Could Serum CA125 Aid the Clinical Assessment of Volume Status of End-Stage Kidney Disease Patients Treated by Peritoneal Dialysis Patients?

We reviewed the results from PD patients attending for their routine outpatient peritoneal membrane testing after starting PD, who had serum CA125 measured by an immuno-assay sandwich assay (Roche Cobas, Roche diagnostics, Sussex, UK), and cardiac assessment by echocardiography. Clinical assessments were made in all cases. All patients used standard single bag glucose-based PD dialysates or 7.5% icodextrin (Baxter Health Care, Deerfield, IL, USA) for overnight exchanges with continuous ambulatory peritoneal dialysis (CAPD) or for patients using nightly intermittent peritoneal dialysis with a dry day, or an overnight PD cycler with a day-time exchange (CCPD). Peritoneal solute transfer rates (PSTRs) were calculated from a 2.0 L 4-h 22.7 g/L dextrose exchange, and body composition was measured by multifrequency bioimpedance, after peritoneal dialysate had been drained (InBody 720, Biospace, South Korea), using a standard protocol, to measure ECW [ 15 ]. To compare ECW between patients, we used the ratio of ECW to ICW and to total body water (TBW), and that predicted from ICW [ 16 ]. Dialysis adequacy was calculated using standard equations [ 17 ]. All laboratory analyses, including NT-proBNP were conducted in a UK accredited laboratory (Medical Laboratory Accreditation – ISO 15189). Cardiac function was assessed using transthoracic two-dimensional echocardiography (EPIQ, Philips, Eindhoven, Netherlands). Patient demographics, medications and Stoke-Davies comorbidity were obtained from the Hospital patient records [ 18 ], along with frailty assessment using the clinical frailty score (CFS) [ 19 ]. This retrospective audit was reviewed by and complied with UK National Health Service (NHS) Research Authority Ethics Service (NRES) guidelines for clinical audit and service development and did not require formal ethical review. In keeping with NRES protocols, all data were appropriately anonymised prior to analysis. In keeping with the Helsinki accord, and following NRES protocols, all data were appropriately anonymised prior to analysis. Data are presented as integer, percentage, mean ± standard deviation, median (interquartile range), or percentage. Standard statistical tests, ANOVA and Kruskal-Wallis were used with appropriate post hoc corrections made for multiple testing, along with Chi-square adjusted for small numbers. Spearman and Pearson univariate analysis were used, along with receiver operator curves (ROCs). Variables associated with serum CA125 on univariate and ROC analysis were entered into multivariable models. In summary, we included in the multivariate model: age, sex, serum NTproBNP, and CRP, ECW/ICW, TBW, ECW/TBW ratios, lower serum sodium and increased ECW/ICW, along with, faster PSTR, greater comorbidity, and reduced left ventricular ejection fraction. Multicollinearity was formally assessed using the variance inflation factor. The analysis did not indicate any significant multicollinearity issues among the included predictors. To avoid redundancy and over representation of closely related measures, and potential collider bias, then highly correlated indices (e.g., ECW/TBW, ECW/ICW) were not simultaneously included in the model. Instead, a set of representative variables was selected to ensure parsimony and interpretability. As serum CA125 was not distributed, log transformed step-backward and ordinal regression models were analysed. Statistical analysis used Prism 11.0 (Graph Pad, San Diego, USA) and SPSS 29 (IBM, Armonk, NY, USA). Statistical significance was taken as p < 0.05.

