Relative tachycardia in patients with sepsis: an independent risk factor for mortality

Q J Med 2007; 100:629 634 doi:10.1093/qjmed/hcm074 Relative tachycardia in patients with sepsis: an independent risk factor for mortality L. LEIBOVICI 1, A. GAFTER-GVILI 1, M. PAUL 1, N. ALMANASREH 2, E. TACCONELLI 3, S. ANDREASSEN 4, A.D. NIELSEN 4, U. FRANK 2 and R. CAUDA 3, for the TREAT Study Group From the 1 Department of Medicine E, Rabin Medical Center, Beilinson Campus, and Sackler Faculty of Medicine, Tel-Aviv University, Israel, 2 Department of Clinical Microbiology and Hospital Hygiene, Freiburg University Hospital. Freiburg, Germany, 3 Department of Infectious Diseases, Agostino Gemelli Hospital, Universitá Cattolica del Sacro Cuore School of Medicine, Roma, Italy, and 4 University Center for Model-based Medical Decision Support, Aalborg University, Aalborg, Denmark. Received 30 July 2006 and in revised form 17 June 2007 Summary Background: Excess activation of the sympathetic nervous system may be a risk factor for mortality in patients with the systemic inflammatory response syndrome (SIRS) or sepsis. Aim: To examine whether excessive tachycardia, relative to the degree of fever is an independent risk factor for death in patients with SIRS. Design: Prospective observational study. Setting: Departments of medicine in three university hospitals in Israel, Germany and Italy. Methods: We collected data for 3382 patients with SIRS, whether community- or hospital-acquired, 91% with sepsis, as part of an ongoing trial. Introduction Excess activation of the sympathetic nervous system is a risk factor for a fatal outcome in patients with congestive heart failure or post myocardial infarction, and blockade of beta-receptors reverses this effect. 1 3 An example of a similar but short-term effect is the protection that beta-blockers afford to people with cardiovascular risk factors who undergo non-cardiac surgery. 4 We hypothesized that excess activation of the sympathetic nervous system might be a risk factor Results: Overall 30-day mortality was 12% (408/ 3382). The pulse/temperature ratio was significantly higher in patients who died than in survivors: mean SD 2.55 0.57 vs. 2.40 0.48 bpm/ C (p < 0.0001). Excessive tachycardia was significantly associated with a mortality in a logistic model accounting for other strong predictors of mortality (OR 1.54, 95%CI 1.10 2.17). Patients with septic shock were the only group for whom this association did not hold. Discussion: Our data are compatible with the hypothesis that some patients with sepsis experience an excess activation of the sympathetic nervous system, leading to a fatal outcome. for a fatal outcome in patients with systemic inflammatory response syndrome or sepsis as well. The hypothesis is supported by several reports in the literature. 5 9 A simple way to test the hypothesis is to examine whether tachycardia disproportionate to fever is an independent risk factor for a fatal outcome in febrile patients. We used a large, prospective database on patients suspected of harbouring a bacterial infection to address this question. Address correspondence to Dr L. Leibovici, Department of Medicine E, Beilinson Campus, Petah-Tiqva 49100, Israel. email: leibovic@post.tau.ac.il! The Author 2007. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

630 L. Leibovici et al. Methods We analysed a multi-national database of patients for which data were collected in the following institutions and departments: Rabin Medical Center, Beilinson Campus (Israel), six departments of internal medicine (240 beds); University Hospital of Freiburg (Germany), two gastroenterology, two nephrology, and two intensive care wards (94 beds); and A. Gemelli University Hospital (Italy), three infectious disease wards (60 beds). Patients were enrolled as part of a two-phase study (observational and interventional) designed to evaluate the effectiveness of TREAT, a decision support system for antibiotic treatment of common bacterial infections in in-patients. 