Trailblazers in critical care: ten studies that changed the way we practice

Critical care medicine is a relatively new area of specialization. Several landmark studies have been published over the years that have transformed the way we practice and added a new dimension to our approach to patient care. I have summarized the groundbreaking literature that has captivated our attention over the years and continues to influence the way we practice. I have added tables with salient findings of each study; the primary outcomes are marked in bold.

The ARDS-net low tidal volume (ARMA) study


If you walked into an ICU in the 1990s, you would see Servo 900 series ventilators, the ever-reliable workhorses, blowing tidal volumes ranging between 600–800 ml at 12–15 breaths/min in many patients. The all-encompassing goal of mechanical ventilation those days was to drive up arterial oxygenation and normalize carbon dioxide levels.  High tidal volumes of 10–15 ml/kg was commonly employed to offer recruitment of collapsed alveoli and improve gas exchange.1  The first groundbreaking randomized controlled trial came from Sao Paulo, Brazil.  Amato et al. studied 53 patients, randomized to receive tidal volumes of 6 ml vs. 12 ml/kg body weight. They observed a spectacular reduction in the 28-day mortality with a low tidal volume ventilation strategy (38% vs. 71%)2 and paved the way for the historic ARDS-net trial.3

The study

In this multicentric randomized controlled trial (RCT), patients with acute lung injury and acute respiratory distress syndrome (ARDS), (based on pre-Berlin definitions) were enrolled. The study was conducted across 10 university hospitals in the USA for a 3-year period between 1996-99. Patients were ventilated with a tidal volume of 12 ml/kg predicted body weight at a plateau pressure of 50 cm H2O or less, and compared with 6 ml/kg predicted body weight at a plateau pressure of 30 cm H2O or less. The trial was stopped early at the fourth interim analysis after enrolment of 861 patients. 

Main findings

The study revealed that ventilation with low tidal volumes of 6 ml/kg reduced mortality and the duration of mechanical ventilation.  

Outcome6 ml/kg12 ml/kgP value 
Death before discharge home with unassisted breathing 31%39.8%0.007
Unassisted breathing at 28 days 65.7%55%< 0.001
Ventilator-free days at 28 days12±1110±110.007
The ARMA trial

How did the study change practice? 

The study was hailed as a landmark in the field of critical care practice. However, criticisms were levelled against the authors for comparing extremes of tidal volumes compared to a more conventional standard of care. Although slow to take up, the findings of the ARMA study led to a paradigm shift in the ventilation strategy of ARDS patients.

The TRICC trial 


There are some clinicians who are still ardent believers of maintaining a hemoglobin level of 10 gm/dl in most patients. In fact, three decades ago, “top up” transfusions were fairly common in patients who clearly seemed to suffer no adverse effects due to a low hemoglobin level. However, the Canadian Critical Care Trials group challenged this dogma in their epoch-making trial.

The study 

The TRICC trial enrolled 838 critically ill patients with hemoglobin levels <9.0 gm/dl within 72 hour of ICU admission. Eighty-two percent of patients were on mechanical ventilation. Of these, 418 patients received a restrictive transfusion strategy; red cells were transfused if the hemoglobin levels dropped below 7 g/dl, with maintenance of levels between 7–9 g/dl. In the liberal group, red cells were transfused if the hemoglobin levels dropped below 10 g/dl, aiming for a target hemoglobin of 10–12 g/dl. The overall 30-day mortality (the primary outcome) was not significantly different; however, the hospital mortality was lower with a restrictive strategy. The 30-day mortality was also significantly lower in less severely ill patients (APACHE II score <20) and in the younger age group (<55 years old). 

Outcome Restrictive Liberal P value
30-day mortality 18.7%23.3%0.11
Hospital mortality 22.2%28.1%0.05
APACHE II <208.7%16.1%0.03
<55 years old5.7%13.0%0.02
Cardiac disease 20.5%22.9%0.69
The TRICC trial

How did the study change practice?

The TRICC trial was a well-conducted study that came up with important new findings that would change transfusion practice among critically ill patients. However, a large number of patients were excluded as the treating clinicians considered transfusion necessary. The study was underpowered as the final sample size did not reach the planned target. However, it clearly demonstrated that a transfusion trigger of 7 gm/dl with a range 7–9g/dl would have no adverse impacts in a heterogenous group of critically ill patients. The authors advised caution with a low transfusion threshold in critically ill patients with myocardial ischemia. 

