TOPLINE:

Different medications prescribed in primary care showed an increase before the diagnosis of pancreatic cancer. Prescribing rates for insulin rose at 19 months before the diagnosis, with an earlier increase seen in women than in men.

METHODOLOGY:

• Researchers in England conducted a retrospective cohort study using linked primary care data from Clinical Practice Research Datalink Aurum and cancer registry data from the National Cancer Registration and Analysis Service and analysed patients diagnosed with pancreatic cancer between 2011 and 2018.
• They included 12,990 patients aged 18 years or older at diagnosis (median age at diagnosis > 72 years; 49.8% women; 92.7% White) to assess primary care prescription patterns before the diagnosis of pancreatic cancer.
• Prescription records for eight medication categories potentially treating pancreatic cancer features, such as back pain, abdominal pain, nausea, vomiting, and new-onset diabetes, were analysed over 5 years pre-diagnosis, excluding the final 2 months before diagnosis.
• The medication categories were antiemetics, antireflux medications, insulin, other hypoglycaemic agents, opioids, non-opioid analgesics, neuropathic analgesics, and non-steroidal anti-inflammatories.
• The frequency and rate of prescribing and first-time prescriptions were evaluated according to the medication category; inflection points when prescription rates increased above baseline were identified.

TAKEAWAY:

• During the 5-year prediagnostic period, approximately half of patients received one or more prescriptions for antireflux medications (52.6%), non-opioid analgesics (52.5%), opioid analgesics (47.9%), and non-steroidal anti-inflammatory drugs (44.6%); however, insulin was least frequently prescribed (6.3%).
• Prescription rates increased gradually for all medication categories except non-steroidal anti-inflammatory drugs; the largest increases were observed for antireflux drugs (from 24.7% at 5 years to 40.0% at 1 year pre-diagnosis), non-opioid analgesics (from 27.6% to 32.8%), and opioid analgesics (from 22.7% to 27.1%), with marked increases at 5-7 months before diagnosis.
• First-time insulin prescribing more than doubled from 0.8% to 1.9% from 2 years pre-diagnosis to the year prior to diagnosis. Insulin prescribing increased at 19 months pre-diagnosis (95% CI, 14.2-23.8), with an earlier rise seen in women vs men (25 vs 11 months).
• Neuropathic analgesics showed the earliest increase at 35 months before diagnosis (95% CI, 15-55). Other hypoglycaemic agents showed increased prescribing at 13 months pre-diagnosis (95% CI, 7.7-18.5), antireflux medication prescribing rose at 7 months pre-diagnosis (95% CI, 5.6-8.4), opioid analgesic prescribing rose at 7 months pre-diagnosis (95% CI, 4.4-9.6), and antiemetic prescribing rose at 5 months pre-diagnosis (95% CI, 2.9-7.1).

IN PRACTICE:

"[The study] findings suggest that prescribing for conditions such as diabetes, dyspepsia, nausea and vomiting, and pain could indicate underlying pancreatic cancer or related diseases. Despite some uncertainty around inflection points, these findings highlight possible extended diagnostic windows in which investigation of patients presenting with some clinical features, or being prescribed specific medications, could contribute to earlier cancer diagnosis," the authors wrote.

SOURCE:

This study was led by Cecilia H. Morel, MSci, MRes, Queen Mary University of London, London, England. It was published online on April 16, 2026, in the British Journal of General Practice.

LIMITATIONS:

The population-level analysis did not capture individual predictive values. The proportion of patients driving these changes could not be examined. Residual confounding from factors and comorbidities could not be excluded. Only primary care prescriptions were captured, and the study did not consider prescriptions from secondary care or the private healthcare or prescriptions for over-the-counter medications.

DISCLOSURES:

This study received funding from a Cancer Research UK programme grant and was supported by Barts Charity. One author disclosed receiving support from the Wellcome Trust PhD programme Health Data in Practice: Human-Centred Science. The authors declared having no competing interests.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.