10,11 Data were collected during the following time periods: between June and December 2002 in Israel and Germany, and between March and September 2003 in Italy (observational phase); and between May and November 2004 at all the three sites (randomized controlled trial). Research ethics committees in the three sites approved the study protocols. Inclusion and exclusion criteria We included patients who fulfilled the systemic inflammation response syndrome diagnostic criteria; 12,13 patients with a focus of infection; patients with shock compatible with septic shock; patients with febrile neutropenia; patients prescribed antibiotics (not for prophylaxis); and patients from whom blood cultures were drawn. We excluded HIV-positive patients with a current (suspected or identified) opportunistic disease and/or AIDSdefining illness currently or within the past 6 months; solid-organ or bone-marrow transplant recipients; children <18 years; suspected travel infections or tuberculosis; and pregnant women. Patients fulfilling inclusion criteria were prospectively identified by daily chart review. Within hours of admission (for a community-acquired infection) or of suspicion (in hospital-acquired ones) we collected data on: demography (e.g. age, sex, place of infection acquisition); background conditions (e.g. diabetes mellitus, chronic obstructive pulmonary disease, malignancy, chronic heart failure, chronic and acute renal failure, acute coronary syndrome, immunodeficiency); predisposing conditions (e.g. recent surgery) and devices (e.g. urinary catheter, intravenous catheter); presence of chills, temperature, pulse rate, systolic and diastolic blood pressure; focal signs and symptoms (e.g. cough, vomiting, rash); all available routine laboratory data (e.g. blood count, creatinine, urea, electrolytes, liver function tests). At follow-up, 30 days after recruitment, we collected data on survival, final diagnosis, duration of hospital stay, fever days, duration of stay in the intensive care unit, treatment, adverse events and all microbiological results. Definitions and outcomes As a measure of relative tachycardia, we used the ratio of heartbeats per minute divided by the temperature in degrees Celsius. We used the highest quartile of the ratio to designate excessive tachycardia (heart rate/temperature ratio 42.71 bpm/ C). The measures used were the first obtained on admission to hospital (in community acquired cases); and closest to the start of the episode in hospital acquired cases. Mortality was defined as all-cause mortality at 30 days following the first encounter with the patient for the infection. Septic shock was defined as sepsis with hypotension despite adequate fluid resuscitation, with the presence of perfusion abnormalities, including, but not limited to: lactic acidosis, oliguria, or an acute alteration in mental status. 13 Functional capacity was measured on a scale of 0 3: 0, full functional capacity; 1, limited; 2, limited in daily life activities; 3, bedridden. Statistical analysis Categorical explanatory variables were compared with the outcome of mortality using Fisher s exact test or the w 2 test, and continuous explanatory variables were compared using Student s t test or the Mann-Whitney U test, as appropriate. To compare the odds ratios between strata, we used the Breslow- Day test. To look for significant and independent predictors of a fatal outcome, we used a stepwise logistic regression analysis. Variables were entered into the model if their association with mortality was significant on univariate analysis at p < 0.05. Results The database included 3529 patients. Data on first measurement of heart rate and temperature, and on 30-day survival, were available in 3382 patients (96%), and the present report pertains to these patients. Temperature was measured orally in 2253 patients (mean 38.3 1.0 C), per rectum in 510 patients (mean 38.3 1.0 C), and at other sites in 619 (mean 38.2 1.2 C). The median age was 69 years, range 18 104, and 1546 (46%) were women. Sepsis criteria were fulfilled in 91% of patients. The 30-day mortality was 12% (408 of 3382 patients). The heart rate/temperature ratio was

