html xmlns=”http://www.w3.org/1999/xhtml” xml:lang=”en”>

Resuscitation of the Injured Patient: The Enemy of Good Is Better

Ronald V. Maier, MD

Many will ask whether there is anything new in the field of resuscitation of hemorrhagic shock. Even though there is much that we know, we do not always seem to practice it. This chapter presents a bit of my philosophy of resuscitation in the field and is based both on work that I have been part of and on contributions from many groups of investigators. I hope the chapter provides some insights that you will be able to use as we move the field forward.

At the turn of the last century, the father of American surgery, Dr. William Halsted, introduced a radical change by championing the German system of surgical education. He integrated dedicated faculty and a commitment to education as key tenets of the curriculum. The other major component of Halsted’s legacy was his incorporation of the scientific approach into surgical practice. According to Ira Rutkow’s superb American Surgery history text, “so-called teaching hospitals… consisted mainly of operating room work but no integration of the fundamental sciences.”¹

Halsted, in his attack on the previous system, advocated a scientific tone as the major corollary to his educational efforts. He created the “new” American surgery, based as much on pathology, physiology, and, now, molecular biology and genomics as it had been on anatomy in the past. He took us from the operating theater where every case was a tour de force, frequently never seen before or experienced again, to an approach based on a rational scientific method, thus joining our practice in the operating room to the new science of surgery.

Rounding out the philosophy is an old quote from my mentor and good friend C. James Carrico, past president-elect of the American College of Surgeons, president of the American Association for the Surgery of Trauma, and a phenomenal leader. As a mentor, he used to say to me, “Maier, remember that the enemy of good is better.”

No adage can be taken as an absolute. But the core concept of Carrico’s statement is that we need to use a rational approach, and this chapter illustrates the way that I adopted this idea. Here I review some of our efforts to translate science into treatment in the resuscitation of the patient, while asking, “Does our practice reflect the enemy of good being better?”

Resuscitation of Hemorrhagic Shock

In considering pertinent examples of surgical progress over the last 100 years, we encounter this axiom: “Control the hemorrhage and resuscitate with intravenous fluids.” This is primarily attributed to Walter B. Cannon, describing care of the injured during the first World War in Europe and detailed as part of a 6-article series in JAMA in 1918.² It’s a simple axiom dealing with resuscitation of the injured patient, but we are still asking how we should control the hemorrhage and resuscitate the patient; we still don’t know what to use, when, or in what sequence.

In 1918, when an injury consisted of a single bullet wound with a very focused area of bleeding and the resuscitation fluids weren’t very effective, it made sense to stop the bleeding first so you didn’t have to give much resuscitation fluid. Of course, I would argue that we’re still there: We still don’t have great fluids to give and we should probably still take the same approach, now called “damage control resuscitation,” once again illustrating that little is new.

Overresuscitation (“Supranormal Resuscitation”)

A prime example of “the enemy of good is better” is the intentional overresuscitation of patients in shock. Many of the younger physicians practicing today have never seen the constellation of symptoms that are the consequence of this practice: gross anasarca, acute respiratory distress syndrome (ARDS), and abdominal compartment syndrome. But this postresuscitation presentation, this iatrogenic disease, was common in our ICUs, very common. When the common treatment of trauma was essentially to drown the patient in salt water, we all created these diseases to a variable degree.

How did this practice come about? In the 1960s, Dr. Shoemaker in Los Angeles and others observed that hypovolemic shock distorts homeostasis and that occult hypovolemia and hypoperfusion may be present despite good urine output and adequate blood pressure³. Despite resuscitation, the patient may have ongoing hypoperfusion, leading to hypovolemic shock, which distorts normal physiological responses. Occult hypoperfusion persists, leading to multiple organ failure, coagulopathy, and an “oxygen debt” that must be repaid. Therefore, we should resuscitate to “supranormal” oxygen delivery (>600 mL/min) by maximally driving cardiac output.⁴

Note that this link between shock, multiple organ failure and coagulopathy was recognized over fifty years ago. Shoemaker postulated that hypoperfusion was responsible for the hypocoagulable state following trauma and that physicians need to correct that to prevent further disruptions in homeostasis. That was a good idea, but then he made the good idea “better” by including in this paradigm the concept of “oxygen debt.” I have never understood how a living organism calculates an oxygen debt and how it stores that debt so that it can be repaid. But that was the crux of the proposal: that we need to overresuscitate very aggressively, drive oxygen delivery, and eliminate that debt as soon as possible.

