saem poster new cpr sepsis 2012

1
Sepsis is a common encounter disease in ED with high mortality. While several clinical prediction rules (CPR) including Mortality in the ED sepsis score (MEDS), Systemic inflammatory response syndrome (SIRS), and Confusion Urea Nitrogen Respi- ratory Rate Blood Pressure 65 Years or Older (CURB-65) existed to facilitate clinicians in early recognition of risk of mortality for sepsis, most are of suboptimal performance. To derive a novel CPR for mortality of sepsis utilizing clinically available and objective predictors in ED. We derived a novel clinical prediction rule for mortality owing to sepsis utilizing history of malignancy, absence of chills, hypothermia, fewer RBC counts, and wider RDW with better performance of existing CPRs. Further validation study is merited. Kuan-Fu Chen MD PhD 1 , Chun-Kuei Chen MD 2 , Sheng-Che Lin MD 2 , Peng-Hui Liu MD 2 , Jih-Chang Chen MD 2 , Te-Fa Chiu MD 2 , Yi-Lin Chan MD 2 , 1 Chang-Gung Memorial Hospital, Keelung City, Taiwan; 2 Chang-Gung Memorial Hospital, Taoyuan county, Taiwan; Contact Information: Kuan-Fu Chen, MD, PhD; phone: 886-975-360714; email: [email protected] INTRODUCTION OBJECTIVES CURRENT CONCLUSIONS RESULTS Comparison of a Novel Clinical Prediction Rule, MEDS, SIRS, and CURB-65 in the Prediction of Hospital Mortality for Septic Patients Visited the Emergency Department METHODS STUDY DESIGN Retrospectively cohort study consisted of all adult septic patients visited the ED at a tertiary hospital during 2010 with two sets of blood culture ordered by physicians. MEASUREMENT Basic demographics, ED vital signs, symptoms and signs, underlying illnesses, laboratory findings, microbiological results and discharge status were collected via electronic chart review. Microsoft Access SQL & Stata were applied. STATISTICAL ANALYSIS Multivariate logistic regressions were utilized to obtain a novel CPR using predictors with < 0.05 p- value tested in univariate analyses. Performance of the novel CPR were compared with existed CPR, i.e. MEDS, SIRS, and CURB-65, using AUC. Definition of sepsis could be difficult in retrospective studies, however subgroup analyses did not reveal huge deviation of the results. Many predictors could not be obtained or clarified owing to the nature of this retrospective cohort study. LIMITATIONS RESULTS Table 1 baseline characteristic of patients Figure 1. Performance of the CPRs Figure 2. Mortality rate by novel CPR DISCUSSIONS RESULTS Of 7111 included patients, 7% died in hospital, 53% had diabetes, 49% were older than 65 years of age, 20% had malignancy, and 20% had positive blood bacterial culture tests. Predisposing factors including history of malignancy, liver disease, male gender, no history of diabetes , and older than 65 years of age were found to be associated with mortality (all p < 0.05). Variables with clinical significant was highlighted. Table 5. Final predictors in the novel CPR Predictors OR SEM [95% C.I.] p-value History of malignancy 2.00 0.21 (1.63-2.46) <0.001 No complaints of chills 2.28 0.31 (1.75-2.97) <0.001 Hypothermia (<36°c) 2.12 0.32 (1.57-2.85) <0.001 Low RBC counts (< 4 million/uL) 2.45 0.27 (1.97-3.04) <0.001 RDW > 14.5% 3.27 0.36 (2.63-4.05) <0.001 Table 1. Basic Demographic overall Mortality survivor Median IQR Median IQR Median IQR p-value or N or % or N or % or N or % Age 65 (49-78) 