Relative tachycardia in SIRS 631 Number of patients 800 700 600 500 400 300 200 100 0 1 2 3 4 Quartiles of heart rate/temperature ratio 30-day mortality Alive Dead Figure 1. Mortality in patients divided according to quartiles of the heart rate/temperature ratio. significantly higher in patients who died than in survivors (meansd 2.55 0.57 vs.2.40 0.48 bpm/ C, p < 0.0001). Of 843 patients with excessive tachycardia (heart rate/temperature ratio 42.71 bpm/ C), 148 (18%) died vs. 260 of the other 2539 (10%): p < 0.0001, odds ratio (OR) 1.87, 95%CI 1.50 2.32 (Figure 1). The difference was even more evident at 14 days after presentation, when 92/843 patients (10.9%) with excessive tachycardia had died vs. 12/2539 (4.8%) of the rest: p < 0.0001, OR 2.40, 95%CI 1.82 3.18. There was no significant difference in the mean heart rate/temperature ratio between the three sites (Israel, Italy and Germany). A description of the group and factors associated with 30-day mortality on univariate analysis are detailed in Table 1. Factors associated with mortality on univariate analysis (p < 0.05) were entered into a stepwise logistic regression analysis. Lower respiratory tract and any bacterial infection were entered separately into the model. Excessive tachycardia was included in the last model, significantly associated with a fatal outcome, OR 1.54 (95%CI 1.10 2.17) (Table 2). A logistic model in which the heart rate/temperature ratio was entered as a continuous variable looked similar. In this model, an increment of 1 bpm/ C in the heart rate/ temperature ratio was associated with an OR for mortality of 1.89 (95%CI 1.38 2.58). In a third model, we included an interaction member for temperature and heart rate instead of the ratio, and the interaction was significantly associated with a fatal outcome (OR 1.11, 95%CI 1.00 1.26), while the rest of the model was similar to the one shown in Table 2. In 1265 patients, O 2 saturation was available at onset of infection. In a logistic model for 30-day mortality similar to that shown in Table 2 but including O 2 saturation, the OR for mortality related to excessive tachycardia was 1.58 (95%CI 1.15 2.16). We used a stratified analysis to test the association between excessive tachycardia and a fatal outcome in strata defined by the levels of all variables detailed in Table 1. The only difference between strata was found for patients with septic shock compared to those without septic shock. Excessive tachycardia was not associated with a fatal outcome in patients suffering septic shock: the fatality rate in those with excessive tachycardia was 17/47 (45%), while in patients without excessive tachycardia the rate was 21/42 (50%). In patients without septic shock, the fatality rates were 16% (119/762) in patients with excessive tachycardia vs. 10% (230/ 2400) in patients without excessive tachycardia. The differences in ORs between strata were significant (Breslow-Day p ¼ 0.01). We constructed a logistic model in which components of the definition of septic shock (i.e. disseminated intravascular coagulopathy, blood pressure, and level of consciousness) were not entered, to allow septic shock to enter the model. In this model, we tested the two-way interactions between excessive tachycardia and septic shock, bacterial infection, congestive heart failure, acute coronary syndrome and age. Septic shock entered the final model with an OR of 5.22 (95%CI 2.29 11.88). The only interaction that was included in the final model was between excessive tachycardia and septic shock: OR 0.25, 95%CI 0.07 0.81. The OR for a fatal outcome associated with excessive tachycardia in this model was 1.66, 95%CI 1.23 2.24. Table 3 shows values of the heart rate/temperature ratio in survivors vs. patients who died, for strata of important confounders. The ratio was significantly higher in patients who died than in survivors in all the quartiles of temperature, showing that low temperature (or hypothermia) does not explain the association we describe. The same is shown for systolic blood pressure and urea, the two variables in the database associated with dehydration. Of note, the association of excessive tachycardia and mortality was no stronger in patients with bacterial infections, in patients with congestive heart failure, acute coronary syndrome, diabetes or in elderly patients (data not shown). Discussion In this large group of patients with SIRS (the large majority with sepsis), tachycardia out of proportion to temperature on one measurement