So we intentionally overresuscitated trauma patients, and it took us 10 to 15 years to figure out that this was the greatest iatrogenic disease the world had seen in many years. In 1994, more than 30 years after Shoemaker’s initial postulate, Hayes et al⁵ published a randomized controlled trial in which matched patients in the ICU were assigned to treatment with resuscitation targeted to higher oxygen delivery versus control resuscitation. Amazingly, we were killing 50% of the patients with this great new idea. In the supranormal resuscitation group, intra-abdominal hypertension doubled, and abdominal compartment syndrome, multiple-system organ failure, and mortality all increased 2 to 3 times.

Although careers have been made describing the treatment and management of abdominal compartment syndrome, like most diseases it is better avoided in the first place. We’ve done that, and now we rarely see significant intra-abdominal hypertension syndrome in our ICUs. The enemy of good is better and we proved it very well.

Hypotensive Resuscitation

In the same year that Hayes et al⁵ published their study, Dr. Mattox’s group from Houston reported on their randomized, controlled trial of immediate versus delayed resuscitation following penetrating torso trauma.⁶ They observed a statistically significant survival advantage in the group that received limited initial volume resuscitation, and they raised the question of whether the best treatment was to not give the patients any intravenous fluids—another example of “the enemy of good is better”?

The Houston trial has been extensively analyzed; it was very difficult to conduct, and enormous credit should be given to the investigators for accomplishing the study. Although the study definitely showed an improvement in this patient population, it only enrolled patients with penetrating torso trauma, similar to Dr. Cannon’s study in 1918. In interpreting the results, we must consider some important limitations (Table 11-1). When the investigators looked at the group of patients who received prehospital fluid, the only patients who suffered a worse outcome were those who had tamponade at the time of infusion. So even within the penetrating trauma group, only a small subset of patients were harmed by prehospital fluid—those who already had tamponade present.

Despite this, many trauma surgeons extrapolated the findings to every injured patient with the premise, “If it’s good for the patient shot in the chest, it must be good for the patient with pelvic fractures from blunt trauma.” And that’s not at all what the study tested or what the authors proposed.

A new term, hypotensive resuscitation, evolved from these studies. I do not like this term, as it implies that we want to make patients hypotensive so that they’ll do better. To me, the semantics imply a lot when these words are extrapolated to situations other than penetrating torso trauma, to patients with different mechanisms of injury, different injury patterns and different physiologic requirements.

Looking at the data from Houston, one could have instead come up with a conclusion similar to that known for 30 years in treatment of leaking aortic aneurysm: for acute vascular disruption, hypotension should not be aggressively corrected with intravenous fluids until surgical control has been accomplished.

If a patient has a hole in the aorta, the best thing to do is give a limited volume of fluid to maintain critical organ perfusion, but don’t give large volumes, because blowing the clot off before surgical control will certainly worsen the hemorrhage. And that is what Mattox et al⁶ showed: The patient with a bleeding major artery secondary to a gunshot wound does better with very little prehospital fluid.

We can further illustrate how a single, focused study should not be extrapolated broadly. Part of the equation that this study did not address (and, again, it was not designed to) is the impact on associated traumatic brain injury. This is an important circumstance in which no resuscitation may, in fact, be a very bad idea. An epidemiological study including all trauma patients treated in Denver in 1992 evaluated the cause and timing of death in the 289 injury-related fatalities.⁷ Even though hemorrhage represented a significant portion of early mortality, brain injury was almost as common, even early on. However, after 48 hours, people died primarily of their brain injuries and multiple organ failure.

We know that the injured brain does not like hypoxia and that enhancing secondary brain injury results in long-term consequences. In a seminal study, Chesnut et al⁸ demonstrated that a single episode of hypotension (systolic blood pressure <90 mm Hg) doubled mortality and increased morbidity. Patients whose initial hypotension was not corrected by the time of arrival to the emergency department had worse outcomes than those whose hypotension was corrected. I would argue that when we fail to resuscitate someone with a blood pressure of 50, that patient remains hypovolemic, hypotensive, and hypoxic and the brain is going to suffer. You cannot generalize the Mattox approach to an overall trauma population that frequently sustains blunt injury and traumatic brain injury.

