Editor’s Note:
Half of all global child deaths under the age of 5 are caused by infections, and half of all sepsis patients worldwide are children, especially newborns. These alarming statistics highlight the significant challenge of infection treatment in pediatric clinical practice. At the recent IDWeek 2023 congress in the United States, a predictive model for severe infection in infants under 90 days of age, based on clinical data and inflammatory biomarkers, was reported. This predictive model may aid in the early identification of newborns at high risk of severe infection, allowing timely clinical intervention to reduce the risk of disease progression. Dr. Mei Zeng from the Children’s Hospital of Fudan University/National Children’s Medical Center provides an introduction and commentary on this research.
Research Overview
Severe bacterial infection (SBI) is a life-threatening disease in infants under 90 days of age, and timely differentiation from non-bacterial infections (NBI) is crucial. Recently recommended low-threshold inflammatory biomarker combinations still need further validation. This study aimed to develop a predictive model for SBI in infants under 29 days of age and those aged 29-90 days. The study used detailed clinical data and high-threshold inflammatory biomarkers.
This retrospective cohort study was conducted in a tertiary referral hospital in northern Israel from 2010 to 2019. The study included previously healthy full-term infants under 90 days of age who were admitted to the pediatric ward for febrile illnesses. SBI was defined as bacteremia, meningitis, or urinary tract infection, while NBI was defined as the absence of bacterial infection signs. Detailed clinical and laboratory data were collected for both age groups.
A total of 1,250 patients met the inclusion criteria, with 515 (41.2%) in the under 29 days age group (young age group) and 735 (58.8%) in the 28-89 days age group (older age group).
In the young age group: There were 95 cases (18.4%) with SBI, and risk factors for SBI included diarrhea, absolute neutrophil count (ANC) >75mm3, and C-reactive protein (CRP) >3mg/dL; family members having viral infection symptoms were protective factors for SBI in infants in the young age group.
In the older age group: There were 111 cases (15.1%) with SBI, and risk factors for SBI included ANC >75mm3, CRP >3mg/dL; while being male, family members having viral infection symptoms, and diarrhea were protective factors for SBI in infants in the older age group.
Based on the above predictive factors, such as local infection or abnormal urine analysis in the physical examination, a predictive model called “NeoSBIscore” was developed. The model had a sensitivity of 94.7% in the young age group, with a negative predictive value (NPV) of 97.0%, and a sensitivity of 92.8% in the older age group, with an NPV of 94.7%. In both age groups, only cases of urinary tract infection or bacteriuria were missed.
Dr. Mei Zeng Commentary
Infections are a significant cause of death in infants and young children. Global Burden of Disease (GBD) analysis shows that infectious causes account for 49.2% of deaths in children under 5 years of age, with preterm complications (17.7%), lower respiratory infections (13.9%), and diarrhea (9.1%) being common causes of death. These diseases are also important causes of neonatal sepsis, with data indicating that half of sepsis cases occur in children, especially newborns. Early identification of severe infections and reducing their progression to sepsis is a major challenge in pediatric clinical practice. This is because the diagnosis and treatment of severe infections and sepsis are a race against time.
In recent years, while molecular diagnostic techniques like metagenomic next-generation sequencing (mNGS) have improved the speed of pathogen diagnosis to some extent, they still do not provide real-time or instant reporting. Moreover, molecular diagnostics may not be readily available in some hospitals, particularly at the primary care level, and can be costly. Utilizing various biomarkers and clinical indicators for comprehensive assessment is an important method for early diagnosis, such as the SOFA and qSOFA scores for sepsis. The study mentioned above developed a predictive model for severe infection risk in infants under 90 days of age based on clinical symptoms and biomarkers like ANC and CRP, demonstrating good potential for clinical application with sensitivities and negative predictive values (NPV) both exceeding 93%.
In fact, research has been ongoing since the 1980s and 1990s to develop and validate clinical predictive models for severe bacterial infections (SBIs). However, there are variations in the definitions of SBIs in different studies, with common inclusion criteria being urinary tract infections, bacteremia, and bacterial meningitis. Some studies also include bacterial diarrhea, pneumonia, and cellulitis. These predictive models typically rely on clinical presentation, age, urinalysis, white blood cell counts (and/or ANC and its ratio), and cerebrospinal fluid (CSF) analysis. These predictive models often have high NPVs, as the overall incidence of SBI is relatively low. Some studies take the “reverse approach,” identifying low-risk populations for SBI and conducting comprehensive assessments and empirical treatment in infants outside the low-risk category.
Many predictive models, such as PECARN, IMHC, FYIRC, and PROS, have shown that age is a crucial predictive factor for SBI. Therefore, the American Academy of Pediatrics (AAP) released clinical practice guidelines in 2021 for the evaluation and management of febrile infants aged 8-60 days, categorizing infants into three age groups (8-21, 22-28, and 29-60 days) based on evidence-based medicine.
The change in the AAP guidelines mentioned above is a typical example of how such predictive models influence clinical practice. As biomarkers for infectious diseases are continually evolving, more precise predictive models with different implications are likely to emerge in the future. The NeoSBIscore model from the study has good diagnostic sensitivity, but there is still a certain rate of underdiagnosis in patients with urinary tract infections. Further validation is expected before the model can be applied in clinical practice.
References:
[1] Halima Dabaja-Younis, et al. Predictive model for serious bacterial infections in infants less than 90 days of age based on clinical data and a combination of high-threshold inflammatory biomarkers. IDWeek 2023, abstract 99.
[2] GBD 2019 Under-5 Mortality Collaborators. Global, regional, and national progress towardsSustainable Development Goal 3.2 for neonatal and child health: all-cause and cause-specific mortality findings from the Global Burden of Disease Study 2019. Lancet. 2021;398(10303):870-905.
[3] Perin J, Mulick A, Yeung D, et al. Global, regional, and national causes of under-5 mortality in 2000-19: an updated systematic analysis with implications for the Sustainable Development Goals [published correction appears in Lancet Child Adolesc Health. 2022 Jan;6(1):e4]. Lancet Child Adolesc Health. 2022;6(2):106-115.
[4] Rudd KE, Johnson SC, Agesa KM, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet. 2020;395(10219):200-211.
[5] Pantell RH, Roberts KB, Adams WG, et al. Evaluation and Management of Well-Appearing Febrile Infants 8 to 60 Days Old [published correction appears in Pediatrics. 2021 Nov;148(5):]. Pediatrics. 2021;148(2):e2021052228.
