Document resource
WHAT IS ALREADY KNOWN ON THIS TOPIC Biological and psychosocial factors, such as age, immune and endocrine function, psychological conditions, genetic background and primary tumour, have been associated with chronic cancer pain in men and women. However, clinical evidence on sex differences in opioid efficacy, treatment management and adverse drug reactions in cancer pain remains limited and inconclusive.WHAT THIS STUDY ADDS In this post-hoc analysis of a multicentre randomised trial, overall pain reduction with WHO step III opioids was similar in men and women. Nevertheless, dose escalation, opioid switching, co-administration of non-steroidal anti-inflammatory drug and adverse drug reactions differed by sex, with women experiencing higher rates of some adverse drug reactions and men more frequently requiring dose increases for pain control.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Recognition of sex-specific differences in opioid management may support more individualised analgesic strategies in cancer care. Prospective studies are warranted to further inform clinical decision-making and optimise opioid prescribing according to sex.Introduction Opioids are commonly used to treat cancer pain, but some issues related to their use remain unresolved. The impact of the primary tumour type, the presence of metastases, comorbidities and the age of the patients on analgesic efficacy has yet to be extensively assessed. However, concerning opioid drugs, the preliminary studies that led to the approval of their use by the Drug Agencies, dating back many years, were rarely designed to assess differences in response, efficacy and toxicity between men and women. Furthermore, their use is a ‘biased history’, as the pre-clinical and clinical studies required for approval were often performed on middle-aged male subjects only. 1 This has resulted in a lack of data regarding opioid responses in women and men, especially in advanced age. Recently, some biological mechanisms related to sex differences have been identified as potentially influencing the onset and development of pain. For instance, we learnt that the immune cells play a key role in pain transmission and the development of chronic pain,2–4 and that sex hormones play a central role in modulating pain.2 5 However, clinical studies published in this field that focus on pain characteristics and treatment results are few, and those that exist tend to concentrate on only specific aspects. For example, the risk of inducing nausea during extended-release oxycodone treatment is significantly higher in women.6 In another study,7 patients with moderate to severe cancer pain of different ages and sexes had significantly different plasma fentanyl concentrations and consequent pain control. Furthermore, the prevalence and severity of pain in patients aged >70 years were found to be significantly higher among women.8As far as we know, there are no studies targeted at the pain control in women and men affected by severe cancer pain. We then decided to explore this clinical area, which has involved our working group for many years. The objective of the present analysis has been to evaluate the differences between men and women requiring a treatment with a WHO step III opioid9 for cancer pain, in terms of baseline clinical characteristics, analgesic response to opioids, treatment changes and safety profile. We used data from our previously published longitudinal study that evaluated the analgesic response to four different opioids (oral morphine, oral oxycodone, transdermal fentanyl and transdermal buprenorphine) randomly assigned to cancer patients.10Methods Study design and patients This is a post-hoc analysis of the phase IV, randomised, open-label, longitudinal CERP trial (ClinicalTrials.gov ID NCT01809106), aimed to evaluate the efficacy of three WHO step III opioids (oral oxycodone, transdermal fentanyl and transdermal buprenorphine) compared with oral morphine in cancer patients with moderate to severe pain, who require a WHO step III opioid and who had never received one.10 Eligible patients had a confirmed locally advanced or metastatic tumour, persistent moderate to severe cancer pain, required strong opioids for