Continuing Medical Education Activity in Academic Emergency Medicine

Transcription

1 CONTINUING MEDICAL EDUCATION Continuing Medical Education Activity in Academic Emergency Medicine CME Editor: Corey Heitz, MD Authors: Halden F. Scott, MD, Sara J. Deakyne, MPH, Jason M. Woods, MD, and Lalit Bajaj, MD, MPH Article Title: The Prevalence and Diagnostic Utility of Systemic Inflammatory Response Syndrome Vital Signs in a Pediatric Emergency Department If you wish to receive free CME credit for this activity, please refer to the website: Accreditation and Designation Statement: Blackwell Futura Media Services designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 CreditÔ. Physicians should only claim credit commensurate with the extent of their participation in the activity. Blackwell Futura Media Services is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. Educational Objectives After completing this exercise the participant will be better able to discuss the utility of using triage SIRS criteria to identify pediatric patients at risk of critical illness. Activity Disclosures No commercial support has been accepted related to the development or publication of this activity. Faculty Disclosures: CME Editor: Corey Heitz, MD has no relevant financial relationships to disclose. Authors: Halden F. Scott, MD, Sara J. Deakyne, MPH, Jason M. Woods, MD, and Lalit Bajaj, MD, MPH have no relevant financial relationships to disclose. This manuscript underwent peer review in line with the standards of editorial integrity and publication ethics maintained by Academic Emergency Medicine. The peer reviewers have no relevant financial relationships. The peer review process for Academic Emergency Medicine is double-blinded. As such, the identities of the reviewers are not disclosed in line with the standard accepted practices of medical journal peer review. Conflicts of interest have been identified and resolved in accordance with Blackwell Futura Media Services s Policy on Activity Disclosure and Conflict of Interest. Instructions on Receiving Free CME Credit For information on applicability and acceptance of CME credit for this activity, please consult your professional licensing board. This activity is designed to be completed within an hour; physicians should claim only those credits that reflect the time actually spent in the activity. To successfully earn credit, participants must complete the activity during the valid credit period, which is up to two years from initial publication. Follow these steps to earn credit: Log on to Read the target audience, educational objectives, and activity disclosures. Read the article in print or online format. Reflect on the article. Access the CME Exam, and choose the best answer to each question. Complete the required evaluation component of the activity. This activity will be available for CME credit for twelve months following its publication date. At that time, it will be reviewed and potentially updated and extended for an additional twelve months.

2 ORIGINAL CONTRIBUTION The Prevalence and Diagnostic Utility of Systemic Inflammatory Response Syndrome Vital Signs in a Pediatric Emergency Department Halden F. Scott, MD, Sara J. Deakyne, MPH, Jason M. Woods, MD, and Lalit Bajaj, MD, MPH Abstract Objectives: This study sought to determine the prevalence, test characteristics, and severity of illness of pediatric patients with systemic inflammatory response syndrome (SIRS) vital signs among pediatric emergency department (ED) visits. Methods: This was a retrospective descriptive cohort study of all visits to the ED of a tertiary academic free-standing pediatric hospital over 1 year. Visits were included if the patient was <18 years of age and did not leave before full evaluation or against medical advice. Exclusion criteria were trauma diagnoses or missing documentation of vital signs. Data were electronically extracted from the medical record. The primary predictor was presence of vital signs meeting pediatric SIRS definitions. Specific vital sign pairs comprising SIRS were evaluated as predictors (temperature heart rate, temperature respiratory rate, and temperature-corrected heart rate, in which a formula was used to correct heart rate for degree of temperature elevation). The primary outcome measure was requirement for critical care (receipt of a vasoactive agent or intubation) within 24 hours of ED arrival. Results: There were 56,210 visits during the study period; 40,356 met inclusion criteria. Of these, 6,596 (16.3%) visits had fever >38.5 C, and 6,122 (15.2% of included visits) met SIRS vital sign criteria. Among included visits, those with SIRS vital signs accounted for 92.8% of all visits with fever >38.5 C. Among patients with SIRS vital signs, 4993 (81.6%) were discharged from the ED without intravenous (IV) therapy and without 72-hour readmission. Critical care within the first 24 hours was present in 99 (0.25%) patients: 23 patients with and 76 without SIRS vital signs. Intensive care unit (ICU) admission was present in 126 (2.06%) with SIRS vital signs and 487 (1.42%) without SIRS vital signs. SIRS vital signs were associated with increased risk of critical care within 24 hours (relative risk [RR] = 1.69, 95% confidence interval [CI] = 1.06 to 2.70), ICU admission (RR = 1.45, 95% CI = 1.19 to 1.76), ED laboratory tests (RR = 1.41, 95% CI = 1.37 to 1.45), ED IV medication/fluid administration (RR = 2.54, 95% CI = 2.29 to 2.82), hospital admission (RR = 1.52, 95% CI = 1.42 to 1.63), and 72-hour readmission (RR = 1.31, 95% CI = 1.01 to 1.69). SIRS vital signs were not associated with 30-day in-hospital mortality (RR = 0.37, 95% CI = 0.05 to 2.82). SIRS vital signs had a low sensitivity for critical care requirement (23.2%, 95% CI = 15.3% to 32.8%). The pair of SIRS vital signs with the strongest association with critical care requirement was temperature and corrected heart rate (positive likelihood ratio = 2.74, 95% CI = 1.87 to 