72 (56-81) 64 (49-77) <0.001 > 65 years 3422 48.8% 117 53.9% 1752 44.8% 0.009 Male 3795 54.1% 139 64.1% 2118 54.2% 0.005 Nursing home 207 3.0% 3 1.4% 103 2.6% 0.256 Diabetes 3721 53.1% 91 41.9% 1978 50.6% 0.013 Asthma 308 4.4% 6 2.8% 188 4.8% 0.166 COPD 352 5.0% 11 5.1% 226 5.8% 0.659 CHF 286 4.1% 8 3.7% 127 3.3% 0.725 CKD 293 4.2% 6 2.8% 121 3.1% 0.783 Hemodialysis 480 6.8% 19 8.8% 220 5.6% 0.055 CVA 910 13.0% 16 7.4% 430 11.0% 0.094 Liver disease 689 9.8% 29 13.4% 363 9.3% 0.046 Malignancy 1385 19.8% 103 47.5% 774 19.8% <0.001 Chemotherapy 596 8.5% 51 23.5% 359 9.2% <0.001 Terminal stage 15 0.2% 1 0.5% 9 0.2% 0.502 Immunosuppressed 372 5.3% 10 4.6% 197 5.0% 0.776 Table 2. Vital signs at ED overall Mortality survivor Median IQR Median IQR Median IQR p-value or N or % or N or % or N or % Triage level 2 (2-3) 2 (2-2) 2 (2-3) <0.001 Body temperature 37.7 (36.8-385) 37 (36.3-38) 37.7 (36.8-38.5) <0.001 Pulse rate 104 (90-119) 108 (93-123) 103 (90-118) 0.001 Respiratory rate 20 (18-22) 22 (19-25) 20 (18-22) <0.001 Systolic BP 139 (117-159) 121 (99-148) 140 (118-160) <0.001 Diastolic BP 83 (71-94) 73 (59-88) 83 (72-95) <0.001 Mean arterial pressure 102 (87-115) 90 (73-107) 102 (89-115.) <0.001 Pulse pressure 54 (40-70) 46 (31-64) 55 (40-71) <0.001 Pulse Oximeter 96 (93-98) 93 (86-97) 96 (93-98) <0.001 Table 3. Laboratory Data overall Mortality Survivor Median IQR Median IQR Median IQR p-value or N or % or N or % or N or % Hemoglobin 12 (10.2-13.6) 10.1 (8.7-11.7) 12.1 (10.4-13.7) <0.001 Hematocrit 36.1 (31.2-40.4) 30.6 (26.8-35.6) 36.4 (31.6-40.7) <0.001 RDW 13.9 (13-15.4) 15.8 (14.3-17.6) 13.8 (13-15.2) <0.001 Platelet 204 (148-270) 168 (94-266) 205 (152-270) <0.001 WBC 11 (7.7-14.9) 11.9 (7.7-16.4) 10.9 (7.7-14.8) 0.018 Lymphocyte 10.5 (6-16.9) 9 (5-15) 10.8 (6.1-17) <0.001 Band 0 (0-1) 0.5 (0-3) 0 (0-1) <0.001 C-RP 66.25 (22.5-154.) 111.22 (45.8-198.) 65.14 (21.3-150.) <0.001 PCT 0.615 (0.09-3.06) 2.54 (0.44-23.4) 0.58 (0.09-2.32) 0.003 BUN 17.9 (11.2-32.5) 33.8 (18.6-63.7) 17.2 (11-30.2) <0.001 Creatinine 0.95 (0.72-1.41) 1.16 (0.81-2.86) 0.93 (0.72-1.36) <0.001 AST 34 (24-64) 62 (36-134) 33 (23-60) <0.001 BNP 451.5 (119-949) 445 (149-949) 458 (106-990) 0.509 Ammonia 71 (49-107) 79 (53-125) 69 (48-103) 0.045 Total bilirubin 1 (0.6-1.9) 1.7 (0.8-4.1) 0.9 (0.6-1.8) <0.001 aPTT 33 (28-39.9) 39.6 (31.5-53.7) 32.5 (27.8-38.8) <0.001 PT 13 (11.8-15.3) 15.9 (13.4-22.4) 12.8 (11.7-14.7) <0.001 Table 4. Miscellaneous overall Mortality Survivor Median IQR Median IQR Median IQR p-value or N or % or N or % or N or % SIRS 2 (1-3) 2 (2-3) 2 (1-3) <0.001 MEDS 5 (2-8) 8 (5-10) 5 (2-8) <0.001 SIRS >2 2354 33.6% 97 44.7% 1288 33.0% <0.001 Fever 6548 93.4% 198 91.2% 3719 95.2% 0.01 Chills 2030 29.0% 45 20.7% 1210 31.0% 0.001 ICU admission 458 6.5% 40 18.4% 178 4.6% <0.001 Bacteremia 1384 19.7% 59 27.2% 710 18.2% 0.001 Utilizing clinical available parameters might be able to facilitate clinicians in risk stratifying patient with sepsis in ED. RBC RDW found in our study as well as others indicated possible microcirculatory alterations behind the severe sepsis. Patients with chills had a significantly higher survival rate could be confounded by the fact of unmeasurable disease severity Our study indicated that hypothermia might not be protected in septic patients, which might again not support induced hypothermia therapy for sepsis.

