Patterns of C-reactive protein ratio response to antibiotics in pediatric sepsis: A prospective cohort study

Accepted Manuscript Patterns of C-reactive protein ratio response to antibiotics in pediatric sepsis: A prospective cohort study

Vanessa Soares Lanziotti, Pedro Póvoa, Arnaldo Prata-Barbosa, Lucas Berbet Pulcheri, Ligia S.C.F. Rabello, José Roberto Lapa e Silva, Marcio Soares, Jorge I.F. Salluh PII: DOI: Reference:

S0883-9441(17)31063-8 doi:10.1016/j.jcrc.2017.11.018 YJCRC 52777

To appear in: Please cite this article as: Vanessa Soares Lanziotti, Pedro Póvoa, Arnaldo Prata-Barbosa, Lucas Berbet Pulcheri, Ligia S.C.F. Rabello, José Roberto Lapa e Silva, Marcio Soares, Jorge I.F. Salluh , Patterns of C-reactive protein ratio response to antibiotics in pediatric sepsis: A prospective cohort study. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Yjcrc(2017), doi:10.1016/ j.jcrc.2017.11.018

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Title: Patterns of C-Reactive Protein ratio response to antibiotics in pediatric sepsis: a prospective cohort study. Vanessa Soares Lanziotti MD, MSc1,2,3; Pedro Póvoa MD, PhD5,6; Arnaldo Prata-Barbosa, MD, PhD1,4; Lucas Berbet Pulcheri, MD7; Ligia S.C.F. Rabello,

SC RI PT

MD, MSc1,2; José Roberto Lapa e Silva, MD, PhD2; Marcio Soares MD, PhD1,2; Jorge I.F. Salluh, MD, PhD1,2.

Affiliations:

NU

1- D’Or Institute for Research and Education - Rio de Janeiro (RJ), Brazil.

2- Postgraduate program in Internal Medicine- Federal University of Rio de

MA

Janeiro (Universidade Federal do Rio de Janeiro – UFRJ) – Rio de Janeiro

ED

(RJ), Brazil.

3- Institute of Pediatrics and Child Care Martagão Gesteira – Pediatric

PT

Intensive Care Unit - Federal University of Rio de Janeiro (Universidade

CE

Federal do Rio de Janeiro- UFRJ)- Rio de Janeiro (RJ), Brazil.

4- Department of Pediatrics, School of Medicine, Federal University of Rio de

AC

Janeiro (Universidade Federal do Rio de Janeiro – UFRJ) – Rio de Janeiro (RJ), Brazil.

5- NOVA Medical School, New University of Lisbon - Lisboa, Portugal.

6- Polyvalent Intensive Care Unit, Hospital de São Francisco Xavier, Centro Hospitalar de Lisboa Ocidental – Lisboa, Portugal. 7- Rios D’Or Hospital, Rede D’Or São Luiz- Rio de Janeiro (RJ), Brazil. 1

ACCEPTED MANUSCRIPT Email addresses for all authors:

Vanessa Soares Lanziotti- [email protected] Pedro Póvoa – [email protected]

SC RI PT

Lucas Berbet Pulcheri – [email protected] Ligia S.C.F. Rabello – [email protected]

Arnaldo Prata-Barbosa- [email protected]

NU

José Roberto Lapa e Silva- [email protected]

MA

Marcio Soares – [email protected]

ED

Jorge Ibrain Figueira Salluh – [email protected]

Corresponding author:

PT

Vanessa Soares Lanziotti

CE

Address: Rua Diniz Cordeiro, 30 – Botafogo. Rio de Janeiro (RJ), Brazil. Zip code 22281-100

AC

Phone: +55-21-2538-3541 Fax: +55-21-2538-3555 Email: [email protected]

Financial Support: Drs. Salluh & Soares are supported in part by individual research grants from CNPq and FAPERJ.