4.01). Conclusions: Systemic inflammatory response syndrome vital signs are common among medical pediatric patients presenting to an ED, and critical illness is rare. The majority of patients with SIRS vital signs were discharged without IV therapy and without readmission. Patients with SIRS vital signs had a statistically significant increased risk of critical care requirement, ED IV treatment, ED laboratory tests, admission, and readmission. However, SIRS vital sign criteria did not identify the majority of patients with mortality or need for critical care. SIRS vital signs had low sensitivity for critical illness, making it poorly suited for use in isolation in this setting as a test to detect children requiring sepsis resuscitation. From the Department of Pediatrics, Section of Emergency Medicine, Children s Hospital Colorado, University of Colorado School of Medicine (HFS, LB), Aurora, CO; Research Informatics, Research Institute, Children s Hospital Colorado (SJD), Aurora, CO; and Children s National Health System, The George Washington University School of Medicine (JMW), Washington, DC. Received June 16, 2014; revisions received September 30 and October 15, 2014; accepted October 15, This study was presented at the Pediatric Academic Societies Annual Meeting, Washington, DC, May The authors have no relevant financial information or potential conflicts to disclose. Supervising Editor: Michelle L. Macy, MD, MS. Address for correspondence and reprints: Halden F. Scott, MD; Halden.scott@childrenscolorado.org. 382 ISSN by the Society for Academic Emergency Medicine 382 PII ISSN doi: /acem.12610

3 ACADEMIC EMERGENCY MEDICINE April 2015, Vol. 22, No ACADEMIC EMERGENCY MEDICINE 2015;22: by the Society for Academic Emergency Medicine Sepsis is a leading cause of pediatric death, with mortality rates between 8 and 10%. 1,2 Challenges in diagnosing compensated septic shock in children are an important barrier to timely, guideline-compliant care. 3 Although international guidelines for management of pediatric septic shock have been established, triage protocols for the early detection of septic shock in the emergency department (ED) are not included. 4 Systemic inflammatory response syndrome (SIRS) is part of the formal definition of sepsis and has been successfully used in the adult Surviving Sepsis campaign to initiate sepsis protocols in adults, although SIRS may be excessively sensitive for detecting septic shock in the adult ED setting. 5,6 SIRS has been suggested for use as a screening strategy to identify pediatric patients at risk for septic shock requiring resuscitative care; however, the prevalence of SIRS in children presenting to EDs has been estimated to be as high as 20%, and its effectiveness as a screening test for detecting critically ill children with sepsis is unknown. 7 Consensus pediatric SIRS definitions require the presence of two or more of the following, one of which must be abnormal temperature or leukocyte count: core temperature <36 or >38.5 C, tachycardia (or bradycardia in infants), tachypnea, leukocyte count elevated or depressed for age, or >10% immature neutrophils (Data Supplement S1, available as supporting information in the online version of this paper). 8 A modification to this approach includes use of a formula to correct heart rate for temperature elevations, but this has not been included in formal definitions. 9,10 Additionally, EDs have adopted screening approaches incorporating only the vital sign criteria from SIRS, because leukocyte counts are not known at the time of triage. 10,11 Prevalence and clinical outcomes of patients with SIRS vital signs are important to understand to appropriately triage and allocate resources. Their use to define sepsis suggests that they may define a cohort of patients more likely to require critical care, laboratory testing, and intravenous (IV) treatments. To date, studies have not examined the rates of such clinical interventions among children with SIRS vital signs presenting to pediatric EDs. This study sought to determine the prevalence of SIRS vital signs, to determine the severity of illness of patients with SIRS vital signs, and to examine the test characteristics of SIRS vital signs for identifying critical illness among pediatric medical ED visits. The hypothesis was that children with SIRS vital signs would have higher rates of hospitalization and critical illness when compared with all other children without SIRS vital signs presenting for medical (nontrauma) complaints to the ED. Additionally, the study sought to determine whether specific combinations of SIRS vital sign criteria improved the predictive value of SIRS vital signs. METHODS Study Design This was a retrospective descriptive cohort study. To describe the epidemiology of SIRS vital signs, 12 months of visits were evaluated. This study was approved by the Colorado Multiple Institutions Review Board under expedited review status, and informed consent was waived. Study Setting and Population All visits to the ED of a tertiary academic free-standing pediatric hospital between April 2011 and March 2012 were screened for inclusion in this study. This period corresponds to the year immediately preceding implementation of vital sign SIRS-based screening alerts in the electronic medical record and enhanced sepsis care strategies. There were no education programs or protocols related to sepsis and SIRS at the time of this study. Visits were included for analysis if the patient was <18 years of age and the patient did not leave before full evaluation or against medical advice. Exclusion criteria were as follows: visits with an isolated trauma diagnosis as identified by an injury or poisoning ICD-9-CM code ( , exclusive of the SIRS ICD-9-CM code 995.9x, which was included) and missing documentation of any of the SIRS vital signs (heart rate, respiratory rate, or temperature). Study Protocol Visits were identified in the electronic health record (Epic, Verona, WI). Deidentified clinical data, billing codes, and demographic data from these visits were electronically extracted from the medical record using Crystal Reports 2008 (SAP SE, Waldorf, Germany). Data were exported and loaded into SAS 9.3 for analysis. Clinical data included both ED and inpatient clinical course. The data elements included vital signs, medication and fluid administration, hospital admission, transfer, and discharge, 72-hour readmissions to an inpatient service, and vital status. Vital status was ascertained based on data within the electronic health record, and vital status data up until April 30, 2012, was included to fully ascertain 30-day mortality. Billing data were used to identify ICD-9 codes and ventilator charges. Demographics included patient age at encounter, sex, and insurance status. When multiple visits by the same subject were made within 72 hours, only the first visit was included as an index case. Additional visits within the study time period were evaluated as separate visits. Measures The primary predictor was presence of SIRS vital signs recoded in the ED medical record. SIRS vital signs were defined as any two concurrent vital signs meeting SIRS criteria during the ED encounter, one of which was temperature. 8 SIRS vital signs were analyzed as predictors, because they represent the only data known on a

4 384 Scott et al. SIRS IN THE PEDIATRIC ED patient in triage before laboratory results are available, which matches ED practice environments. Each vital sign comprising SIRS was individually analyzed and vital sign pairs that comprise SIRS were evaluated as predictors (temperature heart rate, temperature respiratory rate). In addition, a temperature heart rate correction was performed following published practices (for each 1 C above 38.5 C, 10 beats/min was subtracted from the heart rate); corrected heart rate SIRS represents patients whose heart rate met SIRS criteria after temperature correction. 9,12 Patients meeting hypothermia and bradycardia criteria were described separately from patients meeting fever and tachycardia criteria, but all alterations in heart rate (tachycardia and bradycardia) and temperature (hypothermia or fever) were considered together in analysis. We additionally examined the presence of a heart rate meeting SIRS criteria on final ED vital signs as a predictor variable and compared it both to the entire population and to the population of patients with more than one set of vital signs recorded. The primary outcome measure was requirement for critical care within 24 hours of ED arrival. Critical care was defined as receipt of a vasoactive agent infusion or intubation, to avoid the potential influence of institutionspecific criteria for intensive care unit (ICU) admission. Secondary outcomes were ICU admission from the ED, admission, 30-day in-hospital mortality, 72-hour readmission to an inpatient service, ED laboratory evaluation (including point-of-care tests), and ED IV therapy. Complex chronic conditions were identified by presence of validated ICD-9 codes in the visit s billing record. 13 Data Analysis Data were analyzed using SAS 9.3. Data were summarized using standard descriptive statistics. All results were summarized including 95% confidence intervals (CIs) and p-values where appropriate, and p-values < 0.05 were considered significant. Proportions were compared using chi-square tests. Relative risks (RRs) for outcomes, including the primary outcome of critical care requirement, in patients with SIRS vital signs compared to those without SIRS vital signs were calculated. Test characteristics of SIRS vital signs for the primary outcome of critical care requirement, including sensitivity, specificity, positive likelihood ratio (LR+), and positive and negative predictive value (PPV, NPV), were assessed. To assess the predictive ability of SIRS vital signs when adjusted for potentially confounding variables, multivariate logistic regression was used. The covariates complex chronic condition and age were empirically chosen a priori and forced to remain in the model because of data availability, accessibility to clinicians, and use in published sepsis triage strategies Multivariate logistic regression analysis was planned for the predictor any SIRS vital signs and for the most predictive pair of vital signs and used to evaluate the primary outcome of critical care requirement and ICU admission. A subgroup analysis was planned of patients with complex chronic conditions, and by age groups. Characteristics of patients with the outcome of critical care requirement were described and compared between those with and without SIRS vital signs. RESULTS In the 1-year study period there were 56,210 eligible complete visits to the ED. The following visits were excluded: 13,135 with trauma diagnoses, 951 missing SIRS vital sign values, and 1,768 with ED visits or admission in the preceding 72 hours. The resulting 40,356 visits comprised the data set used for analysis (Figure 1). In the study population, 6,596 (16.3%) patients had fever >38.5 C; 6,122 (15.2%) met SIRS vital sign criteria. Thus, 92.8% of patients with fever > 38.5 C had SIRS vital signs. Characteristics of patients with SIRS are compared to all patients in Table 1. Sex, payer status, and presence of chronic complex conditions were similar in the SIRS vital signs and non-sirs vital signs populations. Children in the 1- to 4-years age groups comprised a significantly larger proportion of the SIRS vital signs population when compared to the non-sirs vital sign population, and those in the 10- to 14-years Assessed for eligibility (n = 56,210) Excluded (n = 15,854) Trauma diagnosis (n = 13,135) SIRS vital signs missing (n =951) ED visit in preceding 72 hours (n = 1,478) ED + admission in preceding 72 hours (n =290) Analyzed (n = 40,356) Vital sign SIRS (n = 6,122) No vital sign SIRS (n =34,234) Figure 1. Study flow diagram. SIRS = systemic inflammatory response syndrome.