The SAFE study


Albumin, among the first proteins to be identified and isolated from plasma, has been in widespread clinical use since the 1940s. A major shift in the clinical use of albumin occurred following a 1998 Cochrane meta-analysis that included 1419 patients from 30 RCTs. This meta-analysis revealed increased mortality rates with the use of albumin in critically ill patients.4 Subsequently, the Australian and New Zealand Intensive Care Society (ANZICS) Clinical Trials Group conducted the SAFE trial to compare normal saline with 4% albumin as resuscitative fluid in critically ill patients.5

The study 

The SAFE study included a heterogenous group of 6997 critically ill patients. They were randomized to receive 4 percent albumin or normal saline as resuscitation fluid for 28 days. There were 3497 patients in the albumin group and 3500 in the saline group. There was no difference between groups in the 28-day mortality, the primary outcome of the study. 

Outcome 4% albumin Normal saline P value 
28-day mortality 20.9%21.1%0.87
New organ failure47.3%46.7%0.85
ICU days 6.5±6.66.2±6.20.44
Hospital days15.3±9.615.6±9.60.30
Ventilation days 4.5±6.14.3±5.70.74
Renal replacement therapy0.5±2.30.4±2.00.41
The SAFE study

On subgroup analysis a higher relative risk of death was noted with 4% albumin in trauma patients; this was due to a higher mortality among patients with traumatic brain injury. The use of 4% albumin suggested benefit in patients with severe sepsis. There was no difference in the relative risk of death in patients with ARDS. The study concluded that the use of 4% albumin as resuscitation fluid led to similar clinical outcomes compared to normal saline. The use of 4% albumin may be detrimental in patients with trauma, particularly with traumatic brain injury.

How did the study change practice?

The SAFE study contradicted the findings of the 1998 Cochrane meta-analysis and showed that albumin resuscitation may be equally safe compared to normal saline. However, the study established that normal saline may be equally effective as a resuscitation fluid. The ratio of the volume albumin to that of saline administered during the initial four days of treatment was 1:1.4. This finding challenged the conventional dogma that the volume of crystalloids administered must be 2–3 times higher for equivalent efficacy. 

The NICE-SUGAR study


Sugar is sweet, but high levels may not be nice in ICU patients. Tight glucose control within a narrow range of 80–110 mg/dl was widely recommended after the original study by Van den Berghe et al. from Leuven, Belgium. This study revealed a significantly reduced ICU mortality from 8 to 4.6% among surgical patients with maintenance of tight glucose control. 

The study

The NICE-SUGAR study was conducted across 42 ICUs in Australia, New Zealand, and North America.6 In the intervention arm, the blood glucose levels were maintained between 81–108 mg/dl, while a higher level of up to 180 mg/dl was allowed in the control arm. A total of 6104 patients underwent randomization, with 3054 in the intensive glucose control arm and 3050 in the conventional arm. The study revealed a significantly lower 90-day mortality with a more conventional blood glucose target of 180 mg/dl among critically ill patients.

OutcomeIntensive control ConventionalP value 
90-day mortality 27.5%24.9%0.02
28-day mortality 22.3%20.8%0.17
Hypoglycemia(Blood sugar 40mg/dl or less)6.8%0.5%<0.001
The NICE-SUGAR study

There was no significant difference between groups in the number ventilation, ICU, or hospital days, the onset of new organ dysfunction or the requirement for renal replacement therapy.

How did the study change practice?

The NICE-SUGAR study clearly demonstrated that aiming for “tight” control of blood glucose levels may result in significant hypoglycemia and worsen clinical outcomes. It quelled the initial enthusiasm towards maintaining lower blood glucose levels in critically ill patients.     

The PROSEVA study


Ventilation in the prone position has been in use for more than four decades.7 Placing the patient prone had been previously known to improve oxygenation with possible alleviation of ventilator-induced lung injury.6 However, there was no strong evidence that it improved clinical outcomes until the PROSEVA trial. 

The study

The PROSEVA trial included patients from 27 ICUs in France and Spain. All patients had severe ARDS with a PaO2/FiOratio of less than 150, while on an FiOof 0.6 or more, and a PEEP of 5 cm of H2O or higher. In the intervention arm, patients remained in the prone position continuously for at least 16 hours; in the control arm, they remained semi-recumbent in the supine position. A total of 237 patients underwent prone positioning, while 229 were maintained supine, semi-recumbent. 

