CCP Podcast 057: Papers of the Month

 November 10

by Jonathan Downham

In this podcast episode, Gavin Denton and I look at three recent papers, breaking them down to try to help us and you understand some of their conclusions and how they reached them.

Our notes can be found below.

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

Simons et al The Lancet 2016

Clinical Question

Continuous bright light therapy, delivered through a ceiling-mounted lighting system, in addition to regular lighting and daylight, would reduce the incidence and duration of ICU-acquired delirium.


Single-centre, randomised, controlled trial.


ICU of the Jeroen Bosch Hospital in ’s-Hertogenbosch, Netherlands.


Inclusion criteria: All consecutive ICU patients admitted to the ICU between July 1, 2011, and Sept 9, 2013, were screened for eligibility. Patients had to be 18 years or older with an expected ICU stay of at least 24 h.

Exclusion criteria: Those whose anticipated life expectancy was less than 48 h or who could not be assessed for delirium (eg, severe hearing or visual impairment, unable to understand Dutch, or severe mental impairment).


A bright lighting system was installed into the ceiling of every ICU room. 800–1000 lux bluish-white light at the level of the patients’ heads, in line with previous reports that had showed effects with lighting therapy.

Mean illuminance was calculated for each patient per hour for the duration of his or her stay in the ICU. The lighting system was controlled from a central module, located in the nursing station, which was only accessible by the investigators.

For patients in the DLA group, lighting level and colour temperature rose from 0700 h onwards to peaks at 0900 h.

This light intensity was maintained until 1130 h.

To allow patients a period of rest, in accordance with normal daily practice in the ICU, the lighting level was decreased until 1330 h.

From 1330 h onwards, the level was again increased and was maintained until 1600 h, after which a gradual fall occurred over 1 h.

At 2230 h the light was switched off automatically.


Patients in the control group were exposed to the standard lighting settings of 300 lux and 3000 K. The light could be turned on and off in the room and could be changed to a bright setting of 1000 lux for procedures.

Cumulative incidence of delirium was defined as the presence of delirium (at least one positive CAM-ICU screening) on at least 1 day during ICU stay.

The number of delirium-free and coma-free days in 28 days was calculated by subtracting the number of days patients had delirium or were comatose from 28.

Patients were deemed to be comatose when the CAM-ICU could not be scored and the Richmond Agitation Sedation Scale score was lower than –3 for the whole day.

All days without both conditions were defined as delirium-free and coma free days.

When a patient had been free from delirium for 48 h since a positive CAM-ICU score, the delirium was judged to have resolved and the 2 days were recorded as delirium free.

To detect differences in the effects of DLA and standard lighting settings on circadian rhythm, concentrations were measured of the melatonin metabolite 6-sulfatoxymelatonin and the stress hormone cortisol in urine. 6-sulfatoxymelatonin is a reliable proxy for melatonin secretion, and in healthy individuals, concentrations rise in the evening and peak during the night.

Cortisol concentrations are normally low during the night but rise in the early morning and peak at the time of awakening.


Primary outcome: The primary outcome measure was the cumulative incidence of ICU-acquired delirium.

Secondary outcomes: Secondary outcome measures were:

  • duration of ICU-acquired delirium (measured as the number of delirium-free and coma-free days in 28 days),
  • duration of mechanical ventilation,
  • length of stay in the ICU and in hospital overall, and
  • mortality in the ICU and during the overall hospital stay.


Delirium occurred in 137 (38%) of 361 patients in the DLA group and 123 (33%) of 373 in the control group (odds ratio 1⋅24, 95% CI 0⋅92–1⋅68, p=0·16 ).

The median numbers of delirium-free and coma-free days in 28 days were similar in the two groups. No association between the cumulative illuminance and the cumulative incidence of delirium.

When patients with early delirium were excluded, the median time to development of delirium was 4·4 days (IQR 2–6) in the DLA group and 4·0 days (2–7) in the control group (p=0·84). Among all patients with delirium, the median number of delirium-free and coma free days in 28 days was 20 (IQR 8–24) in the DLA group and 17 (7–25) in the control group (p=0·96).

Patients with delirium were:

  • significantly older,
  • more severely ill, and
  • more had history of cognitive disturbances, alcohol abuse, and smoking than patients without

Patients with delirium had significantly longer stays in the ICU and hospital than those without, but mortality did not differ.

The percentage of patients who received mechanical ventilation, the duration of mechanical ventilation, use and types of sedatives, and use and cumulative amount of administered haloperidol did not differ significantly between the DLA and control groups.

Absence of sedation had no effect on delirium incidence (DLA group 17 [16%] of 105 vs control group 11 [10%] of 110, odds ratio 1⋅74, 95% CI 0⋅77–3⋅91, p=0·13) or the median number of delirium-free and coma-free days in 28 days (p=0·98).

