination of left ventricular function by emergency physician echocardiography of hypotensive patients. Acad Emerg Med 2002;9,186-193[CrossRef][ISI][Medline]



This article has been cited by other articles:


Home page
 ChestHome page
S. Trzeciak and M. E. Chansky
"Vital" sign?
Chest, October 1, 2006; 130(4): 933 - 934.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jones, A. E.
Right arrow Articles by Kline, J. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jones, A. E.
Right arrow Articles by Kline, J. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS