115  Aspire Presentation

115.1 Summary

• Methods of Identifying Inpatients with Hypercapnia for ResearchResearch In Progress 3/2/23
• Hypercapnic Respiratory Failure: ’Syndrome’
• Proposed framework:Sufficient vs Component Causes (Rubin)
• COPD, Very Severe
• How common are admissions with hypercapnia?
• Admissions with Hypercapnic RF → High Risk of Readmission
• Who are “all-comers admitted with hypercapnia”?
• 5% dx’d obesity; 6% dx’d OSA
• What is the role of sleep-disordered breathing?
• Needs: initial decompensation → treatment
• Computable phenotypes: algorithms to identify specific patient cohorts
• Why might this help in Hypercapnic RF? (why not just use PaCO2 > 45 mmHg?)

115.2 Slide outline

115.2.1 Slide 1

• Methods of Identifying Inpatients with Hypercapnia for ResearchResearch In Progress 3/2/23
• Mentorship Team:
• Krishna Sundar MD (PCCM/Sleep)
• Ram Gouripeddi MBBS MSc (Bioinformatics)Jeanette Brown MD PhD (PCCM)
• Rob Paine III MD (PCCM)
• Wayne Richards BSc (Bioinformatics)
• Brian Locke, MD
• 3rd Year Pulmonary and Critical Care Fellow
• University of Utah
• University of Utah T32 PI: Paine
• Masters of Science in Clinical Investigation
• ATS ASPIRE Program ### Slide 2
• Hypercapnic Respiratory Failure: ’Syndrome’
• Definition: PaCO2 too high (> 45 mmHg)
• Key physiologic principles:
• Occurrence of CO2↑ : Ventilatory Control or Actuator Frailty?
• Metabolic parabola means those with hypercapnia are more prone to decompensation
• Many patients are first identified during decompensation:
• higher risk for morbidity
• also, more available data
• Acute management effective, but what should happen after the admission?
• Build up here CO2 in blood rises ### Slide 3
• Proposed framework:Sufficient vs Component Causes (Rubin)
• Obesity
• Untreated
• Sleep Apnea
• COPD
• Loop diuretic
• Opiates
• Sufficient cause: a set of factors that will cause disease when present
• End-stage COPD with hyperinflation
• Component cause: a factor that, if not present, no disease would occur
• Untreated OSA, metabolic alkalosis, chronic opiate use
• Hypothetical patient with hypercapnic respiratory failure ### Slide 4
• COPD, Very Severe
• PaCO2 52 mmHg
• BMI 45 kg/m2
• AHI 50 event/hr
• PaCO2 50 mmHg
• COPD, mod severe
• AHI 50 events/hr
• AHI 10 events/hr
• Met. Alk loop diur
• COPD
• Trial Data
• +Guidelines
• Wait 2-4 wks
• OHS
• Observational Data
• +Guideline
• Start Immediately
• Overlap Sydnrome
• Undifferentiated
• No tx data
• Limited epi data
• Swap any combination of:
• Muscular weakness
• Obesity
• Opiate use
• Lung disease
• Etc.
• ???
• Post-Admission for Hypercapnic Respiratory Failure Management ### Slide 5
• How common are admissions with hypercapnia?
• 2022: Liverpool AUS, 1 regional hospital services the entire district
• Identified by initial ABG (w/n 24h) PaCO2 over 45, excluded iatrogenic causes/sedation. N891 people, 1135 blood gasses (repeat hosp.)
• Normalized rates of hypercapnia to population demographics
• 150 per 100,000 person/year
• Acidosis in 55%.
• For Comparison: ### Slide 6
• Admissions with Hypercapnic RF → High Risk of Readmission
• Admitted to hospital (Floor or ICU) with billing code for Hypercapnic Respiratory Failure ### Slide 7
• Who are “all-comers admitted with hypercapnia”?
• Liverpool AUS: ABG w/n 24h of admission.
• Diagnostic codes: 52% Charlson Comorbidity 5+; Prevalence-based estimates: NMD under-represented
• Chung Y, Garden FL, Marks GB, Vedam H. Causes of hypercapnic respiratory failure and associated in-hospital mortality. Respirology. 2022
