Cancer Survivors Show Widening Gap Between Two Common Kidney-Function Tests, Chinese Study Finds

For millions of people who have survived cancer, a routine blood test quietly determines whether they can start a particular chemotherapy, whether a drug dose must be reduced, and how closely their kidneys need to be watched. That test almost always rests on an estimate rather than a direct measurement: the estimated glomerular filtration rate, or eGFR, a calculated proxy for how well the kidneys filter waste from the blood. A new cross-sectional study from three health-examination centers in China suggests that for adults with a history of cancer, the two most common ways of calculating that estimate can diverge far more often than they do in people without cancer, raising fresh questions about how kidney function should be assessed in this growing population.

The research, published in BMC Cancer by Lei Wang of Chongqing University Three Gorges Hospital, Hao Sun of the Second Affiliated Hospital of Shandong First Medical University and Taizhou People’s Hospital, and colleagues, examined a question that sits at the intersection of oncology and nephrology, a field sometimes called onco-nephrology. Clinicians can estimate filtration rate from serum creatinine, a muscle-derived waste product, or from cystatin C, a small protein produced steadily by nearly all nucleated cells. Because creatinine levels are influenced by muscle mass, diet, and other non-kidney factors, while cystatin C is thought to be less dependent on body composition, the two markers can yield different answers. When the cystatin C-based estimate falls substantially below the creatinine-based one, a situation researchers call large negative discordance, it may flag hidden kidney impairment that the creatinine number alone would miss.

Whether a history of cancer makes such discordance more likely had remained unclear. Cancer and its treatments can reshape body composition, deplete muscle, and alter inflammation, all of which can perturb creatinine independently of true filtration capacity. At the same time, some malignancies and therapies may affect cystatin C production or clearance. The Chinese team set out to quantify the association in a real-world screening population rather than in hospitalized patients, drawing on adults who attended single-visit health examinations with linked questionnaire data at centers in Chongqing, Shandong, and Jiangsu between January and July 2026.

The analytic cohort was deliberately restricted to people whose kidneys looked healthy by the usual screening standard: participants needed both a creatinine and a cystatin C measurement, an ascertainable cancer history, complete covariate data, plausible laboratory values, and a creatinine-based eGFR of at least 60 milliliters per minute per 1.73 square meters, the conventional threshold below which chronic kidney disease is generally staged. The investigators recalculated eGFR from the original measurements using creatinine alone, cystatin C alone, and both markers together. The primary outcome, large negative discordance, was defined as a cystatin C-based estimate at least 30 percent lower than the creatinine-based estimate. Prevalence ratios were estimated with modified Poisson regression using robust variance, and the team layered on an unusually thorough set of sensitivity analyses: center-specific and pooled models, alternative discordance definitions, continuous differences between the estimates, migration across GFR categories, subgroup analyses, overlap weighting, and an age-spline analysis.

Among 3,491 participants, 257, or 7.36 percent, reported a physician-diagnosed history of cancer. The headline finding was stark at the crude level: large negative discordance occurred in 14.79 percent of participants with a cancer history compared with 4.79 percent of those without, a difference the authors report as statistically significant with a P value below 0.001. In other words, roughly one in seven cancer survivors whose creatinine-based estimate suggested preserved kidney function had a cystatin C-based estimate at least 30 percent lower, a pattern seen in fewer than one in twenty of their cancer-free counterparts.

Adjustment for other characteristics softened but did not erase the signal. In the fully adjusted pooled model, cancer history remained associated with a higher prevalence of large negative discordance, with a prevalence ratio of 1.53 and a 95 percent confidence interval of 1.08 to 2.17, corresponding to a P value of 0.018. The association was also visible on a continuous scale: cancer history was linked to an adjusted relative eGFR difference that was 2.73 percentage points lower, with a confidence interval spanning -4.46 to -1.00 and a P value of 0.002. When the team relaxed the discordance threshold to 20 percent or used an absolute difference of -15 milliliters per minute per 1.73 square meters, similar patterns emerged, lending the result a degree of robustness across how the phenomenon is defined.

