Intranasal Dexmedetomidine for Pain Management in Older Patients: A Cross-Over, Randomized, Double-Blinded, Active-Controlled Trial

The purpose of this study was to determine whether IN DEX could decrease pain and anxiety better than subcutaneous (SC) opioids in elderly patients during daily nursing care.

The INDEX trial was designed as a randomized, superiority, cross-over, double-blind, active-controlled trial. It was stopped early because of recruitment failure and a shortage of staff during the COVID pandemic. A descriptive analysis is provided below.

The trial was conducted from November 2018 to July 2020 at Geneva University Hospitals (Geneva, Switzerland). Written informed consent was obtained from each participant, in accordance with the Declaration of Helsinki [29].

Inclusion criteria for patients were admission to the Geneva University Hospitals’ palliative care or geriatric rehabilitation divisions, age ≥ 65 years, requirement of opioids for pain relief on a regular basis, and undergoing, at least once a day, a nursing procedure (lasting from 15 to 30 min) that caused sufficient discomfort (either pain, anxiety, or both) as to require the administration of extra opioid doses. Patients were not considered if they were at risk of bronchial aspiration, respiratory or cardiac complications, or if they had a contraindication for nasal drug administration (for details, see Table 1).

The population targeted in this study were elderly patients, who may be considered as a “vulnerable” population. To avoid risks of coercion and anticipate severe cognitive disorders at end-of-life that would make a participant incapable of judgment, information was provided as soon as possible to patients and their relatives. Before each of the two sessions of the study, patients were asked, by a research nurse, if they were still willing to participate, and the inclusion/exclusion criteria were rechecked.

Each participant underwent two nursing care sessions in a cross-over design. Both nursing care sessions took place in the morning and were spaced out over a 24- or 48-h time period. The choice of this delay was based on the elimination half-life of IN DEX, which is 2–3 h [5].

All participants received their regular ongoing opioid treatment (i.e., morphine or hydromorphone). Oral intake was allowed before nursing care. All patients received both study drugs: either IN DEX and SC placebo [NaCl 0.9%], or IN placebo [NaCl 0.9%] and SC opioid. Study drugs were administered 45 min before the start of the nursing care sessions, so that a 15- to 30-min nursing care would be covered by the peak effect of both SC opioids and IN DEX.

Subcutaneous midazolam 0.05 mg/kg (of actual body weight) was available, if required, as a rescue treatment if the total Elderly Caring Pain Assessment (ECPA) score reached 5 points, and the nursing care session was not about to end. Oxygen supplementation was provided if oxygen saturation fell below 85%. Two nurses or assistant nurses were responsible for nursing care interventions, while a research nurse or the first author oversaw assessments of outcomes.

The timing of measurement of the primary, secondary, and safety outcomes is described in Table 2.

The ECPA score, assessed by a research nurse or physician, enables the measurement of both pain and anxiety related to nursing care. The ECPA is a validated tool to assess pain and anxiety and comprises eight items divided in two parts, based on the behavior of non-communicating elderly patients [30]. One to five minutes are needed to complete each part of this evaluation. The first part of the scale (score: 0–16) is used for pain assessment before nursing care, while the second part (score: 0–16) assesses pain during nursing care. The total score ranges from 0 to 32. Our primary outcome was the number of patients with an ECPA total score > 5.

We considered that IN DEX would make a clinically meaningful improvement if it could decrease the proportion of patients who would experience pain defined as an ECPA score > 5 by 20%. We hypothesized that the proportion of patients who would experience pain when receiving SC opioids would be 50%, and that this proportion would decrease to 30% with IN DEX. We also hypothesized that the intra-patient correlation was 0.50. To reach a power of 80% to show a difference of 20% at a significance level of 5% using a two-sided McNemar’s chi-squared test, a total of 52 patients was needed. As a 25% drop-out was expected (either because of withdrawal of consent or death before the beginning of the study or between the two sessions due to their own evolution), the sample size was increased to 66 patients. Because of a shortage in healthcare providers and challenges in recruiting during the COVID pandemic, recruitment was stopped after having randomized 24 patients.

Randomization was generated in blocks of four by the Geneva University Hospitals’ Pharmacy to ensure blinding and concealment of allocation. Two randomization lists were generated, corresponding to each of the two opioids allowed as control treatment (i.e., morphine and hydromorphone). Randomization determined the order of administration of the two interventions. Patients were randomized to receive either IN DEX in the first nursing care session and SC opioids in the second, or SC opioids in the first and IN DEX in the second nursing care session.