• All adult ED visits @ single Toronto ED w ABG <7.35/45+ OR VBG <7.34/50+ & CEDIS resp code for resp symptoms. Charts reviewed for 12 months.
• Giulia Cavalot, Vera Dounaevskaia, Fernando Vieira, Thomas Piraino, Remi Coudroy, Orla Smith, David A. Hall, Karen E. A. Burns & Laurent Brochard (2021) One-Year Readmission Following Undifferentiated Acute Hypercapnic Respiratory Failure, COPD: Journal of Chronic Obstructive Pulmonary Disease, 18:6, 602-611
• All patients with dyspnea or pulm disease admitted to hospital received capillary blood gas (some screening with VBG). Stratified by pH > 7.35 or pH < 7.35
• Hospital specializing in lung disease (unclear referral pattern):
• Vonderbank S, Gibis N, Schulz A, Boyko M, Erbuth A, Gürleyen H, Bastian A. Hypercapnia at Hospital Admission as a Predictor of Mortality. Open access emergency medicine : OAEM 2020; 12: 173-180. ### Slide 8
• 5% dx’d obesity; 6% dx’d OSA
• Chung Y, Garden FL, Marks GB, Vedam H. Causes of hypercapnic respiratory failure and associated in-hospital mortality. Respirology. 2022
• Giulia Cavalot, Vera Dounaevskaia, Fernando Vieira, Thomas Piraino, Remi Coudroy, Orla Smith, David A. Hall, Karen E. A. Burns & Laurent Brochard (2021) One-Year Readmission Following Undifferentiated Acute Hypercapnic Respiratory Failure, COPD: Journal of Chronic Obstructive Pulmonary Disease, 18:6, 602-611
• 10.8% OSAS
• Vonderbank S, Gibis N, Schulz A, Boyko M, Erbuth A, Gürleyen H, Bastian A. Hypercapnia at Hospital Admission as a Predictor of Mortality. Open access emergency medicine : OAEM 2020; 12: 173-180.
• OSA/OHS – 19.3%
• Retrospective Reviews: mirrors what rate these are diagnosed (not what rate they present)
• What is the role of sleep-disordered breathing? ### Slide 9
• What is the role of sleep-disordered breathing?
• COPD (66%)
• No COPD (33%)
• AHI Median [IQR]
• 31.9 [14.3, 45.6]
• 66.0 [48.0, 83.8]
• AHI > 15 present
• 66%
• 94%
• AHI > 30 present
• 51%
• 81%
• Overall prevalence of severe OSA ( AHI > 30 ) in survivors over 50%
• & No prior OSA Dx
• N36 survivors of hypercapnia; n16 agreed to PSG. 56% with AHI > 30 /hr ### Slide 10
• Needs: initial decompensation → treatment
• Presentation With Hypercapnia
• Hypercapnia is common, highly morbid, and likely increasing in frequency (obesity, opiates, multimorbidity)
• Correctly Recognized?
• Causes Correctly Diagnosed?
• Is there data to guide treatment?
• Is it logistically possible?
• Does it happen?
• Inpt Hypercapnia is frequently missed in practice
• Many patients do not receive definitive testing
• We don’t know what to do for many (most?) patients with hypercapnia
• Device qualification restrictions.
• Requires coordination between Inpt  Outpt. Consider CHF (CMS f/u after discharge)
• Marik, JICM 2012: 75% miss rate.
• Nowbar, AJM 2004
• OSA in 5-15% vs 50-80%
• No good estimate? (NMD/Rest, COPD, OHS, OVS)
• No good estimate? ### Slide 11
• Computable phenotypes: algorithms to identify specific patient cohorts
• Problems with retrospective EHR-based research:
• Missing Data
• Time/Event determination (immortal time)
• Observational nature (residual confounding; multiple comparisons)
• Disease Identification
• Contrast: Cohort study with assessment and verification of diagnostic labels. Particularly, respiratory diagnoses (Asthma, COPD, OHS, OSA)
• Precise definitions (phenotypes) can help with patient identification
• Simple: Inclusion criteria to a study
• Complex: use additional data elements (demographics, meds), natural-language processing, etc. ### Slide 12
