Continuous subcutaneous infusion for pain control in dying patients: experiences from a tertiary palliative care center

Pain is a common clinical problem in palliative care [1]. Eighty percent of patients with advanced cancer and 67% of patients with cardiovascular disease or chronic obstructive pulmonary disease will experience moderate to severe pain at the end of their lives [2]. A better understanding of pain mechanisms has improved pain therapy in recent years [3]. Especially, cancer patients in palliative care are adequately treated with low or moderate doses of opioids and only about 10 % need daily morphine doses over 600 mg [4]. However, with progressive deterioration at the end of life, continuous subcutaneous infusions (CSCI) become increasingly important to continue pain control in the dying [5]. The use of an ambulatory infusion pump (AIP), a small, portable pump that delivers medication via a thin subcutaneous catheter, has several advantages for pain control: It ensures a steady infusion of drugs, with reliable absorption if inserted in unaffected tissue, and allows combinations of drugs to be administered parenterally in a manner that is more convenient than repeated and painful injections. Thus, there is no need for an intravenous access [6, 7]. However, despite encouraging data in palliative care in earlier phases, there is a lack of prospective studies on the use of CSCI in the imminently dying.

Low-dose add-on peroral methadone in combination with other opioids for pain is proposed to be a useful alternative to methadone therapy for better pain control at the end of life [8‐12]. The addition of methadone is reported to improve pain relief in complex pain situations but, so far, only peroral or intermittent parenteral administration of low-dose add-on methadone has been studied, routes that are often not feasible in the imminently dying patient [13‐15]. Alternative routes of methadone administration are therefore needed and there is a need for further exploration of the effects, and possible adverse effects, of CSCI in this patient group [13, 14].

Whereas patient-reported outcomes are encouraged, this is not possible in imminently dying patients, as the majority of them are not able to complete a questionnaire for obvious reasons, such as, terminal delirium, extreme tiredness or lowered level of consciousness. For this reason, very few studies on symptoms and symptom control have been performed in the very last days of life, a period which is important both for the patient and for their families. A way to overcome this problem is to use validated instruments allowing proxy measures, e.g. assessments made by staff.

The primary aim of the study was to report the effects on pain intensity and occurrence of adverse effects, e.g. sedation, confusion, and respiratory depression, when prescribing an AIP for CSCI in imminently dying patients. A secondary aim was to specifically study the effects of the addition of low-dose methadone to a CSCI comprising another opioid in this patient group.

The methods and results sections are reported, when applicable, based on the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria [16].

During the study period, 321 patients were prescribed CSCI and a total of 93 were included in the study intended to be followed prospectively. Of those not included, 88% were assessed as too ill to give formal consent and 12% abstained (Fig. 1). Mean age of the included patients was 76.3 years (median 77 years), 57% were women, (Table 1). Indications for CSCI were: inability to swallow due to deterioration of general condition in 49 (53%) patients, uncontrolled pain with oral pain medication in 27 (29%) patients, bowel obstruction in 6 (6%) patients and unspecified reasons in 11 (12%) patients. On day 0 (initiation of CSCI), the mean pain score was 2.2 (median 2), with 5 (5%) patients rating no pain (Table 2). Two thirds of the reported pain mechanisms were combined nociceptive and neuropathic. For other characteristics, see Table 1.

Three patients, one of whom received methadone, experienced local erythema which disappeared within one day after changing the injection site. In all three cases, the needle had been in place for at least 5 days.

In this study on symptom relief during the very last days of life, we report that the use of AIP for CSCI for pain control in dying patients contributed to improved analgesia with no clinically significant change in adverse effects. In addition, CSCI with low-dose methadone in combination with other regular opioids was prescribed at the individual physician’s discretion and resulted in improved analgesia for patients with the most severe pain.

Several studies and guidelines report on the treatment of severe pain in patients in palliative care in less advanced stages [3, 22‐24]. However, few studies investigate pain and symptom management in the imminently dying, due to the well-known problems related to symptom assessment [25, 26]. The short median survival time of four days confirms that the patients in this study were at the very end of life. In this setting, the most common reason for initiating CSCI was general deterioration causing impaired oral intake. Another major reason was to provide better pain relief by converting to parenteral drug delivery and, when judged necessary, by adding parenteral methadone. The need for better pain relief may be due to mixed nociceptive and neuropathic pain in two thirds of the patients in this study, a combination of pain mechanisms often difficult to treat [14].

Significant reduction in pain was seen for the entire group of patients who received CSCI with opioids for pain and who could be followed for at least three days. This was regardless of whether the pain was measured as the proportion of patients with severe pain or as the median and average pain scores based on the Likert scale for pain in the IPOS [17]. Advantages in CSCI administration of opioids in relation to pain control include a more stable serum concentration of the drug with avoidance of end-of-dose interval breakthrough pain and less adverse effects occurring at high peak concentrations following intermittent injections which allows a more adequate opioid dose titration [6, 27, 28]. In the imminently dying patients a rotation from oral intake to parenteral routes might be especially beneficial, as the oral route, including swallowing and absorption, might be unreliable in the last days and hours of life.