5 ACADEMIC EMERGENCY MEDICINE April 2015, Vol. 22, No and 15- to 17-years age groups comprised smaller proportions. Univariate analysis of ED testing and treatment patterns and clinical outcomes are shown in Table 2. These demonstrate a significantly increased RR of critical care requirement, ICU admission, ED laboratory tests, repeat vital signs, IV therapy, and admission and readmission in patients with SIRS vital signs compared to patients without SIRS vital signs. However, the overall proportion of patients with critical care and ICU admission outcomes was small in both SIRS vital signs and non- SIRS vital signs populations. Although the proportion of patients requiring critical care was minimally higher in the SIRS compared to the non-sirs vital signs groups, in absolute numbers, the majority of these patients were in the non-sirs vital signs group (Table 2). Test characteristics of SIRS vital signs for predicting critical care requirement are shown in Table 3. SIRS vital signs had a low sensitivity and PPV for critical care requirement, an outcome that occurred in only 0.25% of visits in our data set. Further analysis of specific combinations of SIRS vital signs demonstrates that the pair of temperature and corrected heart rate vital signs had the highest LR+ for predicting critical care requirement. There were three patients in the study population who met SIRS criteria with bradycardia, of whom one required critical care and two were admitted to the ICU. There were 79 patients who met SIRS criteria with hypothermia, of whom none required critical care and five were admitted to the ICU. Given the small numbers of patients in these categories, additional analyses of these subgroups were not undertaken. In multivariate analysis, including age category and chronic condition as categorical covariates, the presence of SIRS vital signs was significantly correlated with the primary outcome, critical care requirement (adjusted odds ratio [AOR] = 1.74, 95% CI = 1.08 to 2.80), and with the secondary outcome, ICU admission (AOR = 1.55, 95% CI = 1.27 to 1.90), as shown in Table 4. In multivariate analysis of the most predictive pair of vital signs, temperature and corrected heart rate, Table 1 Characteristics of Study, Any SIRS Vital Sign Pair, and Non-SIRS Vital Sign Populations Variable Population (N =40,356) Any SIRS Vital Sign Pair Present (n =6,122) No SIRS Vital Sign Pair Present (n =34,234) p-value Age (yr) <0.0001* <1 7,516 (18.62) 979 (15.99) 6,537 (19.10) ,679 (36.37) 3,234 (52.83) 11,445 (33.43) 5 9 8,634 (21.39) 1,283 (20.96) 7,351 (21.47) ,040 (14.97) 476 (7.78) 5,564 (16.25) ,487 (8.64) 150 (2.45) 3,337 (9.75) Male 21,019 (52.08) 3,186 (52.04) 17,833 (52.09) Chronic condition 6,470 (16.03) 1,033 (16.87) 5,437 (15.88) Medicaid/self-pay/CHP+ 29,911 (74.12) 4,585 (74.89) 25,326 (73.98) Data are reported as n (%). Percentages represent percentage of population listed in the column heading. p-values calculated with chi-square. CHP = Children s Health Program; SIRS = systemic inflammatory response syndrome. *Statistically significant. Table 2 Prevalence of Clinical Outcomes and Resource Utilization in Study, SIRS, and Non-SIRS Vital Signs Populations With Unadjusted RR (95% CI) Variable Population (N =40,356) Any SIRS Vital Sign Pair Present (n =6,122) No SIRS Vital Sign Pair Present (n =34,234) RR SIRS vs. Non-SIRS ED laboratory tests 16,338 (40.48) 3,293 (53.79) 13,045 (38.11) 1.41* ( ) At least two sets of ED vital 12,419 (30.77) 3,202 (52.30) 9,217 (26.92) 1.94* ( ) signs obtained ED IV medication/fluid 1,569 (3.89) 490 (8.00) 1,079 (3.15) 2.54* ( ) Hospital admission 4,125 (10.22) 881 (14.39) 3,244 (9.48) 1.52* ( ) ICU admission 613 (1.52) 126 (2.06) 487 (1.42) 1.45* ( ) Ventilation 79 (0.20) 14 (0.23) 65 (0.19) 1.20 ( ) Vasoactive agent 28 (0.07) 11 (0.18) 17 (0.05) 3.62* ( ) 24-hour critical care (vent/vasoactive) 99 (0.25) 23 (0.38) 76 (0.22) 1.69* ( ) 72-hour readmission 374 (0.93) 71 (1.19) 303 (0.88) 1.31* ( ) 30-day in-hospital mortality 16 (0.04) 1 (0.02) 15 (0.04) 0.37 ( ) All data reported as n (%) except RR (95% CI). ICU = intensive care unit; RR = relative risk; SIRS = systemic inflammatory response syndrome. *Statistically significant.