• Why might this help in Hypercapnic RF? (why not just use PaCO2 > 45 mmHg?)
• ABGs are variably obtained in clinical practice: less reliable in usual care than enrollment criteria
• People who get ABGs are different from people who don’t
• Symptoms are variable, the diagnosis is often missed.
• Pre-test probability varies widely by demographic and comorbidity; optimal evidence may vary too
• Pre-test Odds LR (Strength of Evidence) Post-test Odds
• Study Aim
• Efficacy Trial
• Determine Burden of Disease
• Improve Processes of Care
• Ideal Characteristic
• Select a population with severe disease, homogeneous in characteristic that predicts treatment responsiveness
• Capture all patients who suffer from the adverse consequences of hypercapnic respiratory failure
• Identify all patients who will benefit from instituting certain management pathways
• Example
• Severe COPD, no other cause of CO2, severe air trapping, PaCO2 > 52 mmHg
• More sensitive definition?
• (If only use PaCO2, you will miss patients and skew toward people with classic/severe presentation) ### Slide 13
• Gap & Hypothesis
• Methods used in the literature will identify different populations
• Those populations will differ in ways important to prognosis or treatment response (such as comorbidities, management)
• Data from EHR’s are sufficiently granular to identify pathophysiology-based endotypes, which is currently presumed to matter clinically, but not reflected in current epidemiologic data.
• These endotypes may predict risk and heterogeneity in treatment responsiveness, and thus would be useful to identify ### Slide 14
• Study
• Definition Used
• Aim
• Meservey et al 2020. USA (Vermont)
• DOI: 10.1007/s00408-019-00300-w
• 18 y/o+ admitted (either ICU or floor) with diagnostic code for hypercapnic respiratory failure
• What features of patients admitted for hypercapnic respiratory failure predict readmission?
• Chung et al, AUS, 2021 DOI: DOI: 10.1164/rccm.202108-1912LE
• Identified by initial ABG PaCO2 over 45, excluded iatrogenic causes/sedation.
• What is the population prevalence of hypercapnia from any cause? Also, subsequent publication evaluating comorbidity prevalence.
• Vonderbank et al 2020. Germany
• DOI: 10.2147%2FOAEM.S242075
• All patients with dyspnea or pulm disease admitted to hospital received capillary blood gas (some screening with VBG) ; PaCO2 45 mmHg.
• Is hypercapnia predictive of mortality at a specialty hospital?
• Mini-Systematic Review: (Reviewed in singlicate)
• EMBASE search for human cohort studies: “hypercapnic respiratory failure” “hypercarbic respiratory failure” “hypoventilation” “respiratory acidosis” and all EMTREE subtypes.
• AS Review – AI-assisted systematic review
• 9455 → 10 studies evaluating prevalence, comorbidity profile, post-discharge prognosis discharge of patients with hypercapnic respiratory failure
• van de Schoot, R. et al. An open source machine learning framework for efficient and transparent systematic reviews.
• Nat Mach Intell
• DOI:10.1038/s42256-020-00287-7 ### Slide 15
• Prevalence of OSA
• Study
• Definition Used
• Aim
• Thille et al 2018 France
• DOI: 10.1007/s00134-017-4998-3
• ICU; pH<7.35 & PaCO2 45+; treated with NIV or IMV; survivors
• What is the prevalence of undiagnosed OSA among survivors of Hypercapnic RF?
• Ouanes-Besbes et al. 2021, Tunisia PubMed: 33171053
• Admitted to ICU 2015-2018, PaCO2 > 45mmHg, pH < 7.35. No prior OSA syndrome diagnosis.
• Adler et al. 2017, Switzerland