Further confirmation of this is provided by data presented by Josh Brown and colleagues⁹ at the American Association for the Surgery of Trauma meeting in 2012. Using the Glue Grant epidemiological database, the researchers evaluated patients who received less than or greater than 2 L of crystalloid resuscitation during the prehospital portion of their care. In the group of patients without prehospital hypotension, there was a negative impact on survival. Again, the enemy of good is better—if we have a patient with a normal blood pressure and a brain injury, he doesn’t need volume, and giving him 2 L of Ringer’s decreases his chance of survival.

However, in the other population evaluated by Brown et al, patients who actually had prehospital hypotension, the curve reverses. Those who received less than 500 mL of crystalloid before admission to the hospital experienced a negative impact on their survival, whereas those who received 2 L of Ringer’s lactate, as originally proposed by my mentors Shires and Carrico, showed the best survival; that is, prehospital hypotension requires volume, not drowning. Again, from the Glue Grant data, giving limited volume (approximately 2 L) to the correct patient, the trauma patient with prehospital hypotension, is beneficial.

This should be our approach regarding crystalloid resuscitation of hemorrhagic shock, and of course everyone will say that’s what we should be doing, but somehow we are still arguing about it.

Hemostatic Resuscitation (1:1:1 Ratio)

The new data coming out of the military experience in Iraq and Afghanistan have contributed enormously to our understanding of traumatic hypovolemia. However, fortunately, there are very few of these patients in civilian practice. We can assume that a 21-year-old male who has had both legs blown off and has sustained major groin and abdominal injuries from a bomb is going to need massive blood transfusion as part of his initial resuscitation.

There’s no doubt that stopping bleeding is good for the injured patient. In both military¹⁰ and civilian¹¹ practice, there is a strong correlation between lowering the patient’s blood requirement and decreasing his mortality significantly. Techniques for prompt hemorrhage control include direct pressure, judicious use of tourniquets, staged damage control surgery, and angioembolization in interventional radiology.¹²

Trauma-induced coagulopathy is a significant problem very early after injury. Brohi et al¹³ in the United Kingdom, and others, have shown that there is a significant chance, up to 25%, that a coagulopathy is present at the time the patient presents to the emergency department. For each level of Injury Severity Score, the presence of coagulopathy as measured by an increased international normalized ratio is significantly associated with increasing mortality.¹⁴ All of this argues that since the coagulopathy happens early and exerts a significant negative impact, if we’re going to treat it we need to treat it early.

Military experience, led by Dr. John Holcomb and others in the Middle East conflicts, furthered the concept of damage control or hemostatic resuscitation to directly address the early coagulopathy of trauma.¹⁵ In patients receiving massive transfusions at a combat support hospital, the ratio of blood products (ratio of plasma to red blood cells [RBCs]) affected overall mortality.¹⁶ As the ratio decreases from 1:8 (very little plasma) to near 1:1, mortality decreases significantly. Similar findings have been shown in civilian practice. A single-center, retrospective review from New Orleans observed that hemostatic resuscitation with an increased ratio of plasma to RBCs decreased mortality significantly in the cohort receiving a massive transfusion (>10 U of packed red blood cells [PRBCs]), with a suggestion of benefit in the cohort that received fewer than 10 U of PRBCs.^17,18 A multicenter, retrospective review of 466 massively transfused civilian trauma patients from 16 level I trauma centers demonstrated a reduction in 30-day mortality from almost 60% to 40% in patients receiving a plasma to RBC ratio or platelet to RBC ratio of 1:2 or greater, primarily in patients with truncal hemorrhage. The benefit was largely seen within the first 6 hours of treatment. Hospital-free, ICU-free, and ventilator-free days were all significantly improved in the groups receiving a ratio less than 1:2.¹⁹

In the Glue Grant database, we also looked at patients with either less than 1:1.5 or greater than 1:1.5 plasma to RBC ratios and we came up with very similar findings, with less blood being transfused, a significant decrease in mortality, and a decreased incidence of ARDS in the group receiving a greater than 1:1.5 ratio, all supporting the concept of hemostatic resuscitation.²⁰ We then used the Glue Grant to see how we were translating knowledge to clinical practice over the 8 years of the grant. We were a little dismayed, because over the 8-year period, the overall ratio of plasma to PRBCs had barely changed. However, we did find that although the ratio didn’t change when measured at 24 and 48 hours after injury, at the 6-hour time point there was a significant increase in the ratio of fresh frozen plasma (FFP) to RBCs, supporting the need for early intervention.