28-day mortality 16%32.8%<0.001
90-day mortality 23.6%41%<0.001
Successful extubation at 90 days 80.5%65%<0.001
The PROSEVA study

The study found a significant reduction in the 28- and 90-day mortality with prone ventilation. The number of ventilation-free days was significantly more in patients who underwent prone positioning. There was no difference in the ICU length of stay, need for non-invasive ventilation, or tracheostomy between groups.  

How did the study change practice?

After several years and many RCTs that failed to show benefit, the PROSEVA study established improved survival among patients with severe ARDS, with a PaO2/FiOratio of less than 150. In today’s world, most centers would prone ventilate severely hypoxic patients before considering more advanced modalities of care, including extracorporeal support. 

Targeted temperature management (TTM) post cardiac arrest


Physicians through the ages have been fascinated by the impact of induced hypothermia in attenuating neurological injury secondary to hypoxia. In 1958, the first human study on induced hypothermia after cardiac arrest suggested improved survival with reducing body temperature to 31–32 °C.8 Two RCTs had previously investigated the effect of lowering body temperature to 32–34°C in post-cardiac arrest setting.9,10 These patients who suffered cardiac arrest with an initial shockable rhythm and remained unconscious, revealed improved neurological function and survival with induced hypothermia. However, the question whether clinical benefit was due to the effect hypothermia or prevention of fever remained unanswered. 

The study 

It is against this background that Nielsen et al. conducted a multicentric, international RCT. Unconscious survivors of out-of-hospital cardiac arrest were assigned to targeted temperature management at 36 vs. 33°C. A total of 939 patients were included, 473 to the 33°C and 466 to the 36°C arm.11 The all-cause mortality until the end of the study, with a mean follow-up period of 256 days, was similar between groups.

 Outcome33°C36°CP value 
Death at the end of the study period 50%48%0.51
Cerebral Performance Category score (180 days): Severe cerebral disability, coma or vegetative state, or brain death 54%52%0.78
Modified Rankin score (180 days):       Moderately severe disability, severe disability, or death52%52%0.87
Death at 180 days48%47%0.92
The TTM study

How did the study change practice?

It was remarkable that basic life support measures were established within a median duration of 1 min in the TTM study; it remains unclear whether in the real-world setting of more delayed resuscitation, a lower target temperature level would be of benefit. 

However, the TTM study strongly suggested that maintenance of normothermia in the post cardiac arrest period may be equally effective in optimizing clinical outcomes as reducing the temperature to lower levels. 

The ANZICS trial and the death of dopamine 


Dopamine was extensively used in ICUs several years ago. It was considered to be a good choice as an initial inotrope in cardiogenic and septic shock and presumed to augment renal function through a unique, dose-dependent mechanism. Besides, peripheral intravenous administration was considered safer, and hence, it was often used in low acuity settings. The ANZICS Clinical Trials Group conducted a randomized controlled trial that pretty much laid to rest most of the clinical benefits attributed to dopamine.12

The study

In a multicentric placebo controlled RCT conducted across ICUs in Australia, 324 patients were included. The inclusion criteria were 2 or more features of the systemic inflammatory syndrome (SIRS) over a 24 h period with early renal dysfunction based on urine output and creatinine levels. The investigators found no difference in peak creatinine levels between groups through the study period.

OutcomeDopaminePlaceboP value 
Peak serum creatinine (mg/dl)2.77 (1.63)2.82 (1.66)0.93
Renal replacement therapy21.7%24.5%Not significant 
ICU stay 13 (14)14 (15)0.67
Hospital stay 29 (27)33 (39)0.29
The dopamine trial

The number of survivors to ICU discharge (108 vs 105; p=0·61) and hospital discharge (92 

vs 97 patients; p=0·66) was not significantly different between groups.

How did the study change practice?

This study shelved the use of dopamine as a “renal protective” strategy in critically ill patients. Except in a few cardiac units, it is unusual to see dopamine being methodically titrated to protect the kidneys. 