the first time and were over 18 years of age. Exclusion criteria were cerebral tumours, leukaemia, concurrent radiotherapy, first-line chemotherapy within 7 days before randomisation, non-pharmacological analgesic treatment and pre-existing renal failure. A complete list of inclusion and exclusion criteria is provided in the online supplemental materials. The study protocol was reviewed and approved by the Independent Ethics Committee of the Fondazione IRCCS Istituto Nazionale dei Tumori, Milan (ID: INT 36/10; approval date: 17 September 2010). The trial was approved by the institutional review boards of each centre, and patients gave written informed consent before any study-related activities were carried out. No patients or members of the public were involved in the design, conduct, reporting or dissemination of this study.SP110.1136/bmjonc-2025-001024.supp1Supplementary dataProcedures The planned follow-up was 28 days with six visits on days 1 day, 3 days, 7 days, 14 days, 21 days and 28 days. At baseline, the oncological history (primary tumour site, presence of metastases, previous and current cancer treatments), comorbidities and Karnofsky performance status were assessed. Pain characteristics collected included the average pain intensity (API) and worst pain intensity (WPI) experienced over the previous 24 hours, as measured on the 0–10 Numeric Rating Scale.The initial opioid dose was based on the European Association for Palliative Care11 recommendations. During the follow-up, any adjustment to pain therapy was allowed to achieve optimal pain control, including dose change, addition of an adjuvant analgesic drug and switching to a different opioid. In the case of persistently unsatisfactory analgesia or severe toxicity, opioid treatment could be discontinued permanently.At each visit, pain characteristics, pain treatment adjustments and adverse drug reactions (ADRs) were recorded. ADRs were measured using a patient-administered questionnaire to record the presence and severity of symptoms, measured using a four-point verbal rating scale (none, mild, moderate, severe) derived from the Therapy Impact Questionnaire.12Outcomes The objective of this post-hoc analysis was to evaluate the difference between men and women for each of the following endpoints related to pain reduction, treatment changes and safety profile.The pain reduction endpoints were the proportion of responders (Rs) according to Corli’s criteria13 and according to Farrar’s criteria.14 Corli’s criteria defined Rs as patients reporting a difference in API ≥30% between the first and last visit and a final API <4, while all patients not defined as Rs were classified as non-Rs (NRs). Farrar’s criteria14 classified patients in three categories: Rs were defined as patients achieving a reduction in pain intensity ≥30, NRs as patients who had reported no change or an increase in pain and the remaining patients, who did not achieve a 30% reduction in pain, were defined as partial Rs.The endpoints related to the treatment changes were the longitudinal trajectories of opioid dose over time, the proportion of patients with an Opioid Escalation Index (OEI) >5%, where OEI represents the daily mean percentage increase in opioid dose, the proportion of patients requiring a switch to another opioid, and the proportion of patients requiring adjuvant medication to optimise the around-the-clock regimen.The safety profile was evaluated based on the proportion of patients experiencing at least once one of the following ADRs: drowsiness, confusion, nausea, vomiting, constipation, dry mouth, hallucinations, muscular pain, gastralgia, dysuria, dyspnoea and itch.Statistical analysis The analysis was performed on patients who provided informed consent, were randomised in the CERP trial, had the baseline and at least one post-baseline pain assessment. For each patient, the follow-up was defined from the first opioid administration until the end of the planned 28-day period, a switch or a premature discontinuation of the study for any reason.Descriptive analyses were used to present patient level data. The χ2 test (or Fisher’s exact test, where appropriate) and t-test (or the Mann-Whitney test, where appropriate) were used to compare