Advertisement

Premature Ending of a Medication Study in Dying Patients: Lessons Learned

Open ArchivePublished:August 09, 2019DOI:https://doi.org/10.1016/j.jpainsymman.2019.07.027
      To the Editor:
      In this letter, we describe the premature discontinuation of a medication study in dying patients. Despite a feasibility assessment, recruitment rates were far below expectations. Suggestions for future studies are proposed.
      Morphine is the most frequently used opioid for the treatment of pain at the end of life. However, the active metabolites of morphine start to accumulate when renal function decreases. The accumulation of M3G is associated with neurotoxic adverse effects like delirium, allodynia, and hyperalgesia.
      • Ashby M.
      • Fleming B.
      • Wood M.
      • Somogyi A.
      Plasma morphine and glucuronide (M3G and M6G) concentrations in hospice inpatients.
      By contrast, the central effects of circulating metabolites of oxycodone are negligible.
      • Lalovic B.
      • Kharasch E.
      • Hoffer C.
      • et al.
      Pharmacokinetics and pharmacodynamics of oral oxycodone in healthy human subjects: role of circulating active metabolites.
      Theoretically, oxycodone for the treatment of pain in dying patients, with a diminished renal function, should therefore result in a reduced occurrence of the neurotoxic adverse effects like delirium in comparison to morphine. To investigate this hypothesis, a randomized, controlled, multicenter trial was designed to compare the prevalence of delirium between oxycodone and morphine, administered by continuous subcutaneous infusion (CSCI), for the treatment of pain in dying patients with a diminished renal function. The study population consisted of residents of hospices and somatic or psychogeriatric wards of nursing homes, ≥18 years, eligible for start of CSCI of an opioid for the treatment of pain in the dying phase. Patients with a diminished renal function, defined as an estimated glomerular filtration rate (eGFR Cockroft-Gault) of <50 mL/minute/1.73 m2, were randomized between morphine and oxycodone. Based on a reported percentage of delirium in terminal patients of 86% and a difference of 15% as considered clinically relevant, 117 patients per group were needed (α 5%, 1 − β 80%).
      Recruitment started on June 1, 2018, and by February 1, 2019, 39 possible eligible patients were identified. There were 27 patients classified as not eligible. Of one patient eligibility was not established. Of the noneligible patients, eight patients declined consent including 2 who died during the consent procedure. Consent was given by 11 patients, 3 of whom were not randomized because they had no indication for CSCI anymore and died without pain. Of the eight randomized patients, three patients were excluded afterward based on eGFR values >50 mL/minute/1.73 m2, one patient was no longer considered in the terminal phase anymore, and one patient died four hours after the start of the study treatment (see Fig. 1). Only three of all identified patients completed study treatment, which was 52 patients below expected. Because this gap seemed insurmountable, the study was terminated prematurely.
      Figure thumbnail gr1
      Fig. 1Patient flowchart. CSCI = continuous subcutaneous infusion; eGFR = estimated glomerular filtration rate (ml/minute/1.73 m2).
      Evaluation of the recruitment process revealed two factors mainly responsible for the poor results. First, six of the participants died before, during, or immediately after the period of obtaining informed consent, at least before any measurement could be completed. Before the start of this study, not much was known about the time span between start of CSCI with opioids and the moment of dying. Therefore, a survey among Dutch elderly care physicians was held and the time span was estimated at several days. However, during the study, the time span appeared to be rather less than 24 hours. This discrepancy might be explained by the extra critical evaluation of the indication for CSCI because of the study. Once the indication was set, study participation did not cause any delays.
      Second, the renal function as established using eGFR in three of the 11 (27%) included patients appeared to be above the threshold of 50 mL/minute/1.73 m2. Because the eGFR is based on serum creatinine levels and elderly and patients in the dying phase are known for lower protein intake and lower muscle mass, eGFR levels are often overestimated in these patients. A recent publication revealed a prevalence of about 50% of decreased eGFR levels in terminal patients, assessed three days (median, IQR 1–4) before death.
      • Masman A.D.
      • Tibboel D.
      • Baar F.P.
      • et al.
      Prevalence and implications of abnormal laboratory results in patients in the terminal phase of life.
      We hypothesize that the level of serum creatinine continues to decrease, resulting in a complete overestimation of renal function. Because renal insufficiency was a key element in the study, removal of this exclusion criterion was no option. Alternative tests like cystatin-C were considered; however, extra costs in this regard were not factored in, or logistically complicated because of transportation to an external laboratory, or not yet sufficiently standardized.
      Considering that halfway the planned study period the number of patients who received study treatment was three, and that extension of the study to extra centers still would not lead to a substantially higher number of participants, the study was ended prematurely. No conclusion could be drawn concerning the association between delirium and prescription of morphine versus oxycodone in dying patients with diminished kidney function. However, this issue remains of utmost clinical importance in palliative care. We recommend to repeat the study in an earlier phase (prognosis 2–4 weeks) to answer the question whether the accumulation of morphine metabolites induces clinical relevant adverse effects. In addition, we recommend a study on renal function in terminal patients assessing multiple biomarkers such as cystatin-C.

      Disclosures and Acknowledgments

      This work was supported by funding from The Netherlands Organization for Health Research and Development (ZonMw), grant number 844001314 , The Hague, The Netherlands and Envida, Maastricht, The Netherlands.
      Ethical approval: MEC approval was given by METC azM/UM Maastricht, The Netherlands, Dec 6, 2017, number NL62110.068.17/METC172020.
      This study was registered at EudraCT, number 2017-002192-25.
      Data are available at request (corresponding author).
      There are no conflicts of interest to report.

      References

        • Ashby M.
        • Fleming B.
        • Wood M.
        • Somogyi A.
        Plasma morphine and glucuronide (M3G and M6G) concentrations in hospice inpatients.
        J Pain Symptom Manage. 1997; 14: 157-167
        • Lalovic B.
        • Kharasch E.
        • Hoffer C.
        • et al.
        Pharmacokinetics and pharmacodynamics of oral oxycodone in healthy human subjects: role of circulating active metabolites.
        Clin Pharmacol Ther. 2006; 79: 461-479
        • Masman A.D.
        • Tibboel D.
        • Baar F.P.
        • et al.
        Prevalence and implications of abnormal laboratory results in patients in the terminal phase of life.
        J Palliat Med. 2016; 19: 822-829