1.1 Background/rationale

Danish hospitals treat 8.000 hip fracture patients every year.¹ The global incidence of hip fractures continues to rise with the expanding older population, making them a significant health issue worldwide.² Hip fractures cause severe pain and require adequate pain relief treatment.^2,3 However, patients who suffer hip fractures are typically older and more fragile, with the fractures occurring at a median age of 81. This reduces the capacity to tolerate analgesic drugs.^1,4 This demand for sufficient pain relief, along with a reduced tolerance for drugs, makes it challenging to provide analgesic treatment for older hip fracture patients. Furthermore, research indicates that commonly used opioids often lead to chronic usage and addiction.^5,6 Consequently, there is a high demand for better alternatives.

The analgesic properties of perioperative methadone have been the subject of several studies.^7–13 However, there is limited evidence about using methadone for older and fragile individuals. Methadone's long half-life makes it an attractive option for pain relief as it provides better daily coverage and requires fewer administrations.⁸ Furthermore, methadone is primarily excreted through the bile, making it a safe option for patients with impaired kidney function.^14,15 Doses in the literature range from 0.10 mg/kg to 0.20 mg/kg and concern younger healthy patients.^7–13 Thus, the maximum dose for the older and fragile remains to be discovered.

In a novel attempt to identify this maximum dose, the present study utilized the Bayesian Continual Reassessment Method (CRM) to increase the dose of methadone during close monitoring of toxicity. Dose-limiting toxicity (DLT) was defined as respiratory depression, as this is a clear indicator of opioid toxicity. This adaptive model offers the advantage of making real-time adjustments if we observe a DLT.

1.2 Aim

This study aims to determine the maximal tolerable dose (MTD) of perioperative methadone in older hip fracture patients and to assess the feasibility of this protocol for future clinical trials.

1.3 Objectives

This study evaluates methadone's tolerability in terms of toxicity, side effects, and PACU stay duration in older hip fracture patients. It also examines its impact on postoperative pain management and opioid consumption.

1.4 Hypothesis

It was hypothesized that an increased dose of methadone would entail a higher risk of respiratory depression, side effects, and prolonged stay in the PACU. Additionally, it was proposed that the maximal tolerable dose is either 0.10 mg/kg, 0.15 mg/kg, or 0.20 mg/kg.

2.1 Study design [4]

The study's design is an interventional cohort study conducted as an adaptive platform trial. It is a non-controlled, single-blinded phase IV trial using the Bayesian CRM. Please refer to the published statistical analysis plan to further elaborate on this method.¹⁶

2.2 Setting [5]

The study was conducted at a University Hospital, specifically in the emergency and orthopedic departments. Between January 10, 2023, and March 21, 2023, patients with hip fractures were consecutively included in the study from the emergency department (ED). An orthopedic physician managed the screening and enrolment process.

2.3 Participants [6]

Patients with hip fractures were consecutively included in the study based on the following criteria:

Inclusion criteria.1.Patients diagnosed with an acute hip fracture on x-rays in the ED (collum femoris fractures, pertrochanteric fractures, and subtrochanteric fractures, ICD-10-codes: DS720-722).2.Age ≥60 years.3.Patients must be able to ask for supplementary analgesics if needed and give informed consent.4.Patients must read and speak Danish.

Exclusion criteria.1.Polytrauma:•Defined as multiple fractures or multi-trauma patients.2.Previous allergic reactions or hypersensitivity towards methadone hydrochloride or sodium chloride.3.Health Conditions Preventing Treatment:•Chronic obstructive pulmonary disease (Gold classification C + D)•History of acute asthma attacks or atopic skin conditions•Cor pulmonale•Raised intracranial pressure or recent head injury•Pheochromocytoma•History of paralytic ileus•QT interval prolongation on ECG upon admission (≥500 ms)•Myasthenia gravis•Known liver disorders•Hypotension (<100 mmHg systolic blood pressure at admission)4.Concurrent administration:•With MAO inhibitors or within two weeks of suspending treatment with these medicinal products.•Of sedatives, e.g., Benzodiazepines or related drugs.5.Inclusion in other studies.6.Impaired cognitive function:•e.g., dementia.7.Current drug addiction: e.g., opioid addiction or intravenous addiction.

