Dynamic light application therapy to reduce the incidence and duration of delirium in intensive-care patients: a randomised controlled trial

Simons K et al. Lancet Resp Med 2016; 4(3): 194-202.

Clinical Question

  • In adult ICU patients, does the use dynamic light therapy compared to normal lighting, reduce the incidence and duration of ICU-acquired delirium?


  • Randomised controlled trial
  • Consecutive recruitment
  • Single-centre
  • Unblinded
  • Intention to treat analysis
  • Power calculation: 1000 patients needed to detect 10% absolute reduction in delirium incidence (90% power, 2-sided alpha value of 0.05).
    • Assumed incidence rate of delirium of 40% in control group


  • Dutch teaching hospital
  • 16 bed mixed medical and surgical ICU; 12 of which had direct daylight access
  • Data collected: July 2011 – September 2013


  • Inclusion criteria:
    • ≥18 years old
    • Expected ICU stay > 24 hours
  • Exclusion criteria:
    • Life expectancy < 48 hours
    • Unable to assess for delirium (severe sensory impairment; severe mental impairment; unable to understand Dutch language)
  • Participant numbers
    • 1374 eligible, 734 randomised (361 Intervention, 373 control), data available for 711
    • 400 excluded as expected LOS < 24h
  • Comparing baseline characteristics (intervention vs control):
    • Age: 66.3 vs 64.4
    • Men: 58% vs 59%
    • APACHE II: 22.7 vs 22.4
    • Medical patients: 73% vs 65%
    • History of cognitive disturbance or alcohol abuse: 17% vs 14%


  • Dynamic lighting level and colour temperature sequence:
    • Rose from 0700h onwards to peaks of 1700 lux (illuminance) and 4300K (colour temperature) at 0900h
    • Maintained until 1130h
    • Decreased to 300 lux and 3000K until 1330h
    • Increased to 1700 lux and 4300K until 1600
    • Gradual reduction to 300 lux over 1h at 1600
    • Light turned off automatically at 2230h
  • If emergency lighting required, 1000 lux light could be used which would automatically revert to previous setting after 30 minutes


  • Standard lighting setting of 300 lux and 3000K. The light could be turned off and on in the room and could be changed to bright settings of 1000 lux for emergencies and procedures

For both intervention and control group

  • Identical lighting system (Philips Dynalite DUS804C, Philips Lighting) was installed in every ICU room


  • Primary outcome: Cumulative incidence of ICU-acquired delirium (comparing intervention and control)
    • No significant difference (38% vs 33%, Odds ratio 1.24, 95% CI 0/92-1.68, p=0.16)
  • Secondary outcomes: comparing intervention vs. control group
    • No statistically significant difference in any of the outcome measures
      • Delirium free days: 26 vs 27 (P=0.29)
      • Duration of mechanical ventilation: 20h vs 19hr (P=0.67)
      • Length of ICU stay: 4 days vs 4 days (P=0.82)
      • ICU mortality: 12% vs 14% (P=0.42)
      • Length of hospital stay: 15 days vs 16 days (P=0.84)
    • No significant difference on further analysis in:
      • Use of sedative agents
      • Use of haloperidol
      • Hospital mortality
      • Excretion profile of 6-sulfatoxymelatonin or cortisol (biochemical markers of circadian rhythm)
  • Mean cumulative daytime lighting levels between 0600h and 1800h were significantly higher in the intervention group compared with the control group: 5366 vs 2793 lux (P<0.0001). Outside this period, mean lux was similar in both groups

Authors’ Conclusions

  • In critically ill patients the use of continuous lighting therapy was not effective in reducing the cumulative incidence or the duration of ICU-acquired delirium compared to normal lighting.


  • RCT
  • Separation of intervention and control light profile achieved
  • CAM-ICU score is recognised and validated for use of assessing delirium. This was measured three times each day


  • Single centre
  • Main exclusion criteria was that the patient was expected to spend < 24h on ICU. Duration of invasive mechanical ventilation in both groups 19-20h which is short
  • Incidence of delirium in control group lower than that used for power calculations

The Bottom Line

  • Dynamic light therapy at supranormal levels conferred no benefit above natural lighting. Delirium on the ICU remains common and difficult to treat. I will continue my usual practice of daily sedation breaks, early mobilisation and orientation therapy on the ICU

External Links


Summary author: @avkwong
Summary date: 6th February 2017
Peer-review editor: @stevemathieu75


  • Duncan Chambler

    Adrian, thanks for your summary and critique. This is a fascinating theory, which unfortunately this trial has not supported.

    From the data you’ve presented (I admit I haven’t had a chance to read the paper yet), it appears they did not recruit the number of patients they intended to. As the incidence of delirium was 38% vs 33%, the Absolute Risk Increase (ARI) is 5% (which is smaller than the 10% they deemed clinically relevant in their power calculation). By recruiting only 75% of their target, the 95% CI for this ARI are -1.62% to 12.14% (i.e. the null hypothesis cannot be rejected), but if they had continued to 1000 patients and if the incidence remained the same in each group, then the 95% CI would have been 0.81% to 9.19%. From this, we might have concluded that dynamic light therapy is harmful, causing delirium in 1 patient for every 20 treated (i.e. NNH 20).

    Therefore, I think this trial is unable to conclude that dynamic light therapy has no effect on delirium but I completely agree with you that this trial demonstrates it does not have a clinically meaningful positive impact on delirium, unfortunately.

  • Agree with power calculations.

    I think its an interesting theory though and as you say, its disappointing results. Given the drive for non-pharmacological interventions e.g. orientation therapy for delirium. My unit is located in the basement of the hospital with little natural light (Unlike Steve’s unit). I do think the lack of light not only affects patients, but also staff. It would be very interesting to do a bit and after if/when we move to a new unit build.

  • Pingback: Nishikimi – The Bottom Line

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.