Background At diagnosis, multiple myeloma (MM) typically presents with a constellation of clinical features, including anemia, renal impairment, hypoalbuminemia, and hyperglobulinemia. However, the temporal evolution of these clinical manifestations in the pre-diagnostic phase remains poorly investigated. We hypothesize that some sub-clinical abnormalities may precede overt MM onset, potentially creating a permissive microenvironment that facilitates malignant plasma cell clonal expansion along the disease susceptibility continuum. This prediction model could serve as a foundation for health surveillance strategies in individuals at risk of developing MM.

Methods Study Design: Prospective cohort study using UK Biobank data (n=501,967) with median follow-up of 12.8 years. MM cases (n=1,371) were identified through ICD-10 codes and cancer registries.

Exposures: Eight biomarkers were analyzed: Anemia markers (hemoglobin, reticulocyte count, red blood cell width), Renal function markers (creatinine, cystatin C), and Protein markers (albumin, albumin/globulin ratio, total protein).

Statistical Analysis: Cox proportional hazards models with standardized biomarkers, adjusted for age, sex, and ethnicity. Trajectory analysis examined biomarker patterns across time-stratified onset groups. Multivariate analysis assessed independent predictive value.

Results Biomarker Analysis: All eight biomarkers demonstrated statistically significant associations with the future risk of MM onset. Protein markers demonstrated the strongest effects: total protein elevation (HR: 1.57, p<2×10⁻¹⁶), albumin/globulin ratio reduction (HR: 0.60, p<2×10⁻¹⁶), and hypoalbuminemia (HR: 0.78, p<2×10⁻¹⁶). Anemia markers included hemoglobin reduction (HR: 0.74, p<2×10⁻¹⁶) and increased red blood cell width (HR: 1.16, p<2×10⁻¹⁶). Renal markers showed elevated cystatin C (HR: 1.12, p<2×10⁻¹⁶) and creatinine (HR: 1.04, p=0.015).

Temporal Evolution: Time-stratified analysis revealed distinct biomarker patterns across MM onset groups (0-3 years, 4-7 years, 7-11 years, and 11+ years after baseline). Protein markers showed the most significant temporal variation (total protein F=18.4, p=1.1×10⁻¹¹; albumin/globulin ratio F=11.5, p=2.1×10⁻⁷), followed by anemia markers (hemoglobin F=10.9, p=4.6×10⁻⁷). Participants developing MM within 0-3 years showed the most pronounced abnormalities, with progressive normalization in longer-term onset groups. This temporal gradient of biomarker abnormalities, extending beyond a decade before diagnosis, suggests that these markers could serve as early predictors of MM risk.

Multivariate Model Performance: Six biomarkers remained independently significant in multivariate analysis (protein abnormalities: 3/3, anemia markers: 2/3, renal markers: 1/2). The integrated model achieved good discrimination with C-index of 0.737 (95% CI: 0.722-0.752), indicating clinically meaningful risk stratification capability to identify individuals at high risk of developing MM using routine laboratory parameters.

Conclusions This large-scale prospective cohort study identified routine laboratory biomarkers as preclinical indicators of MM, with distinct temporal evolution patterns spanning throughout a >10-year pre-diagnostic phase. The identification of independent predictors with good discriminative performance (C-index=0.737) provides a foundation for developing risk stratification algorithms. The temporal hierarchy of biomarker abnormalities suggests optimal screening strategies could focus on protein markers for ultra-early detection (11+ years) and incorporate anemia markers for intermediate-term risk assessment (3-7 years).

Our findings challenge the conventional understanding that anemia, renal impairment, and hypoalbuminemia are solely consequences of MM. We observed that these abnormalities could emerge in the preclinical phase of MM, potentially predating the clinical diagnosis by several years. This temporal relationship suggests a complex interplay in MM pathogenesis, possibly involving bidirectional causality or shared underlying mechanisms that contribute to both conditions prior to overt disease manifestation. The disease susceptibility gradient concept—where biomarker abnormalities are most pronounced in imminent cases and gradually normalize in longer-term cases—challenges traditional binary disease models and supports MM as a slowly evolving systemic disorder.

This content is only available as a PDF.
2025
Sign in via your Institution