In this study, it was not recorded whether the same opioid was used when switching to CSCI or whether an alternative opioid was used. There may exist a cross-tolerance between opioids, meaning that a different opioid has a better analgesic effect than expected from equianalgesic tables [29]. However, a review by Schuster et al. 2018 confirmed the stated findings in the Cochrane review from 2004, that although widely practiced, robust evidence for the benefit of opioid rotation is still lacking [30, 31].

Patients who required the addition of methadone for analgesia had higher pain scores and opioid doses at initiation of CSCI. Thus, the use the co-prescription of low-dose methadone with another opioid for CSCI via AIP was mainly initiated in patients with complex pain of high intensity, in order to improve pain control without large dose escalations of the regular opioid. This is in line with the study by Mercadante et al., that described how addition of methadone to another opioid in patients with cancer pain may reduce the need for opioid dose escalation [32]. As shown in this study, the apparent beneficial analgesic effect adding of low-dose methadone to another opioid for administration by CSCI is promising and reflects observations reported for oral administration [8‐12].

There was a statistically significant decrease in alertness (p < 0.001). However, the clinical impact is probably minor as the change was 1.0 in a 10-grade scale, i.e. a 10 % deterioration, in acutely dying patients. Besides this, there were neither any increase in intensity of other adverse symptoms nor any serious adverse effects that demanded specific interventions. Moreover, we observed no significant differences over time or between MET and NMET regarding performance status, anxiety levels or doses of the anxiolytic midazolam. These observations agree with the general observation that adverse effects such as sedation, constipation, and respiratory depression are associated with the pharmacology of opioids as a class, and similar reactions are expected regardless of route [28].

Notably, those in the subgroup MET had higher total opioid doses, still they survived marginally longer. The MET group’s longer survival from introduction to death may possibly be explained by their more complex pain situation, possibly resulting in an earlier introduction of CSCI.

Particularly interesting is that the prevalence of delirium in the main study group did not change over time and did not differ between patients younger or older than 75 years of age. To the best of our knowledge, differences in adverse effects following introduction of CSCI for pain in different age groups has not been described before, but our findings are consistent with previous studies indicating that a steady infusion of subcutaneous drugs may be better tolerated [6, 33].

Problems with local skin irritation at the site of CSCI has been associated with methadone but are reported to resolve with site rotation [15, 34‐36]. In our prospective study, only low-doses of methadone were infused and we found only three cases of skin irritation, two in the NMET group and one in the MET group. The needle had been in the same place for at least five days which, therefore, may be as likely an explanation for the dermal erythema as the drugs in the CSCI. We therefore suggest that the risk of skin irritation should not be considered a major limiting factor for the use of methadone in CSCI.

There was a significant difference in mean age between the MET group, 68 years, and the NMET group, 78 years, (p < 0.05). We described a similar difference in a study from 2020 on a different sample [13]. To the best of our knowledge this has not been studied before and we do not know the underlying reason.

We do recognize some limitations of the study. First, as accounted for in the methods section, all patients receiving a CSCI were not included, usually due to illness and of ability to provide informed consent, as well as some patients refraining from participation. Second, as methadone, according to clinical guidelines, was mainly prescribed to patients with the most severe pain, the MET and NMET groups differed significantly at baseline as regards pain intensity and MEDD. Still, such comparisons of subgroups are of interest, as they reflect a clinical reality. Comparable control groups are difficult to achieve in a population of dying patients in need of parenteral drug administration, and randomization with control groups is seldom possible, for ethical reasons. Third, due to the severe illnesses of the patients, observer ratings had to be used. By definition, observer assessment may not directly reflect the patient’s experiences, meanwhile, they allow studies in the last days and hours of life, on an important patient group that is often excluded from studies. The assessments were based on subjective judgements by the registered nurses who performed the registrations. Consequently, the assessments came to a certain extent to depend on each individual’s level of knowledge, skills, and personal attitudes. Finally, this is a single-center study reflecting the experiences at one particular setting.

Strengths include the prospective design until death and the use of validated instruments for symptom assessments. Another strength is that the proxy assessments were performed by trained personnel, which may have contributed to improved assessments of adverse effects [37].

Regardless of age, CSCI via AIP for pain seems effective in reducing pain in dying patients without any substantial increase of adverse effects such as delirium or respiratory depression. An addition of low-dose methadone may be beneficial for CSCIs for patients with severe cancer pain at the end of life. The effectiveness of low-dose methadone in combination with the patient’s regular opioid needs further investigation, preferably with a randomized controlled trial.