The OMAMA study is an open-label, randomized, multi-center study to examine whether magnesium hydroxide is non-inferior to macrogol/electrolytes in the prevention of OIC. This study will be conducted in eleven hospitals throughout The Netherlands. The trial is designed in accordance with the Declaration of Helsinki and the principles of Good Clinical Practice (GCP) set by the International Conference of Harmonization (ICH) [18, 19]. It has been approved by the Medical research Ethics Committee of Amsterdam UMC. The study is registered in the Dutch Trial Register, identifier 80508; EudraCT number 2022–000408-36. Patient recruitment and data collection have started in October 2022.

Both patients from the outpatient clinic and hospitalized patients from eleven academic and non-academic hospitals will be recruited. We will ask patients with metastatic or locally advanced incurable cancer to participate in this study at the moment that they will start to use slow release or transdermal opioids for pain. Additional inclusion criteria are that the patient is ≥ 18 years old and able to complete a Dutch questionnaire. Previous treatment with opioids is allowed, if discontinued for more than four weeks before the start of the study.

Exclusion criteria for participation include patients with contra-indications for laxatives; use of laxatives during the previous four weeks; severely impaired renal function, defined as a serum creatinine > 180 umol/l; and an estimated life expectancy < 3 months.

Patients will be asked to participate in the study by the physician who prescribes the opioid for pain and will give them a package containing the patient information leaflet (PIL) and the informed consent form. In case of a patient in the outpatient clinic, the patient will be given the time to read the PIL directly after the consult. The local investigator will then give additional information, answer all questions and ask for informed consent. In case of a hospitalized patient, the patient will receive the PIL and informed consent form on the day that opioid use will start. The local investigator will then meet the patient the next day to give additional information, answer all questions and ask for informed consent. When informed consent is given, the PIL is signed twice by the patient and the investigator. After signing informed consent patients are free to withdraw from the trial at any moment.

Patients will be randomized after fulfilling eligibility criteria and written informed consent. The randomization will be 1:1 using computerized block randomization with varying block sizes of 2 and 4 in random order. Randomization will be stratified by center and presence of constipation at the start of the study, defined as a score ≥ 30 on the Bowel Function Index (BFI, see paragraph ‘study measurements’). Randomization will be performed by the local researcher after baseline assessments.

Patients will be treated with either macrogol/electrolytes or magnesium hydroxide in a standard dosing regimen. Macrogol/electrolytes is started at a dose of 1 sachet (containing 13,125 g of macrogol) once daily and magnesium hydroxide at a dose of 724 mg t.i.d., both orally. Macrogol/electrolytes is regarded as the standard treatment (usual care) to which magnesium hydroxide will be compared. The dose of macrogol/electrolytes and magnesium hydroxide may be increased to 2 sachets daily or 1448 mg t.i.d., respectively, or may be decreased or discontinued because of intolerance or side effects during the study period. To determine whether a change of the dosage is needed, a contact moment with the local investigator by telephone will take place on day 7.

Demographic and baseline data will be extracted from the medical record. This includes sex, age, performance status, comorbidities, treatment with systemic and/or radiotherapy during the study period and the type and dose (converted to morphine equivalents) of the opioid which the patient will use.

The primary outcome measure is the proportion of patients with a score of < 30 on the Bowel Function Index (BFI), measured on day 14. The BFI is a clinician-administered, patient-reported questionnaire to assess clinically significant constipation, validated in patients receiving opioids for chronic non-malignant and malignant pain [20,21,22]. The BFI consists of three questions, assessing ease of defecation, feeling of incomplete bowel evacuation and personal judgement of the patient regarding constipation, each during the last 7 days and each rated on a scale of 0 (best possible outcome) to 100 (worst possible outcome). A total score ≥ 30 (mean of the three separate scores) indicates clinically significant constipation. A change of the total score of > 12 is regarded as clinically meaningful. The BFI has been used in a large ‘real world’ observational study in cancer patients receiving opioids [1] and in randomized trials of opioid antagonists for OIC [23,24,25,26]. It has been recommended as the assessment tool of choice for OIC [7, 27]. The BFI will be administered by the local investigator on day 0 (at the start of the study, after giving informed consent and before the randomization) and by telephone on day 14.

As a secondary outcome measure, the local investigator will assess the Rome IV criteria for opioid-induced constipation on day 0 and by telephone on day 14 (see Table 1). At least two of the six criteria have to be fulfilled in order to diagnose constipation.

Another secondary outcome measure is the 5-level EuroQol-5 Dimension questionnaire (EQ-5D-5L) completed on day 0 and 14. The EQ-5D-5L is a self-administered, generic assessment tool developed by the EuroQol Group to assess the quality of life and to determine cost-effectiveness. This measurement tool consists of five questions on dimensions of health (i.e., mobility, self-care, pain/discomfort, usual activities, and anxiety/depression) [28].

Other secondary outcome measures include a pain score (Numeric Rating Scale, ranging from 0 to 10, on days 0 and 14), a questionnaire about patient satisfaction with the laxative on day 14 (one question on a four-point Likert scale, ranging from unsatisfied to very satisfied), a questionnaire about side effects of the laxatives on day 14 and a questionnaire about medical consumption (the modified iMTA Medical Consumption Questionnaire (iMCQ)) on day 14 [29]. An overview of the measurement tools can be found in Table 1.

Lastly, patients will report daily laxative use in a medication diary to evaluate protocol adherence and also daily use of escape medication.