Authors’ Conclusions

DLA delivered greater light exposure during patients’ ICU stays than did normal lighting, but the additional exposure did not reduce the cumulative incidence of ICU-acquired delirium. Furthermore, no differences were found for any of the secondary endpoints.

Some reasons why this study finds differently to those studies done in other environments.

First, in contrast to non-critically ill patients, most patients in the study were sedated and had their eyes closed during the acute disease phase. Since light exerts its effect on the biological clock by modulation of retinal input through photosensitive ganglion cells, closed eyes could have prevented some biological effects.

Second, sedatives can disturb the normal circadian rhythm, which might have counterbalanced the effects of lighting therapy. Again, therefore, effects of lighting therapy might only have been expected after the acute disease phase was over and the patient was awake. For patients who received no sedatives during their ICU stay there were also no differences in outcomes between study groups.

Is there a direct relationship between improved sleep and prevention of delirium? This is not known to be clear. It could be that disturbed circadian rhythms and delirium might be caused by brain inflammation which can occur in sepsis, trauma or MI for example. So ICU patients might be less susceptible than non-ICU patients to external cues such as daylight exposure.

Concentrations of melatonin or its urinary metabolite 6-sulfatoxymelatonin vary widely within and between ICU patients, and are affected to notable degrees by mechanical ventilation, adrenergic drugs, and sepsis.


High adherence rate

Adequately powered

Regular measurements and fixed lighting schedule.


Single centre study.

Blinding not possible

Did not/could not measure sleep- difficult to measure.

CAM-ICU single determination sensitivity is low.



Video laryngoscopy versus direct laryngoscopy for adult patients requiring tracheal intubation.

Lewis et al. Cochrane Database of Systematic Reviews, 2016.


 Clinical question.

In patients requiring oral tracheal intubation, does the use of video laryngoscopy compared to traditional direct laryngoscopy techniques, reduce complications and failed intubation rates?


  • Systematic review of adult oral intubation literature.
  • Search of Medline, Embase, gov databases.
  • PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) use for reporting the search process.
  • Two reviewers searched titles and abstracts, and a third person was used as arbiter where there was disagreement on exclusion/inclusion.
  • Cochrane risk of bias tool applied to each selected paper. Assessment of heterogeneity and publication bias also made.
  • Only randomised controlled trials with human subjects and no mannequins selected.
  • Study period between 1970 and February 2015.
  • Primary outcome included failed intubation, and episodes of hypoxia.
  • 8 secondary endpoints were included.
  • 64 studies included in the final analysis including 7044 patients. 3 of 64 studies were pre-hospital. 17 studies were of a cross over design, comparing views with different scopes followed by a randomised intubation with a given scope. Airtrach studies that did not apply a camera to device were excluded.



  • Intubation in elective and emergency surgery, emergency department and critical care patients.



  • Adults over 16 years.



  • Video laryngoscopes (VL) of either classic or hyper-angulated type versus traditional direct laryngoscopy (DL).



Primay outcomes.

  • Failed intubation. 8 studies combined, VL vs DL statistically significant difference in favour of VL. OR 0.35, CI 0.19 to 0.65; 4127 participants. Moderate quality evidence.


  • 3 studies, VL vs DL no statistically significant difference. OR 0.39, 95% CI 0.10 to 1.44; 1319 participants. Low quality evidence.

Secondary outcomes.

  • First pass success. 35 studies, VL vs DL no statistically significant differences between groups OR 27, CI 0.77 to 2.09 4731 participants. Moderate quality of evidence.


  • 2 studies. VL vs DL, statistically significant difference in favour of DL. OR 1.09, CI 0.65 to 1.82; 663 participants. Very low quality of evidence.


  • Airway trauma. 22 studies, VL vs DL, statistically significant difference in favour of VL. OR 0.68, CI 0.48 to 0.96; 3110 participants.


  • Time to intubation. Unable to combine due to heterogeneity.


  • Type of VL. The C-MAC (most similar geometry to a Macintosh blade) had the lowest intubation failure rate of all VL devices.


  • Unable to combine data.


  • Failed intubation in predicted difficult airway. 9 studies. VL vs DL, statistically significant difference in favour of VL. OR, 0.28, CI 0.15 to 0.55; 830 participants. This held true for simulated difficult airway, OR, 0.18, CI 0.04 to 0.77; 810 participants.


 Author’s conclusion.

In patients with a predicted or difficult airway, VL may decrease the rate of failed intubation compared to DL.



  • Significant effort to identify the relevant literature.
  • Only included human studies.
  • Clearly points out bias and heterogeneity of evidence.



  • Very heterogeneous studies.
  • Does not allow for clear delineation of hyper-angulated vs traditional geometry VL devices which may have significant performance differences.


Bottom line.