6 386 Scott et al. SIRS IN THE PEDIATRIC ED Table 3 Test Characteristics of SIRS Vital Signs as a Diagnostic Test for Critical Care Requirement (Defined as Intubation or Vasoactive Agent Use) Variable n (%) Sensitivity (95% CI) Specificity (95% CI) PPV (95% CI) NPV (95% CI) LR+ (95% CI) Patients with at least a single individual ED vital sign meeting SIRS criteria Temperature 6,596 (16.3) ( ) ( ) ( ) ( ) 1.48* ( ) HR 8,627 (21.4) ( ) ( ) ( ) ( ) 2.18* ( ) Respiratory rate 33,252 (82.4) ( ) ( ) ( ) ( ) 1.04 ( ) Patients with at least one pair of ED vital signs meeting SIRS criteria Temperature + HR 3,955 (9.8) ( ) ( ) ( ) ( ) 2.27* ( ) Temperature corrected HR 3,135 (7.8) ( ) ( ) ( ) ( ) 2.74* ( ) Temperature + respiratory 5,956 (14.8) ( ) ( ) ( ) ( ) 1.51* ( ) rate Any pair 6,122 (15.2) ( ) ( ) ( ) ( ) 1.53* ( ) Last ED vital signs with HR meeting SIRS criteria Last ED HR meets SIRS (all 3559 (8.82) ( ) ( ) ( ) ( ) 1.15 ( ) patients) Last ED HR meets SIRS 639 (5.2) ( ) ( ) ( ) ( ) 2.13* ( ) among all patients with at least two sets of recorded ED vitals (n =12,419) HR = heart rate; LR+ = positive likelihood ratio; NPV = negative predictive value; PPV = positive predictive value; SIRS = systemic inflammatory response syndrome. *Statistically significant. Table 4 Adjusted OR (95% CI) of Clinical Outcomes, Adjusted for Age and Presence of Chronic Condition Variable Critical Care ICU Admission ED vital-sign SIRS 1.74* ( ) GF = * ( ) GF = 0.09 Temperature and corrected heart rate 3.23* ( ) GF = * ( ) GF = 0.10 Hosmer and Lemeshow goodness-of-fit statistic for each model. GF = goodness of fit; ICU = intensive care unit; SIRS = systemic inflammatory response syndrome. *Statistically significant with age and chronic condition as covariates, the AOR for the outcomes of critical care and ICU admission were higher than using any pair of SIRS vital signs as a predictor (Table 4). Although the models with critical care as an outcome were well calibrated according to the Hosmer and Lemeshow goodness-of-fit statistic, there were many influential observations and significant variability unaccounted for in all four models. Patients with the primary outcome of critical care are described in Table 5. Patients with SIRS vital signs were more likely to require critical care due to a need for vasoactive agents, while patients without SIRS vital signs were more likely to require critical care due to a need for ventilation. SIRS vital signs were the first recorded ED vital signs in 82.1% of patients with SIRS vital signs. Among patients with SIRS vital signs, 52.3% had more than one set of vital signs recorded in the ED, compared to 26.9% of patients without SIRS vital signs (p < ). Of patients with more than one set of vital signs recorded, 5.15% demonstrated heart rates meeting SIRS criteria on their final recorded vital signs. These patients were more likely to require critical care than patients whose final heart rates did not meet SIRS criteria (Table 3). Subgroup analysis by age and chronic condition was performed to evaluate RR of clinical outcomes in patients with vital sign SIRS. The age subgroups did not demonstrate RRs that were different from those in the overall population for the outcomes evaluated: admission, critical care, 72-hour readmission, and 30-day in-hospital mortality. SIRS vital signs performed similarly in patients with complex chronic conditions compared with the general population: SIRS vital signs were associated with increased risk of admission (RR = 1.58, 95% CI = 1.42 to 1.75), but not 30-day in-hospital mortality (RR = 0.38, 95% CI = 0.05 to 2.86). Unlike the general population, SIRS vital signs were not associated with increased risk of critical care (RR = 1.53, 95% CI = 0.88 to 2.66) or 72-hour readmission (RR = 0.76, 95% CI = 0.47 to 1.25) in patients with complex chronic conditions. Only 10.7% of the study population had white blood cell counts measured. In patients with SIRS vital signs, 16.3% had white blood cell counts recorded, while in patients without SIRS vital signs, 9.7% of patients had

7 ACADEMIC EMERGENCY MEDICINE April 2015, Vol. 22, No Table 5 Characteristics of Subjects With a 24-Hour Critical Care Requirement (Intubation or Vasoactive Agents), Described by SIRS and Non-SIRS Vital Sign Populations Variable SIRS Vital Signs (n =23) Non-SIRS Vital Signs (n =76) p-value Age (yr) 4.0 (2.0 10) 4.0 (0 10.5) Hours of ICU stay 76 (52 144) 88 (41 268) Hours of hospital stay 210 ( ) 165 (69 290) Chronic complex condition 16 (69.6) 55 (72.4) Ventilation 14 (60.9) 65 (85.5) 0.010* Vasoactive agents 11 (47.8) 17 (22.4) 0.018* 30-day in-hospital mortality 0 (0) 2 (2.6) Data are reported as median (interquartile range) or n (%). Percentages represent percentage of population listed in the column heading. p-values calculated with Wilcoxon rank-sum