Note: Female patients were not tested for pregnancy as they were ≥60 years old.

2.3.1

2.3.1 Patient recruitment

According to hospital protocol, every patient presenting with a hip fracture to the emergency department is evaluated by an orthopedic physician within 4 hours of arrival. During this initial consultation, the physician provides the patient with both oral and written information regarding the study. After a consideration period of 2 hours, the orthopedic physician proceeds with the screening and enrollment procedures, ensuring that eligible patients are appropriately informed and enrolled in the study. Simultaneously, the physician plans the necessary surgical intervention for the hip fracture. The flowchart in Fig. 1 details each step of the recruitment process, including the number of patients lost at each stage, from initial consultation to final enrollment.

Close

Fig. 1 Flowchart of the enrolment process. All patients lost at each step are reported.

Export to PPT

3.1 Primary outcome

The main focus of the study was respiratory depression, defined as having a respiratory rate of less than ten breaths per minute and a peripheral oxygen saturation of less than 94 %, even with 4 L of oxygen per minute. As previously indicated, this was defined as the dose-limiting toxicity (DLT). Data was collected from observation charts completed by primary caregivers in the post-anesthesia care unit (PACU) and in the orthopedic ward. In the PACU, registered nurses (RNs) continuously monitored peripheral oxygen saturation and respiratory rate until patients were transferred to the ward. Respiratory depression was evaluated upon arrival at the orthopedic ward and at 6, 24, and 72 hours after surgery.

3.2 Secondary outcomes

The RNs at the orthopedic ward consistently completed observation charts, which were used for data collection. REDCap hosted at OPEN at OUH was utilized to manage the charts electronically. These observation charts were the data source for our secondary outcomes.

Length of stay at the PACU: The length of stay at the PACU was recorded in hours for each patient, sourced from observation charts completed by RNs. Discharge from the PACU adhered to national guidelines outlined by the Danish Society for Anesthesiology and Intensive Medicine (DASAIM). Patients were discharged based on meeting criteria related to peripheral oxygen saturation, arousal, and pain level or following assessment by an anesthesiologist.

Amount of antidote needed: The number of times an opioid-specific antidote (e.g., Naloxone) was required was recorded. The data source was the observation charts filled in by the RNs at the orthopedic ward.

Postoperative opioid consumption: The mean consumption of rescue morphine equivalents for each group was recorded within specific time frames: Upon arrival and within the first 6, 24, and 72 hours after surgery. Morphine equivalent doses were calculated for various opioid types. The data source was the medical charts, with RNs recording the amounts on observation charts.

Opioid-related side effects: PONV was recorded binomial as present or not at 6 and 24 hours after surgery. The data source was the patient's statement, entered on the observation chart.

Postoperative pain assessment: Patients were asked to assess their pain intensity at the hip using the VRS, a validated scale for hip fracture patients.¹⁷ The scale gives patients six choices; they must choose the one that best describes their pain. The choices are 0 (no pain), 1 (slight pain), 2 (moderate pain), 3 (severe pain), 4 (extreme pain), and 5 (unbearable/worst imaginable pain). They were asked to assess pain intensity upon arrival at the ward and 6, 24, and 72 hours after surgery. The data source was the patient's statement, entered on the observation chart.

3.3 Bias [9]

The assessment of pain is highly subjective, necessitating single-blinding in this trial. Patients are kept unaware of the methadone dose to prevent reporting bias. Another potential bias is that RNs on the ward might withhold rescue medication due to fear of inducing an opioid overdose. Thus, the RNs in the ward are also unaware of the exposure group. There is also a risk of reporting bias in the assessment of DLT, where occurrences might be overemphasized or underreported. However, in this trial, a certified RN anesthetist records DLT on a standard, non-trial-specific observation sheet. Additionally, the RN anesthetist is not involved in the study or patient treatment after they leave the PACU.

3.4 Study size [10]

The trial's sample size was calculated using the formula introduced by Kuen Cheung et al.¹⁸. Given the fragility of this population, the target toxicity level was defined as 0.10, with an accuracy of 0.6, and an odds ratio of 2. As a result, a sample size of 40 individuals would be adequate.