The study medications were provided by our pharmacy in prepacked sealed kits for each randomization number and for each session, labeled in a blinded manner. Two types of kits, sequentially and clearly labeled with different colors, were prepared, one for each type of opioid used (morphine and hydromorphone). Kits for both nursing care sessions were prepared in separate zip plastic bags. Each session package contained one of the study treatments.

A vial for SC administration contained either 3 mL of 20 mg/mL of morphine (or 10 mg/mL of hydromorphone) or NaCl 0.9% with similar packaging. The dosage administered corresponded to 10% of each patient’s opioid daily dose.

A vial for IN administration contained 1.2 mL of 100-µg/mL solution of DEX or NaCl 0.9%, with similar packaging. A 1-mL syringe capped with an IN mucosal atomization device (MAD300; Teleflex, Wayne, PA, USA) was used for IN administration. The investigational product administered to patients in a supine position by the IN route corresponded to 0.125 mL/10 kg of DEX (1.25 µg/kg of actual body weight) or NaCl 0.9%.

During the 20-month period of recruitment, 523 patients were assessed for eligibility, of whom 24 were randomized. Three randomized patients did not complete the two trial interventions. One patient showed signs of impending death and did not receive the first nursing care session in the IN DEX/SC opioid sequence, and two patients were excluded before starting the second session in the context of hypotension and septic shock in the SC opioid/IN DEX sequence. Finally, a total of 21 patients were included in the intention-to-treat analysis (Fig. 1).

The trial was conducted from 14 November, 2018, to 27 July, 2020, and was stopped because of recruitment difficulty during the pandemic.

Baseline characteristics of participants are presented in Table 3. Of the 21 patients analyzed, ten received the IN DEX/SC opioid sequence, and 11 received the SC opioid/IN DEX sequence. Two patients received hydromorphone, and 19 morphine, as SC opioids. Sex distribution was balanced (12 women and 9 men), age ranged from 65 to 93 years, and median actual body weight was 57 kg. Eight participants had a diagnosis of cancer. Their Palliative Performance Scale ranged from 20 to 50%. Nine patients had one or more wounds and five presented with a recent fracture, for which only a conservative treatment was decided. Hypoactive delirium was present in four participants.

All participants were able to rate their current level of pain with the VAS before completed interventions. Pain intensity at baseline ranged from VAS 0 to 8 cm (median 3) and their mOAA/S score range from 4 to 6 (median 5). The mean heart rate was 94 ± 17 and, mean oxygen saturation was 94 ± 2.3.

Outcome measures are summarized in Table 4.

The major finding of this study is that IN administration of study drugs was easy and convenient in all participants, even in those in whom a confusional state was detected. Furthermore, our results, while limited, suggest that IN DEX could be an alternative to SC opioids in older patients. Indeed, there were no severe adverse events reported, and the IN DEX sedative effect was mild, enabling the patient collaboration during nursing care and minimal monitoring.

Emerging literature highlights the potential role of DEX in palliative care. A case series [34], case reports [35, 36], a small prospective (26 adults of mean age > 65 years) [37], and a nine-patient pediatric cohort study [38] all suggest the potential interest of IN, IV, or SC infusions of DEX for the treatment of intractable pain, opioid-induced hyperalgesia, terminal intractable delirium, and dystonia [34‐38]. The adequate dose of IN DEX to be administered was a challenging choice. One case report suggested that both 1 and 1.5 µg/kg of IN DEX were safe and effective to induce light sedation in the management of complex wound dressings for cutaneous thoracic metastatic bilateral breast cancer in a 45-year-old woman [39]. The recommended doses of IN DEX varied from 0.5 to 1.5 µg/kg for adults, while the doses commonly reported in children varied from 0.5 to 4 µg/kg [13, 15, 20, 23‐25, 40, 41]. Intranasal DEX bioavailability was estimated to be 40.6% (95% CI 34.7–54.4) for atomization [42]. It must be underlined that age-adjusted dosing is not recommended, although caution is warranted as sedative effects, hypotension, and bradycardia might be more pronounced in elderly patients. Based on the literature review, previous local experience and observations, the dose chosen for DEX treatment in the present study was 1.25 μg/kg of actual body weight. Interestingly, in studies published recently (2019–22), and therefore after the present trial had been initiated, IN DEX 100, 50 µg, or 1 µg/kg was administered before anesthesia, or after spinal/or general anesthesia in elderly healthy subjects, with a relatively good tolerance profile [43‐45].

