Checkpoint inhibitors raised thyroid risk in bowel cancer

Immune checkpoint inhibitors have reshaped cancer treatment, but by design they unleash immune activity, and endocrine glands are a frequent casualty. In colorectal cancer specifically, several regimens are now in use — checkpoint inhibitors alone, combined with tyrosine kinase inhibitors, or combined with chemotherapy and anti-angiogenic antibodies — and it has not been clear whether they carry comparable endocrine risk. Direct head-to-head trials comparing these combinations are scarce. A network meta-analysis can borrow strength across trials to make indirect comparisons, which is the practical route to informing monitoring protocols for patients on these drugs.

The authors searched for randomised controlled trials of immune checkpoint inhibitor therapy in colorectal cancer published up to 22 November 2025, then conducted a network meta-analysis estimating risk ratios with 95% confidence intervals and formally assessing risk of bias. Six RCTs met the inclusion criteria. Outcomes covered hypothyroidism, hyperthyroidism, thyroiditis, diabetes mellitus, adrenal insufficiency, and adverse events graded 1–2 and 3–4. Regimens compared included pembrolizumab, checkpoint inhibitor plus tyrosine kinase inhibitor, and checkpoint inhibitor plus chemotherapy plus an anti-angiogenic antibody, all relative to conventional therapy. SUCRA rankings were used to order regimens by relative risk. The review was pre-registered on PROSPERO.

Relative to conventional therapy, checkpoint inhibitor-based regimens carried a higher thyroid-related toxicity burden overall. Pembrolizumab and the checkpoint inhibitor plus tyrosine kinase inhibitor combination significantly increased hypothyroidism risk, while hyperthyroidism was significantly higher with the tyrosine kinase inhibitor combination and with checkpoint inhibitor plus chemotherapy plus anti-angiogenic antibody. Grade 1–2 adverse events were consistently increased across all checkpoint inhibitor-based treatments. For thyroiditis, diabetes mellitus, adrenal insufficiency and grade 3–4 adverse events, the effect estimates were imprecise with wide confidence intervals, meaning no firm conclusion could be drawn — though SUCRA rankings tended to place the tyrosine kinase inhibitor combination towards the higher-risk end for thyroiditis and diabetes.

This extends what is already recognised about checkpoint inhibitor endocrinopathies into the specific setting of colorectal cancer and, importantly, differentiates between regimens rather than treating immunotherapy as one category. The thyroid findings are the firm part; everything concerning rarer or more severe endocrine events sits in a zone of genuine uncertainty because six trials simply do not generate enough events to estimate low-frequency outcomes precisely. SUCRA rankings should be read as hypothesis-generating orderings rather than proof that one combination is riskier than another. Larger pooled datasets or dedicated safety studies would be needed to resolve the diabetes and adrenal insufficiency questions.

If you or someone you know is being treated with immunotherapy for colorectal cancer, this supports why clinicians monitor thyroid function during treatment — the increased risk of both under- and overactive thyroid is the clearest signal in the data. The lower-grade adverse events being consistently increased is a reminder that these are common rather than exceptional. Crucially, the absence of a clear result for diabetes, adrenal insufficiency and severe toxicity reflects imprecise evidence, not demonstrated safety. This is a synthesis of trial data intended to inform monitoring practice, and treatment decisions belong with an oncology team.