the distribution of categorical and continuous variables between sexes, respectively. Given the exploratory purpose of the analysis, any potential issues associated with multiple testing, such as type I error inflation, could be considered acceptable.To control for the potential predictive effect of the opioid type, differences between men and women were assessed according to the type of opioid received.Linear mixed model for repeated measures will be used to analyse the dose trajectories over time, including the dose as dependent variable, sex, time and the interaction between sex and time as fixed effect and the patient as the random effect.To evaluate the impact of baseline patient characteristics on the response to analgesic treatment, univariable logistic regression models were initially used. A stepwise selection approach was then applied to select the covariates for inclusion in a multivariable logistic regression model. Treatment arm and sex were forced in the multivariable model because of their relevance for the analysis irrespective of statistical significance. Results are presented as OR and 95% CI (95% CI).In addition, to explore the effect of primary tumour location, the differences between men and women in terms of pain reduction were assessed in subgroups of patients based on primary tumour location.Statistical analyses were conducted using SAS V.9.4.Results Between May 2011 and July 2014, 520 patients were allocated to one of the study arms, applying a 1:1:1:1 randomisation, stratified by centre. Out of these, 22 were excluded because they did not receive any of the study treatments past the baseline visit; therefore, 498 were included in this analysis, 277 men (55.6%) and 221 women (44.4%). Table 1 shows the demographic and clinical characteristics of the patients. At the start of the WHO Step III opioid treatment, women were younger than men (mean age, 65.1 years, SD 12.7 vs 68.3 years, SD 10.8, p=0.0027), but the mean time from primary tumour diagnosis to randomisation was longer for women (40.3 months, SD 56.6) compared with men (23.0 months, SD 39.7, p<0.001). The most common primary tumour sites were lung, bronchus or pleura (126 patients, 45.5%), followed by colon/rectum (34 patients, 12.3%) in men and breast (64 patients, 29.0%), followed by lung, bronchus or pleura (57 patients, 25.8%) in women. Statistically significant differences were found between the sexes regarding the type of previous antitumour treatment. More women received surgery (123 women, 55.7% vs 107 men, 38.6%, p<0.001), chemotherapy (149 women, 67.4% vs 153 men, 55.2%, p=0.0057), biological therapy (37 women, 16.7% vs 28 men, 10.1%, p=0.029) and hormone therapy (48 women, 21.7% vs 28 men, 10.1%, p<0.001). More than half of the men were recorded as having a cardiovascular comorbidity (146 men, 52.7% vs 92 women, 41.6%, p=0.014). Significant sex differences were also observed in reports of worriedness (108 women, 48.9% vs 108 men, 39.0%, p=0.027) and depression (77 women, 34.8% vs 72 men, 26.0%, p=0.032).Table 1Demographic and baseline characteristics by sex at the start of the WHO step III opioid treatmentMenn=277Womenn=221P valueAge (years)0.0027 Mean (SD)68.3 (10.8)65.1 (12.7) Median (Q1–Q3)69.3 (62.6–75.9)66.2 (55.5–74.8) Min–max28.0–91.121.2–100.3Time from primary diagnosis to randomisation (months)<0.001 Mean (SD)23.0 (39.7)40.3 (56.6) Median (Q1–Q3)9.6 (2.6–22.7)19.2 (4.6–53.5) Min–max0.0–379.00.0–372.8Primary site: n (%)<0.001 Lung/bronchus/pleura126 (45.5)57 (25.8) Colon/rectum34 (12.3)23 (10.4) Breast1 (0.4)64 (29.0) Pancreas21 (7.6)18 (8.1) Urinary tract25 (9.0)6 (2.7) Stomach/oesophagus/duodenum10 (3.6)8 (3.6) Head/neck32 (11.6)10 (4.5) Sex specific tumour28 (10.1)35 (15.8)Metastatic disease: n (%)229 (82.7)195 (88.2)0.083Prior antineoplastic therapy: n (%)215 (77.6)177 (80.1)0.50Surgery: n (%)107 (38.6)123 (55.7)<0.001Chemotherapy: n (%)153 (55.2)149 (67.4)0.0057Biological therapy: n (%)28 (10.1)37 (16.7)0.029Hormone therapy: n (%)28 (10.1)48 (21.7)<0.001Radiotherapy: n (%)83 (30.0)82 (37.1)0.093Non-cancer comorbidities: n (%)190 (68.6)130 (58.8)0.024Dysmetabolic/hormonal diseases: n (%)60 (21.7)35 (15.8)0.10Cardiovascular diseases: n (%)146 (52.7)92 (41.6)0.014Respiratory