Both the diagnosis and management of infectious diseases in hospitalized patients remain prominent clinical challenges and are associated with a substantial financial burden for healthcare systems. This is particularly true in the case of sepsis and septic shock.
For example, though the in-hospital sepsis mortality rate decreased by 17.0 percent from 2016-2019, with the emergence of COVID-19, it rose by 38% from 2019-2021. The ensuing aggregate annual hospital costs for sepsis now exceed $52.1 billion.1 A recent review of patient safety practices (PSPs) focusing on sepsis prediction and recognition found that current paradigms for early sepsis identification had no effect on clinical processes, length of stay, or mortality compared to usual care.2 However, these PSPs do not reflect the emerging understanding of the complexity and heterogeneity within the host response to infection. A more comprehensive and complete appreciation of these factors has the potential to improve outcomes.
There are two long-established, complementary, diagnostic approaches in cases of suspected infection. One focuses on identifying the pathogen through traditional methods such as culture, immunoassay, PCR, and/or genomic sequencing. The other aims to characterize the host response to infection. Traditional measures of the host response include clinical findings—such as fever, redness or swelling, and anatomical abnormalities revealed through imaging techniques—and laboratory measurements—such as changes in white blood cell counts and inflammatory markers. Neither approach is perfect. Up to 60% of pathogen identification attempts fail, even among severely ill patients3 and the presence of a pathogen may not actually mean that it is responsible for the patients’ syndrome (e.g., a colonizer).4 Similarly, lab-measured host markers are often not specific for infection and can be noted in a variety of non-infectious conditions such as malignancy or trauma.
Nonetheless, host response tests are driving improvements in the care of infected and critically ill patients. For example, the serum biomarker procalcitonin is routinely used to guide antibiotic therapy5 and remains a cornerstone of many antimicrobial stewardship programs. Interleukin-6, likewise, helps to identify COVID-19 patients at high risk for respiratory failure.6,7 Monocyte Distribution Width (MDW), a novel and recently validated hematological host response parameter, has been shown to identify occult sepsis and, as a corollary, to reduce time to antibiotic administration—a key determinant of survival in sepsis.8,9
The next generation of host response diagnostics leverages ‘omic-scale measurements and machine learning methods to identify biomarker panels with not only greater accuracy but also greater actionability. Including multiple biomarkers and their computational integration in these panels allows for the simultaneous readout from multiple biological pathways, improving diagnostic discrimination. Reliance on this strategy facilitated the identification of a three-protein panel (IP-10, CRP, and TRAIL), which evolved into the MeMed BV test.10 The concentrations of these three host proteins are used to derive the MeMed BV score. Importantly, the score helps differentiate bacterial from viral infection, and when properly implemented, the MeMed BV test reduced inappropriate antibiotic use by ~25%.11,12
Sepsis, defined as a dysregulated immune response to infection, can result in life-threatening organ dysfunction. It is a highly heterogeneous syndrome comprised of multiple different pathologies. In the current model of care, nearly all suspected sepsis patients are treated similarly, which fails to account for that heterogeneity and potentially represents a barrier to further improving sepsis outcomes.
One exciting approach to address this challenge leverages the identification of sepsis subphenotypes. By applying machine learning algorithms to large scale data, sepsis subphenotypes have been detected. Transcriptomics have revealed immunocompromised and immunocompetent profiles linked with differential risks for mortality and a heterogeneous response to corticosteroids.13 Using only clinical data, another machine learning algorithm identified four sepsis subphenotypes with differing outcomes and responses to fluid resuscitation.14 Reconciling these and many other subphenotyping schemes will be a challenge for the field but one with the potential to change the way sepsis is diagnosed and managed. In short, future management of sepsis will require clinical and hospital leaders to understand that sepsis care must be individualized in hopes of reducing the mortality and morbidity related to this process.
Implementing biomarker panels has some challenges. Single biomarkers are reported as a concentration or as discrete measured values with reference ranges with straightforward interpretation. Biomarker panels, though, aggregate multiple measurements and typically report a unified (often semi-quantitative) score that can provide insights into a patient’s likely clinical trajectory. As a result, clinical interpretation of emerging biomarker panels is more nuanced, leading to an increased training burden to support effective adoption. When evaluating potential vendors, consider whether the diagnostic manufacturer can provide training and educational materials that clinicians require. Furthermore, hospital leaders may seek customization in these tools to help address the needs of their specific organizations, such as diagnostic accuracy measures (e.g., positive and negative predictive values) to tailor tools to the patient populations served, allowing clinical leaders to engage more collaboratively by clearly presenting the of over-diagnosing or missing real cases in the local patient population.
