Program Schedule

669
DRIP – Drug Resistance In Pneumonia: Derivation and Prospective Multi-center Validation of a Scoring Model to Predict Drug-Resistant Pathogens

Session: Poster Abstract Session: Approach to Clinical Infections
Friday, October 10, 2014
Room: The Pennsylvania Convention Center: IDExpo Hall BC
Posters
  • DRIP IDSA poster.pdf (1.2 MB)
  • Background:

    Predicting risk for community-acquired pneumonia due to drug-resistant pathogens (CAP-DRP) is challenging.  Healthcare-associated pneumonia (HCAP) criteria have limited predictive value. However, a more accurate model has not yet been validated in a U.S. cohort.

    Methods:

    In a retrospective derivation cohort of 200 culture-positive CAP and HCAP cases, previously identified risk factors for CAP-DRP were analyzed by logistic regression.  A novel prediction tool, the DRIP score, was derived from 10 of these factors, weighted according to AOR (Table 1).  A validation cohort of 200 culture-positive cases was identified prospectively at four U.S. centers (May 2013 - May 2014).  The performance of DRIP, HCAP and other prediction models was evaluated in this group (Table 2).

    Results:

    In the derivation group CAP-DRP prevalence was 26%.  Prior antibiotic use (60 days) (AOR 7.3, p<0.01), tube feeding (AOR 25.6 p=0.02), long term care (AOR 4.1 p=0.05), and prior CAP-DRP (AOR 7.2 p=0.079) were associated with CAP-DRP.  The DRIP score best differentiated high and low risk at a threshold of  ≥4 points with a PPV of 73.1, NPV 90.5, accuracy of 86% and AUROC of 0.844.  This was superior to HCAP (AUROC 0.604, accuracy 72%), and all other predictive models (Fig. 1).  In the validation cohort, DRIP was again most predictive: AUROC of 0.831, PPV 67.1, NPV 85.5, accuracy 78.5% (Fig. 2).  CAP-DRP prevalence was 34.5%.  Compared to HCAP, the DRIP model would reduce over treatment by 46% with similar rates of under treatment.

     

    Conclusion:

    In this prospective multi-center U.S. study, the DRIP score was more predictive of CAP-DRP than HCAP and other models and has potential to decrease antibiotic over-utilization.  Further validation is needed.

    Table 1.  DRIP Score

    Points

    Antibiotic use <60 d

    2

    Long term care

    2

    Tube feeding

    2

    Prior CAP-DRP (1 yr)

    2

    Hospitalization <60 d

    1

    Chronic pulmonary disease

    1

    Poor functional status

    1

    Gastric acid suppression

    1

    Wound care

    1

    MRSA colonization (1 yr)

    1

     

    Table 2. Validation Cohort, n=200

     

     

    Model

    Sens %

    Spec %

    PPV %

    NPV %

    AUROC

    Accuracy %

    DRIP

    73.9

    80.9

    67.1

    85.5

    0.83

    78.5

    HCAP

    76.8

    64.9

    53.5

    84.2

    0.71

    69.0

    Schreiber

    49.3

    77.9

    54.0

    74.5

    0.66

    68.0

    Shorr

    79.7

    54.2

    47.8

    83.5

    0.75

    67.0

    Niederman

    86.9

    57.3

    51.7

    89.3

    0.79

    67.5

    Shindo

    81.2

    59.5

    51.4

    85.7

    0.77

    67.0

    Aliberti

    87.0

    55.7

    50.8

    89.0

    0.73

    66.5

    Park

    75.4

    71.8

    58.4

    84.7

    0.79

    73.0

    Figure 1: Derivation cohort

    Figure 2: Validation cohort

    Brandon Webb, MD1,2, Edward Stenehjem, MD MSc3, Kristin Dascomb, MD, PhD3, Holenarasipur R. Vikram, MD, FACP, FIDSA4, Neera Agrwal, MD5, Kenneth K. Sakata, MD6, Kathryn Williams, MD7, Bruno Bockorny, MD8, Kavitha Bagavathy, MD8, Shireen Mirza, MD8, Mark Metersky, MD9 and Nathan Dean, MD10,11, (1)Infectious Disease, University of Utah, Salt Lake City, UT, (2)Clinical Epidemiology and Infectious Disease, Intermountain Healthcare, Murray, UT, (3)Clinical Epidemiology and Infectious Diseases, Intermountain Medical Center, Murray, UT, (4)Division of Infectious Diseases, Mayo Clinic Hospital, Phoenix, AZ, (5)Hospital Internal Medicine, Division of Internal Medicine Mayo Clinic Hospital, Phoenix, AZ, (6)Internal Medicine, Mayo Clinic, Phoenix, AZ, (7)Internal Medicine, Mayo Clinic in Arizona, Phoenix, AZ, (8)Internal Medicine, University of Connecticut Medical Center, Farmington, CT, (9)Pulmonary and Critical Care Medicine, University of Connecticut School of Medicine, Farmington, CT, (10)University of Utah, Murray, UT, (11)Intermountain Medical Center, Murray, UT

    Disclosures:

    B. Webb, None

    E. Stenehjem, None

    K. Dascomb, None

    H. R. Vikram, None

    N. Agrwal, None

    K. K. Sakata, None

    K. Williams, None

    B. Bockorny, None

    K. Bagavathy, None

    S. Mirza, None

    M. Metersky, None

    N. Dean, None

    Findings in the abstracts are embargoed until 12:01 a.m. EDT, Oct. 8th with the exception of research findings presented at the IDWeek press conferences.

    Sponsoring Societies:

    © 2014, idweek.org. All Rights Reserved.

    Follow IDWeek