diseases: n (%)25 (9.0)10 (4.5)0.051Neurological/psychological diseases: n (%)14 (5.1)11 (5.0)0.97Digestive system diseases: n (%)11 (4.0)12 (5.4)0.44Karnofsky Performance Status (%)0.53 Mean (SD)66.5 (17.1)67.5 (16.9) Median (Q1–Q3)70.0 (60.0–80.0)70.0 (60.0–80.0) Min–max20.0–100.030.0–100.0Self-reported psychological status High emotional tension (a lot/extremely): n (%)70 (25.3)71 (32.1)0.092 High worriedness (a lot/extremely): n (%)108 (39.0)108 (48.9)0.027 High irritation (a lot/extremely): n (%)60 (21.7)57 (25.8)0.28 High depression (a lot/extremely): n (%)72 (26.0)77 (34.8)0.032API at baseline0.62 Mean (SD)6.0 (1.4)6.0 (1.3) Median (Q1–Q3)6.0 (5.0–7.0)6.0 (5.0–7.0) Min–max4.0–10.04.0–10.0WPI at baseline0.80 Mean (SD)7.9 (1.5)8.0 (1.5) Median (Q1–Q3)8.0 (7.0–9.0)8.0 (7.0–9.0) Min–max4.0–10.04.0–10.0Pain duration (months)0.0063 Mean (SD)2.8 (3.3)3.9 (5.7) Median (Q1–Q3)2.0 (1.0–3.0)2.0 (1.0–4.0) Min–max0.0–36.00.0–38.0Type of pain: n (%)0.41 Only nociceptive234 (84.8)178 (80.5) Only neuropathic0 (0.0)1 (0.5) Both nociceptive and neuropathic40 (14.5)39 (17.6) Insufficient information to classify the pain2 (0.7)3 (1.4) Missing10For categorical variables, χ2 test (or Fisher test if appropriate) was performed, while for continuous variables, Student’s t-test was performed.Sex-specific tumours include prostate tumours for men and gynaecologic tumours for women.API, average pain intensity; Min–max, minimum–maximum values; N, number of subjects; Q1–Q3, first–third quartile; WPI, worst pain intensity.No statistical difference between sexes was found in terms of mean basal API (6.0, SD 1.4 in men vs 6.0, SD 1.3 in women, p=0.62) and WPI (7.9, SD 1.5 in men vs 8.0, SD 1.5 in women, p=0.80). The mean duration of pain before the three WHO step III opioid treatment was longer for women than for men (3.9 months, SD 5.7 vs 2.8 months, SD 3.3, p=0.0063).Out of 498 patients included in the present analysis, 122 patients, 125 patients, 127 patients and 124 patients were randomised to oral morphine, oral oxycodone, transdermal buprenorphine and transdermal fentanyl, respectively. All patients started the treatment with their assigned opioid, and 347 (69.7%) patients completed the entire planned follow-up, ranging from 73 (59.8%) patients in the oral morphine group to 94 (74.0%) patients in the transdermal buprenorphine group. Reasons for discontinuation are detailed in online supplemental table S1. There were no statistically significant differences between the sexes (Fisher’s test p values: oral morphine 0.45; oral oxycodone 0.72; transdermal buprenorphine 0.36; transdermal fentanyl 0.56).A detailed flowchart summarising at each visit the number of patients still receiving the planned treatment and the reasons for interruption for discontinued patients is provided as online supplemental figure S1.The efficacy of the four opioids in reducing pain did not appear to differ statistically between men and women. The proportion of API-Rs according to Corli’s criteria ranged from 70.8% (51 patients in the oral oxycodone group) to 74.6% (50 patients in the oral morphine group) in men and from 72.2% (39 patients in the transdermal fentanyl group) to 81.4% (48 patients in the transdermal buprenorphine group) in women. Further details are provided in online supplemental table S2.Changes in opioid dose are described in online supplemental table S3 and presented graphically in figure 1. Across all treatment groups, median opioid doses remained largely stable over time, often clustering at fixed dose levels, whereas mean doses progressively increased. This pattern suggests the coexistence of a large subgroup of patients maintaining stable dosing and a smaller subset requiring substantial dose escalation. The longitudinal models for dose change over time are presented in online supplemental table S4.Figure 1Equivalent doses (mg/day) of opioid treatments by visit. OMMED, oral morphine milligram equivalent doseIn patients treated with oral morphine, a significant overall effect of sex was observed, with women receiving higher mean opioid doses than men over the follow-up period (F-test p=0.039), while dose trajectories over time were similar between sexes. In patients treated with transdermal fentanyl, opioid