Background Opioid-induced constipation (OIC) is common, although there is a lack of consensus on diagnostic criteria, and as a result a lack of consistency around the prevalence (with resultant impact on patient care). This study primarily aims to compare different methods of assessing OIC in a heterogeneous group of patients with cancer.

Methods Data was collected on 100 consecutive patients with cancer that were receiving regular opioid analgesics. Constipation was assessed using a simple question ( ‘are you constipated’), the EAPC criteria, the so-called ‘Camilleri’ criteria (for OIC), the Rome IV criteria (for OIC), and assessment by a specialist palliative care professional. The degree of OIC was assessed using the bowel function index (BFI).

Results The median age was 62 (range 37–86); 46% were female, 54% were male. Cancer diagnoses were as follows: 34% gastrointestinal, 13% gynaecological, 12% head and neck, 11% urological, 11% breast, 10% lung, 5% haematological, 2% skin, 2% malignancy of unknown origin.

The prevalence of constipation using the different diagnostic criteria was:

• Simple question – 35% (with 17% ‘unsure’)

• EAPC criteria – 29% (with 6% ‘unsure’)

• Camilleri criteria (for OIC) – 63% (with 6% ‘unsure’)

• Rome IV criteria (for OIC) – 71%

• Assessment by specialist palliative care professional – 68% had constipation; of which 44% had OIC.

Of the patients identified as having OIC from the Rome IV criteria, 30% thought they were constipated, and the mean BFI was 65. In contrast, the mean BFI for patients that did not meet the Rome IV criteria was 32.

Conclusions The prevalence of OIC depends on the diagnostic criteria employed. It appears that many patients with OIC do not realise that they are in fact constipated (based on objective criteria).