Infectious Etiologies and Patient Outcomes in Pediatric Septic Shock

Transcription

1 Journal of the Pediatric Infectious Diseases Society ORIGINAL ARTICLE Infectious Etiologies and Patient Outcomes in Pediatric Septic Shock Stefanie G. Ames, 1 Jennifer K. Workman, 2 Jared A. Olson, 4 E. Kent Korgenski, 6 Susan Masotti, 5 Elizabeth D. Knackstedt, 3 Susan L. Bratton, 2 Gitte Y. Larsen 2 1 Department of Pediatrics, 2 Division of Critical Care, Department of Pediatrics, and 3 Division of Infectious Disease, Department of Pediatrics, University of Utah, 4 Department of Pharmacy, and 5 System Improvement, Primary Children s Hospital, and 6 Pediatric Clinical Program, Intermountain Healthcare, Salt Lake City, Utah Background. Septic shock remains an important cause of death and disability in children. Optimal care requires early recognition and treatment. Methods. We evaluated a retrospective cohort of children (age <19) treated in our emergency department (ED) for septic shock during to investigate the association between timing of antibiotic therapy and outcomes. The exposures were (1) receipt of empiric antibiotics in 1 hour and (2) receipt of appropriate antibiotics in 1 hour. The primary outcome was development of new or progressive multiple system organ dysfunction syndrome (NP-MODS). The secondary outcome was mortality. Results. Among 321 patients admitted to intensive care, 48% (n = 153) received empiric antibiotics in 1 hour. These patients were more ill at presentation with significantly greater median pediatric index of mortality 2 (PIM2) scores and were more likely to receive recommended resuscitation in the ED (61% vs 14%); however, rates of NP-MODS (9% vs 12%) and hospital mortality (7% vs 4%) were similar to those treated later.early, appropriate antibiotics were administered to 33% (n = 67) of patients with identified or suspected bacterial infection. These patients had significantly greater PIM2 scores but similar rates of NP-MODS (15% vs 15%) and hospital mortality (10% vs 6%) to those treated later. Conclusions. Critically ill children with septic shock treated in a children's hospital ED who received antibiotics in 1 hour were significantly more severely ill than those treated later, but they did not have increased risk of NP-MODS or death. Keywords. anti-bacterial agents; critical care; pediatrics; septic shock severe sepsis. Septic shock is an important cause of death and long-term morbidity in US children [1]. The prevalence of pediatric septic shock has increased, but prior to programs for early recognition and treatment, mortality had remained unchanged [2 4]. Delay in treatment of shock and subsequent development of acute multiorgan failure increases risk of death [4]. Early resuscitation of shock to mitigate or prevent progression of organ failure was demonstrated to decrease mortality in adults in 2001 [5]. Early recognition of shock followed by timely treatment of the underlying cause and rapid reversal of cardiovascular insufficiency using protocol-based care is now accepted as standard medical practice. These goals are the cornerstone of the Surviving Sepsis Campaign (SSC) [6]. An important aspect of early goal directed therapy is the rapid administration of antibiotics. The SSC endorses administering antibiotics within 1 hour of shock recognition [6]. However, rapid antibiotic therapy is a nuanced factor because not all patients with septic shock have bacterial infections, and among those with treatable bacterial infections, the initial coverage may be inadequate. Delaying treatment with appropriate antibiotics is associated with increased mortality in adults [7]. Our children's hospital emergency department (ED) has an ongoing quality improvement program for early recognition and treatment of pediatric septic shock [8]. This program has been in place since 2007, and adherence to recommended interventions including rapid antibiotic administration has increased over time [8]. In the present study, we evaluate whether administration of empiric and appropriate antibiotics in the ED within 1 hour from identification of shock is associated with decreased development of new or progressive