• DOI: 10.1164/rccm.201608-1666OC
• Recruited at ICU discharge after surviving: primary admission for Resp failure, PaCO2 > 47.5 mmHg and requiring NIV or IMV
• What is the prevalence of comorbidities, including OSA, among survivors of Hypercapnic RF? ### Slide 16
• Evaluation of Subgroup by Compensation Status
• Study
• Definition Used
• Aim
• Compensated Hypercapnia
• Wilson et al., 2021. USA (Michigan)
• PubMed: 33951397
• 18+, hospitalized w/ ABG showing CO2 over 45 and pH 7.35-7.45
• Prognostic Significance of Compensated Hypercapnia
• Bulbul et al. 2014. Turkey.
• doi: 10.4103/1817-1737.128851
• Hospitalized adults with PaCO2 over 45, no acidosis, and breathing room air
• What comorbidities occur with inpatient, compensated hypoventilation
• Marik 2016. USA (Virginia)
• DOI: 10.1002/osp4.27
• Age 18-90; BMI ≥ 40 kg m2; daytime PaCO2 > 45 mmHg, admission HCO3 > 28 mEq L−1.
• Excluded: intrinsic lung disease; thoracic MSK/NMD; 20+ pack year smoking, COPD
• What is the prevalence of undiagnosed Obesity Hypoventilation among hospitalized adults?
• Respiratory Acidosis
• Cavalot et al 2021, Toronto DOI: 10.1080/15412555.2021.1990240
• ABG with pH < 7.35 and PaCO2 > 45 mmHg OR VBG pH 7.34 and PvVO2 > 50 mmHg & presence of respiratory symptoms (CEDIS codes).
• Exclude CF, NM dz, ILD, thoracic dz, lung ca, CNS dz, Overdose, or trach
• 1 year re-admission rate for patients presenting with acute respiratory acidosis ### Slide 17
• Methods: Data source
• De-identified, patient-level data (not PHI, though recommended to be treated as such)
• Federated data from medical center EHRs
• 70 Healthcare Organizations; mostly (but not exclusively) AMCs
• 80+ gb of data.
• 1.2 Billion ‘facts’ in medications alone.
• Center for High Performance Computing protected environment
• Data Available
• Data unavailable
• Demographics
• Diagnosis
• Medications
• Procedures
• Labs
• Cancer registry (NAACCR)
• Allergies
• Death Records (in progress)
• Some notes
• Diagnostic reports
• DICOM image objects
• Providers
• Departments/clinics ### Slide 18
• Methods: Data Request
• We requested data on all inpatient encounters where the individual met one of the following criteria
• ICD group, Hypercapnic Respiratory Failure ICD Code. Patients with an inpatient diagnostic code for any of the following:
• · J96.12 (chronic respiratory failure with hypercapnia),
• · J96.02 (acute hypercapneic respiratory failure),
• · J96.22 (acute and chronic respiratory failure with hypercapnia),
• · J96.92 (respiratory failure unspecified with hypercapnia),
• · E66.2 (morbid obesity with hypoventilation)
• ABG Group, First ABG showing Hypercapnia. Patients who have a value over 45 on the first first of any of the following lab codes obtained during the hospitalization
• · LOINC 2019-8 Carbon Dioxide [ Partial Pressure ] Arterial Blood over 45[BL3]
• · LOINC 32771-8 Carbon Dioxide [ Partial Pressure] adjusted to patient’s actual temperature in arterial blood over 45
• · LOINC 11557-6 Carbon Dioxide [ Partial Pressure] in Blood over 45
• And:
• Age > 18 at admission
• Admission between 1-1-2018 and 9-28-2022
• Only first admission for per-individual that met the criteria
• Cross-sectional analysis:
• no dependence on cross-encounter data (to limit data quality issues) ### Slide 19
• Methods: Data Processing
• Mechanisms of Data Missingness:
• E.g. Diagnoses
• For a given encounter, the institution must have submitted some data (any data) for each of the data elements requested.
• Data validation measures:
