Pneumonia Update 2016 Management of the Hospitalized Patient October, 2016 Scott A. Flanders, M.D., MHM Professor of Medicine Director, Hospital Medicine Program Associate Chair for Quality and Innovation University of Michigan Disclosure of Financial Relationships Scott A. Flanders, MD Has disclosed relationships with entities producing, marketing, re-selling, or distributing health care goods or services consumed by, or used on, patients. Consultant NONE Research and Grant Support CDC Foundation AHRQ Blue Cross Blue Shield, MI Pneumonia Guideline Groups * ATS / IDSA GL panel member Advisory Boards NONE Speakers Bureau NONE Board Member NONE * non-financial Key Topics to Cover Values and Preferences Pneumonia Diagnosis: CXR / Etiologies / Procalcitonin Antibiotics Steroids Healthcare-Associated Pneumonia Which pneumonia patients are at risk for resistant organisms? 1. Recommend treatments shown to improve clinical outcomes 2. Reduce unnecessary antibiotic use -Use lowest number / spectrum of antibiotics -De-escalate rapidly and shorten durations 1
A National Priority First attempt to characterize the annual human toll of antibiotic resistance. Clostridium difficile diagnoses: U.S. Hospitals 1993-2011 Lessa, et al. NEJM 2015 450,000 US cases (2011) 66% healthcare associated 29,000 deaths Current Challenges Conditions Driving Antibiotic Use Antibiotic Use in U.S. Hospitals The Big Three CDC: trend analysis over 550 hospitals; 2006-2012 55% of patients on antibiotics Overall use stable; shift to more broad spectrum use Baggs, J, et al., JAMA IM, 2016. 30%-50% of use is felt to be inappropriate Dellit TH, Clin Infect Dis. 2007. Fridkin S, MMWR, 2014 #1 Urinary Tract Infection (UTI) #2 Pneumonia #3 Skin and Soft Tissue Infection (UTI and pneumonia account for over half of all inpatient antibiotic use) JAMA. 2014; 312(14): 1438-46 2
Key Topics to Cover Is CXR the Gold Standard? Pneumonia Diagnosis: CXR / Etiologies / Procalcitonin Antibiotics Steroids Healthcare-Associated Pneumonia Which pneumonia patients are at risk for resistant organisms? French study of 320 patients with suspected CAP Excluded very severe cases (ICU, CRB65 >3) Performed CXR in all Asked ED docs to assess: Probability of CAP Antibiotic plan Site of care decisions Perform Chest CT! Ask docs again Claessens YE, AJRCCM. 2015 Is CXR the Gold Standard? 120 pts with CXR : 30% with CT infiltrates 190 pts with + CXR: 30% with NO CT infiltrates ED docs modified probability of CAP in 60% 80% in line with adjudicated diagnosis Modify treatment, drug or site of care : 61%! CT for Everyone!? Caveats Antibiotics after CT: 15% stopped, 45% started CT differed often with clinical / CXR disconnect CXR negative but high WBC, CRP, age, crackles CXR positive but no concerning findings Alternate diagnoses: CHF, COPD Claessens YE, AJRCCM. 2015 Takehome: CXR is one part of the assessment and is imperfect 3
Case #1 A 64 yo woman with CHF is admitted with 2 days of fever, nonproductive cough, and shortness of breath. She has been on levofloxacin for 24 hrs per PCP. T 38.8, BP 110/75, HR 90, sats 89% RA. WBC 11 k. CXR shows a patchy RLL infiltrate. She is started on Ceftriaxone / Azithro in the ED and admitted to you. Her respiratory panel is positive for adenovirus and her serum procalcitonin (x2) is 0.05mcg / L. Now what?: 1) Wait for blood and sputum cultures before deciding 2) Stop antibiotics 3) Continue antibiotics for possible bacterial co-infection 4) Stop antibiotics and start Lasix for CHF exacerbation Traditional thinking: CAP Etiologies S. Pneumoniae H. Influenzae Moraxella catarrhalis Klebsiella Atypicals (legionella, Mycoplasma, C. pneumoniae) S. Aureus CAP Etiologies CDC study in Chicago / Nashville (5 hospitals) All hospitalized adults with + CXR (>2000) Screened 18 hrs/ d, 7 d / week! 2010-2012 3-10 week f/u Control group without CAP in Nashville 80% Ward, 20% ICU Jain S. NEJM 2015 4