One of the unresolved issues is that this change in ratio represents an amount of plasma that should do nothing for coagulopathy in the average patient. It is such a small difference in volume and component products that it should have no effect. And yet our study found that even though the Injury Severity Score increased over the 8 years of the study, the number or percentage of patients who developed massive transfusion decreased dramatically.

The results could be interpreted as meaning that a small increase in the amount of FFP given in the first 3 to 6 hours post injury has a dramatic effect that decreases bleeding and cuts massive transfusion rates in half. Or these results may simply be a marker of improvements in the system of care and hospitals that achieve those goals, in the same way that the simultaneous military development of better tourniquets and wider application of tourniquets reduced massive transfusion requirements. As with all historical control studies, during the study period a lot of improvements were happening simultaneously. This is always a problem with using historical databases. But if the improvements are attributable to the plasma, such a small volume should not have much impact on subsequent incidence of massive transfusion.

A study published in Critical Care Medicine in 2012 looked at all the published studies that dealt with the impact of plasma to RBC ratio in massive transfusion.²¹ The authors identified 11 observational studies, with no randomized controlled trials:

• 3 studies found a survival benefit to a 1:1 ratio versus higher or lower ratios.
  
• 6 studies were not 1:1 but concluded that higher ratios were associated with improved survival.
  
• 2 studies showed no advantage to a 1:1 ratio.
  
• 2 studies demonstrated increased multisystem organ failure with a 1:1 ratio.

The authors noted that there was insufficient evidence to support use of a 1:1 ratio, and they identified survivor bias as a significant confounder of the results. In short, the current studies are observational and they are a scattergram. You can pull data saying that a higher ratio helps, it hurts, or it does nothing, just like most retrospective studies. What is critical is that there are no randomized controlled trials. It is hoped that the PROPPR (Pragmatic, Randomized Optimal Platelet and Plasma Ratios study http://www.clinicaltrials.gov/ct2/show/NCT01545232), which is presently recruiting patients, will answer this, but to date no other studies exist. Thus, there is insufficient evidence to statistically prove superiority of a 1:1:1 ratio.

One of the major problems with all of these studies is the possibility of a survivor bias; if you are alive you receive more plasma, so it looks like the plasma helped you, but that’s just because you are alive longer, possibly for some other reason than plasma transfusion. Recently, Dr. Sperry and his group addressed the question of survivor bias.²² Using the Glue Grant database, the investigators divided the patients into groups of those receiving a low plasma to RBC ratio versus those receiving a high ratio within the first 6 hours after injury. The investigators then examined the subsequent course and found that benefit of hemostatic resuscitation was established by the 6-hour time point; those who had an initial high ratio had a better survival, whereas those who had a low ratio had a worse survival. This addresses to a large degree the argument of bias from survival and indicates that early achievement of high ratio is associated with improved outcomes.

So enter the current discussions. If early blood and plasma are good, then “better” would be to provide them on the way to the hospital. There have been proposals to give blood before the patient arrives at hospital, to give FFP early, to infuse plasma in the ambulance, and others.

Let’s examine these proposals and see whether we are again tempting the Enemy of Good. Once again, older data appear to be lost. In the epidemiological study from Denver,²³ the highest independent risk factor for developing severe postinjury complications was receiving a significant blood transfusion, with a relative risk of posttraumatic multiple organ failure of 7.9-fold in the subgroup that received more than 6 U of PRBCs in 24 hours. What these data don’t answer is whether the patients were injured more, received more blood, and therefore developed more multiple organ failures or whether the blood itself contributed to the postinjury complications. This question is still debated.

The CRIT Study in ICU patients²⁴ helped to change care in all of our ICUs. Patients who receive transfusions have a much higher mortality. And if you match the patients, you find that their mortality is higher not because they are sicker but because they received a transfusion. A study using cardiac surgery data had excellent matching because of a longstanding tradition encouraging a high transfusion threshold, exclusive of physiological changes.²⁵ In this study, the dose response for mortality from 0 to 5 U of PRBCs transfused is a step-dose response. It’s like giving increasing boluses of a toxin to your patients. Blood, if it were a product under review by the US Food and Drug Administration, would never be approved.

As an additional genomic example, transfusion down-regulates the molecular pathway for insulin signaling. This means that glucose-mediated insulin signaling for intracellular glucose uptake is blocked by the transfusion. A poisoning of the cells occurs; they can’t get energy through glucose, they can’t make adenosine triphosphate, and they release lactate through ongoing anaerobic metabolism. This explains in part the delay in lactate clearance in these patients.