Abbreviations: 2

ACCEPTED MANUSCRIPT CRP- C-reactive protein CRPr – C-reactive protein ratio ICU – Intensive Care Unit

SC RI PT

IEC - Institutional Ethics Committee IQR – Interquartile Range LOS – Length Of Stay

NU

MV – Mechanical Ventilation

MA

PELOD - Pediatric Logistic Organ Dysfunction

AC

CE

PT

ED

PICU – Pediatric Intensive Care Unit

Abstract (192 words)

Purpose: Evaluate sequential C-Reactive Protein (CRP) measurements and patterns of CRP-ratio response to antibiotic therapy during first 7 days in Pediatric Intensive Care Unit (PICU) of septic children. 3

ACCEPTED MANUSCRIPT Methods: Prospective, cohort study of children (1 month-12 years) admitted at 3 PICUs, with diagnosis of sepsis with <72hrs course. CRP-ratio was calculated in relation to D0_CRP value. Children were classified according to an individual pattern of CRP-ratio response: fast – CRP_D4 of therapy was <0.4 of D0_CRP; slow – continuous but slow decrease of CRP; non – CRP

D0_CRP followed by secondary rise ≥0.8. Results:

103

septic

children

SC RI PT

remained ≥0.8 of D0_CRP; biphasic – initial CRP decrease to levels <0.8 of

(age-median:

2yrs;

54%

male)

were

prospectively included (infection focus: 65% respiratory, 12.5% central

NU

nervous system). Overall PICU mortality was 11.7%. 102 children could be classified according to a predefined CRP-ratio response pattern. Time-

MA

dependent analysis of CRP-ratio and CRP course of the different patterns were significantly different. Besides, PICU mortality rate was significantly

ED

different according CRP-ratio response patterns : fast response 4.5%; slow response 5.8%; non-response 29.4%; biphasic response 42.8%.

PT

Conclusions: In pediatric sepsis, CRP-ratio serial evaluation was useful in

CE

early identification of patients with poor outcome. Key words: C-reactive protein; sepsis; biomarkers; children, pediatric

AC

intensive care unit; outcome

4

ACCEPTED MANUSCRIPT Text (2598 words)

Introduction Sepsis is a major cause of admission to pediatric intensive care units (PICU)(1,2).In the last decade, a number of initiatives were implemented

SC RI PT

aiming to achieve a better understanding of sepsis concepts (3,4), as well as to improve morbidity and mortality, namely through higher diagnosis awareness and early antibiotic therapy and specific guidelines for treatment of pediatric sepsis (5, 6). However, the assessment of the response to treatment

NU

is still based on clinical judgment which is frequently insufficient for an early recognition of outcomes in children with sepsis(1, 2) .Thus, biomarkers have

MA

been proposed to aid physicians in the clinical decision-making process (7- 9). C-reactive protein (CRP) is a biomarker well acknowledged for its diagnostic

ED

value in pediatric sepsis, with low cost, widely available and easy to interpret (9,10). However, there is still little evidence in pediatric patients on the use of

PT

CRP as a surrogate marker of outcomes and most available studies are

CE

limited to neonates (10,11), infants (12,13) and adults (14-16). Since CRP has 1st order elimination kinetics, relative changes could

AC

be more informative than absolute. Using CRP-ratio, that is the relative CRP changes in relation to its initial concentration, patients could be classified according to different patterns of response that were shown to be associated with different outcomes (15). Besides, CRP levels are not modified by frequent therapies or interventions as use of systemic corticosteroids and renal replacement therapy (17-19).

5

ACCEPTED MANUSCRIPT The aim of the present study was to assess serial CRP measurements and the outcomes associated with patterns of CRP-ratio response to antibiotic therapy during the first 7 days in the PICU, in children with sepsis.

SC RI PT

Materials and Methods Study design, subjects and setting

We conducted a prospective observational cohort study in three PICUs at tertiary hospitals in Rio de Janeiro, Brazil (Instituto de Puericultura e

NU

Pediatria Martagão Gesteira- IPPMG- Federal University of Rio de Janeiro, Copa D’Or Hospital and Rio’s D’Or Hospital). Pediatric patients (1 month to 12

MA

years) with sepsis diagnosis with <72 hrs of hospital admission were consecutively included between July 2013 and July 2016. The present study

ED

was strictly observational and did not interfere with clinical management or clinical decision-process. The Institutional Ethics Committees (IECs) approved study (D’Or

Institute for Research

PT

the

and

Education

IEC

–

No

CE

178.663/December 2012 and IPPMG/UFRJ IEC – No 190.152/ January 2013) and as no interventions were performed, the need for informed consent was

AC

waived.

