Changes in routine blood test results over time did a better job of pointing to cancer than single abnormal results in some cases, according to a University of Oxford study of 275,205 adults who went to their family doctor in England with unexplained weight loss. The study was published September 17, 2026, in PLOS Medicine.
What that means for you: if you are losing weight without trying, it is reasonable to ask your doctor to look at how your routine blood counts have moved over past tests, not only whether today’s results are flagged as abnormal. It does not mean healthy people should start ordering repeat blood work to hunt for cancer. Whether you need more testing is a call for your doctor.
Key takeaways
- In people with unexplained weight loss, trends across repeat blood tests beat a single abnormal result for 34 test and cancer pairings.
- Simply adjusting results for a patient’s age and sex made both single results and trends more accurate.
- This was tested only in people who already had weight loss. It is a way to sort people who have a symptom, not a screening test for everyone.
What the study found
Researchers led by Brian Nicholson used anonymized English primary care records from 2000 to 2018. They pulled every result for 26 common blood tests taken before each patient’s visit for weight loss and linked the records to the national cancer registry.
Of the 275,205 patients, 13,798 (5.0%, or about 1 in 20) were diagnosed with cancer.
The team compared two ways of reading the same tests. One was the usual approach: is the latest result out of range, such as low albumin or high platelets? The other looked at the direction each test had moved over the 1, 3, 5, and 10 years before the weight loss visit.
After adjusting for age and sex, the trend approach beat the single-result approach on 34 occasions. The best trend reached an accuracy score (called AUC) of 0.82, very likely between 0.81 and 0.83. In plain terms, if you picked one person with cancer and one without at random, that model ranked the person with cancer as higher risk about 82% of the time.
Specific patterns lined up with specific cancers:
- Bowel cancer and lymphoma: trends in mean cell volume, the average size of red blood cells.
- Lung cancer: trends in white blood cell count and neutrophils, a type of white cell.
- Prostate cancer: trends in red blood cell count, hematocrit (the share of blood made up of red cells), a liver enzyme called AST, and the platelet-to-lymphocyte ratio.
For prostate cancer, this adds to other work on combining blood markers, such as a blood test plus MRI that catches more aggressive prostate cancers.
Dr. Kumar’s take
The useful idea is simple. A result can sit inside the normal range and still be moving the wrong way. The press release describes trends as a way of “identifying cancer-related changes in blood test results that do not appear abnormal.” A blood count that has slid steadily for three years and just barely stays in range is not the same as one that has held flat, even though the lab flags neither.
The part most likely to be oversold is who this applies to. Everyone in this study had already walked in with unexplained weight loss. These numbers do not carry over to healthy people, where cancer is far rarer. This is a triage tool for deciding who needs a closer look, not a screening test.
The authors’ own order of priorities matters too. The first and biggest fix was adjusting results for age and sex. Nicholson called that “relatively simple” and said it “could easily be performed at the laboratory.” That is a change for labs and health systems to make. Patients do not need to chase it. Trends added more on top, but only for some tests and some cancers.
There are real limits. This was a look back at existing records, not a trial. It could not account for bias in who got tested in the first place, and the trend analysis only covered people with at least two tests in each time window. Nobody has yet shown that acting on these trends finds cancer earlier or saves lives.
What it means for you
- If you have lost weight without trying, see your doctor. That symptom deserves attention on its own.
- It is fair to ask whether your past blood counts show a steady drift, even if each one was labeled normal.
- A single “normal” result is not proof that nothing is wrong when a symptom like weight loss is present.
- If you feel well and your weight is stable, this study gives no reason to add extra blood tests.