The CRASH-2 trial


Hemorrhage is a leading cause of death following trauma. A profound, pathological, fibrinolytic response may follow major trauma and surgery, and exacerbate bleeding. A meta-analysis with the use of tranexamic acid, an antifibrinolytic agent, in patients undergoing elective surgery had revealed a significantly reduced requirement for blood transfusion.13Against this background, the CRASH-2 collaborators set out to investigate the effect of tranexamic acid following major trauma.14

The study

The multicentric, placebo controlled RCT was carried out in 274 centers across 40 countries. Patients with significant hemorrhage and hypotension or those who were deemed to be at a high risk of significant hemorrhage were included, within 8 hours of injury. In the intervention arm, tranexamic acid was administered in a dose of 1 g intravenously over 10 min, following by 1 g over 8 hours; a matching placebo was administered in the control arm. In the final analysis, 10060 patients were included in the tranexamic acid arm and 10067 patients in the control arm. The use of tranexamic acid was associated with a significant reduction in hospital mortality at 4 weeks; death due to bleeding was also significantly less with tranexamic acid therapy. 

 Tranexamic acid ControlP value 
Hospital mortality within 4 weeks 14.5%16%0.0035
Death due to bleeding 4.9%5.7%0.0077
The CRASH-2 trial

How did the study change practice?

The absolute reduction in mortality was just 1.5%; besides, the study also found that tranexamic acid use was not associated with a reduction in the transfusion of blood products. The absolute risk reduction in mortality with the use of tranexamic acid in trauma patients was very small. Besides, it did not reduce the amount of blood products administered. A subsequent exploratory analysis revealed that tranexamic acid should be administered within the first 3 hours to bleeding trauma patients. Later use is less effective and possibly harmful.15

The ADRENAL trial

The use of corticosteroids in septic shock has been the subject of enduring enthusiasm among clinicians for several decades. From the mega doses used in the 1980s, to the modest dosing from the 1990s, several studies have addressed the efficacy of corticosteroids in septic shock, with contrasting results. Venkatesh et al. asked the question whether hydrocortisone reduces mortality in patients with septic shock.16

The study

This RCT was conducted across five countries in 69 medical-surgical ICUs. Adult patients with a documented or suspected focus of infection with two or more criteria of the systemic inflammatory response syndrome (SIRS) and requiring pharmacological support for maintenance of blood pressure for at least 4 hours were included in the study. Hydrocortisone was administered as a continuous infusion of 200 mg/day. The control group received a matching placebo. The final analysis included 1853 patients who received hydrocortisone and 1860 patients who received placebo. 

The 90-day mortality, the primary endpoint, was not different between groups. There was a statistically significant, reduced time to resolution of shock by one day, reduced time to cessation of mechanical ventilation and a lower requirement for blood transfusion.

OutcomeHydrocortisonePlaceboP value 
90-day mortality 27.9%28.8%0.50
28-day mortality 22.3%24.3%0.13
Time to shock resolution (days)3 (2–5)4 (2–9<0.001
Time to ICU discharge (days)10 (5–30)12 (6–42)<0.001
Days alive and out of ICU (days)58.2±34.856.0±35.40.047
Time to cessation of ventilation (days)6 (3 to 18)7 (3 to 24)<0.001
Blood transfusion 37%41.7%0.004
The ADRENAL study

There was no significant difference in the incidence of recurrent shock, days alive and out of hospital, re-initiation of mechanical ventilation, use of renal replacement therapy, days alive and free of renal replacement therapy, and new-onset bacteremia or fungemia.

How did the study change practice?

The ADRENAL study was by far the largest study on the use of steroids in septic shock. It adds to the body of knowledge corticosteroids may reduce the time to resolution of shock, but may not improve survival. 

The CHEST trial


The long-standing controversy over the use of synthetic colloids versus crystalloid for intravascular volume resuscitation was convincingly resolved by the ANZICS investigators in late 2012.17 They conducted a multicentric randomized controlled trial across 32 hospitals in Australia and New Zealand to answer the question whether hydroxyethyl starch or normal saline may be better a resuscitation fluid among ICU patients. 

The study 

A total of 7000 patients were randomized to receive 6% HES (130/0.4, Voluven) or normal saline as resuscitation fluid until 90 days after randomization, ICU discharge or death. The final analysis included 3315 patients in the 6% HES group and 3336 patients in the normal saline group. 

 HES 6%Normal Saline P value 
90-day mortality 18%17%0.26
Renal replacement therapy 7%5.8%0.04
Adverse events 4.6%3.3%0.006
The CHEST trial

Patients in the risk (R) and injury (I) categories of the RIFLE score were higher in the normal saline group; however, there were more patients with failure (F) in the HES 6% group. There was no difference between groups in the 28-day mortality, death in ICU, or in hospital. The duration of mechanical ventilation, renal replacement therapy, stay in ICU and hospital were not different between groups.