3.5 Intervention/exposure

Methadone hydrochloride was used as the investigational drug in this study. There were three exposure groups with dosages of 0.10 mg/kg, 0.15 mg/kg, and 0.20 mg/kg. Each patient's dosage was determined based on their total body weight, with weight measured in the emergency department when possible. If weight data were not available, self-reported weight was used.

The primary investigator allocated each patient to an exposure group, as detailed in the section Treatment assignment, and then calculated the individual dosage. An ED RN was responsible for preparing the investigational drug. The prepared syringe was placed beside the patient awaiting surgery at the infusion stand. The syringe did not specify the exposure group, ensuring only the preparing RN and the primary investigator knew the allocation. Ten minutes before incision, the investigational drug was intravenously administered with a single administration by the RN anesthetist, who remained unaware of the exposure group. Participants and RNs in the orthopedic ward were kept unaware of the exposure group, rendering this study single-blinded. A gradual upward titration of methadone was employed using the CRM, adjusting dosages according to the occurrence of DLT.¹⁶ Refer to Fig. 2 for the flowchart depicting the allocation.

Close

Fig. 2 Flowchart illustrating the cycles of treatment assignment. DLT = Dose Limiting Toxicity (i.e., respiratory depression), general anesthesia, Spinal = spinal anesthesia, recommended dose = the dose suggested by the algorithm when running our script in R studio.

Export to PPT

All patients received analgesic treatment according to the hospital guidelines. This included a preoperative peripheral nerve block (femoral nerve block), paracetamol (1g x4), a long-acting opioid (Contalgin 5mg x2), and a rescue opioid as needed. The investigational medicine was given in addition to standard care to confirm its safety when combined with other treatments. Epidurals are not routinely used in our region for this patient group as they can delay early mobilization and are not part of the standard treatment for hip fractures.

3.6 Statistical methods [11 + 12]

4.1 Participants [13]

From 10 to 01–2023 to 21-03-2023, 89 hip fracture patients were screened for inclusion at the hospital. 32 patients were confirmed eligible and included in the study. Nine patients received 0.10 mg/kg (5 patients in general anesthesia and 4 in spinal anesthesia), and 21 received 0.15 mg/kg (9 in general anesthesia and 12 in spinal anesthesia). One patient received 0.20 mg/kg by mistake, as the report of DLT, unfortunately, was delayed for the patient prior. Thus, the algorithm changed the dose to 0.15 mg/kg when the occurrence of a DLT was included (see Fig. 2). The trial continued using 0.15 mg/kg and never reached 0.20 mg/kg again. Another patient was withdrawn per the request of the attending anesthesiologist. The statistical analyses disregard the patient receiving 0.20 mg/kg and the patient withdrawn by the anesthesiologist (see Fig. 1).

Ten correct cycles of treatment assignment were executed throughout the study, ignoring the incorrect one. See Fig. 2 for a flowchart of the treatment assignment. Stopping rule no. 3 (see Statistical methods → Primary analysis) was triggered when 30 patients had been included, and the inclusion of further patients stopped immediately. The stopping rule was triggered because the 95 % credibility interval of the toxicity level for the 0.10 mg/kg group was smaller than 20 points, and more than ten patients had been included.

4.2 Descriptive data [14]

Descriptive statistical analyses were conducted on the patient cohort, confirming its similarity to critical demographics and clinical factors in the general population. The mean age was 82, and 37 % were male. Furthermore, as shown in Table 1, no significant differences were found in the population characteristics between the exposure groups.

4.3 Outcome data [15 + 16]

4.3.1

4.3.1 Primary outcome data

4.3.1.1

4.3.1.1 Dose-limiting toxicity (DLT)

In the study, three cases of DLT, i.e., respiratory depression, were observed. They were all in the 0.15 mg/kg exposure group, and all underwent general anesthesia. Participants who underwent spinal anesthesia in the 0.15 mg/kg group did not manifest any instances of DLT (see Fig. 2 for flowchart). The mean posterior estimate of toxicity for 0.10 mg/kg = 0.07 (± SD 0.04, CI 0.02; 0.17), 0.15 mg/kg = 0.13 (± SD 0.06, CI 0.04; 0.26), and the predicted estimate for 0.20 mg/kg = 0.19 (± SD 0.07, CI 0.08; 0.35). Thus, 0.10 mg/kg is the only dose with an estimated toxicity below our cut-off toxicity of 0.10 and a 95 % credibility interval below the clinical practice level of 0.24. Fig. 3 shows the distribution of participants and DLT. Fig. 4 visualizes the estimated posterior probability of DLT for the different doses. Every instance of DLT occurred in the PACU within a few hours of methadone administration. No cases of DLT occurred later than 3 hours after administration of methadone.