Upload: kuan-fu-chen

Post on 21-Apr-2015

673 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: SAEM Poster New CPR Sepsis 2012

Sepsis is a common encounter disease in ED with high mortality. While several clinical prediction rules (CPR) including Mortality in the ED sepsis score (MEDS), Systemic inflammatory response syndrome (SIRS), and Confusion Urea Nitrogen Respi- ratory Rate Blood Pressure 65 Years or Older (CURB-65) existed to facilitate clinicians in early recognition of risk of mortality for sepsis, most are of suboptimal performance.

To derive a novel CPR for mortality of sepsis utilizing clinically available and objective predictors in ED.

We derived a novel clinical prediction rule for mortality owing to sepsis utilizing history of malignancy, absence of chills, hypothermia, fewer RBC counts, and wider RDW with better performance of existing CPRs. Further validation study is merited.

Kuan-Fu Chen MD PhD1, Chun-Kuei Chen MD2, Sheng-Che Lin MD2, Peng-Hui Liu MD2, Jih-Chang Chen MD2, Te-Fa Chiu MD2, Yi-Lin Chan MD2, 1Chang-Gung Memorial Hospital, Keelung City, Taiwan; 2Chang-Gung Memorial Hospital, Taoyuan county, Taiwan; Contact Information: Kuan-Fu Chen, MD, PhD; phone: 886-975-360714; email: [email protected]

INTRODUCTION

OBJECTIVES

CURRENT CONCLUSIONS

RESULTS

Comparison of a Novel Clinical Prediction Rule, MEDS, SIRS, and CURB-65 in the Prediction of Hospital Mortality for Septic Patients Visited the Emergency Department

METHODS STUDY DESIGN Retrospectively cohort study consisted of all adult septic patients visited the ED at a tertiary hospital during 2010 with two sets of blood culture ordered by physicians. MEASUREMENT Basic demographics, ED vital signs, symptoms and signs, underlying illnesses, laboratory findings, microbiological results and discharge status were collected via electronic chart review. Microsoft Access SQL & Stata were applied. STATISTICAL ANALYSIS Multivariate logistic regressions were utilized to obtain a novel CPR using predictors with < 0.05 p-value tested in univariate analyses. Performance of the novel CPR were compared with existed CPR, i.e. MEDS, SIRS, and CURB-65, using AUC.

•  Definition of sepsis could be difficult in retrospective studies, however subgroup analyses did not reveal huge deviation of the results.

•  Many predictors could not be obtained or clarified owing to the nature of this retrospective cohort study.

LIMITATIONS

RESULTS

Table 1 baseline characteristic of patients

Figure 1. Performance of the CPRs Figure 2. Mortality rate by novel CPR

DISCUSSIONS

RESULTS Of 7111 included patients, 7% died in hospital, 53% had diabetes, 49% were older than 65 years of age, 20% had malignancy, and 20% had positive blood bacterial culture tests. Predisposing factors including history of malignancy, liver disease, male gender, no history of diabetes , and older than 65 years of age were found to be associated with mortality (all p < 0.05).

Variables with clinical significant was highlighted.

Table 5. Final predictors in the novel CPR Predictors OR SEM [95% C.I.] p-value History of malignancy 2.00 0.21 (1.63-2.46) <0.001

No complaints of chills 2.28 0.31 (1.75-2.97) <0.001

Hypothermia (<36°c) 2.12 0.32 (1.57-2.85) <0.001

Low RBC counts (< 4 million/uL) 2.45 0.27 (1.97-3.04) <0.001

RDW > 14.5% 3.27 0.36 (2.63-4.05) <0.001

Table 1. Basic Demographic overall   Mortality   survivor    

  Median IQR Median IQR Median IQR p-value   or N or % or N or % or N or %  

Age 65 (49-78) 72 (56-81) 64 (49-77) <0.001 > 65 years 3422 48.8% 117 53.9% 1752 44.8% 0.009

Male 3795 54.1% 139 64.1% 2118 54.2% 0.005 Nursing home 207 3.0% 3 1.4% 103 2.6% 0.256 Diabetes 3721 53.1% 91 41.9% 1978 50.6% 0.013 Asthma 308 4.4% 6 2.8% 188 4.8% 0.166 COPD 352 5.0% 11 5.1% 226 5.8% 0.659 CHF 286 4.1% 8 3.7% 127 3.3% 0.725 CKD 293 4.2% 6 2.8% 121 3.1% 0.783

Hemodialysis 480 6.8% 19 8.8% 220 5.6% 0.055 CVA 910 13.0% 16 7.4% 430 11.0% 0.094 Liver disease 689 9.8% 29 13.4% 363 9.3% 0.046 Malignancy 1385 19.8% 103 47.5% 774 19.8% <0.001