Definitions and Data Collection Sepsis was diagnosed according to the international pediatric sepsis consensus definitions (3). Demographic, clinical, laboratory and outcome data were prospectively collected, using standardized case report forms, including age, gender, comorbidities, primary infection site, presence of nosocomial infection, use of vasopressors, dialysis and mechanical ventilation (invasive

6

ACCEPTED MANUSCRIPT and non invasive). Comorbidity was defined as the presence of underlying disease and/or chronic comorbid conditions, such as genetic syndrome, encephalopathy, heart disease and immunodeficiency. Pediatric Logistic Organ Dysfunction (PELOD) score (20) was calculated daily using clinical and laboratory data routinely collected during the first week of PICU stay. Empiric

SC RI PT

antibiotic and supportive therapies were started upon PICU admission according to local guidelines and international guidelines for sepsis and septic shock (5).Patients were followed until death or hospital discharge. An additional follow-up for 90-day mortality was also performed for all patients,

NU

after PICU or hospital discharge, through phone calls or checking hospital

MA

electronic data system.

CRP was sampled every other day from admission (D0) to D6 of PICU

ED

stay (D0, D2, D4, D6). CRP-ratio was calculated in relation to the D0 CRP value. Patients were retrospectively classified according to predefined CRP-

PT

ratio patterns of response to antibiotics (15,16) : 1) fast response – when CRP at D4 of therapy was ≤0.4 of D0 CRP concentration; 2) slow response –

CE

characterized by a continuous but slow decreased of CRP (CRP at D4 >0.4

AC

and at D6 ≤0.8 of D0 CRP); 3) non-response – when CRP remained always >0.8 of D0 CRP; 4) biphasic response – characterized by an initial CRP decrease to levels ≤0.8 of D0 CRP followed by a secondary rise >0.8 of D0 CRP. Comparisons between survivors and non-survivors and between the four different CRP-ratio patterns were performed.

7

ACCEPTED MANUSCRIPT Data processing and statistical analysis Data entry was performed by a single investigator (VSL) and consistency was assessed with a rechecking procedure of a 10% random sample of patients. Data were screened in detail for missing information, implausible and outlying values. Continuous variables were reported as mean

SC RI PT

± standard deviation (SD) or median (25% to 75% interquartile range, IQR) according to data distribution. Comparisons between groups were performed using the unpaired and paired t-test, or the Mann-Whitney U test and Wilcoxon signed-rank test (for comparison between 2 groups) and ANOVA or

NU

Kruskal Wallis test (for comparison between 4 groups) for continuous variables according to data distribution. The Fisher’s exact test was used to

MA

carry out comparisons between categorical variables. Time-dependent analysis of CRP was performed via General Linear Model (GLM) univariate

ED

repeated-measures analysis using a split-plot design approach. The SPSS version 20.0 software package (Chicago, IL, USA) and Prism 7.0 (Graphpad,

PT

USA) were used for statistical analysis. In all cases, statistical significance

AC

Results

CE

was defined as a two-tailed test with an alpha of 0.05.

Main characteristics of the study population During the study period, 103 out of the 112 patients that fulfilled the eligibility criteria were include (Figure 1); 9 patients were excluded since they had no sequential CRP measurements (Figure 1). The main characteristics of the studied population are depicted in Table 1. The median age was 2 years old, the lung was the main site of

8

ACCEPTED MANUSCRIPT primary infection (65%), followed by central nervous system (12.5%) and abdominal (6.8%). 71% of the patients required vasopressors during PICU stay and 69% required invasive mechanical ventilation. The overall PICU and 90-day mortality rates were the same (11.7%). 14.5% (N=15) of infections were microbiologically documented (54%

SC RI PT

Gram-negative bacteria, 19%, Gram-positive bacteria, 19% virus and one Mycobacterium tuberculosis). Detailed data of microbiological documentation is shown in Table 2.

NU

C-reactive protein, organ failure course and CRP-ratio patterns of

MA

response to antibiotic therapy in survivors and non-survivors

Time course of CRP concentrations during the first week of antibiotic

ED

therapy is shown in Figure 2 A. Time dependent analysis of CRP values from D0 to D6 between survivors and non-survivors showed a significant decrease

PT

in survivors whereas it remained almost unchanged in non-survivors

CE

(P<0.0001). As early as D2, this difference was already present, and by D4 of antibiotic therapy, this divergent evolution of CRP values was markedly

AC

different between survivors and non-survivors (P= 0.01). Time dependent analysis of the course of PELOD values from D0 to D6 between survivors and non-survivors are shown in Figure 2 B. PELOD score values were significantly different (GLM P< 0.0001) and the difference is marked since D2 of antibiotic therapy, with significantly lower values in survivors. Patients included in the study were divided according to the four patterns of the CRP-ratio course during antibiotic therapy, as previously