632 L. Leibovici et al. was significantly associated with an increased fatality rate. This association was independent of other strong predictors of a fatal outcome, but held true only in patients without septic shock. This association might be an epiphenomenon, even if emerging as independent on all the regression models we have tried. However, it is compatible with the hypothesis that some patients with SIRS experience an excess activation of the sympathetic nervous system, and that this by itself leads to a fatal outcome. The b1-selective blocker esmolol protected the myocardium in septic rats, 5 and the immunosuppression induced by catecholamines and their metabolic effects can be attenuated by beta-blockers. 7 9 Table 1 Characteristics of patients and factors associated with mortality Characteristic Whole group (n ¼ 3382) Survivors (n ¼ 2974) Deaths (n ¼ 408) Age (years) 69 (18 104) 68 (18 104) 78 (18 99) 0.0001 Underlying disorder Malignancy 21 19 35 0.0001 Acute coronary syndrome 5 4 8 0.003 Diabetes 23 22 27 0.02 Heart failure 17 15 27 0.0001 Chronic renal failure 14 12 22 0.0001 Endotracheal tube 4 3 12 0.0001 Aspiration 7 6 14 0.0001 Functional capacity 0 (0 2) 0 (0 2) 2 (0 2) 0.0001 Source of fever 0.0001 Lower respiratory tract 28 27 38 Urinary tract 16 17 12 Skin/soft tissues 8 9 6 Intra-abdominal 4 4 5 Other bacterial infections 15 14 22 Non-infectious 9 9 7 Other 20 20 10 Bacteraemia 10 10 15 0.001 Hospital-acquired 22 19 33 0.0001 On presentation Heart rate (bpm) 90 (20 177) 90 (36 170) 98 (20 177) 0.0001 Temperature ( C) 38.2 (26.7 42.0) 38.3 (26.7 42.0) 38.1 (33.0 41.0) 0.008 Ratio pulse rate/temperature (bpm/ C) 2.37 (0.41 4.60) 2.37 (0.41 4.60) 2.53 (0.53 4.32) 0.0001 Systolic blood pressure (mmhg) 125 (30 224) 127 (30 195) 121 (50 224) 0.02 Diastolic blood pressure (mmhg) 70 (0 128) 70 (0 128) 69 (24 122) 0.001 Respiratory rate (/min) 18 (5 45) 18 (5 40) 20 (10 45) 0.0001 Stupor or coma 10 9 23 0.0001 ARDS 1 1 2 0.014 DIC 1 1 3 0.004 Septic shock 3 2 10 0.0001 Dyspnoea 27 25 48 0.0001 Acute renal failure 12 10 22 0.0001 O 2 saturation a (%) 96 (67 100) 96 (67 100) 95 (72 100) 0.01 Glucose (mg/dl) 118 (40 1054) 116 (41 1054) 138 (40 975) 0.0001 Haematocrit 36 (17 55) 37 (21 51) 34 (17 55) 0.0001 Neutrophils (10 9 cells/l) 8.6 (0.0 51.0) 8.3 (0.0 51.0) 10.9 (0.0 36.0) 0.0001 Albumin (mg/dl) 3.5 (1.3 5.8) 3.5 (1.3 4.9) 2.9 (1.3 5.8) 0.0001 Creatinine (mg/dl) 0.9 (0.1 11.9) 0.9 (0.1 6.9) 1.1 (0.2 11.9) 0.0001 Urea (mg/dl) 38 (1 550) 35 (6 399) 66 (1 550) 0.0001 p Categorical variables are presented as percentages, continuous variables as median and range. a Data available for only 1244 patients. ARDS, adult respiratory distress syndrome; DIC, disseminated intravascular coagulation.

Relative tachycardia in SIRS 633 We expected to support this hypothesis by looking at subgroups of patients in which this association might be stronger, but found no such groups. This might be due to a limitation in the way data was collected. We did not gather detailed data on coronary disease, only on the presence of an acute coronary event and decompensated congestive heart failure. We do not have details on medications (other than antibiotic drugs), to test Table 2 Variables associated with mortality on logistic regression analysis: final model (intercept 5.12, Hosmer and Lemeshow w 2 3.01, 8 df, p ¼ 0.93) Variable Table 3 whether beta-blockers are protective and interact with the measurement of excessive tachycardia; or whether stopping beta-blockers was deleterious and led to tachycardia. We also did not collect data on fluid resuscitation and use of amines during treatment. We believe the association presented here should be further investigated. This simple index (pulse/ temperature) is a strong predictor of a fatal outcome, and should be tested in additional studies for its predictive value (in addition to other factors). The interaction between heart disease, beta-blockers, excessive tachycardia and a fatal outcome in patients with SIRS or sepsis should be better defined in prospective observational studies. Further trials in animals should address the question whether betablockers are truly protective and improve outcome in sepsis. If these studies support our hypothesis, a randomized controlled trial of beta-blockers in septic patients with excessive tachycardia (without shock) should be undertaken. 