In a retrospective cohort study of adults with rheumatoid arthritis, researchers found that the prognostic relevance of the extracellular water-to-total body water ratio depends on patient age rather than functioning as a uniform risk marker across all adult age groups. Elevated ratios were linked to higher risks of adverse long-term outcomes specifically among older individuals, while showing no independent prognostic discrimination in younger patients.

Aging is characterized by a progressive decline in physiological reserve and altered body fluid distribution, yet clinicians often observe broader deterioration in overall patient condition that is not fully captured by conventional disease-specific metrics. Although rheumatoid arthritis represents a systemic chronic inflammatory condition where long-standing inflammation contributes to biological decline, contemporary mortality risks also reflect aging-related factors such as frailty and sarcopenia.

The retrospective cohort investigation was conducted at the rheumatology outpatient department of a tertiary care university hospital in Niigata, Japan, including 186 adult outpatients diagnosed with rheumatoid arthritis by treating rheumatologists according to standard classification criteria. Baseline bioelectrical impedance analysis was performed between April and June 2017 within a prospectively designed clinical study, and electronic medical records were retrospectively reviewed through December 2025. Among the 186 participants, 141 individuals (75.8 percent) were female, the median cohort age was 68 years with an interquartile range of 59 to 75 years, and patients were followed for a median duration of 95.2 months with an interquartile range of 52.7 to 102.0 months.

Whole-body composition and fluid distribution were assessed at baseline using a multi-frequency bioelectrical impedance device, specifically the InBody S20, with patients positioned in the supine position in accordance with manufacturer protocols. The extracellular water-to-total body water ratio was analyzed both as a continuous variable and categorized into quartiles, where the cohort exhibited a median baseline ratio of 0.399 with an interquartile range of 0.389 to 0.406. The highest quartile, defined by values exceeding 0.406, was classified as elevated extracellular water-to-total body water ratio to facilitate clinical interpretability and maintain model stability.

The primary study endpoint was defined as a composite of all-cause mortality and transition to do-not-attempt-resuscitation or best supportive care status during the follow-up window. To ensure consistency and avoid misclassification, transition to best supportive care or do-not-attempt-resuscitation status was recorded only when medical records explicitly documented formal orders or care transitions in the context of advanced clinical deterioration. The authors treated this transition as a distinct clinical milestone reflecting advanced decline and shifting care goals rather than a direct surrogate for death, noting that treating clinicians did not have access to baseline bioimpedance measurements during decision-making. Over the follow-up period, 17 participants experienced composite outcome events, comprising 11 deaths and 6 care transitions with no overlapping cases.

Statistical analysis utilizing Cox proportional hazards models revealed that advancing chronological age was the strongest independent predictor of the composite outcome, carrying a hazard ratio of 2.53 per 10-year increase (95 percent confidence interval, 1.47 to 4.34; p < 0.001). In contrast, models adjusting for age without interaction terms showed that the fluid ratio alone had no independent association with outcomes, yielding a hazard ratio per 0.01 increase of 1.10 (95 percent confidence interval, 0.88 to 1.37; p = 0.40). Crucially, incorporating an interaction term between age and the fluid ratio revealed a statistically significant interaction (hazard ratio 1.22, 95 percent confidence interval, 1.05 to 1.42; p = 0.010), demonstrating that the prognostic association of the fluid ratio varied significantly as age advanced.

Stratified Kaplan-Meier analyses based on the median cohort age of 68 years further clarified this divergent pattern between age brackets. Among younger individuals under 68 years of age, the extracellular water ratio failed to discriminate event-free survival (log-rank p = 0.53), with zero events occurring among 9 patients in the highest quartile and 3 events among 76 patients in lower quartiles. Conversely, among older patients aged 68 years and above, an elevated fluid ratio was significantly associated with reduced event-free survival (log-rank p = 0.0052), with events occurring in 10 of 36 patients with high ratios compared to 4 of 65 patients with lower ratios. Age-stratified Cox regression using Firth correction demonstrated that high ratio values in older participants conferred a hazard ratio of 3.10 (95 percent confidence interval, 1.21 to 8.63; p = 0.018).

To investigate clinical trajectory differences, researchers performed competing-risk cumulative incidence function analyses evaluating first hospitalizations as either planned elective admissions or non-planned urgent admissions. Among older participants aged 68 and older, cumulative incidence patterns differed markedly by fluid ratio status: high ratio values were significantly associated with an increased incidence of non-planned emergency admissions (Gray’s test p = 0.014), whereas planned admissions showed no statistically significant difference between groups (Gray’s test p = 0.78). In contrast, younger individuals exhibited numerically higher rates across both admission categories without distinct divergence by admission type.