Results from 571 adult patients who had electively started PD attending for outpatient peritoneal membrane testing, between October 2013 and December 2025, were reviewed, mean age 53.8 ± 10.5 years 62.4% male, 44.0% diabetic, median duration of PD treatment 2 (2–3) months ( Table 1 ). Two hundred and 16 patients were excluded; 197 with missing echocardiography data, 8 acute starters, 5 who had suffered an episode of peritonitis post-catheter insertion, and 3 patients each who had no bioimpedance or admitted to hospital between catheter insertion and first assessment of peritoneal membrane assessment. On clinical examination, no patient had signs of left ventricular failure. Serum CA125 was higher in women vs. men (25 (16–38) vs. 17 (11–29) IU/L, p < 0.05. However, there was no difference with ethnicity White 23 (13–38), Black 16 (11–29), Asian 22 (15–34) IU/L). Patient demographics, peritoneal dialysis prescriptions Patients divided into terciles according to serum CA125 levels. Data presented as integer, percentage, mean ± SD, or median (interquartile range). PD, peritoneal dialysis; CAPD, continuous ambulatory peritoneal dialysis; NIPD, nocturnal intermittent peritoneal dialysis with a dry day; CCPD, night-time cycler with a day-time exchange; CrCl, 24-h urinary creatinine clearance; UF, ultrafiltration; Na, sodium; PSTR, peritoneal solute transfer rate; HbA1c, glycated haemoglobin; NT-proBNP, N-terminal brain natriuretic. * p < 0.05, **0.01, *** <0.001 vs. lowest tercile. Patients were divided into terciles according to serum CA125 levels. Patients in the lowest tercile group were more likely to be male, younger, heavier weight, less frail and with lower comorbidity ( Table 1 ). In terms of PD prescription, fewer patients in the lowest tercile group were treated by CAPD, and used less icodextrin, but there were no differences in dialysis adequacy or residual kidney function. Patients in the highest tercile group had faster PSTR, lower serum sodium, albumin and creatinine, but higher NT-proBNP values ( Fig. 1 ). Patients divided according to serum CA125 tercile and log N-terminal brain natriuretic peptide. Data presented as media, interquartile and 10–90% confidence limits. * p < 0.05, *** <0.001 vs. lowest tercile. Bioimpedance measurements demonstrated that patients in the highest tercile had greater ECW, whether expressed as a ratio to ICW, ECW, or predicted ( Table 2 ). Whereas more patients were prescribed antihypertensive medications in the lowest tercile group, in particular more angiotensin converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, and alpha blockers. There was a trend for increasing left ventricular mass and reducing tricuspid annular plane systolic excursion between the lowest to the highest CA125 tercile, but this was not statistically significant. However, the left ventricular ejection fraction was lower with increasing CA125 group. Cardiovascular assessment of patients, including blood pressure measurements, antihypertensive medication prescription, bioimpedance and echocardiography Patients divided into terciles according to serum CA125 levels. Data presented as integer, percentage, mean ± SD, or median (interquartile range). SBP, systolic blood presume; DBP, diastolic blood pressure; ECW/ICW, extracellular water, intracellular; TBW, total body water; ht, height; ACEI¸ angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCB, calcium channel blocker; BB¸ β-blocker; BP, blood pressure; meds, medications; BP med classes, number of classes of antihypertensive medications; inhibitor; LVEDD, left ventricular end diastolic dimension; LVSD, left ventricular end systolic dimension; PWd, posterior wall diameter; LV, left ventricular; RV, right ventricular; TAPSE, tricuspid annular plane systolic excursion; R, right; AV, aortic valve; LV, left ventricular; LVH, left ventricular hypotrophy. * p < 0.05 **0.01, *** <0.001 vs. lowest tercile. On univariate analysis serum CA125 concentrations were associated with a number of factors, including those associated with volume, volume and blood pressure management, peritoneal solute transport, residual urinary volume, patient size – female sex, weight, body surface area, serum creatinine, patient comorbidity, frailty, and inflammation as assessed by CRP ( Table 3 ). Unlike many other biomarkers, there was no association with residual kidney function ( r = −0.05 p = 0.21). ROC analysis demonstrated that the area under the curve (AUC) for the highest ECW/ICW quartile and serum CA125 the AUC was 0.66 (95% confidence interval [CI] 0.61–0.71), for %ECW excess AUC 0.64 (95% CI: 0.59–0.69), and for those patients with a cardiac ejection fraction <40%, AUC 0.67 (95% CI: 0.57–0.74), all p < 0.001. Variables associated with increasing serum CA125 on univariate or ROC analysis were then entered into multivariable models, avoiding collider bias, with models checked for collinearity. Variables associated with serum CA125 Univariate Spearman analysis (rho). On multivariable analysis log serum CA125 was independently associated with markers of volume overload; lower serum sodium and increased ECW/ICW, along with female sex, faster PSTR, greater comorbidity, and reduced left ventricular ejection fraction ( Table 4 ). On ordinal regression, then increasing serum CA125 was associated with lower serum sodium, female sex, faster PSTR, reduced cardiac ejection fraction and greater comorbidity, but not ECW/ICW ratios ( Table 5 ). Multivariable step-backward regression model of variables associated with log serum CA125 Sex male vs. female sex, PSTR, comorbidity score, serum sodium, echocardiogram left ventricular ejection fraction, extracellular to intracellular body water (ECW/ICW). StE Beta, standard error of beta; Stand Beta, standardised beta; t, t statistic; 95% CI, 95% confidence limits of beta; Col, collinearity; VIF, variable inflation factor; PSTR, peritoneal solute transfer rate. Model r 2 0.233, adjusted r 2 0.224. Multivariable ordinal regression model of variables independently associated with serum CA125 terciles Female sex, peritoneal solute transfer rate (PSTR), serum sodium, comorbidity, echocardiogram left ventricular ejection fraction (low ejection fraction). Standard error of beta (StE Beta), 95% confidence limits of beta (95% CI). Model Nagelkerke adjusted r 2 0.231.