9

ACCEPTED MANUSCRIPT described. Out of the 103 patients, 102 could be classified: 43% were classified as fast response (N=44), 33.5% as slow response (N=34), 16.5% as non-response (N=17) and 7% as biphasic response (N=7). The timedependent analysis of CRP-ratio of the four different patterns was significantly different (P=0.023), as shown in Figure 3 A. The time-dependent analysis of

SC RI PT

PELOD values in the four different patterns of CRP-ratio response to antibiotics is shown in figure 3 B. We observed that PELOD course on the first week of antibiotic therapy were significantly different in the different patterns (P= 0.027). Patients classified as fast response had a continuous and faster

NU

improvement of organ dysfunctions (decrease of PELOD value) as compared to patients who were classified as slow response during the first week of

MA

antibiotic therapy. Patients classified as biphasic response had slightly increased in PELOD values by D2, followed by a slight decrease. Patients

ED

classified as non-response had an initial increase in PELOD values (D2)

PT

followed by later decrease in the values.

We also compared the absolute PELOD changes during the first week

CE

of antibiotic therapy on the four different patterns of CRP-ratio response that

AC

are shown in figure 4A. We collapsed patients with initial CRP decrease as responders (fast plus slow response) and with no CRP decrease as nonresponders (non-response plus biphasic) and performed a comparison between them (figure 4B) where we can observe a clear graphic difference, with an evident negative and continuous PELOD variation on responders. However, the curves were not significantly different.

10

ACCEPTED MANUSCRIPT CRP-ratio patterns of response presented significant different PICU mortality rates (Figure 5): 4.5% of mortality in patients with the fast response; 5.8% in patients with slow response; 29.4% in non-response patients and 42.8% in biphasic response patients (P= 0.001). When we collapsed the patterns of CRP-ratio response with good outcome (responders- fast response

SC RI PT

and slow response patterns) and bad outcome (non-responders- non-response and biphasic response) the mortality rate were, respectively, 5.1% and 33% (P= 0.0001). We also analyzed additional outcomes as ventilator-free days, hospital and PICU length of stay (21)

in the different CRP-ratio response

NU

patterns to antibiotics. Ventilator-free days median was calculated for each different pattern and patients with fast response, slow response, non-response

MA

and biphasic response patterns presented a ventilator-free days medians [IQR 25-75] of 22.5 [18.75-26.25]; 23.5 [13.75-27.5]; 16 [0-20.37] and 19 [0-20.5],

ED

respectively, and we observed a gradient from responders to non-responders (P=0.020). Regarding PICU length of stay (LOS) it was 9 [6-13.25] in the fast

PT

response patients; 10 [7-17.5] in the slow response patients; 14 [9-18] in non-

CE

response patients and 13 [10-27] in biphasic response patients (P=0.057). We also analyzed the rate of nosocomial infections according to four

AC

CRP-ratio patterns of response: fast response 11.3% (N=5); slow response 14.7% (N= 5); non-response 41.1% (N=7); biphasic response 42.8% (N=3) (P=0.02). Nosocomial infections rates were significantly higher on the patterns with poor outcomes (non-response and biphasic response). Analyzing specifically patients with biphasic response pattern we found that the secondary rise of CRP was coincident with the day of nosocomial infection in all biphasic patients who presented nosocomial infection (N=3).

11

ACCEPTED MANUSCRIPT

Discussion In the present study, we described the course of CRP in pediatric septic critically ill patients and observed an association between individual patterns of CRP-ratio response and PICU mortality rates. Patients with a

SC RI PT

substantial decrease in CRP values and with CRP-ratio patterns of fast response and slow response presented a significant lower mortality rate when compared to those patients with no decrease on CRP sequential values and with CRP-ratio patterns of non-response and biphasic response.