dose escalation differed significantly by sex, with males showing a greater increase in dose over time compared with females (sex-by-time interaction F-test p<0.001; overall sex effect F-test p=0.0026). No statistically significant differences were found between males and females regarding opioid switching, as reported in online supplemental table S5. However, among patients treated with transdermal fentanyl, the mean time to the switch was statistically shorter in females (12.9 days, SD 4.6) compared with males (23.2 days, SD 7.2, p=0.0050). Among the patients switched from buprenorphine, switching occurred due to inadequate analgesia in 12 out of 14 (85.7%) men and due to an adverse reaction in five out of seven (71.4%) women (p=0.017). No statistically significant differences were found between men and women regarding the need for an adjuvant therapy with non-steroidal anti-inflammatory drugs (NSAIDs) at least once during the follow-up period.Figure 2 shows for each type of opioid the profile of the use of NSAIDs as adjuvant therapy at each visit for men and women among patients who completed the planned 28 day follow-up.Figure 2Use of NSAIDs as adjuvant therapy by sex according to opioid treatment, in patients that continued the treatment for six visits. NSAIDs, non-steroidal anti-inflammatory drugs.Online supplemental table S6 presents the results of the univariable and multivariable models for the analgesic response, as defined by Corli. Treatment arm sex, age, presence of metastases, feeling of tension and WPI at baseline were selected for inclusion in the multivariable model. In the multivariable analysis, the only covariate that showed a statistically significant effect was the feeling of tension. Patients who felt tense at baseline had a significantly lower chance of responding to the analgesic treatment (OR 0.64, 95% CI 0.41 to 1.00, p=0.0495).The evaluation of the response to opioids in men and women according to primary tumour location in the colon/rectum, pancreas and lung is shown in table 2. Although the interaction test between sex and tumour site did not reach statistical significance (p=0.38), analyses were performed separately for the selected tumour sites due to their clinical relevance. The lack of statistical differences between men and women was confirmed both in patients with pancreatic and lung tumours.Table 2Response to opioid treatment by sex according to primary tumour siteColon/rectumn=57Pancreasn=39Lungn=183Menn=34Womenn=23Menn=21Womenn=18Menn=126Womenn=57API at baseline Mean (SD)5.7 (1.4)5.7 (1.1)5.9 (1.3)5.7 (1.2)6.1 (1.4)6.4 (1.6) Median (Q1–Q3)5.5 (5.0–7.0)6.0 (5.0–6.0)6.0 (5.0–7.0)6.0 (5.0–7.0)6.0 (5.0–7.0)6.0 (5.0–7.0) Min–max4.0–8.04.0–8.04.0–8.04.0–8.04.0–10.04.0–10.0 Mann-Whitney test, p value0.850.620.44API at last visit Mean (SD)2.9 (2.4)1.7 (1.7)3.7 (2.1)3.2 (2.9)2.8 (2.0)2.8 (1.9) Median (Q1–Q3)2.0 (1.0–4.0)2.0 (0.0–2.0)3.0 (2.0–5.0)3.0 (1.0–4.0)3.0 (1.0–4.0)2.0 (1.0–4.0) Min–max0.0–10.00.0–7.00.0–8.00.0–8.00.0–7.00.0–7.0 Mann-Whitney test, p value0.0420.340.87API difference from baseline Mean (SD)2.9 (2.3)4.0 (2.1)2.2 (2.2)2.5 (2.5)3.3 (2.2)3.6 (2.6) Median (Q1–Q3)3.0 (2.0–4.0)4.0 (3.0–5.0)2.0 (1.0–3.0)3.0 (0.0–4.0)3.0 (2.0–5.0)3.0 (2.0–5.0) Min–max−2.0 to 7.0−2.0 to 7.0−1.0 to 8.0−2.0 to 7.0−1.0 to 10.0−2.0 to 10.0 Mann-Whitney test, p value0.0290.440.43API response, Farrar criteria: n (%) Non responder5 (14.7)2 (8.7)4 (19.0)5 (27.8)11 (8.7)6 (10.5) Partial responder4 (11.8)0 (0.0)5 (23.8)0 (0.0)20 (15.9)7 (12.3) Full responder25 (73.5)21 (91.3)12 (57.1)13 (72.2)95 (75.4)44 (77.2) Fisher’s exact test, p value0.210.110.78†API response, Corli criteria, n (%) Non-responder9 (26.5)2 (8.7)9 (42.9)5 (27.8)34 (27.0)14 (24.6) Responder25 (73.5)21 (91.3)12 (57.1)13 (72.2)92 (73.0)43 (75.4) χ2 test, p value0.17*0.330.73WPI at baseline Mean (SD)7.9 (1.6)7.6 (1.6)7.8 (1.6)7.5 (1.5)8.0 (1.4)8.1 (1.4) Median (Q1–Q3)8.0 (7.0–9.0)7.0 (6.0–9.0)8.0 (7.0–9.0)7.5 (7.0–8.0)8.0 (7.0–9.0)8.0 (7.0–9.0) Min–max5.0–10.04.0–10.05.0–10.05.0–10.04.0–10.05.0–10.0 Mann-Whitney test, p value0.460.630.89WPI at last visit Mean (SD)4.6 (2.8)3.0 (2.3)4.7 (2.2)4.6 (3.3)4.3 (2.7)4.3 (2.6) Median (Q1–Q3)4.0 (2.0–6.0)3.0 (2.0–4.0)5.0 (3.0–6.0)4.0 (3.0–7.0)4.0 (3.0–6.0)4.0 (3.0–6.0) Min–max0.0–10.00.0–8.01.0–8.00.0–10.00.0–10.00.0–10.0 Mann-Whitney test, p value0.0540.850.85WPI difference from baseline Mean (SD)3.2 (2.8)4.6 (2.7)3.1 (2.4)2.9 (3.0)3.7 (2.9)3.8 (3.2) Median (Q1–Q3)3.0 (2.0–5.0)4.0 (3.0–7.0)3.0 (2.0–5.0)3.0 (1.0–5.0)3.0 (2.0–6.0)3.0 (2.0–6.0) Min–max−2.0 to 8.0−1.0 to 10.0−1.0 to 8.0−2.0 to 8.0−3.0 to 10.0−3.0 to 10.0 Mann-Whitney test, p value0.0640.800.99WPI response, Farrar criteria: n (%) Non-responder6 (17.6)2 (8.7)2 (9.5)4 (22.2)17 (13.5)8 (14.0) Partial responder7 (20.6)0 (0.0)7 (33.3)4 (22.2)25 (19.8)8 (14.0) Full responder21 (61.8)21 (91.3)12 (57.1)10 (55.6)84 (66.7)41 (71.9) Fisher’s exact test, p value0.0220.550.64†*Fisher’s exact test†χ2 testAPI, average pain intensity; Min–max, minimum–maximum values; N, number of subjects; Q1–Q3, first–third quartile; WPI, worst pain intensity.At baseline, men and women with colorectal tumours had similar API levels. However, at the final evaluation, a greater reduction in API levels was observed in women compared with men (median 4.0, first–third quartile 3.0–5.0 vs median 3.0, first–third quartile 2.0–4.0, p=0.029). No statistically significant difference was found in terms of analgesic response when considering the API. However, when considering the WPI, a higher proportion of women achieved a full response according to Farrar’s criteria compared with men (21 R women, 91.3% vs 21 R men, 61.8%, p=0.022). No significant difference in response was observed using Corli’s definition for the API (21 R women, 91.3% vs 25 R men, 73.5%, p=0.17).There was a statistically significant difference in the safety profile between men and women treated with transdermal buprenorphine only, as shown in online supplemental table S7 and figure 3. Of these patients, 53 (77.9%) men and 55 (93.2%) women experienced at least one ADR (p=0.016).Figure 3Adverse events by sex according to random arm.Discussion The main purpose of this analysis was to determine whether third-step opioid treatment in patients with severe cancer pain produces different analgesic effects in men and women. A total of 498 cancer patients, including 277 men and 221 women, were initially allocated to four treatment groups according to the randomisation of the original study, which compared responses to four different opioids. 10 For the purposes of the present post hoc analysis, each treatment group was further stratified by sex to allow separate evaluation of the outcomes between women and men.Several sex-related differences were observed at baseline. Women were younger, had a longer interval from primary tumour diagnosis to study entry (40.3 months vs 23.0 months) and experienced pain for a longer period before starting strong opioids (3.8 months vs 2.8 months). Previous antitumour treatments were more frequent in women, whereas comorbidities were more frequent in men. Women reported more frequently feelings of worry and depression.Baseline pain intensity was similarly severe in both sexes (mean API 6.0 in both groups; mean WPI 7.9 in men vs 8.0 in women). At the last evaluation, overall pain reduction was similar in women and men, and R rates by Corli’s and Farrar’s criteria showed no consistent sex-related differences.However, the analgesic effect of opioids cannot be adequately assessed only in terms of the difference between initial and final pain intensity. The number and the nature of therapeutic adjustments over time also influence analgesic response. In this context, important differences between women and men emerged, which are discussed below.Opioid dose escalation No significant sex differences were observed for oxycodone and buprenorphine. In patients treated with transdermal fentanyl, men experienced a greater dose increase than women, despite similar initial doses. Conversely, women treated with oral morphine required a larger dose escalation over time. These findings align with preclinical studies showing increased tolerance and hyperalgesia in rats receiving fentanyl. 