multiple organ dysfunction syndrome (NP-MODS) compared with those treated later. We also describe infectious etiologies in children presenting to our ED with septic shock as well as the choice and timing of empiric antibiotic therapy. Received 27 August 2015; accepted 15 December Correspondence: J. K. Workman, MD, Division of Pediatric Critical Care, PO Box , Salt Lake City, UT (jennifer.workman@hsc.utah.edu). Journal of the Pediatric Infectious Diseases Society 2017;6(1):80 6 The Author Published by Oxford University Press on behalf of The Journal of the Pediatric Infectious Diseases Society. All rights reserved. For permissions, please journals.permissions@oup.com. DOI: /jpids/piv108 METHODS Study Design and Setting We evaluated the relationship between timing of antibiotic administration for treatment of septic shock and outcomes of children requiring admission to the pediatric intensive care unit (PICU). We retrospectively identified patients presenting 80 JPIDS 2017:6 (March) Ames et al

2 to the ED with septic shock between 2008 and 2012, and we reviewed the hospital course and clinical outcomes. The study was conducted at Primary Children's Hospital (PCH), a tertiary care pediatric hospital with ED visits per year and 2000 PICU admissions per year. The University of Utah Institutional Review Board approved the study and waived requirement for informed consent. Participant Selection Criteria Initial patient selection was from the PCH Pediatric Septic Shock Project Database, which was established with initiation of the PCH ED Septic Shock Guidelines in Patient records were flagged for initial review and possible inclusion in this database if they met any of the following broad criteria: (1) triage in the ED to emergent status, (2) lactate drawn during ED course, (3) PICU admission within 12 hours of presentation to ED, (4) repeat visits to the ED within 48 hours of initial visit, and (5) having a rapid response called on a patient within 12 hours of admission from the ED to an inpatient ward for management of sepsis. After initial screening, patients were retained in this database if their condition met the American College of Critical Care Medicine (ACCM)/Pediatric Advanced Life Support (PALS) definition of septic shock: known or suspected infection accompanied by temperature abnormalities, vital sign abnormalities based on PALS vital sign parameters for age, and physical exam consistent with deficit in end-organ perfusion [9 11]. The ACCM/PALS definition was used because of its usefulness in bedside diagnosis and treatment in real time [12]. We included subjects for the present study if they were 18 years of age and were admitted from the PCH ED directly to the PICU between January 2008 and December Microbiology Definitions Patients with a primary bacterial infection had a culture taken on the day of hospitalization (urine, blood, cerebral spinal fluid [CSF], other sterile body fluid, or a wound culture) that grew pathogenic bacteria. Bacteria were classified as pathogenic if determined to be the cause of septic shock based on review of clinician documentation. A primary respiratory bacterial infection required the following: a respiratory sample from either an endobronchial brush specimen, a tracheal aspirate, or a bronchial alveolar lavage that met Centers for Disease Control and Prevention criteria for a respiratory infection, and presence of an infiltrate on chest imaging [12]. Patients with presumed primary bacterial infections were defined as those with clinical documentation indicating a likely bacterial infection with a full treatment course of antibiotics but no positive culture. These included community-acquired pneumonia, aspiration pneumonia, intestinal and intra-abdominal infections, and life-threatening skin and soft tissue infection. Intra-abdominal infections include perforated bowel from any cause including appendicitis or necrotizing enterocolitis. Patients were determined to have a primary viral infection if bacterial cultures were negative and a viral respiratory panel or viral polymerase chain reaction (PCR) study was positive. Patients without an infectious source identified were