The sample size was determined to achieve 90% power to declare non-inferiority of magnesium hydroxide in case the proportion of patients with a score < 30 on the BFI at day 14 to be equal for magnesium hydroxide and macrogol/electrolytes assuming one-sided testing at a significance level of 2.5% [30]. The sample size was based on two non-randomized studies that showed that approximately 40% of patients starting with opioids and not using laxatives do not develop constipation after 14 days [31, 32]. Prophylaxis with laxatives can only be effective in (part of) the remaining 60% of patients. Treatment with macrogol/elektrolytes is effective in 70% of patients with symptomatic chronic constipation [33, 34]. Based on these data, we estimated the percentage of patients without constipation (BFI < 30) after 14 days of treatment with macrogol/electrolytes to be 82% (40% plus 0,7 × 60%). The main considerations are that the proportion of patients with a BFI < 30 under magnesium hydroxide at the non-inferiority margin should clearly be higher than the estimated 40% of patients without opioid induced constipation and the clinical judgement that more than 15% difference could no longer be translated in a conclusion that magnesium hydroxide is non inferior to macrogol/electrolytes. Based on the previous data, we calculated a sample size of 138 evaluable patients per arm. We will increase the inclusion with approximately 20% to anticipate potential loss to follow-up to N = 165 per arm.

All data will be coded and collected in a secure cloud-based clinical data management platform (Castor EDC). Analysis will be performed using the latest version of IBM SPSS Statistics.

Primary analysis for non-inferiority will be performed both on a protocol and intention-to-treat basis. The primary outcome is the difference in proportion of responders (BFI < 30 at day 14) between the two arms. The primary analysis will use the Newcombe score method to estimate a one-sided 97.5% confidence interval for the difference in proportion of responders (proportion responders under magnesium hydroxide minus proportion responder under macrogol/electrolytes). The null hypothesis (H0: Magnesium hydroxide is inferior to macrogol/electrolytes in the prevention of OIC) will be rejected and non-inferiority will be claimed for magnesium hydroxide if the lower boundary of the 97.5% confidence interval is larger than the prespecified non-inferiority margin of -15%. An additional analysis will be performed in which the risk difference for response is estimated using a generalized linear model for a binomial outcome and an identity link, while adjusting for the stratification factors of center and constipation at the start of the study.

An additional per protocol analysis will be performed using inverse-probability-of-treatment-weighing (IPTW) in the intention-to-treat population. For this IPTW analysis, logistic regression analysis in the intention-to-treat population will be used to predict the probability of a subject being adherent based on predefined confounders. In this analysis, non-adherence is defined as < 80% of the prescribed laxative being taken as determined from the patient diary [35]. The inverse of the predicted probabilities will then be used as weights in an analysis comparing treatment outcomes between the groups and to estimate a per-protocol effect. In this weighted analysis, the risk difference for response and its two sided 95% confidence interval is estimated using generalized linear models for a binomial outcome and an identity link, while adjusting for the stratification factors.

Secondary analyses include the same analyses for constipation judged by professional care givers, based on the Rome IV criteria for opioid-induced constipation. Furthermore, we aim to study baseline predictors of the response to macrogol/electrolytes and magnesium hydroxide in patients starting with opioids, including sex, age, performance status and comorbidity.

Lastly, the cost-effectiveness of macrogol/electrolyte compared to magnesium hydroxide will also be a secondary outcome. The cost-effectiveness will primarily be influenced by the effect of the medication on healthcare resource use and quality of life, since their unit costs are similar. In this analysis, we will compare the cost-effectiveness of macrogol/electrolyte and magnesium hydroxide in patients. The analysis will be performed from a healthcare perspective with a time horizon of two weeks. A decision tree will be developed with branches for initial therapy and responders and non-responders. The costs and quality of life of the patients in the trial will be used as input for this decision tree. This requires data collection on healthcare procedures and quality of life in responders, non-responders and patients with side-effects of laxatives for the follow-up period of the study (two weeks). These healthcare procedures will be extracted from the modified iMCQ after two weeks. The modified iMCQ will be adjusted in such a way that resource use due to gastrointestinal complaints will only be measured in these patients with intensive healthcare resource use. All individual health care procedures will be linked to their unit costs. Reimbursement prices issued by the Dutch Healthcare Authority (NZa) and national reference prices will be used for this assessment as outlined in current Dutch pharmaco-economic guidance. Quality of life will be measured using the EQ-5D-5L at baseline and after two weeks.

A scenario will be added to the analyses for the use of naloxegol and naldemedine as a first-line treatment in case of OIC. Naloxegol and naldemedine are peripheral µ-receptor antagonists that have shown to be effective for the treatment of OIC [36]. There are no studies available yet evaluating the efficacy of naloxegol for prevention of OIC. For naldemedine there is one recent randomized controlled trial available that has evaluated the efficacy of naldemedine for prevention of OIC [10]. As no effect data of first-line treatment with naloxegol or naldemedine will be measured during the trial, the scenario analysis will be an early health-economic analysis. In this analysis we will use effect data based on literature and expert opinion and will perform a threshold analysis to estimate the effect size needed to result in a cost-effective intervention. Bootstrapping will be performed to assess the uncertainty of costs and utility loss of (sub)groups, while deterministic (one-way) and probabilistic sensitivity analyses will be performed to assess the uncertainty of the analyses in the decision tree.

With regards to missing data, we do not plan imputation for the main outcome. If unexpectedly outcomes are missing, we will perform sensitivity analyses using single-value imputation under the most extremes scenarios. Multiple imputation may be considered if important confounders are missing or if missing on the main outcome is higher than currently anticipated.