Procalcitonin Procalcitonin: The Algorithm Microbial toxins and bacterial-specific proinflammatory mediators stimulate release Levels are higher with bacterial infections vs. other Rapid rise / fall correlates with clinical response Correlates with severity of illness / prognosis Not affected by steroids Rapid testing available CHEST 2012 Procalcitonin and Antibiotic Use RCT/Year Patients with Respiratory Infection Setting Christ-Crain, 2004 243 ED Christ-Crain, 2006 302 ED/Inpt. Stolz, 2007 208 ED/Inpt. Briel, 2008 458 Multicenter Nobre, 2008 53 ICU Kristoffeger, 2009 210 ED/Inpt Schuetz, 2009 1359 Multicenter Stolz, 2009 101 Multicenter, ICU Long, 2009 127 ED Burkhardt, 2010 550 ICU Bouadma, 2010 397 ED Long, 2011 156 ED Antibiotic Use 25 20 15 10 5 0 Procalcitonin and Outcomes 14 Trials, 4221 Patients with Respiratory Infections OR=0.94 (0.71-1.23) 5.7% 6.3% Mortality OR=0.82 (0.71-0.97) 19% 22% Treatment Failure PCT Control CID 2012 Cochrane 2012 5
Procalcitonin: Caution Warranted Procalcitonin: Use in Practice 260 CAP pts: Extensive diagnostic testing (blood, sputum, urine antigens, PCT, PCR for viral) 23% bacterial, 16% viral (+/- bacterial), 46%? Bacterial: 23% with PCT < 0.1 mcg / L (many PSI IV) Viral: 23% with PCT > 0.25 mcg / L A low PCT cannot RELIABLY exclude serious bacterial infection Musher, et al. J. Infection 2013 200,000 COPD patients; 5% with PCT testing Weak association with fewer abx starts No association with duration of therapy Lindenauer, In Press, 2016 85% did not get more than 1 PCT test Lindenauer, In Press, 2016 Vaughn, Submitted, 2016 Procalcitonin: Bottom Line Key Topics to Cover It is another less than perfect, but potentially useful diagnostic test Should not be sole variable used to decide whether to use antibiotics or not Most useful when on the fence re giving / stopping antibiotics e.g., infiltrate, positive viral panel AND low PCT It can reduce antimicrobial use if protocols for deescalation are followed Pneumonia Diagnosis: CXR / Etiologies / Procalcitonin Antibiotics Steroids Healthcare-Associated Pneumonia Which pneumonia patients are at risk for resistant organisms? 6
Key Issues in Antibiotic Management Macrolides and Outcomes Meta-analysis; 23 studies with138,000 patients Are macrolides (atypical coverage) necessary? What regimens are effective for aspiration pneumonia? What duration of treatment is ideal? Macrolides vs. Nonmacrolides Mortality: RR=0.78 (0.64-0.95) Asadi, CID, 2012 Macrolides: Immunomodulatory Properties? Antibiotic Therapy IDSA / ATS 2007 Guidelines + Modifications b-lactam* + macrolide (or doxycycline) (*Ceftriaxone, Cefotaxime, Amp / Sul, Ertapenem, Ceftaroline) Or, Respiratory fluoroquinolone Corrales-Medina, Jo Infection, 2011 ICU: ß-lactam+macrolide Or ß-lactam+fluoroquinolone Anti-pseudomonal (many options) or CA-MRSA Rx (Vanco or Linezolid) if risk factors: independent of ICU status 7