The consistency of the age-by-fluid ratio interaction was supported across multiple sensitivity models adjusting individually for clinical disease activity index, serum albumin, appendicular skeletal muscle mass index, body mass index, estimated glomerular filtration rate, prednisolone dosage, disease duration, history of heart failure, and functional disability scores. Additionally, a death-only sensitivity model using Firth-penalized Cox regression confirmed that the interaction remained statistically significant when restricted strictly to all-cause mortality (hazard ratio 1.18, 95 percent confidence interval, 1.04 to 1.34; p = 0.039). Alternative age thresholds of 65, 70, and 75 years also demonstrated consistent directional trends despite smaller subgroup event numbers.

The authors emphasized several key study limitations, including the single-center retrospective design, single-point baseline bioimpedance measurement without longitudinal tracking, and a small absolute number of primary outcome events. Direct assessments of frailty, detailed comorbidity indices, socioeconomic factors, and longitudinal therapy exposure were unavailable, leaving open the possibility of residual confounding. Consequently, the researchers noted that the findings represent prognostic associations rather than evidence of biological causality, emphasizing that the fluid ratio is an exploratory marker reflecting overlapping dimensions of physiological vulnerability that requires validation in independent cohorts before clinical application.

The biological rationale for evaluating body fluid distribution centers on how aging and persistent inflammation alter tissue structure. In bioelectrical impedance analysis, the extracellular water-to-total body water ratio captures both relative extracellular fluid expansion and intracellular water reduction, physiological shifts frequently identified in advanced aging and chronic disease states. While frailty has conventionally been evaluated through clinical scoring indices or quantitative muscle loss, it also encompasses qualitative disturbances in tissue composition and fluid distribution that reflect diminished physiological reserve. Elevated extracellular fluid ratios have previously been documented across several distinct clinical settings, including hemodialysis cohorts, chronic heart failure, cardiovascular surgery, and cancer complicated by sarcopenia. Furthermore, related body composition parameters like phase angle have been linked to adverse outcomes in older populations, while antirheumatic therapies such as biological disease-modifying drugs can influence nutritional status and muscle parameters. Rather than functioning as an independent biological entity, extracellular fluid expansion appears to reflect overlapping dimensions of vulnerability that become increasingly consequential as physiological reserve declines.

To evaluate these physiological associations reliably, the researchers established rigorous analytical and ethical frameworks. The baseline bioimpedance assessments originated from an earlier prospective clinical study approved by the Niigata University Ethics Review Committee, while the secondary retrospective analysis received separate formal ethical approval with a waiver of informed consent due to the use of de-identified records. Functional impairment was systematically captured through the modified Health Assessment Questionnaire, a simplified instrument validated in Japanese populations and widely applied in rheumatoid arthritis cohorts. For statistical modeling, continuous age and fluid ratio variables were rescaled into clinically meaningful units and mean-centered before being entered into interaction equations, preventing interpretive distortion of regression coefficients. Missing covariate information across secondary models was addressed exclusively through complete-case analysis without statistical imputation. Additionally, outcome classification protocols maintained strict differentiation between active therapeutic trajectories and care limitation decisions, explicitly excluding ambiguous end-of-life discussions and requiring unequivocal medical record documentation of formal orders before designating a case as best supportive care.

The investigation also provided structured insight into patterns of clinical deterioration through competing-risk evaluations of hospital admissions. Researchers restricted their competing-risk framework strictly to each patient’s first post-baseline hospitalization, categorizing events as either planned admissions for scheduled interventions or non-planned emergency hospitalizations. This first-event structure maintained a mutually exclusive framework and avoided the additional modeling assumptions required when evaluating recurrent admissions while isolating acute medical deterioration. Complementing cumulative incidence functions, exploratory Fine-Gray models confirmed a consistent directional link between elevated fluid ratios and emergency hospital admissions among older individuals. Because the study utilized a cohort-derived upper-quartile threshold of 0.406 rather than a validated universal standard, the authors cautioned that this cutoff cannot serve as an established clinical threshold. Clinicians frequently observe broad clinical deterioration that eludes standard disease activity scores, and these exploratory findings indicate that non-invasive fluid distribution markers may eventually help clinicians identify older patients facing unstable health trajectories, provided independent multicenter cohorts validate their prognostic utility.

Subject of Research: Prognostic significance of extracellular water-to-total body water ratio across age groups in rheumatoid arthritis

Article Title: Age modifies the prognostic relevance of extracellular water-to-total body water ratio in rheumatoid arthritis: a retrospective cohort study

Article References: Kobayashi, D., Wada, Y., Sudo, M., Hasegawa, E., Sakai, S., Wakamatsu, A., Nozawa, Y., Sato, H., Kuroda, T., & Yamamoto, S. (2026). Age modifies the prognostic relevance of extracellular water-to-total body water ratio in rheumatoid arthritis: a retrospective cohort study. BMC Rheumatology, 10(1), Article 68. https://doi.org/10.1186/s41927-026-00668-2

Image Credits: AI Generated

DOI: 10.1186/s41927-026-00668-2

Keywords: rheumatoid arthritis, bioelectrical impedance analysis, extracellular water, total body water, aging, cohort study, prognostic markers, geriatric rheumatology, modifies, prognostic, relevance, extracellular