Elevation of systemic oxygen delivery in the treatment of critically ill patients, by Michelle Hayes et al. 
Since we are living the era of “less is more” in critical care over the last years I couldn’t find more appropriated article to review. Why? First: because looking to our past can enlighten our future; second: in my opinion, this article is a milestone of the end of “more is more”; and third: since I was a young Padawan at the time, it’s kinda of fun to see how critical patients were managed at the time. Somehow it reminds me a drunken party, a beer bong and everybody yelling: More! More! More! Of course I did it in the past, the beer bong, not the massive infusions of dobutamine, but now I don’t think it’s safe, the dobutamine, not the beer bong!
During the 70’s and 80’s we started to see lot of articles discussing physiological patterns among surviving vs nonsurviving critically ill patients, classical works by Shoemaker   showed us that these hemodynamic/physiological variables were markedly different among those two groups. At the time was thought that organ damage was due to inadequate oxygen delivery. Things like cardiac index (CI), oxygen delivery (DO2) and oxygen consumption (VO2) were higher among surviving patients, therefore why not aim those miracle survival targets in a randomized controlled trial and see the mortality ratio goes down?
They chose a heterogenous group of critically ill patients: high risk surgical patients, septic shock, sepsis syndrome and acute respiratory failure. The defining criteria for those groups of patients were, at best, slightly different than today. For two consecutive years they screened all admitted patients to the intensive care units at two hospitals. If the screen was “positive”, central venous, arterial, urinary and pulmonary artery catheters on the house! There were three hemodynamic targets to be accomplished:
– CI > 4.5 liters per minute per square meter of body-surface area
– DO2 > 600 milliliters per minute per square meter
– VO2 > 170 milliliters per minute per square meter
Before starting the protocol all patients received dopamine at dose of 2 micrograms per Kg per minute and fluid resuscitation with something that reminds me of a cake (blood, albumin or synthetic colloids). Hemoglobin should be kept above 10g/dl and oxygen saturation above 90%. If the patient met the target criteria he wasn’t randomized. From a total of 109 patients, 9 were not randomized and the others were equally divided among intervention and control group.
In order to achieve those targets the treatment group would receive a dobutamine dose from 2 to 200 micrograms per Kg per minute (yeah, 200! T-W-O H-U-N-D-R-E-A-D). The control group would receive dobutamine only if the CI <2.8 liters per minute per square meter of body-surface area. The mean arterial pressure target for both groups was 80 mmHg. The treatment was initiated at ICU admission and continued until death or if the patients got better. The data collection protocol was pretty good and I cannot help to think about the huge load of work nurses and doctors experienced during the intervention period, since almost all decision were made based on hemodynamic parameters. The statistical plan was straight forward with interim analysis planned after each block of 50 patients. The objective was to demonstrate a 15% mortality reduction from a expected mortality of 33%. With 80% power and type I error of 5% they would need 130 patients.
The real deal
The randomization did its job, both groups were homogenous among them but with a heterogenous profile of patients within. Like an ordinary ICU, older patients, some of them really sick…
Needless to say that in time zero all patients were equally treated, but although they might share the syndromic characteristics, in fact, they were pretty different. The non randomized group, which after some fluids got better were younger and less ill and all survived, but not without some catheters floating on their hearts.
During the intervention period they did what they supposed to do! CI and DO2 were significantly higher in the intervention group, but the VO2 was not. Mean arterial pressure and lactate levels also were not different. Here I summarize the patient’s vasoactive drug and hemodynamic profiles:
All patients in intervention group received dobutamine and only 21 in the control group. Despite all hemodynamic support 35 of 50 intervention patients weren’t able to achive all targets simultaneously. And in 24 patients dose increment of dobutamine was limited by complications.
After the second interim analysis the study was stopped. The in hospital mortality was 34% in control group versus 54% in intervention group (the predicted mortality was 34% for both), p=0.04. Some data about subgroup analysis were published, and although statistical useless and with only 100 patients the data is fun to be seen. The septic shock subgroup had in-hospital mortality of 52% in control versus 71% in intervention. The ARDS subgroup, wait a minute, 67% in control versus 81% in intervention! The cause of death among all patients were: Intractable hypotension ( 4 in each group), cardiac event ( 2 control and 4 intervention) and multiple organ failure (9 control and 17 intervention).
We can all say that the intervention strategy were pretty aggressive, in fact, even the control strategy seems aggressive to me. Of course, is easy to say that with all the studies we have now guiding us for hemoglobin, mean arterial pressure, the use of pulmonary artery catheter and so on… The hyperdynamic response to an insult, from trauma to infection, is not protective per se, it seems it is a sign that our patient has a less severe disease. We all love to treat a 20y/o guy with hyperdynamic septic shock but we are all a bit scared the we see that same patient as cold as ice, with marble skin, even if he have good MAP or CI. Maybe all that data of 70’s and 80’s were saying to us that patients with a hyperdynamic response do better than patients without, in fact, that is what I believe. And trying to make a hypodynamic patient sound like a hyperdynamic one only make things worse.
The intervention group had more cardiac deaths, probably related to dopamine and also more multiple organ failure deaths, maybe due to microcirculatory changes caused by higher doses of dobutamine. An interesting topic to be discussed in the future.
Every new generation of intensivists must read this paper, discuss it, and see how it helped to change our concepts and how the apogee of an era started to fall. The authors conclude in the abstract that “aggressive efforts to increase oxygen consumption may have been detrimental”, which is an euphemism. The “may have been” is too much.
1. Hayes MA, Timmins AC, Yau EH, Palazzo M, Hinds CJ, Watson D. Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med. 1994;330(24):1717-1722.
2. Shoemaker WC, Montgomery ES, Kaplan E, Elwyn DH. Physiologic patterns in surviving and nonsurviving shock patients. Use of sequential cardiorespiratory variables in defining criteria for therapeutic goals and early warning of death. Arch Surg. 1973;106(5):630-636.
3. Bland RD, Shoemaker WC, Abraham E, Cobo JC. Hemodynamic and oxygen transport patterns in surviving and nonsurviving postoperative patients. Crit Care Med. 1985;13(2):85-90.