Macrolides and Outcomes Meta-analysis; 23 studies with138,000 patients Macrolides and Pneumonia RCT: Beta Lactam + Macrolide vs. Beta Lactam Beta-lactam + macro vs. Fluoroquinolone Mortality RR = 1.17 (0.91-1.50) Macrolide vs. Nonmacrolide (RCTs only) Mortality RR = 1.13 RR=1 (0.65-1.98) Asadi, CID, 2012 RCT of 580 CAP pts; 6 Swiss hospitals Excluded: Severe Pneumonia, HCAP risk factors Legionella urine antigen performed in all Day7 Stability: HR < 100, SBP > 90, T < 38, RR < 24, Oxygen Sat > 90% RA by day 7 Macrolide: 67% No Macrolide: 59% Garin, et al. JAMA IM, 2014 Macrolides and Pneumonia RCT: Beta Lactam + Macrolide vs. Beta Lactam Hazard Ratio for Instability Atypical pathogens 0.33 (0.13-0.85) PSI Class IV 0.81 (0.59-1.10) No atypical pathogens 0.99 (0.80-1.22) PSI Class I-III 1.06 (0.82-1.36) (No diff at 90 days: mortality, ICU admits, LOS, complications, recurrence) Garin, et al. JAMA IM, 2014 Macrolides and Pneumonia Pragmatic, Cluster, Randomized,Non-Inferiority Trial Beta-lactam Beta-lactam + macro Fluoro Patients 656 739 888 Adherence 93% 88% 93% 90 d Mort. 9% 11% 8.8% Days to oral 4 4 3 LOS 6 6 6 Beta-lactam monotherapy: Non-inferior! (for non-icu patients without legionella) Postma DF, et al. NEJM, 2015 8
Azithromycin: The Downside Excess CV mortality during treatment NEJM 2012 Based on large registries; due to QT prolongation? 1 per 5000 courses: high CV risk patients Fluoroquinolones: The Downside Very broad spectrum drugs Broad GNR coverage Declining activity against Pseudomonas, E. Coli Increasing resistance (S. Pneumo, others) Adverse drug events C. difficile Adverse drug events FDA warnings against use in uncomplicated infections Tendon rupture, neuropathy / CNS effects, etc. A major driver of C. difficile Risk for C. Difficile High Risk Clindamycin Fluoroquinolones Cephalosporins / Carbepenems Low(er) Risk Penicillins, Macrolides, TMP/SMX No risk Tetracyclines Clearance of Colonization? Brown, AAC, 2013 Beta lactam / lactamase, metronidazole Dubberke, AAC, 2015 Ceftriaxone + Doxycycline? Does Doxycycline Protect against C. difficile? (Doernberg, et. al. CID, 2012) Observational, single center study Ceftriaxone + doxy vs. ceftriaxone +(levo or azithro) HR for C. diff = 0.15 27% lower rate of C. difficile for each day of doxy Doxycycline Has some activity against C.difficile Inhibits protein synthesis (and C. diff toxin production) Absorbed in the upper GI tract (less colonic delivery) 9
Antibiotics for Pneumonia: Options Beta lactam + macrolide vs. beta lactam alone More benefit demonstrated in sicker patients May not be necessary in some ward patients (Assuming they do not have legionella) Fluoroquinolones vs. beta lactam + macrolide Ward patients: comparable outcomes to beta + macrolide beta + macrolide favored for sicker patients Important downsides exist Beta-lactam/lactamase+macrolide; beta-lacta + doxy Potentially beneficial if C. difficile is a local problem Less outcome data Case # 2 A 78 y.o. woman is admitted from assisted living with 2 days of somnolence / confusion and 1 day of fever, cough, and sputum production. She was recently started on hydrocodone for acute low back pain. T 38.5, HR 115, BP 110-70, sat 89%. Her CXR shows a RLL infiltrate, concerning for aspiration. In addition to stopping her narcotics, the most appropriate treatment is: 1) Supportive care only 2) Ceftriaxone + Azithromycin 3) Ceftriaxone + Azithromycin + Clindamycin 4) Clindamycin Aspiration Pneumonia 15% of CAP hospitalizations called aspiration pneumonia, but All pneumonia is essentially aspiration pneumonia 50% of healthy adults aspirate during sleep