Indiscriminately giving blood to your patients puts them at higher risk for development of ARDS and other complications and increases their mortality measurably. And yet we are talking about administering blood in ambulances. Personnel in an ambulance aren’t able to determine who needs blood and who does not. So there is the potential that many patients who don’t need blood will receive it, and they will be subjected to toxicity on a curve looking like a postoperatively transfused cardiac patient, potentially dying unnecessarily.

The other part of this proposal involves plasma. In Dr. Holcomb’s early experience in Iraq, although he showed a decrease in mortality as the plasma to RBC ratio increased, it appeared that in the patients who survived, those who received more plasma and survived had more complications.

In a recent publication by Inaba et al²⁶ from the Los Angeles County system, matched patients who received FFP and minimal transfusion were compared with patients who received FFP and massive transfusion as well as with control patients who received neither. Patients who did not require a massive transfusion who then received 1 to 3 U, 4 to 6 U, or more than 7 U of plasma had a doubling of complications compared with matched patients who did not receive plasma. In addition, there was a 10-fold increase in the incidence of ARDS. Again, these data would block an FDA drug application: For every increase in the dose of drug, twice as many patients die or have major complications. Aggressive use of plasma appears to be an independent contributor.

Again, I raise the concern regarding human nature and jumping on bandwagons. In the desire to do better for our patients, we are very eager to move forward with a charge when instead we should be cautious. If we give a lot of blood and plasma in the prehospital setting, we will cause a lot of unnecessary comorbidities.

Inflammation, Genomic Storm

A “genomic storm” is induced by severe blunt trauma.²⁷ A patient with cancer has perhaps 10, 15, or 20 genes that are abnormal. When a patient is hit by a cement truck, 75% of her genes change dramatically in their activity. It is a true genomic storm, which is why it has been very hard to measure 1 or 2 proteins as biomarkers to predict how this patient is going to respond. It is too simplistic an approach for such a massive genomic response.

In the Glue Grant, genomic screening was performed in injured and control patients over a period of 28 days following injury. Normally, inflammatory genes are suppressed, but they become activated following injury and then slowly revert toward normal. The portion of the genome that primarily represents the adaptive immune response, after the car crash, is suppressed and then slowly returns to an up-regulated state as the patient recovers. Following injury, (1) the proinflammatory innate immune response is overexpressed, (2) adaptive immunity (necessary to fight ventilator-associated pneumonia and other nosocomial infections) is suppressed, and (3) overall the predominant changes result in suppression, not activation, and this lasts for up to 28 days.

In a recent study we identified 63 probe sets differentially expressed between patients with an uncomplicated outcome and those with a complicated outcome.²⁸ The patients with uncomplicated courses sustain genomic changes following injury but return quickly to baseline. But patients who have complications have a much more prolonged change in genes, which stay altered much longer. The patient who is able to return to homeostasis, to baseline, quickest avoids complications. In future studies, we need to track this response online and make certain that our interventions are actually returning the patient to homeostasis.

If the patient is not able to return to homeostasis, he develops what Fred Moore has described as the persistent inflammation, immunosuppression, and catabolism syndrome (PICS).²⁹ What Moore has proposed is exactly what the genomic pattern predicts: persistent inflammation, persistent immunosuppression, persistent blockage of glucose utilization, catabolism of muscles for amino acids as an energy source, and breakdown of lean muscle. Burn ICUs have recognized this for years, and we recognize it more and more in our trauma ICUs. This disease entity is increasing complications, preventing patients from recovering, and prolonging ICU resource consumption. The answer is to restore homeostasis.

Scoring Systems for Prediction

There are many scores for predicting outcome in hemorrhagic shock, all attempting to identify the patient who potentially will benefit the most from intervention (Table 11-2). All of these scores are very reasonable and are useful statistically; the receiver operating characteristic curves are 80% to 85%. Nevertheless, a lot of the components are hard to assess rapidly, you can’t obtain them before the patient reaches the hospital, and even once you have the data, the dysregulation syndromes may already have progressed.³⁰ Importantly, when we apply these scores to a patient, not a population but rather an individual, they all fail, and a significant number of patients who do not develop complications are still included in the high-risk category.

Related posts:

Media, Technology, and Education in the ICU Water, and Brain–What Should the Intensivil Know? Paradigms in Sepsis Initiatives in Emergency Preparedness Therapy in the ICU: Practical Management Consideration on Sepsis Resuscitation

Stay updated, free articles. Join our Telegram channel

Join