for continuous variables, Fisher s exact, or chi-square test as appropriate for categorical variables. ICU = intensive care unit; IAR = interquartile range; SIRS = systemic inflammatory response syndrome. *Statistically significant. white blood cell counts tested, a difference that was significant (p < ). Among SIRS vital signs patients, 5,064 (82.7%) were discharged from the ED; 71 of these had IV therapy in the ED at the index visit or readmission to the hospital within 72 hours. This left 4,993 (81.6%) SIRS vital signs patients discharged from the ED without IV therapy and without 72-hour readmission. DISCUSSION This study demonstrated limitations to existing SIRS vital sign criteria as an isolated predictor of critical illness in an ED pediatric population. Contributing to these findings is the rare occurrence of critical illness and 30-day in-hospital mortality among patients presenting to this tertiary care pediatric ED for medical care and the high prevalence of SIRS vital signs. SIRS vital sign criteria have a low sensitivity for critical illness; they did not identify the majority of patients with mortality and critical care. If used as a screening tool, SIRS vital signs have potential to create a false sense of security when patients do not meet SIRS criteria. Additionally, the high prevalence of SIRS vital signs has potential to increase testing and treatment or lead to alert fatigue. It is noteworthy that 92% of patients with fever meet SIRS criteria and that temperature criteria alone demonstrate virtually identical test characteristics to any pair of SIRS vital signs. Even considering that early ED IV fluid and antibiotics might prevent patients from requiring critical care, IV therapies were only administered to 8% of patients with SIRS vital signs, leaving the overwhelming majority of SIRS vital sign patients who required neither IV treatment nor critical care. Given the low PPV of SIRS vital signs for critical illness, the low baseline rates of ED testing and IV treatment in these patients, and the high prevalence of SIRS vital signs in the pediatric ED population, caution should be used in mandating resuscitative protocols or specific testing requirements for pediatric patients with SIRS vital signs. Among existing criteria, the pair of temperature and heart rate vital signs demonstrated the strongest associations with outcomes, particularly temperature-corrected heart rate, which increased PPV without sacrificing NPV compared to other SIRS vital sign pairs. If SIRS-based vital sign screening criteria are employed, use of the temperature-corrected heart rate parameter alone appears to be the optimal choice. Multivariate analysis was undertaken to determine whether SIRS continues to have predictive value when simple, potentially confounding variables that are easily identified are included (age and chronic complex condition). This AOR continued to be significant. This was not intended to develop a fully predictive model, and influence statistics indicate that there is variability unaccounted for by the model, underscoring the need for additional study to better predict critical outcomes among pediatric patients with fever. The findings of this study suggest some areas where modifications may improve the predictive value of vital signs to detect critical illness. This study suggests that fever alone was a significant determinant of the presence of tachypnea and tachycardia meeting SIRS criteria, and improved methods to correct for the effect of fever on heart rate and respiratory rate might improve specificity. The formula used in this study, as in other pediatric sepsis literature, was derived from a study of infants less than 1 year old, but a simple linear correction for heart rate is likely not optimal at all ages, with the 10 beats/min correction representing a proportionally larger correction in older children with slower baseline heart rates. 12 Given the limitations of existing SIRS vital signs to predict patients who will require critical care, research to determine alternate vital sign cut points with optimal test characteristics would be warranted. In addition, non vital sign approaches, including use of laboratory tests and validated, reproducible physical examination characteristics, may offer improvements on SIRS vital signs. These findings are consistent with the limited existing literature about pediatric SIRS, which has demonstrated a high prevalence and low PPV for serious illness. Horeczko and Green 7 demonstrated a prevalence of 18.1% of SIRS in children in a nationally representative ED database and found rates of admission to be higher in patients with SIRS. The higher prevalence of SIRS in Horezcko and Green s sample may reflect the inclusion