Are the elderly unique? Silent Aspiration Elderly with CAP 70% Elderly w/o CAP 10% Pneumonia if reduced host defenses / high bacterial load Marik, Curr Op Pulm Med. 2011 Aspiration Pneumonia A Distinct Clinical Phenotype? At risk : Chronic neurologic / swallowing disorders Altered consciousness / alcohol or drugs Vomiting / witnessed choking Gravity dependent radiographic changes Older, more co-morbidity Higher 1 year mortality / higher recurrence Adds to mortality in CAP and HCAP Taylor, Am J Med. 2013 Komiya, Respirology. 2013 10
Aspiration Pneumonia: Anaerobes El Sohl, et al. Am J Crit Care 2003 95 elderly NH patient with severe aspiration Quantitative bronchial sampling 67 pathogens isolated 50% GNRs 15% staph 15% anaerobes (over half also had GNR); low virulence orgs Poor functional status associated with anaerobes 85% of patients with an anaerobe isolated improved on regimens without dedicated anaerobic coverage Aspiration Pneumonia: Anaerobes When to cover anaerobes Indolent symptoms Severe periodontal disease Putrid sputum Necrotizing / cavitary lesions / post-obstructive Large effusions (is it an empyema?) Elderly + poor functional status? Alcoholics Aspiration Pneum(onitis vs. onia) Aspiration Pneumonia CAP Pneumonitis? Time to Symptom Resolution Jaoude, Emerg Med, 2012 Aspiration Pneumonia Time to Clinical Stability < 2 days > 2 days Patients 102 227 Nursing home 67% 64% ICU transfer 0 23%* LOS (d) 4 6.5* 30 d mortality 9% 37%* β-lactam+macro 73% 70% Added clinda 2% 6% Authors propose stopping antibiotics in aspiration patients whose symptoms resolve in <48 hrs * = significant difference Jaoude, Emerg Med, 2012 11
Aspiration Pneumonia Treatment Most do well with routine CAP treatment Risk stratify for anaerobes If concerned: Ceftriaxone / Levo are active against most oral anaerobes (but may miss B. Fragilis, some clostridia) Ceftriaxone + Metronidazole: all anaerobes Ceftriaxone + Clindamycin: all anaerobes, but also gets your pt C. Difficile Treatment Duration: 5 Days 5 Days Rx and Stop if: T < 37.8 for 48 hrs and: No More than 1 of: SBP < 90 HR > 100 RR > 24 Sat < 90% Control Physician discretion Uranga, JAMA IM 2016 Treatment Duration: 5 Days 5 Days and Stop Control Patients 162 150 30 Day Success 92% 89% Abx Days (Median) 5 10 30 Day Readmission 1.4% 6.6% 70% of the intervention group received 5 days of treatment Uranga, JAMA IM 2016 60% Appropriate Duration: How well do we do? 10 Michigan Hospitals 1400 Pneumonia Patients 53.4% 50% 40% 30.8% 30% 20% 15.9% 10% 0% Appropriate Duration Excess Duration Excluded from logic 12
The First 5 Days Steroids for Pneumonia Siemieniuk R, et. al. Ann Intern Med. 2015 13 RCTS with > 2000 pts Outcome Risk Ratio Absolute Effect NNT Mortality 0.67 (0.45-1.01) 3% fewer 33 (Severe CAP) 0.39 (0.2-0.77) Need MV 0.45 (0.26-0.79) 5% fewer 20 ARDS 0.24 (0.10-0.56) 6.2% fewer 16 Ann Intern Med, 1964 LOS(d) -2.9 (-5.2- -.75) 1 d shorter Stability(d) -1.2 (-2.1- -0.4) 1.2 d sooner Hyperglyc 1.5 (1.01-2.2) 3.5% more 29 Steroids for Pneumonia ISSUES Mortality most affected in severe CAP Driven by one small study stopped early Variable definitions of severe Multiple steroid doses; from 1 dose to 10 days Excluded patients at risk for steroid adverse effects h/o GI bleed, immune suppression, etc. Key Topics to Cover Pneumonia Update Diagnosis: CXR / Etiologies / Procalcitonin Macrolides for everyone: necessary and safe? Steroids for pneumonia Healthcare-Associated Pneumonia Which pneumonia patients are at risk for resistant organisms? Hyperglycemia is associated with higher mortality? Await Ongoing Trials 13