classified as no source identified. Exposures and Outcome The principle exposure was receipt of empiric antibiotics 1 hour after recognition of septic shock in the ED, compared with later administration among all patients admitted to the PICU. Time to antibiotics was determined by time from triage in the ED to initial administration of medication. We additionally compared receipt of appropriate antibiotics in 1 hour among those with infections for which antibiotic therapy is recommended. These patients may have required additional or changed antibiotic therapy from the initial, empiric choice. Antibiotics were considered appropriate if the bacteria isolated were susceptible to the regimen given. For presumed bacterial infections, the antibiotics were considered appropriate when the spectrum of activity of the regimen was at least as broad as that recommended by the Infectious Disease Society of America clinical guidelines for bacterial pneumonia, skin and soft tissue infections, and intra-abdominal infections [14 16]. The primary outcome was development of NP-MODS. Given the relatively low pediatric mortality rate, surrogate measures of disease severity are often used. Sepsis is associated with MODS, and NP-MODS has been used as a surrogate measure in this way [17 19]. New or progressive multiple system organ dysfunction syndrome was defined as the development of 2 or more organ dysfunctions during the hospitalization that were not present on the initial day of admission, or worsening of organ failures present on admission. Organ dysfunction present on hospital day 1 was presumed to be present on arrival. Assessment was based on cardiovascular, pulmonary, renal, neurologic, hematologic, and hepatic function, based on the patient's worst parameter during a specified time period (Supplementary Table 1) [18, 19]. Death was not part of NP-MODS assessment. We assessed NP-MODS on days 1, 2, 5, 8, 12, 16, and 18 (or until hospital discharge, whichever was sooner). Data were abstracted from the Intermountain Enterprise Data Warehouse and by manual chart review. The ED course and compliance with the SSC recommendations was recorded [6]. Additional outcome measures including mortality, PICU and hospital length of stay, and hospital resource utilization were captured. All data were stored within the University of Utah Division of Pediatric Critical Care Data Coordinating Center secure server. Statistical Analysis Standard descriptive statistics were used to summarize patient characteristics. Groups were compared by exposure, using χ 2 test for categorical variables and Wilcoxon rank-sum test for continuous variables. A binomial logistic regression was used to estimate the relationship between receipt of early antibiotics Infections in Pediatric Septic Shock JPIDS 2017:6 (March) 81

3 and odds of NP-MODS (primary outcome) as well as mortality (secondary outcome), adjusting for severity of illness with pediatric index of mortality 2 scores (PIM2). Findings were considered statistically significant if P value was <.05. The Bonferroni adjustment was used for restriction of significance when multiple pairwise comparisons were made between groups. All data were analyzed using SPSS (Chicago, Illinois) software, 19th version. RESULTS Three hundred twenty-one patients met study criteria, and select demographic and clinical features are presented in Table 1. Bacterial infections were identified in 154 (48%) patients with septic shock. Presumed bacterial and viral infections accounted for 51 (16%) and 50 (16%) patients, respectively. The remaining 66 (21%) patients had no source of infection identified. Table 1 reports patient groups by infectious etiology. Those without an infectious pathogen identified were significantly more likely to have a complex chronic condition [20] compared with all other groups. Those with documented bacterial infections had significantly longer median hospital length of stay than those with presumed bacterial, viral, or no infection identified. New or progressive multiple system organ dysfunction syndrome was significantly more common in those with documented bacterial infections compared with children