8 388 Scott et al. SIRS IN THE PEDIATRIC ED of trauma patients and leukocytosis criteria or reflect differences between a pediatric-only ED and the broader range of EDs included in the study. Cruz et al. 9 showed that using temperature-adjusted heart rate as part of a vital sign SIRS screening tool had a 99.9% NPV and 4% PPV for the diagnosis of shock as determined by a treating ED clinician. Brent et al. 14 also demonstrated a high NPV and low PPV using temperature and pulse to identify serious bacterial illness among children presenting to the ED. A prior estimate of the rate of severe sepsis among pediatric ED visits nationally, 0.34%, is similar to the rates of critical illness in our sample of pediatric medical ED visits. 15 Although SIRS is part of the academic definitions of pediatric sepsis intended for research use, it is not recommended for diagnostic use by sepsis clinical guidelines. 8 The American College of Critical Care Medicine guidelines for pediatric septic shock suggest that septic shock resuscitation be initiated in patients displaying abnormalities of mental status and perfusion and do not comment on the use of SIRS criteria within the septic shock diagnosis and treatment algorithm. 4 Yet the subjective nature of determining and recording abnormalities of mental status and perfusion, particularly in triage where many young children are anxious, crying, and exposed to environmental temperature changes, has led institutions to rely on objective and readily recorded SIRS vital sign criteria as a first-line screening tool instead. 10 It is problematic to maintain distinct research and clinical definitions for diagnosis of septic shock. The differences in these cohorts were highlighted in a study by Weiss et al. 16 who demonstrated only 67% of pediatric intensive care patients diagnosed clinically with sepsis met research definitions. The existence of a research definition of sepsis that is more easily operationalized than clinical examination leads it to frequently be adopted as a screening test, particularly with the precedent of doing so in adults, and obscures the fact that a simple screening tool not dependent on pediatric clinical acumen has not been established for pediatric sepsis. LIMITATIONS Our patients represent a single-center sample, and their outcomes may not be generalizable to a setting outside of an academic tertiary care facility. We relied on coded and electronically captured data that may be subject to ascertainment bias, but this bias would be unlikely to differentially affect SIRS compared to non-sirs vital sign patients. Our conclusions are limited to the use of vital sign SIRS only. It is possible that SIRS criteria including the use of abnormal leukocyte count would perform differently in identifying the outcomes studied here. However, leukocyte counts are not routinely obtained on all pediatric ED patients, and it is important to understand the test characteristics of vital sign SIRS since they are usually the only data available at the time of triage. There are many relevant outcomes to consider in pediatric sepsis. Critical care, as defined by vasoactive infusion or invasive ventilation, was chosen because it is data accessible for electronic extraction that represents organ dysfunction in existing sepsis definitions and is a standard of critical care across institutions. 8 Although ICU admission is more subject to institutional standards and bed availability, it may represent significant cases of sepsis that require vigilant monitoring and aggressive fluid resuscitation without meeting our critical care definition. In this study, we found that the RR of ICU admission in SIRS versus non- SIRS was similar to the risk of critical care requirement. Similar to the outcome of critical care requirement, SIRS criteria also failed to identify 79% of patients who required ICU admission. Although it is conceivable that other treatments given preferentially to SIRS patients, such as aggressive fluid resuscitation, may have prevented the development of either of these outcomes, only 8% of SIRS patients received any form of IV treatment in the ED, and it is unlikely that additional fluid resuscitation on the inpatient units would have masked critical illness, since a fluid bolus over any rate faster than 1 hour is not permitted on the inpatient units at this institution and requires ICU care. Data were obtained from the existing electronic medical record, and there was no study-specific standardization of the methods used for obtaining vital signs, reflecting SIRS vital sign effectiveness in a real-world ED setting, rather than efficacy under ideal test conditions. Meeting SIRS vital sign criteria may reflect heart rate elevations due to anxiety rather than hemodynamic compromise. We attempted to account for this by also examining the persistence of SIRS on repeat vital signs when anxiety may have