NU

These data suggest that persistently elevated CRP values are indicative of poor response to antibiotic therapy. Although there are many

MA

reasons for the unfavorable outcomes and the lack of recovery of multi-organ dysfunction in sepsis, one of the potentially modifiable reason is inadequate

ED

initial antibiotic therapy as well as the occurrence of a nosocomial infection (16,22,23). In those with biphasic response, albeit very small absolute

PT

numbers, we observed that the 3 patients that presented nosocomial infection

CE

(43% of 7 patients classified as biphasic) during the first week of PICU admission had the secondary CRP rise coincident with the day of diagnosis of

AC

nosocomial infection. Future studies with larger number of patients should be performed to corroborate this finding. Several studies, especially in adults, have confirmed that sequential CRP measurements are a useful tool to monitor clinical course, evaluate antibiotic efficacy and need for antibiotic changes, and mortality prediction (14-16), having better value as a prognostic predictor than using only the

12

ACCEPTED MANUSCRIPT absolute values (24,25). However, data using the sequential approach is scarce in pediatric sepsis. Considering that CRP is widely available in most of PICUs, with easy access and low cost and that it has been one of the best-known inflammatory biomarkers used in clinical practice for many years, CRP can be confirmed as

SC RI PT

a key biomarker of response to antibiotic therapy when dynamically analyzed. We also performed a time-dependent analysis of PELOD score variation in the different CRP- ratio patterns of response. We found that in the patients with CRP-ratio response patterns of fast or slow, PELOD score presented

NU

marked decreases. On the opposite, in biphasic response patients, PELOD remained almost unchanged over the first 7 days of antibiotic treatment. And,

MA

in patients with non-response pattern, we observed an initial increase followed by a decrease in PELOD average value, although the CRP had no decrease

ED

and outcomes of these patients were worse. This can be explained by the fact that the early reversal of organ failure is directly related to the outcomes (26).

PT

Besides, the resolution of the sepsis induced inflammatory response,

CE

assessed by CRP, seems to be surrogate of the reversal organ failure. Moreover, we evaluate the absolute PELOD changes during the first week of

AC

antibiotic therapy on the 4 different patterns of CRP-ratio response (Figure 4A) and compared responders and non-responders (4B), trying to overcome differences of PELOD values at PICU admission. Although the differences were not statistically significant, we could observe evident graphic differences on the curves, showing more evident negative variations of PELOD on the responders (fast and slow response pattern).

13

ACCEPTED MANUSCRIPT Given the low mortality rates among even critically ill children, it has been strongly recommended potential alternative endpoints for clinical trials (21), such as ventilator-free days, hospital and PICU length of stay, etc. Besides the main outcome (PICU mortality), we compared the secondary outcomes on the four different patterns of CRP-ratio responses and the

SC RI PT

patients classified as non-response and biphasic response had less ventilation free days, longer PICU length of stay and higher mortality rates (PICU and 90 day mortality). We could observe that PICU LOS was higher in patients with no decrease CRP (non-response and biphasic response).

NU

Although P value was on the limit of the statistical significance (P=0.057) for this comparison, considering that literature definition for long stay patients in

MA

the PICU is longer than 12 days (27), we can include patients classified as non response and biphasic response patterns, with median PICU LOS of 14

ED

and 13 days, respectively, and higher mortality, as long stay patients, which does not happen with the patients with substantial decrease of CRP (fast and

PT

slow response). Since literature has also shown that patients with a longer

CE

ICU length of stay have higher mortality (27,28),this finding must be considered, corroborating with the worst outcomes in the nonresponders

AC

patterns that presented longer PICU length of stay and higher mortalities. The present study has some limitations. First, considering the rates of events as mortality and nosocomial infections, we have a relatively small sample size (N=103). However, we believe that this limitation could be attenuated, since we included secondary outcomes as alternative endpoints to better evaluate outcomes. The low percentage of microbiological documentation (14.5%) when compared to literature, that showed rates as

14

ACCEPTED MANUSCRIPT high as 72% of microbiological proof (1,23,27), is also a limitation, mainly regarding to the adequacy of initial antibiotics. Moreover, in this study, we only evaluate one biomarker, CRP. Despite limitations, this is the first study of this type of CRP analysis in pediatric septic patients and we believe it has shown

bedside clinical decision-making process.

Conclusions

SC RI PT

relevant results that could help pediatric intensive care physicians on the

In our study, the CRP-ratio pattern of response to antibiotics during the

NU

first week of therapy in septic critically ill pediatric patients could be useful for the recognition of individual clinical evolution, namely good or poor outcomes.

MA

In patients with persistently elevated or increasing CRP-ratio (non response and biphasic response), physicians should perform a thorough evaluation

ED

trying to identify a cause for such a evolution. At a time where intensive care medicine is moving away in a direction of personalized treatment aiming to

AC

CE

PT

improve outcomes, we believe that these findings may be promising.