Volume control in ESKD patients is a major goal of any dialysis treatment, to prevent episodes of decompensated heart failure, left ventricular hypertrophy and cardiovascular mortality [ 20 ]. For PD patients, ultrafiltration failure to control volume is a major cause of technique failure with transfer to haemodialysis and increased mortality [ 21 ]. As such, biomarkers are required to aid clinical examination in helping to assess PD patients and allowing appropriate changes in dialysis prescription and medications. Current biomarkers include natriuretic peptides, lung ultrasound and bioimpedance. A comparative study reported that although increased lung comets were associated with pulmonary oedema, lung ultrasound was not as reliable in assessing volume status compared to bioimpedance [ 22 ], and measurements can be affected by intra-peritoneal volumes [ 12 ]. However, both bioimpedance and natriuretic peptides can also have confounders, particularly inflammatory states, changes in body composition with bioimpedance and residual kidney function with natriuretic peptides [ 11 , 13 ]. Serum CA125 has been reported to be a clinically useful biomarker aiding the management of patients with both acute decompensated and chronic stable heart failure, as well as patients with preserved left ventricular ejection fraction [ 23 – 25 ]. Pulmonary congestion and the inflammatory response associated with heart failure leads to increased pleural fluid accumulation which then stimulates pleural mesothelial cells to increase the production of CA125. In this study, we reviewed patients newly starting PD to include those with varying amounts of residual kidney function and found that serum CA125 was associated with increased ECW and other markers of ECW. Importantly, unlike NT-proBNP which is affected by residual kidney function, there was no independent association between serum CA125 and 24-h urinary creatinine clearance, or combined urinary and peritoneal urea clearance (Kt/V) in multivariable models. Although many variables had an association with serum CA125 on univariate analysis, these could be grouped in to 4 categories. Firstly, those associated with ECW excess including ECW/ICW, ECW/TBW, percentage ECW excess, lower serum sodium and albumin [ 26 ], and management of ECW excess with greater prescription of icodextrin which has been shown to reduce ECW and increase sodium removal compared to hypertonic glucose dialysates, and greater use of CAPD and CCPD which again increase sodium removal compared to nightly intermittent peritoneal dialysis [ 27 , 28 ]. Previous studies investigating hypoalbuminaemia in PD patients demonstrated that serum albumin increased after increasing ultrafiltration and reducing ECW [ 29 ]. Serum CA125 was also associated with faster PSTR, a recognised risk factor for ECW excess, and PD technique failure. Secondly, CA125 was associated with worse cardiac function, affecting both the left and right side of the heart, as evidenced by a reduced ejection fraction and lower tricuspid annular plane systolic excursion, with may have accounted for the association with lower diastolic and mean arterial blood pressure, and lower requirement for antihypertensive medications. Thirdly, there was an association with age, comorbidity, frailty and inflammation as assessed by CRP. Finally, there was an effect of sex, with greater values in female patients, which accounts for the association with lower weight, body surface area and serum creatinine. Serum CA125 is recognised to fluctuate during the menstrual cycle or pregnancy [ 3 ]. Although most of our female patients were post-menopausal, we did not scan all patients to exclude ovarian cysts, or other possible causes of an increased production including endometriosis. CA125 levels are also reported to be increased in cases of cirrhosis, but less than 1% of patients had cirrhosis. As this was a cross-sectional observational study, we can only report associations and not causality. However, on multivariable analysis as serum CA125 was independently associated with variables including volume overload; lower serum sodium, reduced cardiac function in terms of left ventricular ejection fraction along with increased ECW/ICW and faster PSTR, along with greater comorbidity and female sex, it has the potential to be a biomarker of volume status in PD patients. Importantly, compared to natriuretic peptides and bioimpedance, there was no association with residual kidney function and inflammation. We report on a cross-sectional observational study of ESKD patients shortly after starting PD, and as such cannot comment on whether longitudinal changes in serum CA125 may equally reflect changes in volume status. However, data from cardiology studies would suggest that serial measurements of serum CA125 could potentially aid the management of PD patients in terms of volume control [ 8 , 23 , 24 ]. Whether serum CA125 may have the same prognostic value for PD patients as reported for those with heart failure remains to be determined in future longitudinal studies [ 25 ], As there is a difference between male and female patients, then further studies would be required to determine whether separate target ranges are required for both sexes, so to guide clinicians as to optimum volume management. Until then, serial measurements in individual patients could potentially be used as an aid to clinical assessments.