How did the study change practice?

The era of synthetic colloid use came to an end after the publication of The CHEST and the 6S trial. The latter study showed an 8% absolute increase in mortality at 90 days and a 6% absolute increase in the requirement for renal replacement therapy with the use of HES,18 thus corroborating the findings of the CHEST study. 


1.     Wheeler AP, Bernard GR. Acute lung injury and the acute respiratory distress syndrome: a clinical review. Lancet Lond Engl. 2007;369(9572):1553-1564. doi:10.1016/S0140-6736(07)60604-7

2.     Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med. 1998;338(6):347-354. doi:10.1056/NEJM199802053380602

3.     Ventilation with Lower Tidal Volumes as Compared with Traditional Tidal Volumes for Acute Lung Injury and the Acute Respiratory Distress Syndrome. N Engl J Med. Published online 2000:8.

4.     Reviewers CIGA. Human albumin administration in critically ill patients: systematic review of randomised controlled trials. BMJ. 1998;317(7153):235-240.

5.     A Comparison of Albumin and Saline for Fluid Resuscitation in the Intensive Care Unit. N Engl J Med. 2004;350(22):2247-2256. doi:10.1056/NEJMoa040232

6.     Abroug F, Ouanes-Besbes L, Elatrous S, Brochard L. The effect of prone positioning in acute respiratory distress syndrome or acute lung injury: a meta-analysis. Areas of uncertainty and recommendations for research. Intensive Care Med. 2008;34(6):1002-1011. doi:10.1007/s00134-008-1062-3

7.     Piehl MA, Brown RS. Use of extreme position changes in acute respiratory failure. Crit Care Med. 1976;4(1):13-14. doi:10.1097/00003246-197601000-00003

8.     Benson DW, Williams GR, Spencer FC, Yates AJ. The use of hypothermia after cardiac arrest. Anesth Analg. 1959;38:423-428.

9.     Bernard SA, Gray TW, Buist MD, et al. Treatment of Comatose Survivors of Out-of-Hospital Cardiac Arrest with Induced Hypothermia. N Engl J Med. 2002;346(8):557-563. doi:10.1056/NEJMoa003289

10.   Mild Therapeutic Hypothermia to Improve the Neurologic Outcome after Cardiac Arrest. N Engl J Med. 2002;346(8):549-556. doi:10.1056/NEJMoa012689

11.   Nielsen N, Wetterslev J, Cronberg T, et al. Targeted Temperature Management at 33°C versus 36°C after Cardiac Arrest. N Engl J Med. 2013;369(23):2197-2206. doi:10.1056/NEJMoa1310519

12.   Low-dose dopamine in patients with early renal dysfunction: a placebo-controlled randomised trial. The Lancet. 2000;356(9248):2139-2143. doi:10.1016/S0140-6736(00)03495-4

13.   Henry DA, Carless PA, Moxey AJ, et al. Anti-fibrinolytic use for minimising perioperative allogeneic blood transfusion. Cochrane Database Syst Rev. 2011;(1):CD001886. doi:10.1002/14651858.CD001886.pub3

14.   CRASH-2 trial collaborators, Shakur H, Roberts I, et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet Lond Engl. 2010;376(9734):23-32. doi:10.1016/S0140-6736(10)60835-5

15.   The importance of early treatment with tranexamic acid in bleeding trauma patients: an exploratory analysis of the CRASH-2 randomised controlled trial. 2011;377:8.

16.   Venkatesh B, Finfer S, Cohen J, et al. Adjunctive Glucocorticoid Therapy in Patients with Septic Shock. N Engl J Med. 2018;378(9):797-808. doi:10.1056/NEJMoa1705835

17.   Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl Starch or Saline for Fluid Resuscitation in Intensive Care. N Engl J Med. 2012;367(20):1901-1911. doi:10.1056/NEJMoa1209759

18.   Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl Starch 130/0.42 versus Ringer’s Acetate in Severe Sepsis. N Engl J Med. 2012;367(2):124-134. doi:10.1056/NEJMoa1204242

3 thoughts on “Trailblazers in critical care: ten studies that changed the way we practice”

Leave a Reply