References:

[1] Weiss SL, Fitzgerald JC, Pappachan J, Wheeler D, Jaramillo-Bustamante JC, Salloo A, et al. Global Epidemiology of Pediatric Severe Sepsis: The Sepsis Prevalence, Outcomes, and Therapies Study. Am J Respir Crit Care Med. 2015 May 15;191(10):1147–57.

15

ACCEPTED MANUSCRIPT [2] Ruth A, McCracken CE, Fortenberry JD, Hall M, Simon HK, Hebbar KB. Pediatric Severe Sepsis: Current Trends and Outcomes From the Pediatric Health Information Systems Database*. Pediatr Crit Care Med. 2014 Nov;15(9):828–38.

SC RI PT

[3] Goldstein B, Giroir B, Randolph A. International pediatric sepsis consensus conference: Definitions for sepsis and organ dysfunction in pediatrics*: Pediatr Crit Care Med. 2005 Jan;6(1):2–8.

[4] Kissoon N, Carcillo JA, Espinosa V, Argent A, Devictor D, Madden M, et

NU

al. World Federation of Pediatric Intensive Care and Critical Care Societies: Global Sepsis Initiative*: Pediatr Crit Care Med. 2011

MA

Sep;12(5):494–503.

[5] The Surviving Sepsis Campaign Guidelines Committee including The

ED

Pediatric Subgroup*, Dellinger RP, Levy MM, Rhodes A, Annane D,

PT

Gerlach H, et al. Surviving Sepsis Campaign: International Guidelines for Management of Severe Sepsis and Septic Shock, 2012. Intensive Care

CE

Med. 2013 Feb;39(2):165–228.

AC

[6] Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock: 2016. Intensive Care Med. 2017 Mar;43(3):304– 77. [7] Marshall JC, Reinhart K. Biomarkers of sepsis: Crit Care Med. 2009 Jul;37(7):2290–8.

16

ACCEPTED MANUSCRIPT [8] Kaplan JM, Wong HR. Biomarker discovery and development in pediatric critical care medicine*: Pediatr Crit Care Med. 2011 Mar;12(2):165–73. [9] Póvoa P, Salluh JIF. Biomarker-guided antibiotic therapy in adult critically ill patients: a critical review. Ann Intensive Care. 2012;2(1):32.

Educ Pract. 2010 Apr 1;95(2):55–8.

SC RI PT

[10] McWilliam S, Riordan A. How to use: C-reactive protein. Arch Dis Child -

[11] Patil S, Dutta S, Attri SV, Ray P, Kumar P. Serial C reactive protein values predict sensitivity of organisms to empirical antibiotics in neonates:

MA

29; fetalneonatal-2015-309158.

NU

a nested case–control study. Arch Dis Child - Fetal Neonatal Ed. 2016 Apr

[12] Hengst JM. The role of C-reactive protein in the evaluation and

ED

management of infants with suspected sepsis. Adv Neonatal Care Off J

PT

Natl Assoc Neonatal Nurses. 2003 Feb;3(1):3–13. [13] Downes KJ, Weiss SL, Gerber JS, Klieger SB, Fitzgerald JC, Balamuth F,

CE

et al. A Pragmatic Biomarker-Driven Algorithm to Guide Antibiotic Use in the Pediatric Intensive Care Unit: The Optimizing Antibiotic Strategies in

AC

Sepsis (OASIS) Study. J Pediatr Infect Dis Soc. 2016 May 4;piw023. [14] Coelho LM, Salluh JI, Soares M, Bozza FA, Verdeal JCR, Castro-FariaNeto HC, et al. Patterns of c-reactive protein RATIO response in severe community-acquired pneumonia: a cohort study. Crit Care. 2012;16(2):R53.

17

ACCEPTED MANUSCRIPT [15] Povoa P. C-reactive protein as a marker of ventilator-associated pneumonia resolution: a pilot study. Eur Respir J. 2005 May 1;25(5): 804–12. [16] Póvoa P, Souza-Dantas V, Soares M, Salluh JI. C-reactive protein in

SC RI PT

critically ill cancer patients with sepsis: influence of neutropenia. Crit Care. 2011;15(3):R129.

[17] Salluh JIF, Soares M, Coelho LM, Bozza FA, Verdeal JCR, Castro-FariaNeto HC, et al. Impact of systemic corticosteroids on the clinical course

NU

and outcomes of patients with severe community-acquired pneumonia: A cohort study. J Crit Care. 2011 Apr;26(2):193–200.