• Positive controls: e.g if dx of Pna, expect abx Rx
• Negative controls: <BMI 30 in OHS
• True (patho)physiologic state
• What the clinician recognizes
• What the clinician documents
• What information is structured
• What data is aggregated in TriNetX
• Problem for our study
• Missing data here is part of what we want to study ### Slide 20
• Results: Cohort Description
• Final cohort: 119,816
• ‘Missing’ Data Rates:
• ABG: 30%; VBG 70%
• BMP, CBC elements 3%
• Triage VS: 30% (BP) -70% (SpO2)
• Control outcomes:
• Receive paralytic → IMV code? 80% (OR?)
• AECOPD/Asthma → Systemic Steroid? 88%
• Pneumonia dx code → CAP Antibiotic? 81%
• 56% have critical care services billed
• 52% with acidemia (<7.35 pH); higher likelihood of critical care (OR 1.77)
• CHF (OR 0.82) and OSA (OR 0.93) less likely acidemic
• COPD (OR 1.16), CKD (1.21), and OUD (1.10) more likely ### Slide 21
• Overlap between definitions ### Slide 22
• How do the patients differ?
• Total
• Chung 2021
• Meservey 2020
• Adler 2016
• N119,816
• N104,253
• N30,414
• N18,143
• Age
• 61 (±16)
• 62 (±15)
• 60 (±16)
• Female?
• 43% (51,315)
• 42% (43,975)
• 48% (14,506)
• 40% (7,192)
• Black
• 19% (22,824)
• 19% (20,041)
• 19% (5,917)
• 19% (3,513)
• Latino
• 7% (6,867)
• 7% (6,027)
• 6% (1,565)
• 9% (1,376)
• BMI
• 30 (±8)
• 29 (±7)
• 31 (±9)
• CHF
• 13% (15,655)
• 13% (13,126)
• 23% (6,936)
• 19% (3,435)
• COPD
• 21% (24,961)
• 20% (20,649)
• 31% (9,576)
• 21% (3,897)
• Neuromuscular Disease
• 3% (3,619)
• 3% (2,933)
• 4% (1,173)
• 3% (507)
• OSA
• 20% (24,446)
• 19% (19,698)
• 28% (8,568)
• 19% (3,488)
• Opiate Use Disorder
• 4% (4,354)
• 3% (2,955)
• 4% (1,343)
• 4% (696) ### Slide 23
• How do the physiology and management differ?
• Total
• Chung 2021
• Meservey 2020
• Adler 2016
• N119,816
• N104,253
• N30,414
• N18,143
• Acidosis by ABG or VBG?
• 59% (51,546)
• 60% (47,282)
• 62% (11,741)
• 64% (9,151)
• Bicarbonate
• 25 (±6)
• 27 (±7)
• 24 (±7)
• Critical Care Services
• 40% (47,500)
• 39% (40,176)
• 54% (16,467)
• 100% (18,143)
• CPAP proc. code
• 9% (10,777)
• 8% (8,803)
• 16% (5,002)
• 17% (3,172)
• Procedure code for NIV (not nightly CPAP)
• 12% (14,296)
• 11% (11,328)
• 25% (7,474)
• 23% (4,224)
• Procedure code for IMV
• 39% (46,222)
• 39% (40,337)
• 46% (14,051)
• 89% (16,088)
• Length of Stay
• In Hospital Mortality ### Slide 24
• Who do guidelines apply to?
• Based on diagnostic codes present
• ATS OHS Guideline (2019):
• Obesity with Hypoventilation Dx: 4,839 (4%)
• Obese and no COPD/Asthma/CHF/NMD/OUD Dx: 62,400 (51%)
• Missing data…
• Obese and OSA and no COPD/Asthma/CHF/NMD/OUD dx: 7,538 (6%)
• ATS COPD NIV Guideline (2020): 9,659 (9%)
• Neither guideline: 40-87% ### Slide 25
• Can we identify relevant physiology?
• COPD
• OSA
• OUD
• TODO:
• Principal Component Analysis: most parsimonious set of variables that explains how patients with hypercapnia differ from each other
• Clustering on this vs. expert features (can’t vs won’t breathe; compensation etc.)
• → physiologic endotypes?
• Conditional Probability the Dx is Present ### Slide 26
• Strengths and Limitations
• What does ”big data” get you?
• Less susceptible to idiosyncratic local patterns; maybe increased representativeness
• Power for subgroup investigations
• Data hungry algorithms
• All single-encounter data (cross-sectional) due to data quality
• Can’t evaluate readmissions, mortality, subsequent outpatient management.
• Data quality: contingent on local practices (hard to assess)
• Doesn’t include patients where ABG wasn’t ordered or Dx not given
  
