Patients with bladder cancer who also had color vision deficiency experienced shorter overall survival and a 52% higher 20-year mortality risk compared with matched patients without the visual condition, according to a retrospective cohort study published in Nature Health. Survival among patients with colorectal cancer did not significantly differ by color vision status.

Background

Color vision deficiency (CVD) is among the most common inherited visual disorders, affecting approximately 8% of males and 0.5% of females. Most cases impair the ability to distinguish red hues—the color associated with blood.

Because blood in urine or stool is often the first sign of bladder or colorectal cancer (CRC), researchers hypothesized that patients with impaired color perception may fail to recognize these warning signs and delay seeking care.

Previous reports support this possibility. Case studies have described patients with CVD who misinterpreted rectal bleeding as diarrhea and delayed medical evaluation for weeks or months. Experimental research has also shown that individuals with the condition are less accurate at identifying blood in body-fluid samples than those with normal color vision.

Despite these observations, researchers noted that no prior study had examined whether CVD is associated with survival differences in cancers in which visible bleeding is a common early symptom.

Methods

Researchers used the TriNetX Health Research Network, a federated electronic health records platform containing de-identified data from more than 275 million patients across US and international health systems.

Records from January 2004 through March 2025 were analyzed. Bladder cancer cases were identified using ICD-10 and ICD-O code C67, while CRC cases were identified using codes C18, C19, and C20. CVD was identified using ICD-10 code H53.5.

Prior to propensity score matching, investigators identified:

• 149 patients with bladder cancer and CVD and 371,154 patients with bladder cancer alone

• 216 patients with CRC and CVD and 757,629 patients with CRC alone

Researchers then performed 1:1 propensity score matching to balance age, sex, race, ethnicity, and comorbidities including nicotine dependence, hypertension, diabetes, and hyperlipidemia.

Following matching and eligibility criteria, analyses included:

• 135 patients per cohort in the bladder cancer survival analysis

• 187 patients per cohort in the CRC survival analysis

Primary outcomes were overall survival probability and 20-year mortality risk.

Results

Bladder cancer

Following propensity matching, patients with bladder cancer and CVD demonstrated lower overall survival compared with matched patients without the condition.

Researchers subsequently analyzed long-term mortality. Patients with CVD had 1.52 times the mortality risk over 20 years, corresponding to a 52% higher mortality risk.

The mean age at diagnosis was 72 years in the CVD group and 71 years in the comparison group. Women comprised 7% of each matched cohort.

Colorectal cancer

Survival among patients with CRC did not significantly differ between those with and without CVD.

Because no survival difference was detected, researchers did not conduct a 20-year mortality analysis for CRC.

Interpretation

Researchers proposed several explanations for the divergent findings.

Bladder cancer frequently presents with painless visible blood in the urine and may have few other early symptoms, meaning detection often depends on the patient recognizing this visual sign.

CRC, by contrast, can produce a broader range of symptoms—including abdominal pain, changes in bowel habits, and weight loss—that may prompt evaluation even if bleeding goes unnoticed.

Screening may also play a role. The US Preventive Services Task Force recommends routine CRC screening beginning at age 45, which may allow cancers to be detected through colonoscopy regardless of whether patients recognize bleeding.

Researchers suggested that screening could interrupt the pathway linking CVD to delayed diagnosis, though they noted this explanation remains hypothetical.

They also acknowledged the possibility that a survival difference could exist in CRC but was not detected because the study lacked sufficient statistical power.

Limitations

The investigators highlighted several limitations associated with the use of aggregated electronic health record data.

Because the analysis relied on diagnostic coding, errors could introduce misclassification. In addition, CVD is frequently undiagnosed, meaning some affected patients may have been included in the comparison cohorts, potentially diluting the observed effect.

Cancer staging data were largely unavailable within the TriNetX system, preventing direct confirmation of the proposed mechanism that patients with CVD present with more advanced disease.

The retrospective design also precludes conclusions about causation.

Clinical Implications

Researchers characterized the findings as hypothesis-generating and did not recommend changes to clinical practice.

Future studies could test patients with bladder or CRC for CVD to improve sample sizes and determine whether certain subtypes carry higher risk. Researchers also suggested evaluating whether targeted screening strategies could improve outcomes among high-risk patients with the condition.

“These hypothesis-generating findings should increase clinicians’ suspicion of bladder cancer among patients with CVD and nonspecific signs of malignancy,” the researchers wrote.

Disclosures can be found in the study.

Source: Nature Health