MA

[18] Dahaba AA, Rehak PH, List WF. Procalcitonin and C-reactive protein plasma concentrations in nonseptic uremic patients undergoing

ED

hemodialysis. Intensive Care Med. 2003 Apr;29(4):579–83.

PT

[19] Salluh JI, Lisboa T. C-reactive protein in community-acquired sepsis: you

CE

can teach new tricks to an old dog. Crit Care. 2011;15(5):186. [20] Leteurtre S, Martinot A, Duhamel A, Proulx F, Grandbastien B, Cotting J,

AC

et al. Validation of the paediatric logistic organ dysfunction (PELOD) score: prospective, observational, multicentre study. The Lancet. 2003 Jul;362(9379):192–7. [21] Pediatric Acute Lung Injury Consensus Conference Group. Pediatric acute respiratory distress syndrome: consensus recommendations from the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care

18

ACCEPTED MANUSCRIPT Med J Soc Crit Care Med World Fed Pediatr Intensive Crit Care Soc. 2015 Jun;16(5):428–39. [22] Povoa P, Coelho L, Almeida E, Fernandes A, Mealha R, Moreira P, et al. Pilot Study Evaluating C-Reactive Protein Levels in the Assessment of

SC RI PT

Response to Treatment of Severe Bloodstream Infection. Clin Infect Dis. 2005 Jun 15;40(12):1855–7.

[23] Weiss SL, Fitzgerald JC, Balamuth F, Alpern ER, Lavelle J, Chilutti M, et al. Delayed antimicrobial therapy increases mortality and organ

NU

dysfunction duration in pediatric sepsis. Crit Care Med. 2014 Nov;42(11):2409–17.

MA

[24] Póvoa P, Teixeira-Pinto AM, Carneiro AH, the Portuguese CommunityAcquired Sepsis Study Group (SACiUCI). C-reactive protein, an early

ED

marker of community-acquired sepsis resolution: a multi-center

PT

prospective observational study. Crit Care. 2011;15(4):R169.

CE

[25] Póvoa P, Salluh JIF. Use of biomarkers in sepsis: many questions, few answers. Rev Bras Ter Intensiva. 2013 Mar;25(1):1–2.

AC

[26] Levy MM, Macias WL, Vincent J-L, Russell JA, Silva E, Trzaskoma B, et al. Early changes in organ function predict eventual survival in severe sepsis. Crit Care Med. 2005 Oct;33(10):2194–201. [27] Marcin JP, Slonim AD, Pollack MM, Ruttimann UE. Long-stay patients in the pediatric intensive care unit. Crit Care Med. 2001 Mar;29(3):652–7.

19

ACCEPTED MANUSCRIPT [28] Friedrich JO, Wilson G, Chant C. Long-term outcomes and clinical predictors of hospital mortality in very long stay intensive care unit

AC

CE

PT

ED

MA

NU

SC RI PT

patients: a cohort study. Crit Care Lond Engl. 2006;10(2):R59.

20

ACCEPTED MANUSCRIPT Figure Captions:

Figure 1- Study flow chart.

Figure 2-

SC RI PT

2A- C-Reactive Protein (CRP) of survivors and non-survivors pediatric septic critically ill patients. Time course of CRP concentrations (mg/dL) for survivors and non-survivors pediatric septic critically ill patients during the first week of antibiotic therapy. The difference between the two groups was

NU

statistically significant (P<0.0001).

2B- PELOD course of survivors and non-survivors pediatric septic

MA

critically ill patients. Time course of PELOD values for survivors and nonsurvivors septic critically ill patients during the first week of antibiotic therapy.

ED

The difference between the two groups was statistically significant

PT

(P<0.0001).

CE

Figure 3-

3A- Time-dependent analysis of CRP-ratio in different response

AC

patterns. Data are presented as exactly means. Values of CRP-ratio in the different patterns were significantly different (P=0.023). 3B- Time-dependent analysis of PELOD in different response patterns. Data are presented as means. PELOD values variation in the different patterns were significantly different (P=0.027).

21

ACCEPTED MANUSCRIPT Figure 4- Absolute PELOD changes during the first week of antibiotic therapy. 4A- Absolute PELOD changes during the first week of antibiotic therapy in different CRP-ratio response patterns. 4B- Absolute PELOD changes during the first week of antibiotic therapy in

SC RI PT

patients with CRP decrease (fast and slow response patterns) and CRP no decrease (biphasic and nonresponse patterns).

