Scientia Psychiatrica
Vol. 7 No. 2 (2026): Scientia Psychiatrica

Inflammatory-Metabolic Burden and Depressive Symptom Phenotypes in US Adults: A Survey-Weighted Latent Class Analysis of NHANES 2021⁠–⁠2023

Henry Peter Paul (Department of Behavioral Science, Miami Histrory Museum, Miami, United States)
Susan Miller (Department of Behavioral Science, Miami Histrory Museum, Miami, United States)
Ramakhrishnand Ramakhrishnand (Department of Drugs Development, Avant Medical Center, Avarua, Cook Island)
Theresia Putri Sinaga (Department of Public Health, CMHC Research Center, Palembang, Indonesia)



Article Info

Publish Date
30 Apr 2026

Abstract

Background. Depression is clinically heterogeneous, and total Patient Health Questionnaire-9 (PHQ-9) scores can conceal symptom configurations with different inflammatory and metabolic correlates. Evidence linking immune-inflammatory activity and insulin resistance to depression has mostly addressed total symptom burden or binary thresholds. Objective. To identify latent PHQ-9 symptom phenotypes in US adults and to evaluate whether the systemic immune-inflammation index (SII), the triglyceride-glucose index (TyG), and their interaction were associated with phenotype membership. Methods. This cross-sectional secondary analysis used the National Health and Nutrition Examination Survey (NHANES) August 2021–August 2023 fasting subsample (adults aged at least 20 years with complete PHQ-9, blood count, fasting glucose, and triglyceride data). Survey-weighted ordinal latent class models (two to five classes) were compared. Multinomial regression (modal class membership; jackknife repeated-replication variance) adjusted for age, sex, race and ethnicity, education, body mass index, current smoking, and diabetes. Results. Among 2,427 adults, a three-class solution was chosen: Minimal (53.4%), Somatic-dominant (32.5%), and Mixed-severe (14.1%); entropy was 0.810 and average posterior probability 0.915. In the adjusted sample (n = 2,409), odds ratios for Mixed-severe versus Minimal membership per one-standard-deviation increase were 1.16 (95% confidence interval [CI] 0.94–1.42) for SII, 1.10 (95% CI 0.89–1.36) for TyG, and 1.08 (95% CI 0.93–1.24) for SII×TyG. The global interaction test was not significant (p = 0.470). Continuous-score analyses agreed. Conclusion. PHQ-9 responses identified three well-separated symptom phenotypes. SII, TyG, and their interaction did not independently distinguish phenotype membership. Phenotyping may refine epidemiological description. Single-occasion biomarkers of this kind should not currently assign depressive subtypes.

Copyrights © 2026






Journal Info

Abbrev

SciPsy

Publisher

Subject

Health Professions Medicine & Pharmacology Neuroscience Public Health

Description

Scientia Psychiatrica covers the latest developments in various fields of psychiatric : biological psychiatry, neuropsychiatry, child psychiatry, psychiatry of community, psychotherapy, drugs-related mental illness, psychiatry of geriatric, psychosomatics medicine, psychology, cultural psychiatry, ...