Strong clinical and executive leadership will be imperative to realizing the potential for these emerging solutions to improve patient outcomes. In well-defined syndromes, such as acute myocardial infarction, a single biomarker (troponin) essentially defines the disease15—which is not the case with biomarkers for less well-defined syndromes, like infection and sepsis. Variation in pathology and baseline immune profiles, as well as a dynamic and evolving immune response throughout the patient’s course requires a more intricate diagnostic approach that will utilize multiple tests and strategies to improve outcomes.
Authors:
Andrew F. Shorr, MD, MPH, MBA, MA, Pulmonary and Critical Care Medicine, Medstar Washington Hospital Center, Washington, DC; Ephraim L. Tsalik, MD, MHS, PhD, FIDSA, Chief Scientific Officer, Infectious Disease, Danaher; and Melissa Naiman, PhD, EMT-B, Medical Director, Sepsis & Host Response—Medical Scientific Affairs, Beckman Coulter*
*This work is the opinion of the authors and does not reflect the opinions of Medstar Health or other entities.
1. Owens PL, Miller MA, Barrett ML, Hensche M. Overview of Outcomes for Inpatient Stays Involving Sepsis, 2016–2021. HCUP Statistical Brief #306. April 2024.
2. Cherian J, Segal J, Sharma R, Zhang A, Bass E, Rosen M. Patient safety practices focused on sepsis prediction and recognition: rapid review. In: Making Healthcare Safer IV: A Continuous Updating of Patient Safety Harms and Practices. Agency for Healthcare Research and Quality (US); 2023.
3. Jain S, Self WH, Wunderink RG, et al. Community-Acquired Pneumonia Requiring Hospitalization among U.S. Adults. N Engl J Med. 2015;373(5):415-427. doi:10.1056/NEJMoa1500245
4. Byington CL, Ampofo K, Stockmann C, et al. Community Surveillance of Respiratory Viruses Among Families in the Utah Better Identification of Germs-Longitudinal Viral Epidemiology (BIG-LoVE) Study. Clin Infect Dis. 2015;61(8):1217-1224. doi:10.1093/cid/civ486
5. Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181-1247. doi:10.1007/s00134-021-06506-y
6. Ruan Q, Yang K, Wang W, Jiang L, Song J. Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150 patients from Wuhan, China. Intensive Care Med. 2020;46(5):846-848. doi:10.1007/s00134-020-05991-x
7. Santa Cruz A, Mendes-Frias A, Oliveira AI, et al. Interleukin-6 Is a Biomarker for the Development of Fatal Severe Acute Respiratory Syndrome Coronavirus 2 Pneumonia. Front Immunol. 2021;12:613422. doi:10.3389/fimmu.2021.613422
8. Hinson JS, Sirani N, Imhoff B, et al. 846 Monocyte Distribution Width and Time to Antibiotics for Septic Patients With Atypical Presentations. In: Special Issue:SAEM2023 Annual Meeting Abstracts. Vol 30. Society for Academic Emergency Medicine; 2023.
9. Hinson JS, Sarani N, Smith A, et al. A-155 Unlocking the Complete Blood Count: Derivation of a Single-Panel Laboratory Test That Includes Monocyte Distribution Width to Create a Universal Sepsis Screening Tool. Clin Chem. 2023;69(Supplement_1). doi:10.1093/clinchem/hvad097.140
10. Oved K, Cohen A, Boico O, et al. A novel host-proteome signature for distinguishing between acute bacterial and viral infections. PLoS ONE. 2015;10(3):e0120012. doi:10.1371/journal.pone.0120012
11. Shapiro Ben David S, Rahamim-Cohen D, Yehoshua I, Kivity S, Kalmovich B. Real-World Impact of MeMed BV, a Rapid Host Response Test for Distinguishing Bacterial and Viral Infections, in an Urgent Care Network. In: European Society of Clinical Microbiology and Infectious Diseases; 2024.
12. Mann E, Wilson C, Klein E. Impact of a novel host-response diagnostic on emergency department antibiotic decision making. In: 2024 SAEM Annual Meeting Abstracts. SAEM; 2024.
13. Cano-Gamez E, Burnham KL, Goh C, et al. An immune dysfunction score for stratification of patients with acute infection based on whole-blood gene expression. Sci Transl Med. 2022;14(669):eabq4433. doi:10.1126/scitranslmed.abq4433
14. Seymour CW, Kennedy JN, Wang S, et al. Derivation, validation, and potential treatment implications of novel clinical phenotypes for sepsis. JAMA. 2019;321(20):2003-2017. doi:10.1001/jama.2019.5791
15. Thygesen K, Alpert JS, Jaffe AS, et al. Fourth universal definition of myocardial infarction (2018). Eur Heart J. 2019;40(3):237-269. doi:10.1093/eurheartj/ehy462