Blood Inflammation Markers Track With Hearing Loss in U.S. Adults

Adults with hearing loss show measurably different patterns of circulating immune cells than adults with normal hearing, according to a large serial cross-sectional analysis of U.S. National Health and Nutrition Examination Survey (NHANES) data collected between 1999 and 2020. The study, published in Immunity, Inflammation and Disease, examined six composite systemic inflammatory indices calculated from routine blood counts and found that five of them were statistically associated with hearing loss in middle-aged and older adults. The findings do not establish that inflammation causes hearing loss, but they add a nationally representative data point to a growing body of research linking chronic low-grade inflammation with sensory decline.

The research team, working within the NHANES framework, merged nine survey waves spanning 1999 to 2020, excluding the 2013 to 2014 cycle because audiometric data were not gathered that year. Their final analytic cohort included 7,920 adults aged 40 to 85 years who had both audiometric testing and data on key modifiable risk factors. Participants averaged 58.6 years of age, and 54.4 percent were women. Hearing loss, defined as a four-frequency pure-tone average of at least 20 decibels hearing level in the better-hearing ear, was present in 2,262 participants, or 28.6 percent of the sample.

Hearing was assessed with established NHANES audiometry protocols, conducted in sound-attenuating booths by trained examiners. Thresholds were measured in both ears at frequencies from 0.5 to 8 kilohertz using a modified Hughson Westlake procedure. The researchers calculated a four-frequency pure-tone average across 0.5, 1, 2, and 4 kilohertz for the better-hearing ear, along with separate high-frequency averages across 3, 4, 6, and 8 kilohertz and low-frequency averages across 0.5, 1, and 2 kilohertz. Higher scores indicated worse hearing.

The six inflammatory indices were derived from counts of neutrophils, monocytes, lymphocytes, and platelets, the standard components of a complete blood count. The systemic inflammation response index (SIRI) multiplies neutrophil and monocyte counts and divides by lymphocyte count. The systemic immune-inflammation index (SII) multiplies neutrophil and platelet counts and divides by lymphocyte count. The pan-immune-inflammation value (PIV) combines neutrophil, monocyte, and platelet counts divided by lymphocyte count. The neutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), and lymphocyte-monocyte ratio (LMR) are simpler pairwise ratios. Each index captures a different balance between immune cells that promote inflammation and immune cells associated with its regulation.

In the primary analysis, the researchers used modified Poisson regression with sandwich variance estimation, an approach chosen because the rare disease assumption underlying standard logistic regression is violated in this cross-sectional design. Models were built in three stages: the first adjusted for age and sex, the second added ethnicity, educational attainment, and annual household income, and the third further adjusted for hypertension, diabetes, body mass index, and non-high-density lipoprotein dyslipidemia. Missing values for income, hypertension, diabetes, body mass index, and HDL cholesterol, ranging from 0.2 to 11.6 percent, were handled through multiple imputation using classification and regression trees.

After full adjustment, each unit increase in log-transformed SIRI was associated with a 14 percent higher prevalence of hearing loss, with a prevalence ratio of 1.14 and a 95 percent confidence interval of 1.08 to 1.20. Similar positive associations appeared for SII, PIV, and NLR. In contrast, each unit increase in log-transformed LMR was associated with a 13 percent lower prevalence of hearing loss, with a prevalence ratio of 0.87 and a confidence interval of 0.81 to 0.94. No statistically significant association emerged for PLR. All significant associations survived false discovery rate correction using the Benjamini-Hochberg procedure. Because the exposure and outcome were measured at the same time, these figures describe prevalence differences between groups, not incidence over time.

Descriptive comparisons reinforced the pattern. Individuals with hearing loss were older, reported lower educational attainment and household income, and had higher prevalences of hypertension, diabetes, and non-HDL dyslipidemia than those with normal hearing. Violin plots of the log-transformed indices showed statistical differences between hearing groups for SIRI, SII, PIV, NLR, and LMR, but not for PLR. Moderate-to-high correlations among several indices were expected, since they share overlapping hematological components.

Sensitivity analyses broadly supported the main results. Similar cross-sectional associations for SIRI, PIV, and NLR were observed when hearing loss was defined at high- and low-frequency pure-tone averages. When the outcome was restricted to more severe hearing loss, defined as a four-frequency average above 40 decibels hearing level, only SIRI and LMR remained associated. Formal interaction testing revealed significant age-related modification for three indices: SIRI, PIV, and LMR showed statistically significant interactions between age group, comparing adults aged 40 to 64 with those 65 and older, and the inflammatory index, with interaction coefficients of -0.23, -0.17, and 0.32 respectively, all with p-values below 0.001. This indicates that the associations were attenuated in older adults. Sex interactions were not statistically significant, with interaction p-values ranging from 0.09 to 0.71, although the associations were generally stronger among men than women, a pattern the authors say warrants cautious interpretation.

Further robustness checks included repeating the analyses with untransformed index values, adding self-reported smoking status and alcohol consumption to the fully adjusted models, and conducting survey-weighted analyses that incorporated NHANES sampling weights, strata, and primary sampling units. Including all 9,499 participants regardless of hearing threshold severity, and treating the four-frequency pure-tone average as a continuous outcome in linear regression, produced results consistent in direction with the primary findings. The authors note that their use of modified Poisson regression provides unbiased prevalence ratio estimates for binary outcomes, and that the large sample, formal interaction testing, and multiple sensitivity analyses strengthen the work. They also point out that no previous study has simultaneously examined all six indices in a nationally representative population of middle-aged and older adults.

The study has clear limitations. The cross-sectional design cannot establish causality or the direction of the association between inflammatory indices and hearing loss. Participants from disadvantaged socioeconomic backgrounds were less represented, which the authors suggest may lead to underestimation of the associations, since this group may have higher prevalences of both inflammation and hearing loss. Residual confounding cannot be excluded, partly because NHANES cycles lack standardized data on ototoxic medication use, cardiovascular history, acute infections, hematological disorders, and malignancies, all of which could transiently or fundamentally alter the measured indices. The authors propose that future investigations adopt longitudinal designs, consider gene-environment interactions, and apply causal effect studies to clarify whether systemic chronic inflammation contributes to hearing loss or whether hearing loss and its consequences drive inflammatory changes. They also suggest that inflammatory markers, integrated with audiometric and clinical variables, could eventually serve as components of classification models for hearing loss risk stratification and prognosis across different hearing loss phenotypes, though such applications remain untested.

Subject of Research: Associations between systemic inflammatory blood indices and hearing loss in U.S. adults aged 40 to 85 years.

Article Title: Systemic Inflammatory Indices and Hearing Loss in U.S. Adults Aged 40–85 Years: A Serial Cross‐Sectional Study 1999–2020

Article References: Zheng, Y., Guo, Z., Zhao, L., Peng, X., Zhu, Y., Shan, G., Li, T., & Li, X. (2026). Systemic Inflammatory Indices and Hearing Loss in U.S. Adults Aged 40–85 Years: A Serial Cross‐Sectional Study 1999–2020. Immunity, Inflammation and Disease, 14(10), Article e70576. https://doi.org/10.1002/iid3.70576

Image Credits: AI Generated

DOI: 10.1002/iid3.70576

Keywords: hearing loss, systemic inflammation, NHANES, SIRI, SII, NLR, LMR, cross-sectional study, audiometry, inflammatory indices, middle-aged adults, older adults