The geographic texture of the data, however, complicates any simple interpretation. When Chongqing and Shandong were combined, the adjusted prevalence ratio was 1.22 with a confidence interval of 0.78 to 1.89 and a P value of 0.383, meaning the association was not statistically distinguishable from no association at those two centers. In Jiangsu, by contrast, the prevalence ratio climbed to 2.40 with a confidence interval of 1.37 to 4.18 and a P value of 0.002, a substantially stronger effect. The formal test for a center-by-cancer-history interaction was not significant, with a P value of 0.212, but the authors are careful to warn that this nonsignificant interaction should not be read as proof that the effect is equivalent across centers. Absence of a statistically significant interaction is not the same as presence of equivalence, particularly in a cohort of this size.

Not every analytic approach supported the association. Crude downward migration across GFR categories, in which participants appeared to shift into a worse kidney-function category when cystatin C was used instead of creatinine, was substantially attenuated after multivariable adjustment, and the overlap-weighted and adjusted category-migration analyses were described as less precise or null. The authors are explicit about the limits of their design: this is a cross-sectional study, so it cannot establish that cancer or its treatments cause the discordance, nor can it determine which of the two estimates is closer to a directly measured glomerular filtration rate. The study also cannot establish that participants with discordant estimates actually had impaired kidney function; it shows only that the two markers disagreed more often in people with a cancer history.

Why might cancer survivors show this pattern? The study itself does not identify a mechanism, but the biology offers plausible candidates. Creatinine is generated by muscle, and cancer survivors frequently have reduced muscle mass due to the disease, its treatments, or lifestyle changes, which would lower creatinine and inflate the creatinine-based eGFR, making it look as though the kidneys are filtering better than they are. Cystatin C, meanwhile, is less tied to muscle but can be influenced by inflammation, thyroid function, obesity, and possibly malignancy itself. If cancer history shifts the balance of these non-GFR determinants, the two formulas would part ways more often. Disentangling which marker, if either, is telling the truth in this population would require studies comparing both estimates against measured GFR using tracer or clearance methods, which this study did not perform.

The practical takeaway, according to the authors, is one of caution rather than alarm. Among adults with a creatinine-based eGFR of 60 or above attending routine health examinations at three Chinese centers, cancer history was associated with greater negative discordance between the cystatin C- and creatinine-based estimates in the pooled adjusted analysis, with estimates varying across centers. When a kidney-function estimate is clinically consequential, for example when it will guide drug dosing, contrast imaging decisions, or nephrology referral in an adult with a history of cancer, the findings suggest that relying on a single filtration marker should be done cautiously. Measuring both creatinine and cystatin C, or confirming an uncertain estimate with a direct measurement of filtration rate, may be prudent in exactly the situations where the number will change management. As cancer survival continues to improve worldwide and the population of long-term survivors grows, the humble eGFR calculation, and the assumptions baked into it, may deserve closer scrutiny than a single number on a lab report usually receives.

Subject of Research: Association between cancer history and discordance between creatinine- and cystatin C-based estimated glomerular filtration rate in health-examination adults

Article Title: Cancer history and creatinine-cystatin C eGFR discordance across three Chinese health-examination centers

Article References: Wang, L., Sun, H., Yao, X., Xiang, H., Tan, H., Wei, D., & Gong, X. (2026). Cancer history and creatinine-cystatin C eGFR discordance across three Chinese health-examination centers. BMC Cancer. https://doi.org/10.1186/s12885-026-16998-w

Image Credits: AI Generated

DOI: 10.1186/s12885-026-16998-w

Keywords: cancer history, cystatin C, creatinine, eGFR discordance, estimated glomerular filtration rate, kidney function, onco-nephrology, cross-sectional study, chronic kidney disease, health examination, China, BMC Cancer