Effect of Prophylactic Subcutaneous Scopolamine Butylbromide on Death Rattle in Patients at the End of Life
The SILENCE Randomized Clinical Trial

van Esch. JAMA 2021;326(13):1268-1276. doi:10.1001/jama.2021.14785

Clinical Question

  • In patients near end of life, does subcutaneous scopolamine butylbromide compared with placebo, reduce the incidence of death rattle?


  • The paper defines death rattle as noisy breathing caused by the presence of mucus in the respiratory tract
  • Death rattle occurs in a number of patients that are dying. This may be something that is distressing for relatives. It is not thought to be distressing to the patients due to their reduced conscious level. However, the actual effect is unknown, and patients may fear it if they have previously heard a death rattle
  • Anticholinergics decrease the production of mucus and can be considered for the treatment of death rattle. However there is no evidence to support their use, with 2 RCTs reporting no significant benefit (1, 2)
  • This trial aimed to see if using them prophylactically could prevent the death rattle


  • Multi-centre
  • Randomised
  • Double-blind, placebo-controlled trial
  • Variable block sizes of 2-4
  • Informed consent gained in anticipation of entering the dying phase when patient mentally aware
  • Registered on Netherlands Trial Register
  • Sample size
    • 180 patients required to give 80% power to detect a relative reduction of 50% from a baseline of 39%, with a false positive rate of 5%
    • Assuming that 10% of patients who gave informed consent may not be able to be randomised, the study aimed to gain consent from 200 patients
  • Modified intention to treat analysis (excluded patients who were determined not to be in the dying phases after randomisation or if consent withdrawn)
  • Death rattle graded according to Back grade
    • 0: no rattle
    • 1: audible close to the patient
    • 2: audible standing at the end of the bed
    • 3: audible standing in the door opening


  • 6 hospices in the Netherlands
  • Data collected: April 2017 – December 2019


  • Inclusion:
    • Admission to hospice for end-of-life care
    • Life expectancy of ≥3 days
    • No signs of disturbed consciousness at the time of informed consent
  • Exclusion:
    • Signs of active respiratory tract infection (upper or lower)
    • Tracheostomy or tracheal cannula in situ
    • Use of anticholinergic drug or octreotide
  • At the recognition of the dying phase, the patients were re-assessed for their eligibility based on the following criteria:
    • No active respiratory infection
    • Not using systemic anticholinergic drugs
    • Did not have any death rattle
  • Patients who completed 8 days of treatment were withdrawn from the study because it was assumed that they had not entered the dying phase
  • 1097 patients admitted to hospice, of whom 162 were randomised, and 157 included in primary analysis
  • Comparing baseline characteristics of intervention vs. control group
    • Male: 46% vs 42%
    • Age: 78 vs 75
    • Primary diagnosis
      • Cancer: 84% vs 89%
        • Lung cancer: 18% vs 35%
      • Cardiovascular disease: 10% vs 1%
      • COPD: 1% vs 3%
    • Co-morbidity
      • COPD: 10% vs 23%
    • Smoked in the previous year: 14% vs 33%
    • Urinary catheter in situ at the recognition of dying phase: 63% vs 72%
    • Sedatives given at the recognition of the dying phase: 19% vs 21%


  • Scopolamine butylbromide
    • Subcutaneously 20mg (1ml) four times a day


  • Placebo
    • Saline, 1ml

Management common to both groups

  • Dying phase determined by MDT using clinical judgement based on standardised signs
    • These signs included being bedbound, semi-comatose and no longer able to take oral medications or swallow
  • After the recognition of the dying phase the “Care Program for the Dying” was used
    • The template was expanded to include the Beck Scale and the Vancouver Interaction and Calmness Scale (VICS) for restlessness in addition to the 13 other goals of care that were evaluated every 4 hours
  • Treatment was continued until death, or until occurrence of a death rattle of grade 2 or above at two consecutive time points four hours apart (the primary outcome)
    • At this point the study medication was deemed to have failed and usual care, including open use of anticholinergics as needed, was provided


  • Primary outcome:
      • Occurrence of a ≥ grade 2 death rattle measured at 2 consecutive time points within a 4-hour interval – significantly reduced in Scopolamine group
    • 13% vs 27% (difference 14%, 95% C.I. 2%-27%, p=0.02)
      • NNT 8
      • Fragility index 1 patient
  • Secondary outcomes:
    • Comparing intervention vs. control group
      • No significant difference in
        • Restlessness
          • 28% vs 23%
        • Dry mouth
          • 10% vs 15%
        • Urinary retention
          • 23% vs 17%
      • Significantly reduced in Scopolamine group
        • Cumulative incidence of death rattle at 48 hours
          • 8% vs 17%
        • Time from the recognition of the dying phase to death rattle
          • HR 0.44 (0.20-0.92), p=0.03
  • Post-hoc analysis
    • Duration of dying phase – significantly longer in Scopolamine group
      • 42.8 vs 29.5 hours (difference -13.3; 95% C.I. -17.4 to -6.4), p=0.04
    • In the placebo group, the occurrence of death rattle was lower in patients with the following conditions (compared with the rest of the placebo group)
      • Lung cancer: 11%
      • COPD as a comorbidity: 17%
      • Recent history of smoking: 15%

Authors’ Conclusions

  • Prophylactic subcutaneous scopolamine significantly reduced the occurrence of the death rattle


  • Randomised
  • Double-blinded
  • Multi-centre
  • Registered on Netherlands Trial Registry


  • Final analysis included only 10% of the population that were admitted to the hospices
  • The placebo group included a higher % of patients with lung cancer as their primary diagnosis, COPD as a co-morbidity and a history of smoking in the last year. The death rattle was observed less commonly in these groups of patients compared with the rest of the placebo population. With a low fragility index, the differences found in the primary analysis may be due to differences in baseline characteristics
  • The duration of dying was significantly longer in the scopolamine group. This may be because scopolamine increases the duration of dying, which in my opinion would be an unwanted side effect, or it may be because there were differences in baseline characteristics
    • If there were differences in baseline characteristics this may lead to bias
  • With a relatively small sample size there may have been significant differences in side effects that were not picked up
  • This study was conducted in a hospice setting. The staffing ratios and management of patients that are dying may be different in an ICU setting, which may limit the external validty
  • The inter-observer variability for the assessment of the primary outcome was not stated

The Bottom Line

  • In patients near the end of life, scopolamine compared with placebo significantly reduced the incidence of the death rattle. However, with a very low fragility index and differences in baseline characteristics between the study groups, I would want to see further evidence before this becomes routine practice
  • I would also want future trials to be adequately powered to determine any clinically significant differences in side effects, as these may outweigh any benefits from this treatment

External Links


Summary author: @davidslessor
Summary date: 10/11/2021
Peer-review editor: George Walker

Photo by: Kristina Flour on Unsplash


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.