Our study population reflects the geriatric population encountered in daily practice, palliative units, hospices, or retirement homes, despite the stringent non-inclusion criteria applied. The strength of our study is that it addresses an important topic concerning a population rarely studied in randomized controlled trials. It has been well designed and conducted partly in a difficult period. A cross-over design was chosen for two main reasons. First, because pain experience differs from one individual to another and because the differences can be even larger between patients of the targeted population. Second, expecting a relatively scarce number of patients recruited, the cross-over design was an interesting option. A maximum period of 24–48 h between the two nursing care sessions was decided as rapid changes in a patient’s status may occur near end-of-life. Although, because of the patient’s health conditions, some dropouts occurred between the two sessions, we believe that the cross-over design was a good choice. Because of the specific type and health conditions of the studied population, we expected diverse adverse events, unrelated to the investigational product, to occur and therefore focused only on bradycardia and hypotension. Eventually, bradycardia was not observed with IN DEX during the study. Although this alone does not allow a conclusion on the safety of the IN DEX administration, we did not face serious adverse events.

The main limitation of this study is that we encountered difficulties in recruitment, leading to prolonged duration of the trial and lack of resources, and were therefore unable to enroll the number of patients needed to draw firm conclusions. The problem of difficult recruitment in clinical trials including older patients is well recognized [46]. Our restrictive eligibility criteria certainly led to a reduced recruitment. In fact, of the 523 patients assessed for eligibility, 460 (88%) were not eligible. Although these criteria were settled with the aim to minimize the likelihood of harm to vulnerable patients (e.g., ileus, recent stroke, heart, or respiratory failure), and of frequently reported side effects (nausea/vomiting, respiratory depression), some of these criteria should be revised if a similar study was planned. For example, we restricted inclusion to patients with a weight > 80 kg, as an 80-kg patient would need to receive an IN volume of 1 mL, which is the commonly maximum volume recommended for nasal administration. If we were to plan a future study, we may choose to include these patients and adapt the administration scheme with a maximum dose. Additionally, excluding patient using opioids other than morphine and hydromorphone on a regular basis may not be necessary. We chose SC morphine or hydromorphone as active comparators as these were the opioids most frequently used in our institution, reaching their peak effect after 45 min, with a maximum efficiency lasting up to 90 min. However, 42 patients were not approached because they were using other opioids. Less stringent eligibility criteria could improve participation. In the present trial, about half of the patients approached refused to participate. The fact that, from March 2020 on, their relatives were hardly available for discussion, in the context of visit restrictions during the COVID-19 pandemic, did not help. It is possible that in another context, the acceptance rate would have been greater. However, this a fair consent rate for a drug clinical trial according to our experience in Geneva. A recently published survey highlights that more time is needed to enroll older patients and suggests that it can be dealt with by increasing the number of recruiting sites, planning a longer time for recruitment, widening eligibility criteria, and adapting the information on the study to this population [47]. Finally, a future study should try to involve the next-of-kin in a patient and public involvement approach [48]. Collaboration with the next-of-kin of seriously ill patients would certainly help to better understand what can be implemented in the design and conduct of the trial to make it more attractive and valuable for patients and their families. For example, the selection of an appropriate tool for pain assessment is challenging. Although many exist, there is limited evidence regarding their reliability, feasibility, and clinical utility for older adults [49]. We chose what we considered as a meaningful score of 5 on the ECPA, together with a self-reported measure of pain as the ECPA is user friendly and its internal consistency has an acceptable rating of 0.7, compared with other scales [49]. As an appropriate choice of outcomes and tools for measuring outcomes is essential, we strongly encourage researchers to discuss with patients and their next-of-kin in order to select the most appropriate assessment tools for future studies by conducting a feasibility or pilot study. Finally, as this was a clinical trial, we did not attempt to describe the pharmacokinetics of the study drugs in this specific population.

The present study is too small to draw any conclusion regarding the efficacy and safety of the IN administration of DEX. However, we have collected data on 21 patients from a population rarely included in randomized trials. We believe that recognizing and discussing our difficulties, as well as describing our results is relevant as it will enable the scientific and medical community to design and achieve future multi-centric studies and hamper failure in recruitment. Although SC and oral routes could also be possible suitable routes for DEX administration, in palliative care situations, especially in the community or hospice setting [50, 51], future multi-centric trials should consider IN DEX as an option for procedural pain management.