Base on data available up to February 2015, VL decreases failed intubation where the airway is predicted or found to be difficult.



Efficacy of High-Flow Nasal Cannula Therapy in Intensive Care Units: A Meta-Analysis of Physiological and Clinical Outcomes



Liesching et al

Clinical Question

A meta-analysis to compare the physiological and clinical outcomes of high-flow nasal cannula (HFNC) with standard oxygen (O2) or conventional noninvasive ventilation (NIV) in intensive care units (ICUs).


Meta analysis

Search terms

  1.    “oxygen inhalation therapy”
  2.    “oxygen/administration and dosage”
  3.    search #1 OR #2
  4.    “high flow”  OR “high-flow”
  5.    search #3 AND #4
  6.     “high flow nasal cannula” OR “high flow nasal cannulae”
  8.    search #5 OR #6 OR #7
  9.    filter #8:  Humans

Excluded reviews/paediatrics/non ICU/non comparative/retrospective studies.

Identified 18 prospective studies.

Eleven articles are trials and seven are prospective comparative studies.


In summary, the diagnoses include:

1253 (63.8%) cardiac or cardiothoracic surgeries,

359 (18%) pneumonia,

30 (1.5%) ARDS,

18 (0.9%) sepsis, 17 (0.8%) cardiogenic pulmonary edema, and 327 (16%) others.


Significant differences

Respiratory rate

The pooled t test shows a trend of lower RR in the HFNC group than the standard O2 group (21.6 vs 24.7,P=.059). Excluding CCU patients resulted in a more significant improvement (22.0 vs 25.6,P=.039).

Heart rate

The pooled t test shows a significantly lower HR in the HFNC group than the standard O2 group (89.1 vs 98.4,P=.033). Excluding CCU patients resulted in a more significant improvement (88.4 vs100.0,P=.013).

Arterial  Blood  Gas PaO2

The pooled t test shows a significantly higher PaO2 in the HFNC group than the standard O2 group (104.5 vs 90.0 mm Hg,P=.044). Excluding CCU patients resulted in a more significant improvement (109.4 vs 91.4 mm Hg,P=.015)

Dyspnea  Score

The pooled t test shows a significantly lower dyspnea score in the HFNC and standard O2 groups(2.7 vs 4.3,P=.046). Excluding CCU patients also resulted in a significantly lower dyspnea score (2.8 vs 4.6,P=.045).

Discomfort score

The sample size–adjusted pooled t test shows no significant difference in discomfort score when using HFNC versus standard O2 (1.19 vs 1.44, P 1⁄4 .435). Excluding CCU patients resulted in a significantly lower discomfort score in the HFNC group (0.98 vs1.96, P 1⁄4 .028).

Intubation  or  Reintubation  Rate

The OR calculated from total sample size shows no different intubation rate in the HFNC and standard O2 groups(OR=0.79, 95%CI: 0.39-1.21,P=.269). Excluding CCU patients resulted in a significantly lower intubation rate (OR=0.59, 95%CI: 0.37-0.97,P=.036).

In the table at the bottom of Figure 11, no difference in intubation rates was observed when comparing HFNC with NIV (OR=0.83, 95%CI: 0.62-1.11,P=.216). However, excluding CCU resulted in a significantly lower intubation rate (OR=0.58, 95%CI: 0.35-0.95,P=.032).

No significant differences

Arterial  Blood  Gas:  Oxygen  Saturation

Arterial  Blood  Gas:  PaO2/FIO2

Arterial  Blood  Gas:  PaCO2

Arterial  Blood  Gas:  pH


Pulmonary function

Intensive care unit mortality

ICU length of stay

Hospital length of stay


Authors’ Comments and Conclusions

A high degree of heterogeneity across all the studies.

64% of the sample size is contributed by post cardiac or cardiothoracic surgery.

When excluding CCU patients, a majority of patients were diagnosed with pneumonia. For this subgroup of patients, the RR, HR, PaO2, dyspnea, and discomfort score were significantly improved with HFNC comparing to standard O2.

For patients with pneumonia, the HFNC group also showed better clinical outcomes including a significantly lower intubation rate, a trend of lower ICU mortality, and a shorter ICU length of stay, comparing to standard O2. Comparing with NIV, only when excluding CCU patients, the limited data showed better clinical outcomes including lower intubation rate and ICU mortality in the HFNC group.


Interview Questions for Advanced Critical Care Practitioners

The Content on the website is provided for FREE as is the podcast.

You could help support this work by going to Amazon via this link. This means that I will earn a small commission from any purchases you make with NO extra cost to yourself.

Thank you.

Amazon Link

[table id=1 /]




share this

Related Posts

CCP Podcast: AI with Aarti Sarwal

Why do we use vasopressin?

CCP Podcast: So you want to be an ACCP?