without a source of infection identified and those with viral infections. The positive bacterial cultures were as follows: 72 blood, 25 urine, 12 CSF, 10 stool, and 8 sterile body fluid infections. The remaining infections were respiratory and soft tissue infections. A summary of the isolated organisms is in Supplementary Table 2. The most common bacterial species were streptococcal (n = 42), staphylococcal (n = 25), and Gram-negative species (n = 66), of which 19 patients had Escherichia coli infections. Among those with a viral infection, 47 had a respiratory illness with rhinovirus isolated in 36% (n = 17) and respiratory Table 1. Select Patient Demographic and Clinical Characteristics by Infection Type a Characteristic Bacterial Infection N = 154 Presumed Bacterial Infection N = 51 Viral Infection N = 50 No Infection Identified N = 66 P Value Compares All Groups Demographic Characteristics Age, months 65 (13 156) 106 (33 159) 23 (4 128) 81 (21 163).01 Race.45 White 109 (71) 35 (69) 30 (60) 49 (74) Hispanic 20 (13) 7 (14) 6 (12) 6 (9) Other 25 (16) 9 (18) 14 (28) 11 (17) Male 78 (51) 22 (43) 28 (55) 37 (56).49 Study years (50) 26 (51) 17 (34) 28 (42) (50) 25 (49) 33 (66) 38 (58) Clinical Characteristics Tracheostomy dependence 9 (6) 5 (10) 5 (10) 9 (14).29 CCC [18] 45 (29) 18 (35) 10 (20) 31 (47).01 Number of CCC 0 (0,1) 0 (0, 1) 0 (0, 0) 1 (0, 1).01 Antibiotics given in ED 148 (96) 50 (98) 50 (100) 59 (89).03 Received empiric antibiotics 77 (50) 18 (35) 18 (36) 40 (61).01 1 hour Minutes to initial antibiotics 60 (42, 100) 75 (37, 120) 74 (44, 101) 55 (32, 88).1 Appropriate antibiotics 1 hour 56 (37) 11 (23) Not applicable Not applicable.23 Minutes to appropriate 92 (50, 238) 100 (65, 192) Not applicable Not applicable.65 antibiotics PIM2 ED 1.6 (1.2, 4.0) 1.4 (1.0, 3.1) 1.3 (1.0, 3.8) 1.8 (1.2, 5.1).09 Mechanical ventilation during 61 (40) 16 (31) 9 (18) b 16 (24).02 admission Vasoactive medications during 58 (38) 17 (33) 6 (12) b 20 (30).01 admission Length of PICU stay, days 1.9 (0.9, 6.3) 1.7 (0.9, 3.9) 1.7 (0.7, 2.9) 1.7 (0.7, 4.0).34 Length of hospital stay, days 7.1 (4.0, 13.2) 4.9 (3.2, 10.7) 3.7 b (2.5, 6.2) 4.4 b (2.2, 9.9) <.01 Death 10 (7) 3 (6) 0 4 (6).34 Greatest organ dysfunction on 2 (1,2) 1 (0, 2) b 1 (0, 1) b 1 (1, 2) b <.01 any PICU day NP-MODS 25 (16) 5 (10) 1 (2) b 3 (5) b <.01 Abbreviations: CCC, complex chronic condition; ED, emergency department; IQR, interquartile range; MODS, multiple organ dysfunction syndrome; NP-MODS, new or progressive multiple organ dysfunction syndrome; PICU, pediatric intensive care unit; PIM2, pediatric index of mortality 2. a Data are presented as either n (%) or median (IQR). b Significantly different than bacterial infection, P <.008 for pairwise comparisons. 82 JPIDS 2017:6 (March) Ames et al

4 syncytial virus isolated in 21% (n = 10). Two patients had enterovirus meningitis, and 1 had cytomegalovirus viremia. There were 4 children with bacterial diarrhea for which antibiotic treatment is not recommended. Of the patients with a bacterial infection for which antibiotic treatment is indicated, 145 (72%) of the bacterial pathogens were sensitive to ceftriaxone, which is the empiric antibiotic therapy recommended in the PCH ED septic shock guideline. Table 2 compares demographic and clinical features of patients who received empiric antibiotics in 1 hour (N = 153) to those treated later (N = 168). Those treated in 1 hour were more likely to receive initial resuscitation compliant with the SSC guidelines (61% vs 14%). They also had significantly greater risk of mortality estimated by PIM2 score, but neither mortality (7% vs 4%) nor development of NP-MODS (9% vs 12%) differed significantly from those treated after 1 hour. This nonsignificant relationship was also found among the subset of the cohort for whom antibiotic treatment is indicated, with mortality rates 11% vs 6% (relative risk [RR], 1.65; 95% confidence interval [CI], ) and development of NP-MODS 15% vs 18% (RR, 0.81; 95% CI, ), among those treated in 1 hour vs after 1 hour, respectively. Table 3 compares patients with bacterial infections for which antibiotic treatment is indicated who received early