Figure 5- Pediatric Intensive Care Unit (PICU) mortality in the 4 CRP-ratio

NU

(CRPr) response patterns. Mortality rate was significantly different according

AC

CE

PT

ED

MA

to the CRPr pattern (P=0.001).

22

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 1

23

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

Fig. 2

24

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

Fig. 3

25

MA

NU

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 4

26

SC RI PT

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

Fig. 5

27

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

SC RI PT

Table 1 – Clinical patients’ characteristics and comparison between PICU survivors and non-survivors All Patients PICU PICU Characteristics (n= 103; Survivors Nonsurvivors P value* 100%) (n=91; 88%) (n=12; 12%) Age (years) 2 [0.5-5.2] 2[0,6-4.9] 0.75 [0,5-6.2] 0.602 Gender (male) 56 (54%) 49 (54%) 7 (58%) 0.999 PELOD score – 1 [1-11] 1 [1-10.5] 1.5 [1-11] 0.237 D0 (points) Sites of primary 0.541 infection Respiratory 67 (65%) 59 (65%) 8 (66%) Central Nervous 13 (12.5%) 11 (12%) 2 (17%) System Abdominal 7 (6.8%) 6 (6.5%) 1 (8.5%) Urinary 4 (4%) 4 (4.5%) 0(0%) Other 12 (11.7%) 11 (12%) 1 (8.5%) Microbiological 15 (14.5%) 13 (14.3%) 2 (16.6%) 0.826 documentation Comorbidity 34 (33%) 29 (32%) 5 (42%) 0.524 Vasopressors 71 (69%) 59 (65%) 12 (100%) 0.016 Dialysis 6 (6%) 1 (1%) 5 (42%) 0.001 Non-invasive 38 (37%) 36 (40%) 2 (17%) 0.202 MV Invasive MV 71 (69%) 59 (65%) 12 (100%) 0.016 CRP 6.9 [4.05-21.6] 7 [4.3-21.6] 5.5 [3.7-20.6] 0.751 (D0)(mg/dL) Results expressed as median (25%-75% interquartile range) and number (%). Abbreviations: CRP, C-Reactive Protein; PICU, Pediatric Intensive Care Unit; PELOD, Pediatric Logistic Organ Dysfunction; MV, Mechanical Ventilation. *P value for comparison among patients PICU mortality

28

ACCEPTED MANUSCRIPT

Peritoneal fluid Nasopharyngeal aspirate

NU

N=1 N=1 N=3 N=2 N=1

AC

CE

PT

ED

MA

Mycobacteria Mycobacterium tuberculosis Virus Respiratory syncytial virus Bocavirus

SC RI PT

Table 2 -Microorganisms from the 15 (14.5%) pediatric patients with sepsis Microorganism N=15 Isolated site Gram-negative organisms N= 8 Escherichia coli N=4 Urine (N=3);blood (N=1) Neisseria meningitidis N=2 Blood (N=2) Acinetobacter baumannii N=1 Blood (N=1) Pseudomonas aeruginosa N=1 Cerebrospinal fluid (N=1) Gram-positive organisms N=3 Streptoccoccus pyogenes N=2 Blood (N=1); pleural fluid (N=1) Methicilin-resistant N=1 Blood (N=1) Staphylococcus aureus

29

ACCEPTED MANUSCRIPT

Highlights 

Sequential evaluation of CRP and CRP-ratio was useful in the early identification of pediatric septic critically ill patients with poor outcome.



Identification of CRP-ratio pattern of response to antibiotics during the

SC RI PT

first week of therapy could be useful for the recognition of individual clinical evolution, influencing bedside clinical decision-making process. 

Patients with persistently elevated or increasing CRP-ratio (non response and biphasic response) had worse outcomes (higher mortality rate, longer PICU length of stay, reduced ventilation free

NU

days) and an approach consisting in a thorough evaluation could be



MA

performed trying to identify a cause for such evolution. In time-dependent analysis of PELOD score we found a significant

Considering that CRP is widely available in most of PICUs, with easy

PT

access and low cost and it has been one of the best-known

CE

inflammatory biomarkers used in clinical practice for many years, CRP can be confirmed as a key biomarker of response to antibiotic therapy when dynamically analyzed. This is a relevant aspect to justify its wide

AC



ED

decrease in survivors compared to non-survivors.

implementation especially in low-resource settings.

30