Close

Fig. 3 Number of participants and cases of DLT in each exposure group.

Export to PPT

Close

Fig. 4 The estimated probability of DLT for each dose is illustrated with 95 % credibility intervals. The blue dotted line indicates our target toxicity of 0.10. The black diamond is the estimated toxicity for the 0.10mg/kg dosage below 0.10. The credibility interval for 0.10 mg/kg dosage is also below 0.2. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Export to PPT

Close

Fig. 5 Kaplan-Meier curve visualizing the length of the stay at PACU for each dose.

Export to PPT

4.3.2

4.3.2 Secondary analyses

-Length of stay at the PACU: Patients with 0.15 mg/kg methadone had a 0.34-h (CI -1.48; 0.89) shorter stay at PACU, which was statistically insignificant (p = 0.62). The results are visualized in Fig. 5.

Amount of antidote needed: The limited observations precluded meaningful statistical analysis. Only one patient received an antidote, accounting for two out of the 120 possible observations.-Postoperative opioid consumption: Participants in the 0.15 mg/kg group had a higher mean rescue morphine equivalent consumption, but this was statistically insignificant IRR 72h = 1.19 (CI 0.69; 2.07, p = 0.38). The complete analysis is shown in Table 2 and visualizes the consumption over time in Fig. 6

Close

Fig. 6 Graph visualizing the mean consumption of rescue morphine equivalents at the different time points for the two groups. The y-axis represents milligrams of morphine equivalents, and the x-axis represents time in hours.

Export to PPT

. A third of the enrolled patients required no rescue medication throughout their hospital stay. The mean consumption of rescue morphine equivalent within the first 72 hours was 9.4 mg in the 0.10 mg/kg group and 12.6 mg in the 0.15 mg/kg group.

Opioid-related side effects: The group administered with 0.10 mg/kg showed PONV in 9.1 % of our observations, whereas the 0.15 mg/kg group exhibited PONV in 9.8 % of the observations. Consequently, the analysis did not detect any notable variance in the occurrence of PONV between the groups (p = 1.00), indicating no statistically significant difference.-Postoperative pain assessment: Although the postoperative pain scores of patients receiving 0.15 mg/kg were 0.08 points lower on average on the original Likert scale (CI -0.28 to 0.11), this difference was not statistically significant (p = 0.41). There were 105 successful responses, and 15 responses were missing. 0 (no pain) accounted for 59 of the responses, 1 (slight pain) accounted for 36, and 2 (moderate pain) accounted for only 10. There were no responses higher than 2 (moderate pain).

4.4 Other analyses [17]

5.1 Key results [18]

The data indicate that the MTD is 0.10 mg/kg. Administration of 0.15 mg/kg did not demonstrate significant advantages. Specifically, no statistically significant differences were observed in critical parameters, including duration of stay in the PACU, postoperative opioid consumption, postoperative pain levels, antidote requirement, or occurrence of PONV. The secondary analyses may be underpowered due to a small sample size (n = 30). However, the risk of DLT is higher with the 0.15 mg/kg dosage, where the estimated toxicity exceeds 0.10, and the 95 % credibility interval exceeds the clinical practice level of 0.24. Based on these findings, this study concludes that the MTD for this specific patient group is 0.10 mg/kg.

This MTD differs from similar studies of 0.15 mg/kg – 0.25 mg/kg.^22,23 However, the studies in the literature focused on younger, healthier patients undergoing elective surgery, whereas this trial investigated older, frail patients undergoing acute surgery. Thus, a lower MTD for these patients was anticipated. The findings highlight the importance of customizing dosage regimens to suit distinct patient populations. The methodologies presented in this paper can serve as a valuable tool for streamlining and enhancing the precision of future dose adjustments, offering a straightforward and dependable approach.