appropriate antibiotics in 1 hour (N = 67) to those treated after 1 hour (N = 134). This excludes patients with viral infections and no source of infection identified as well as those with bacterial infections for which treatment is not recommended (n = 4). Those receiving appropriate antibiotics in 1 hour were also more likely to receive initial resuscitation compliant with the SSC guidelines (69% vs 17%). They also had significantly greater risk of mortality estimated by PIM2 score and higher rates of mechanical ventilation. However, when adjusting for severity of illness using PIM2, neither mortality (10% vs 6%) nor development of NP-MODS (15% vs 15%) differed significantly between groups. Figure 1a compares development of NP-MODS in the ICU by hours elapsed to empiric antibiotic treatment among all patients (including patients with viral infections and no source identified). Almost half (n = 153) of patients received empiric antibiotics within 1 hour, 118 (37%) between 1 and 2 hours, 26 (8%) between 2 and 3 hours, 10 (3%) between hours 3 and 4, and 14 (4%) in greater than 4 hours. In total, 93% of patients received antibiotics within 3 hours of presentation. When comparing hourly interval to empiric antibiotics, there were no significant differences in development of NP-MODS. Figure 1b compares development of NP-MODS by hours elapsed until appropriate antibiotic treatment among the subset of patients with infections for which treatment is recommended. Sixty-seven (33%) patients received appropriate antibiotics within 1 hour, 56 (28%) patients between 1 and 2 hours, 16 (8%) patients between 2 and 3 hours, 15 (7%) patients between hours 3 and 4, and 47 (23%) in greater than 4 hours. When comparing hourly interval to appropriate antibiotics, there were no significant differences in development of NP-MODS. DISCUSSION Almost two thirds of children with septic shock admitted to the PICU from the ED had a definite or presumed bacterial infection. Among all children, those receiving empiric antibiotics 1 hour were significantly more severely ill at presentation to the ED but did not differ with respect to NP-MODS or mortality Table 2. Comparison of Demographic and Clinical Features of Patients Who Received Empiric Antibiotics in 1 Hour or Longer a Variable Empiric Antibiotics 1 Hour N = 153 Empiric Antibiotics >1 Hour N = 168 P Value Age years 5.0 (1.3, 12.9) 6.1 (1.1, 13.1).79 Minutes to initial antibiotics 40 (22, 50) 95 (75, 138) <.001 Any CCC 54 (35) 50 (30).30 Infection Type.01 Bacterial 77 (50) 77 (46) Presumed bacterial 18 (20) 33 (12) Viral 18 (12) 32 (19) No infectious source identified 40 (26) 26 (16) ED care compliant with sepsis guidelines 94 (61) 23 (14) <.001 PIM2 in ED 1.9 (1.3, 4.7) 1.3 (1.0, 3.5).001 Mechanical ventilation 55 (36) 47 (28).13 Vasoactive medications 53 (35) 48 (29).24 PICU length of stay 2.1 (0.8, 6.1) 1.6 (0.9, 3.6).10 Hospital length of stay 6.9 (3.0, 12.4) 4.7 (3.0, 9.6).10 Death 10 (7) 7 (4).34 Greatest organ dysfunction on any PICU day 1 (1, 2) 1 (1, 2).04 NP-MODS 14 (9) 20 (12).42 Abbreviations: CCC, complex chronic condition; ED, emergency department; IQR, interquartile range; MODS, multiple organ dysfunction syndrome; NP-MODS, new or progressive multiple organ dysfunction syndrome; PICU, pediatric intensive care unit; PIM2, pediatric index of mortality 2. a Data are presented as either n (%) or median (IQR). Infections in Pediatric Septic Shock JPIDS 2017:6 (March) 83

5 Table 3. Comparison of Demographic and Clinical Features of Patients Who Received Appropriate Antibiotics in 1 Hour or Longer a,b Variable Appropriate Antibiotics 1 Hour N = 67 Appropriate Antibiotics >1 Hour N = 134 P Value Age years 4.7 (1, 13.1) 6.3 (1.4, 13.4).38 Minutes to appropriate antibiotics 43 (26, 53) 153 (95, 329) <.01 Any CCC [12] 25 (37) 36 (27).13 Infection Type.08 Bacterial 56 (84) 95 (71) Presumed bacterial 11 (16) 37 (28) ED care compliant with sepsis guidelines 46 (69) 23 (17) <.001 PIM2 in ED 2.2 (1.4, 5.0) 1.4 (1.0, 2.7) <.001 PIM2 in PICU 1.4 (1.1, 5.7) 1.2 (1.0, 3.4) <.03 Mechanical ventilation 34 (52) 41 (31).005 Vasoactive medications 26 (39) 48 (36).68 ICU length of stay 2.5 (0.9, 