Chemotherapy 596 8.5% 51 23.5% 359 9.2% <0.001 Terminal stage 15 0.2% 1 0.5% 9 0.2% 0.502

Immunosuppressed 372 5.3% 10 4.6% 197 5.0% 0.776  

Table 2. Vital signs at ED overall   Mortality   survivor    

  Median IQR Median IQR Median IQR p-value   or N or % or N or % or N or %  

Triage level 2 (2-3) 2 (2-2) 2 (2-3) <0.001 Body temperature 37.7 (36.8-385) 37 (36.3-38) 37.7 (36.8-38.5) <0.001 Pulse rate 104 (90-119) 108 (93-123) 103 (90-118) 0.001 Respiratory rate 20 (18-22) 22 (19-25) 20 (18-22) <0.001 Systolic BP 139 (117-159) 121 (99-148) 140 (118-160) <0.001 Diastolic BP 83 (71-94) 73 (59-88) 83 (72-95) <0.001 Mean arterial pressure 102 (87-115) 90 (73-107) 102 (89-115.) <0.001 Pulse pressure 54 (40-70) 46 (31-64) 55 (40-71) <0.001 Pulse Oximeter 96 (93-98) 93 (86-97) 96 (93-98) <0.001

Table 3. Laboratory Data overall   Mortality   Survivor    

  Median IQR Median IQR Median IQR p-value   or N or % or N or % or N or %  

Hemoglobin 12 (10.2-13.6) 10.1 (8.7-11.7) 12.1 (10.4-13.7) <0.001 Hematocrit 36.1 (31.2-40.4) 30.6 (26.8-35.6) 36.4 (31.6-40.7) <0.001 RDW 13.9 (13-15.4) 15.8 (14.3-17.6) 13.8 (13-15.2) <0.001 Platelet 204 (148-270) 168 (94-266) 205 (152-270) <0.001 WBC 11 (7.7-14.9) 11.9 (7.7-16.4) 10.9 (7.7-14.8) 0.018 Lymphocyte 10.5 (6-16.9) 9 (5-15) 10.8 (6.1-17) <0.001 Band 0 (0-1) 0.5 (0-3) 0 (0-1) <0.001 C-RP 66.25 (22.5-154.) 111.22 (45.8-198.) 65.14 (21.3-150.) <0.001 PCT 0.615 (0.09-3.06) 2.54 (0.44-23.4) 0.58 (0.09-2.32) 0.003 BUN 17.9 (11.2-32.5) 33.8 (18.6-63.7) 17.2 (11-30.2) <0.001 Creatinine 0.95 (0.72-1.41) 1.16 (0.81-2.86) 0.93 (0.72-1.36) <0.001 AST 34 (24-64) 62 (36-134) 33 (23-60) <0.001 BNP 451.5 (119-949) 445 (149-949) 458 (106-990) 0.509 Ammonia 71 (49-107) 79 (53-125) 69 (48-103) 0.045 Total bilirubin 1 (0.6-1.9) 1.7 (0.8-4.1) 0.9 (0.6-1.8) <0.001 aPTT 33 (28-39.9) 39.6 (31.5-53.7) 32.5 (27.8-38.8) <0.001 PT 13 (11.8-15.3) 15.9 (13.4-22.4) 12.8 (11.7-14.7) <0.001

Table 4. Miscellaneous overall   Mortality   Survivor    

  Median IQR Median IQR Median IQR p-value   or N or % or N or % or N or %  

SIRS 2 (1-3) 2 (2-3) 2 (1-3) <0.001 MEDS 5 (2-8) 8 (5-10) 5 (2-8) <0.001 SIRS >2 2354 33.6% 97 44.7% 1288 33.0% <0.001 Fever 6548 93.4% 198 91.2% 3719 95.2% 0.01 Chills 2030 29.0% 45 20.7% 1210 31.0% 0.001 ICU admission 458 6.5% 40 18.4% 178 4.6% <0.001 Bacteremia 1384 19.7% 59 27.2% 710 18.2% 0.001

•  Utilizing clinical available parameters might be able to facilitate clinicians in risk stratifying patient with sepsis in ED.

•  RBC RDW found in our study as well as others indicated possible microcirculatory alterations behind the severe sepsis.

•  Patients with chills had a significantly higher survival rate could be confounded by the fact of unmeasurable disease severity

•  Our study indicated that hypothermia might not be protected in septic patients, which might again not support induced hypothermia therapy for sepsis.