EBQ:Mortality in Multicenter Critical Care Trials: An Analysis of Interventions with a Significant Effect
PubMed Full text PDF
Clinical Question
What interventions have been shown in multicenter RCT's to affect mortality in adult critically ill patients?
Conclusion
Of all the interventions done in critical care, the authors of this study identified 15 total treatments that increased or decreased mortality as documented in a randomized, controlled trail.
The following interventions decreased mortality:
- Noninvasive Ventilation
- Mild hypothermia after cardiac arrest (32-34C)
- Prone Positioning in ARDS
- Low tidal volume ventilation in ARDS
- TXA in patients with or at high risk of traumatic hemorrhagic shock
- Daily interruption of sedatives in critically ill patients
- Albumin in cirrhotic patients with SBP
The following interventions increased mortality:
- Diaspirin cross linked hemoglobin in traumatic hemorrhagic shock
- Starch in septic shock
- Ventilation with high frequency oscillation
- IV Salbutamol in ARDS
- Glutamine supplementation
- Growth hormone treatment
- Supernormal systemic oxygen delivery
- Intensive insulin therapy (target blood glucose of 81-180)
Major Points
Critically ill patients are often complex, and require numerous advanced treatments and therapies. With these patients having a higher mortality rate vs standard patients, and a lopsided share of healthcare resources, it is important that interventions used be backed by scientific research supporting their use and benefits. Over the years, many novel treatments have been employed in the ICU with hopes of improving survival. Many therapies have gone on to become treatment mainstays, and others went on to later be contradicted by larger or higher quality studies. This study done in 2015 looks at the current body of evidence to determine the mortality effects of major interventions in critically ill patients. Overall 15 interventions that affected mortality were identified in high quality, multi center RCT's; 7 of these treatments decreased mortality, and 8 showed an increase in mortality.
Study Design
- Systematic search & review of the literature involving multi center RCT's of interventions influencing mortality in critically ill patients.
- Search via MEDLINE, Pubmed Scopus, Embase
- Consensus conference of experts participated in an in person meeting to evaluate the methodological robustness of all interventions studies to determine if individual papers should be included in the review
Article Selection
The authors of this review searched MEDLINE/PubMed, Scopus and Embase for trials in peer reviewed journals with the following inclusion and exclusion criteria:
Inclusion Criteria
In order to be included in the review, articles need to see the following 4 criteria:
- Publication in a peer reviewed journal
- mRCT design (m = any trial involving more than one hospital)
- Dealing with nonsurgical interventions in adult critically ill patients
- Statistical significant reduction or increase in unadjusted landmark mortality
[For the purpose of this review, critically ill patients referred to those patients with acute failure of at least one organ and/or need for intensive/emergent treatment, regardless of where they were treated (ICU, ED, etc.)]
Exclusion Criteria
All studies were excluded that fulfilled any of the following:
- Quasi randomized or non randomized methods
- Dealt with surgical interventions
- Involved pediatric population
- Dealt with only the perioperative period
- Were performed out of the hospital
- Showed mortality effect only in a subgroup or after adjusted analysis
- Had low (<50%) agreement levels among surveyed clinicians
Results
A total of 15 interventions were identified from the selected 24 total studies that were included in this paper.
Interventions That Decreased Mortality
Treatment | Absolute Mortality Reduction | Relative Mortality Reduction | NNT |
Albumin | 0.191 | 0.668 | 5 |
Sedation Vacation | 0.134 | 0.232 | 7 |
Mild Hypothermia | 0.142 | 0.258 | 7 |
NIV | 0.193 | 0.675 | 5 |
NIV | 0.2 | 0.714 | 5 |
NIV | 0.101 | 0.598 | 10 |
NIV | 0.214 | 0.548 | 5 |
NIV | 0.142 | 0.871 | 8 |
NIV | 0.12 | 0.828 | 7 |
NIV | 0.197 | 0.64 | 5 |
NIV | 0.122 | 0.836 | 8 |
Prone Position | 0.168 | 0.512 | 6 |
Protective Vent | 0.329 | 0.465 | 3 |
Protective Vent | 0.088 | 0.222 | 11 |
Protective Vent | 0.238 | 0.441 | 4 |
TXA | 0.015 | 0.094 | 68 |
Interventions That Increased Mortality
Treatment | Absolute Mortality Increase | Relative Mortality Increase | NNH |
Supernormal 02 | 0.2 | 0.667 | 5 |
Diaspirin | 0.221 | 0.902 | 5 |
Growth Hormone | 0.221 | 1.163 | 5 |
Tight Glucose Control | 0.026 | 0.104 | 38 |
IV Salbutamol | 0.109 | 0.468 | 9 |
Starch | 0.075 | 0.174 | 13 |
Oscillatory Ventilation | 0.117 | 0.332 | 9 |
Glutamine | 0.052 | 0.191 | 19 |
Characteristics of Trials Selected
- Overall only 7 trials (29%) were blinded
- Non-blinded trials were more likely to show a mortality benefit
- mRCT's showing an increased mortality on average involved more centers and more patients
- Most studies (21/24) investigated medium term mortality (in hospital to 28 day survival)
- Median ARR for interventions that decreased mortality was 0.12
- Median NNT was 7 for those trials that showed a mortality benefit
- NNH was 9 for those trials showing a mortality harm
- Funding was declared in 21 studies - public sources in most cases (67%)
Criticisms & Further Discussion
- Generally these trials were small in size (median patient count below 200, and median center count below 10)
- NIV significant support was highly dependent upon its effect in COPD patients (6/8 trials)
- Need for larger ICU trials
See Also
Funding
Funded in part by departmental funds from the Department of Anesthesia & Intensive Care, IRCCSS San Raffaele Scientific Institute (Milan, Italy)