7.6) 1.6 (0.9, 4.0).12 Hospital length of stay 8.9 (3.8, 14.2) 6.0 (3.7, 11.3).13 Death 7 (10) 6 (6).27 c Greatest organ dysfunction on any PICU day 2 (1, 3) 1 (1,2).005 NP-MODS 10 (15) 20 (15).89 c Abbreviations: CCC, complex chronic condition; ED, emergency department; ICU, intensive care unit; IQR, interquartile range; MODS, multiple organ dysfunction syndrome; NP-MODS, new or progressive multiple organ dysfunction syndrome; PICU, pediatric ICU; PIM2, pediatric index of mortality 2. a Data are presented as either n (%) or median (IQR). b Infections for which antibiotics are not recommended such as shigellosis are excluded n = 4. c Adjusted for PIM2 in ED. compared with those treated later. Among those with an infection treatable by antibiotics, children receiving appropriate antibiotics in 1 hour were also significantly more severely ill at presentation to the ED but did not have greater rates of NP-MODS or mortality compared with those treated later. Patients with bacterial infections had greater illness severity at presentation, were more likely to receive antibiotics in 1 hour and resuscitation in compliance with the SSC guidelines, and had significantly greater NP-MODS compared with children with viral infections or no infectious source identified. They also had significantly longer lengths of hospital stay. This is likely related to the pathophysiology of bacterial infections with greater activation of an inflammatory response [21, 22]. This is also consistent with other studies showing culture-positive bacterial infections associated with increased morbidity [23]. Among those with bacterial infections as the etiology of septic shock, the majority (72%) of isolates were sensitive to empiric therapy. This is similar to a report by Weiss et al [24] who showed that the majority (78%) of pediatric septic shock patients treated in a single children's hospital also received appropriate initial antibiotic therapy. We hypothesized that rapid administration ( 1 hour) of empiric antibiotics would be associated with decreased NP-MODS. We were surprised to find that although children treated in 1 hour presented more severely ill, they were more likely to receive SSC compliant care and had similar risk of NP-MODS and death compared with those treated later. Failure to demonstrate a benefit from timely empiric antibiotics may be due to increased error in measurement of patients likely to benefit from the exposure because only 62% had an infection for which antibiotics would be expected to improve the child's status. We also tested whether rapid appropriate antibiotic administration was associated with decreased risk of NP-MODS. Again, we found that children treated in 1 hour were more severely ill at presentation to the ED compared with those with delayed antibiotics, but there were no statistical differences in development of NP-MODS or mortality by time of antibiotic administration or type of infection (presumed or culture positive bacterial infections). Our findings differ from the majority of reports involving adult patients, showing delay to appropriate antibiotic administration and increased mortality [23, 25 27]. The recent study by Weiss et al [24] of children with septic shock suggested improved mortality and patient outcomes if appropriate antibiotic therapy was delivered within 3 hours of shock recognition. We did not find a similar mortality benefit. Our study has important differences compared with the report by Weiss et al [24]. The majority of our patients received antibiotics in 3 hours (93% vs 60%), and receipt of appropriate antibiotics in 1 hour was almost 3 times greater in our patients (33% vs 12%) [11]. Although both cohorts had similar rates of comorbid conditions and similar rates and infection types, the overall mortality (5% vs 12%, P =.08), support with mechanical ventilation (32% vs 62%, P <.001), receipt of vasoactive medications (31% vs 74%, P <.001), and ICU length of stay (2 vs 9 median days), were all greater in the previous report. However, we included only patients directly admitted to the PICU from the ED, used a slightly different definition of septic shock (ACCM/PALS [9] vs Goldstein et al [10]), and our patients had lower risk of mortality measured by PIM2 scores on presentation to the ICU after treatment in the ED. Thus, our patients 84 JPIDS 2017:6 (March) Ames et al