• Inferences about underlying true physiology much less robust than inferences about coding practices and inclusion criteria due to secondary use data. ### Slide 27
• Take aways:
• Multiple etiologic causes contribute to hypercapnia in many cases and further research is needed to guide how to address high rates of morbidity and mortality.
• Rationale exists that current tools (e.g. PAP; or rehab, acetazolamide, etc) could be utilized
• Different Cohort definitions do select non-overlapping and clinically different subsets
• These cohorts differ in comorbidity distribution and physiology → treatment responsiveness and prognosis?
• Therefore, the interpretability of future research might be improved with agreed-upon definitions
• Next step: How should we identify patients with hypercapnic respiratory failure ### Slide 28
• Next steps: How should we identify patients?
• Request TriNetX dataset including:
• Patients with hypercapnic RF (ICD, PaCO2)
• Patients a reasonable clinician might suspect to have hypercapnic RF
• Other respiratory failure + predisposing condition (e.g. NMD)
• +VBG CO2 over 48 mmHg
• Severe Obesity and elevated bicarbonate
• …etc
• Model: predictive abilities of feature (or their combinations)
• Output: Usable by researchers
• Accuracy vs Usability (NLP?)
• Binary output vs Probability?
• ICD Hypercap code and ABG CO2 > 45
• ICD Hypercap code and only VBG w CO2 > 48
• N19,093
• N10,018
• Age
• 63 (±15)
• 61 (±16)
• Location
• midwest
• 21% (3,919)
• 13% (1,331)
• northeast
• 25% (4,705)
• 21% (2,143)
• south
• 46% (8,806)
• 32% (3,233)
• west
• 9% (1,663)
• 33% (3,311)
• BMI
• 30 (±8)
• 32 (±9)
• CHF
• 26% (4,913)
• 18% (1,789)
• COPD
• 33% (6,300)
• 30% (2,974)
• Neuromuscular Disease
• 3% (655)
• 5% (452)
• OSA
• 25% (4,854)
• 34% (3,381)
• Pre-test Odds LR (Strength of Evidence) Post-test Odds ### Slide 29
• The Issue of The Reference Standard
• No Free Lunch: Except for measuring, all methods rely on assumptions
• Formulation as a missing data problem
• Withold ABG results and perform supervised learning on those with ABGs
• Model propensity to obtain P(ABG) and generated Inverse Probability Weights
• Assumptions not met, but would it be better?
• Classification (Hypercapnia vs not) → Prediction (risk of hypercapnia-morbidity)
• Validate sparse definition to richer data-set
• How well can you do with data elements available in NCUPS Readmission Db?
• Prospective: Also allows for
• Bridges the “documentation to physiology gap”
• Allows for more robust assessment of comorbidity status
• Nowbar 2004 used ABG → TcCO2 more tolerable? ### Slide 30
• Questions? Links
• Hypercapnia Studies (link to github with excel spreadsheet download)
• AS Review Lab: AI assisted systematic review
• https://asreview.nl/
• https://trinetx.com/
• Use cases for well-defined computable phenotype:
• Improved study enrollment
• Improved ascertainment as covariate
• Improved outcome (e.g. ambient air pollution)
• Mentorship Team:
• Krishna Sundar MD (PCCM/Sleep)
• Ram Gouripeddi MBBS MSc (Bioinformatics)Jeanette Brown MD PhD (PCCM)
• Rob Paine III MD (PCCM)
• Wayne Richards BSc (Bioinformatics)

115.3 Learning objectives

• Methods of Identifying Inpatients with Hypercapnia for ResearchResearch In Progress 3/2/23
• Hypercapnic Respiratory Failure: ’Syndrome’
• Proposed framework:Sufficient vs Component Causes (Rubin)
• COPD, Very Severe
• How common are admissions with hypercapnia?

115.4 Bottom line / summary

• Methods of Identifying Inpatients with Hypercapnia for ResearchResearch In Progress 3/2/23
• Hypercapnic Respiratory Failure: ’Syndrome’
• Proposed framework:Sufficient vs Component Causes (Rubin)
• COPD, Very Severe
• How common are admissions with hypercapnia?

115.5 Approach

1. TODO: Outline the initial assessment or decision point.
2. TODO: Outline the next diagnostic or management step.
3. TODO: Outline follow-up or escalation criteria.

115.6 Red flags / when to escalate

• TODO: List red flags that require urgent escalation.

115.7 Common pitfalls

• TODO: Capture common errors or missed steps.

115.8 References

TODO: Add landmark references or guideline citations.

115.9 Slides and assets

• Presentations/2023-3-2 ASPIRE presentation.pptx

115.10 Source materials

• Presentations/2023-3-2 ASPIRE presentation.pptx