15 16 Notably, one study17 observed tolerance and hyperalgesia only in male rats. In a clinical trial,18 daily doses between 100 μg and 400 μg of fentanyl were administered to cancer patients as buccal tablets. At the end of the study, tolerance was observed in 17% of patients receiving a dose of 400 compared with 3% for the lowest doses.Opioid switching Overall, 79 switches were recorded: 40 due to inadequate analgesia, 19 due to adverse events and 20 for both. Women using transdermal fentanyl switched earlier than men (12.9 days vs 23.2 days; p=0.0050). In buprenorphine users, switches were mainly due to inadequate analgesia in men (85.7%) and ADRs in women (71.4%; p=0.017).NSAIDs co-therapy Men on oxycodone received NSAIDs more frequently, while in women, the use of NSAIDs increased with morphine (from 2.9% to 14.7%), but decreased with oxycodone (from 17.9% to 2.6%). The highest NSAIDs supplementation was observed in women treated with fentanyl, suggesting a greater need to reinforce analgesia in this subgroup.A common underlying reason for these therapeutic adjustments is the development of opioid tolerance. In the present study, tolerance may be suggested by the progressive dose increases. This phenomenon may, at least in part, be explained by interactions between opioids and the gut microbiota, leading to microbial translocation and gut-brain interaction.19 20Sex-related factors played a role in this case. Hormonal differences have a key role.9 21 Recent studies emphasise the importance of male gonadal hormones, which can influence opioid receptor sensitivity.22 23 Testosterone and other androgens increase μ-opioid and δ-opioid receptor density in males, in both animals24 and humans.25 This enhances the effectiveness of opioids, but in long-term treatment, μ-opioid agonists reduce androgen activity in males, leading to clinically important hypogonadism.26 27 In summary, opioids’ action in men is double and opposite, since initially increases and finally reduces the analgesic efficacy.As the response to opioid treatment by sex, stratified by primary tumour site, no statistically significant differences were observed between men and women in the lung and pancreatic cancer subgroups. In contrast, a sex-related difference was observed in the colorectal cancer subgroup. Women had lower API at the final evaluation than men, despite similar baseline pain and a smaller dose increase. These findings should be interpreted with caution given the limited number of patients within each subgroup.Psychological factors further contributed to sex-related differences. Women reported higher baseline worriedness (p=0.027) and depression (p=0.032), which can negatively influence pain perception.28 Furthermore, both pain itself and opioid use can worsen depression and other moods.29 30 Multivariable logistic regression confirmed that baseline emotional tension was associated with reduced analgesic response (OR=0.64). Other emotional factors showed no significant impact.Regarding safety, women in the transdermal buprenorphine group experienced more than men (93.2% vs 77.9%; p=0.016), while no sex differences were observed for other opioids.A major strength of this work is that it provides a response to our primary question. Men and women with severe cancer pain achieved comparable pain reduction with WHO step III opioid treatment, but through different therapeutic adaptations.A limitation of the study is the lack of sex-stratified randomisation in the original trial, resulting in a slight numerical imbalance between men and women, although this did not affect the statistical analyses. In addition, this was a post-hoc analysis without predefined hypotheses or a formal sample size calculation for sex-based comparisons. Finally, baseline psychological status was assessed via self-report, which may be subject to under-reporting or over-reporting. Despite these limitations, the findings suggest meaningful sex-related differences in longitudinal opioid management.Conclusion In this post-hoc analysis, the overall reduction in pain achieved with strong opioids was comparable between women and men. However, relevant differences emerged in therapeutic strategies and management aimed at achieving optimal pain control. While analgesic outcomes converged, the paths to effective pain relief differed by sex.These results provide insight for future prospective studies to optimise the opioid selection and personalised pain management strategies in cancer pain care.