Healthcare Associated Pneumonia IDSA / ATS Guidelines: Am J Resp Crit Care 2005 Home Therapy IV Wound Care Nursing care through health agency Hospital or Dialysis Clinic in past 30 days for Dialysis / Any IV therapy Hospitalized 2 days in past 90? days Nursing Home or Long-Term Care Facility At Risk for Multidrug-Resistant Organisms (MDROs) Vanco: 16% to 31% Pip-Tazo: 16% to 27% 2006 2010 Jones, BE. CID, 2015 MRSA: 2.5% to 2% Pseudomonas: 1.9% to 2% Pseudomonal Coverage for Pneumonia 50% 10 Michigan Hospitals 1400 Pneumonia Patients 45% 2.1% 40% 35% 30% 25% 20% 15% 10% 1.0% 21.6% 22.6% 0.6% 44.4% 46.5% 0.0% Positive Pseudomonas No Pseudomonas Jones, BE. CID, 2015 5% 0% 14.0% 10.9% 11.5% All CAP HCAP PNA? 14
HCAP: Diagnostic Performance From 2006-2010: Sensitivity increased (a little) Specificity decreased ( a lot) Diagnostic odds no better Meaning? The percentage of patients covered for MRSA and Pseudomonas without MRSA / Pseudomonas doubled 1/3 of patients with positive cultures did not get coverage Our current approach really sucks Predicting Drug Resistant Organisms An Evolving Story HCAP risk factors not all associated with MDROs Prior MRSA / pseudomonal infection : Important! Treatment for MDROs and Outcomes Severity adjusted outcomes not worse with HCAP Risk stratification for MDROs necessary Low risk: outcomes good with CAP treatment High risk: outcome better? with broad spectrum rx JHM, JGIM: 2012 Predicting Drug Resistant Organisms Risk Scores: %Accuracy ATS HCAP Criteria 69% DRIP 81% Schreiber 68% Shorr 56% Niederman 65% Shindo 67% Aliberti 66% Park 73% (Tested in culture positive patients, NPV >>> PPV) Webb, AAC. 2016 Predicting MRSA and Pseudomonas MRSA More Likely Prior MRSA IVDU Dialysis High MRSA prevalence Prior hospital Less Likely No nasal colonization Pseudomonas More Likely Prior pseudomonas COPD Bronchiectasis Prior fluroquinolones Inhaled steroids Prior hospital Metersky, Resp 2016 Waterer, Resp 2016 15
Putting it All Together Strong Risk Factors for Resistant Organisms Prior Hospitalization in past 90 days / prior broad abx LTAC / SNF if prior antibiotics, poor functional status Critically Ill patients Prior MRSA / Pseudomonas Bronchiectasis / COPD (recurrent abx / chronic steroids) Weaker, Unclear Nursing home (alone) Dialysis (may be MRSA risk) Wound Care / Home Health / Feeding tubes / PPI/ H2 Too heterogeneous (but important) Immunosuppressed Higher Illness Severity Lower Illness Severity A General Approach Evaluate MDRO Risk Factors Quantity Risk 1 2 3 4 5 6 If Broad Spectrum R x Blood/Sputum CX De-escalate if negative at 48-72 7 day treatment is sufficient for most High MDRO Prevalence / Prior MDRO Low MDRO Prevalence Legend: Threshold for broad spectrum Rx Conclusions CXR / Etiologies / PCT CXR is imperfect and most CAP is viral Following PCT protocols will help reduce abx overuse Macrolides Data still supports use in sicker pts Ward patients: benefit unclear? if no legionella Steroids May hold some promise Need further study to identify a target population, dosing Patients at risk for resistant organisms Risk stratify and target broad spectrum treatment De-escalate if cx negative; treat for 1 week THANK YOU! QUESTIONS? 16
De-escalation: HCAP Oral Antibiotic Options Augmentin Respiratory Fluoroquinolone (Levofloxacin, Moxifloxacin) Oral 3 rd generation cephalosporin 17