14 Acknowledgments The TREAT project was funded by the EU 5th framework, Information Society Technologies, contract no. IST-9999 11459. The funding source 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. Possible confounders for the relation between heart rate/temperature ratio and death: stratified analysis. Subgroup Mean SD heart rate/temperature ratio (bpm/ C) p Survivors OR (95%CI) Excessive tachycardia a 1.54 (1.10 2.17) Bacterial infection 2.49 (1.02 6.06) Stupor or coma 1.27 (1.01 1.60) Malignancy 2.66 (1.88 3.78) DIC b 5.16 (1.39 19.1) Dyspnoea 1.83 (1.32 2.53) Creatinine c 1.03 (1.00 1.06) Endotracheal tube 2.12 (1.12 4.01) Functional capacity c 1.41 (1.22 1.64) Age c 1.03 (1.02 1.04) Diastolic blood pressure c 0.67 (0.62 0.74) a Heart rate/temperature ratio 42.71 bpm/ C. b Disseminated intravascular coagulopathy. c Continuous variables, increments of 1 mg/dl for creatinine; 1 class for functional capacity; 1 year for age; 10 mmhg for diastolic blood pressure. Deaths Quartiles of temperature ( C) 1: <37.4 2.37 0.50 2.56 0.62 0.001 2: 537.4, <38.2 2.40 0.47 2.56 0.52 0.003 3: 538.2, <39.0 2.41 0.48 2.50 0.55 0.130 4: 539.0 2.42 0.45 2.56 0.57 0.020 Quartiles of systolic blood pressure (mmhg) 1: <110 2.42 0.50 2.54 0.64 0.034 2: 5110, <126 2.40 0.45 2.52 0.54 0.028 3: 5126, <142 2.39 0.47 2.53 0.50 0.004 4: 5142 2.38 0.48 2.60 0.60 0.001 Quartiles of blood urea (mg/dl) 1: <24 2.43 0.46 2.65 0.66 0.027 2: 524, <38 2.41 0.45 2.56 0.56 0.017 3: 538, <62 2.40 0.50 2.48 0.58 0.157 4: 562 2.36 0.48 2.56 0.55 <0.001

634 L. Leibovici et al. References 1. Brophy JM, Joseph L, Rouleau JL. Beta-blockers in congestive heart failure. A Bayesian meta-analysis. Ann Intern Med 2001; 134:550 60. 2. Squire I. Neurohormonal intervention to reduce sudden cardiac death in heart failure: what is the optimal pharmacologic strategy? Heart Fail Rev 2004; 9:337 45. 3. Freemantle N, Cleland J, Young P, Mason J, Harrison J. Beta blockade after myocardial infarction: systematic review and meta-regression analysis. Br Med J 1999; 318:1730 7. 4. McGory ML, Maggard MA, Ko CY. A meta-analysis of perioperative beta blockade: what is the actual risk reduction? Surgery 2005; 138:171 9. 5. Suzuki T, Morisaki H, Serita R, Yamamoto M, Kotake Y, Ishizaka A, Takeda J. Infusion of the beta-adrenergic blocker esmolol attenuates myocardial dysfunction in septic rats. Crit Care Med 2005; 33:2294 301. 6. Aderka D, Sclarovsky S, Lewin RF, Arditti A, Agmon J, Pinkhas J. Life-threatening ventricular arrhythmias in septicemia. Isr J Med Sci 1984; 20:535 7. 7. Bergmann M, Gornikiewicz A, Sautner T, Waldmann E, Weber T, Mittlbock M, Roth E, Fugger R. Attenuation of catecholamine-induced immunosuppression in whole blood from patients with sepsis. Shock 1999; 12:421 7. 8. Dickerson RN, Fried RC, Bailey PM, Stein TP, Mullen JL, Buzby GP. Effect of propranolol on nitrogen and energy metabolism in sepsis. J Surg Res 1990; 48:38 41. 9. Herndon DN, Barrow RE, Rutan TC, Minifee P, Jahoor F, Wolfe RR. Effect of propranolol administration on hemodynamic and metabolic responses of burned pediatric patients. Ann Surg 1988; 208:484 92. 10. Paul M, Andreassen S, Tacconelli E, Nielsen AD, Almanasreh N, Frank U, Cauda R, Leibovici L, on behalf of the TREAT Study Group. Improving empirical antibiotic treatment using TREAT, a computerised decision support system: cluster randomised trial. J Antimicrob Chemother 2006; 58:1238 45. 11. Paul M, Andreassen S, Nielsen AD, Tacconelli E, Almanasreh N, Fraser A, Yahav D, Ram R, Leibovici L; the TREAT Study Group. Prediction of Bacteremia Using TREAT, a Computerized Decision-Support System. Clin Infect Dis 2006; 42:1274 82. 12. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992; 101:1644 55. 13. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992; 20:864 74. 14. Meldrum DR. Beta-blockade during sepsis: Inspired or insane? Crit Care Med 2005; 33:2433 4.