improved; however, these analyses were limited by the fact that only 52% of SIRS vital sign patients had additional vital signs recorded. Our study was retrospective in nature, and the outcomes in the study reflect clinician decision-making. We can comment on the association between SIRS vital signs and these outcomes, but cannot determine whether these treatments or dispositions were appropriate for the patients. We are unaware of a practice pattern toward more or less aggressive treatment within our center, and it is likely that the large sample size including all providers practicing within the ED would mitigate the influence of an outlier clinician. The 72-hour readmission rate suggests that inappropriate treatment requiring readmission was rare, but it is possible that patients who presented elsewhere for repeat visits had outcomes we did not capture. As the largest pediatric tertiary care center in the region, it is likely that the majority of readmissions would present to this center and be captured in the data. CONCLUSIONS Patients with SIRS vital signs represented 15.2% of complete medical ED visits with vital signs recorded at a tertiary pediatric hospital, and the majority of patients with these vital signs were discharged without intravenous therapy and without readmission. Patients with systemic inflammatory response syndrome vital signs had statistically significant and clinically modest increased risks of critical care, admission, and ED intervention. However, SIRS vital signs have a low sensitivity for critical illness, making the vital signs poorly suited

9 ACADEMIC EMERGENCY MEDICINE April 2015, Vol. 22, No for use in isolation as a screening test for children requiring resuscitation for sepsis. We acknowledge the assistance of Kathleen Grice in preparation of the manuscript. References 1. Hartman ME, Linde-Zwirble WT, Angus DC, Watson RS. Trends in the epidemiology of pediatric severe sepsis. Pediatr Crit Care Med 2013;14: Balamuth F, Weiss SL, Neuman MI, et al. Pediatric severe sepsis in U.S. children s hospitals. Pediatr Crit Care Med 2014;15: Inwald DP, Tasker RC, Peters MJ, Nadel S; Paediatric Intensive Care Society Study Group. Emergency management of children with severe sepsis in the United Kingdom: the results of the Paediatric Intensive Care Society sepsis audit. Arch Dis Child 2009;94: Brierley J, Carcillo JA, Choong K, et al. Clinical practice parameters for hemodynamic support of pediatric and neonatal septic shock: 2007 update from the American College of Critical Care Medicine. Crit Care Med 2009;37: Jones AE, Puskarich MA. The surviving sepsis campaign guidelines 2012: update for emergency physicians. Ann Emerg Med 2014;63: Horeczko T, Green JP, Panacek EA. Epidemiology of the systemic inflammatory response syndrome (SIRS) in the emergency department. West J Emerg Med 2014;15: Horeczko T, Green JP. Emergency department presentation of the pediatric systemic inflammatory response syndrome. Pediatr Emerg Care 2013;29: Goldstein B, Giroir B, Randolph A; International Consensus Conference on Pediatric Sepsis. International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med 2005;6: Cruz AT, Williams EA, Graf JM, et al. Test characteristics of an automated age- and temperature-adjusted tachycardia alert in pediatric septic shock. Pediatr Emerg Care 2012;28: Cruz AT, Perry AM, Williams EA, Graf JM, Wuestner ER, Patel B. Implementation of goal-directed therapy for children with suspected sepsis in the emergency department. Pediatrics 2011;127:e Larsen GY, Mecham N, Greenberg R. An emergency department septic shock protocol and care guideline for children initiated at triage. Pediatrics 2011;127:e Hanna CM, Greenes DS. How much tachycardia in infants can be attributed to fever? Ann Emerg Med 2004;43: Feudtner C, Hays RM, Haynes G, Geyer JR, Neff JM, Koepsell TD. Deaths attributed to pediatric complex chronic conditions: national trends and implications for supportive care services. Pediatrics 2001;107:E Brent AJ, Lakhanpaul M, Ninis N, Levin M, MacFaul R, Thompson M. Evaluation of temperature-pulse centile charts in identifying serious bacterial illness: observational cohort study. Arch Dis Child 2011;96: Singhal S, Allen MW, McAnnally JR, Smith KS, Donnelly JP, Wang HE. National estimates of emergency department visits for pediatric severe sepsis in the united states. PeerJ 2013;1:e Weiss SL, Parker B, Bullock ME, et al. Defining pediatric sepsis by different criteria: discrepancies in populations and implications for clinical practice. Pediatr Crit Care Med 2012;13:e Supporting Information The following supporting information is available in the online version of this paper: Data Supplement S1. Pediatric SIRS criteria. Requires two of four criteria, one of which must be temperature >38.5 or <36 C or alteration in leukocyte count.