6 needed to identify the ideal time goal for early appropriate antibiotic therapy that impacts pediatric patient outcomes. CONCLUSIONS More than half of children presenting to the ED with septic shock requiring intensive care had a definite or presumed bacterial infection, and the recommended empiric antibiotic treatment at our hospital was appropriate in 72% of cases. Those treated with appropriate antibiotics in 1 hour were more ill when presenting to the ED with greater PIM2 scores, but they were also more likely to receive rapid resuscitation compliant with SSC guidelines. Despite increased illness severity on presentation, patients treated in 1 hour did not have an increased disease progression as quantified by NP-MODS or death. Because the natural history of untreated septic shock is development of multiple organ failure and death, our findings support the premise that early recognition and treatment of shock, including administration of timely, appropriate antibiotics, remains pivotal. Supplementary Data Supplementary materials are available at Journal of the Pediatric Infectious Diseases Society online. Figure 1. (a) Empiric antibiotic timing and development of new or progressive multiple organ dysfunction syndrome (NP-MODS). Patients who developed NP-MODS versus time from emergency department arrival to empiric antibiotics in hourly increments among all patients. The percentage of patients who developed NP-MODS (lighter portion) within each hour are reflected numerically above each bar. (b) Appropriate antibiotic timing and development of new or progressive multiple organ dysfunction syndrome (NP-MODS). Patients who developed NP-MODS versus time from emergency department arrival to appropriate antibiotics in hourly increments among patients with an infection for which antibiotics are recommended. This cohort is a subset of the study cohort and appropriate antibiotic administration may have required additional or changed drug from that empirically administered. The percentage of patients who developed NP-MODS (lighter portion) within each hour are reflected numerically above each bar. may have been less severely ill at recognition of shock, but they also likely benefitted from more rapid treatment, as evidenced by lack of disease progression toward NP-MODS and death. More importantly, because the majority of our patients received antibiotics in <3 hours, we had inadequate power to determine a difference at the 3-hour cutoff. Our study has several limitations. First, as a retrospective study evaluating appropriate antibiotic choices, we recognize that it is not always possible to predict the pathogen and appropriate antibiotic choice in real time. Second, because this is a single-center study, results may not be generalizable to other care settings. Third, there is a link between early receipt of antibiotics and the delivery of care consistent with the SSC guidelines, suggesting that clinical assessment of illness severity drives the intensity of care and is likely linked to subsequent development of NP-MODS and mortality. Larger, multicenter studies are Notes Acknowledgments. We thank Dr. Roni Lane (Department of Pediatrics, University of Utah) for outstanding leadership of the Primary Children's Sepsis Work group and careful review of this manuscript. We also thank Tanya Stout (Primary Children's Hospital, Department of Systems Improvement) for contributions to the septic shock database used in this study. Disclaimer. S. G. A. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Financial support. This work was supported in part by the Primary Children's Foundation Clinical Excellence Grants Program, which provided support for data collection and management. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. References 1. Watson RS, Carcillo JA, Linde-Zwirble WT et al. The epidemiology of severe sepsis in children in the United States. Am J Respir Crit Care Med 2003; 167: 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: Kutko MC, Calarco MP, Flaherty MB et al. Mortality rates in pediatric septic shock with and without multiple organ system failure. Pediatr Crit Care Med 2003; 4: Rivers E, Nguyen B, Havstad S et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345: Dellinger RP, Levy MM, Rhodes A et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: Crit Care Med 2013; 41: Kumar A, Ellis P, Arabi Y et al. Initiation of inappropriate antimicrobial therapy results in a fivefold reduction of survival in human septic shock. Chest 2009; 136: Larsen GY, Mecham N, Greenberg R. An emergency department septic shock protocol and care guideline for children initiated at triage. Pediatrics 2011; 127:e Infections in Pediatric Septic Shock JPIDS 2017:6 (March) 85

7 9. Zaritsky AL, Nadkarni V, Hickey RW et al American Heart Association (AHA) guidelines for cardiopulmonary resuscitation (CPR) and emergency cardiovascular care (ECC) of pediatric and neonatal patients: pediatric basic life support. Pediatrics 2006; 117:e Goldstein B, Giroir B, Randolph A. International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med 2005; 6: 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: Kuch BA, Carcillo JA, Han YY, Orr RA. Definitions of pediatric septic shock. Pediatr Crit Care Med 2005; 6: Centers for Disease Control and Prevention. CDC/NHSN Surveillance Definitions for Specific Types of Infections Website. Available at: PDFs/pscManual/17pscNosInfDef_current.pdf. Accessed December 1, Solomkin JS, Mazuski JE, Bradley JS et al. Diagnosis and management of complicated intra-abdominal infection in adults and children: guidelines by the Surgical Infection Society and the Infectious Diseases Society of America. Clin Infect Dis 2010; 50: Bradley JS, Byington CL, Shah SS et al. The management of community-acquired pneumonia in infants and children older than 3 months of age: clinical practice guidelines by the Pediatric Infectious Diseases Society and the Infectious Diseases Society of America. Clin Infect Dis 2011; 53:e Stevens DL, Bisno AL, Chambers HF et al. Practice guidelines for the diagnosis and management of skin and soft tissue infections: 2014 update by the Infectious Diseases Society of America. Clin Infect Dis 2014; 59:e Proulx F, Fayon M, Farrell C et al. Epidemiology of sepsis and multiple organ dysfunction syndrome in children. Chest 1996; 109: Doughty L, Carcillo J, Kaplan S et al. The compensatory anti-inflammatory cytokine interleukin 10 response in pediatric sepsis-induced multiple organ failure. Chest 1998; 113: Lacroix J, Hébert PC, Hutchison JS et al. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med 2007; 356: Feudtner C, Christakis DA, Connell FA. Pediatric deaths attributable to complex chronic conditions: a population-based study of Washington State, Pediatrics 2000; 106: Lisboa T, Seligman R, Diaz E et al. C-reactive protein correlates with bacterial load and appropriate antibiotic therapy in suspected ventilator-associated pneumonia. Crit Care Med 2008; 36: Deis JN, Creech CB, Estrada CM, Abramo TJ. Procalcitonin as a marker of severe bacterial infection in children in the emergency department. Pediatr Emerg Care 2010; 26: Gaieski DF, Mikkelsen ME, Band RA et al. Impact of time to antibiotics on survival in patients with severe sepsis or septic shock in whom early goal-directed therapy was initiated in the emergency department. Crit Care Med 2010; 38: Weiss SL, Fitzgerald JC, Balamuth F et al. Delayed antimicrobial therapy increases mortality and organ dysfunction duration in pediatric sepsis. Crit Care Med 2014; 42: Kumar A, Roberts D, Wood KE et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006; 34: Paul M, Shani V, Muchtar E et al. Systematic review and meta-analysis of the efficacy of appropriate empiric antibiotic therapy for sepsis. Antimicrob Agents Chemother 2010; 54: Varula M, Karisson S, Parvianinen I et al. Community-acquired septic shock: early management and outcome in a nationwide study in Finland. Acta Anaesthesiol Scand 2007; 51: Ferrer R, Martin-Loeches I, Phillips G et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med 2014; 42: JPIDS 2017:6 (March) Ames et al