Cancer antigen 125 (CA125) is a transmembrane glycoprotein produced by human mesothelial cells, and it was initially used as an ovarian cancer marker [ 1 ]. However, serum CA125 concentrations also increase in other medical conditions including gastrointestinal cancer [ 2 ], endometriosis [ 3 ], non-Hodgkin lymphomas [ 4 ], cirrhosis [ 5 ], and pelvic inflammation [ 5 ]. More recently, it has been recognised that CA125 can be produced by pleural mesothelial cells, and is increased in patients with heart failure, leading to reports as to whether it can be used as a clinical biomarker of volume overload and direct patient treatment [ 5 – 9 ]. End-stage kidney disease (ESKD) patients are at increased risk of cardiovascular disease and mortality [ 10 ]. Haemodialysis patients typically gain fluid weight between dialysis sessions, and fluid overload due to ultrafiltration failure, is a common cause of technique failure for peritoneal dialysis (PD) patients. The classic clinical examination signs of volume overload may be masked in dialysis patients due to differences in water and sodium gain, and as such other markers of volume overload are required to support clinical examination. Bioimpedance devices are increasingly being used in clinical practice, but changes in body composition, particularly loss of cell mass, and intracellular water (ICW) may lead to errors in reporting the absolute amount of excess extracellular water (ECW) [ 11 ]. Natriuretic peptides, particularly N-terminal probrain natriuretic peptides (NT-proBNP) are frequently used to aid clinical management in patients presenting with heart failure [ 12 ]. However, in addition to rapid changes with acute myocardial ischaemia, and arrhythmias, natriuretic peptide concentrations are also increased in inflammatory states and in patients with chronic kidney disease due to reduced clearance [ 7 , 13 ]. As such, reports of serum CA125 increasing in parallel with a decline in New York Heart Association (NYHA) classifications of heart failure, and stronger correlation with inferior vena cava diameter, have led to suggestions that serum CA125 might be better marker for right-sided heart dysfunction and venous congestion than NT-proBNP [ 7 ]. As volume overload is a major cause of transfer of PD patients to haemodialysis, we reviewed the association between serum CA125 and other assessments of volume overload in a contemporary cohort of PD patients [ 14 ].

Prof. Andrew Davenport was a member of the journal’s Editorial Board at the time of submission. The remaining author has no conflicts of interest to declare.

No funding was received for this article.

This retrospective audit was reviewed by and complied with UK National Health Service (NHS) Research Authority Ethics Service (NRES) guidelines for clinical audit and service development and did not require formal ethical review or individual patient consent. In keeping with NRES protocols, all data were appropriately anonymised prior to analysis.

Data were collected as part of routine clinical service. B Nikitiuk proposed review of the data, and first analysed the data set. A Davenport drafted the 1st version, both authors critically revised the manuscript, and read and approved the final manuscript.