Searching for the Lethal Corner By Gary Grist RN CCP Emeritus
In November of 1990, I was searching the medical library for new ideas pertaining to the perfusion profession. Having been a perfusionist since 1968, I knew all the common pertinent facts relating to cardiopulmonary physiology and extracorporeal support. But I was still troubled. In my early years, I quickly noticed that despite providing oxygenation and good perfusion using my cardiopulmonary bypass (CPB) pump, there were still some disturbing complications after CPB in a few patients. I wondered why patients had “pump head” or ARDS after surgery. Why did some patients die of organ failure while most patients did well? The universal fall back explanation was that inflammatory response to the pump caused many of the complications. The patient’s own physiologic weaknesses contributed as well. Improvements in circuits, priming and technique had ameliorated many of those complications, but had not removed them entirely. I grudgingly accepted the inflammatory explanation for many years until ECMO came along.
The children’s hospital I worked for started its ECMO program in 1986. Most ECMO patients survived the extended exposure (many days or weeks) to an extracorporeal circuit. So, why didn’t the inflammatory response injure or kill them? It did not make sense. The standard explanation was that most patients could tolerate and even acclimate to the inflammatory mediators if exposed to circuits for an extended time. The 25% who died probably could not tolerate it for unknown reasons. To me that explanation sounded less like a cause and more like a rationalization.
Another thing that did not make sense to me was that some ECMO patients who seemingly were not as sick as others died from an unexpected complication. These patients often gradually deteriorated under conservative treatment until it was determined that ECMO was needed. ECMO was then implemented in a controlled and relaxed environment. In the neonatal population I was working with, the average age before the initiation of ECMO among non-survivors was 4.4 +/- 5.8 days (n = 31). By contrast, other neonatal ECMO patients who seemed sicker and required more rapid ECMO implementation did well and recovered without severe complication. Among these neonates the average age of survivors before the initiation of ECMO was only 2.4 +/- 3.6 days (n = 163); a full two days earlier (1). Did the two day delay while conservative treatments were attempted play a role in the demise of these patients?
A further mystery was that most of the patients who died had a lethal complication not related directly to their pulmonary or cardiac disease. Brain bleeds and brain infarctions were very common as was organ failure or simply failure to improve. The usual explanation was that their shock and/or hypoxia before ECMO had caused the damage. But that did not always make sense because many patients did not experience severe hypoxia or shock before ECMO. They were often placed on ECMO simply because the high ventilator settings or the need for high doses of support medications were hazardous. Failing the ventilator and medications, others, by contrast, who were placed on ECMO emergently due to severe shock or hypoxia subsequently did well and recovered.
The physicians that I worked with felt that each patient was different and that there was no way to predict the outcome with any certainty. I did not believe that. Human beings are complicated biologic machines, but machines nonetheless. And problems with machines are usually predictable. That is why preventative maintenance is performed; we know what is going to break before it breaks and why it is going to break.
Despite all the knowledge about cardiopulmonary physiology, I felt that doctors and perfusionists had missed something very important, some, as yet, unknown concept or principle. So as I wandered through the library’s stacks that November evening I selected a few books and journals at random, not expecting to find any new answers. I opened one journal to an article by Popel that discussed how oxygen was distributed at the capillary level (2). There were terms that were completely new to me: the oxygen pressure field, the Krogh cylinder, perfused capillary density, radial and axial oxygen vectors, and the lethal corner. “What is this stuff?” I asked myself. So I followed the references and was introduced to August Krogh, Ferdinand Kreuzer, Dietrich Lubbers, Niels Lund and Paul Schumacker (3-8).
These articles presented a new concept to me; the oxygen pressure field theory (OPFT). Oxygenation is usually evaluated with single number depending on their source. For example, arterial or venous pO2 or hemoglobin saturation values such as paO2 = 100 mmHg, pvO2 = 35 mmHg, SaO2 = 98%, SvO2 = 70%. But it is the one number that is used to evaluate the patient’s status. OPFT is different. Using the Krogh cylinder as a mathematical model, OPFT explains how oxygen is distributed outside of the vascular system within tissues in a wide field of partial pressures with values ranging from being equal to the arterial pO2 to being as low as zero. Theoretically the area containing zero oxygen can vary in size from inconsequential to large enough to cause tissue death (the lethal corner). This lethal corner can develop despite normal cardiopulmonary function and oxygen delivery. Perhaps everyone else already knew about this concept and they just forgot to tell me! So I went to experienced perfusionists and doctors that I worked with and asked them. They had never heard of OPFT either.
At that time our ECMO program was about four years old and had done about 120 patients with about 75% survival, acceptable results at the time. But I was still nagged by the deaths of the 30 others. As I explained earlier, there was no rational explanation for the seemingly random deaths. Could OPFT solve the mystery? I was in the perfect situation to answer that question. I was the chief perfusionist in an active pediatric ECMO program. Each child on ECMO was closely followed with every blood and physiologic test pertinent to critical care during extracorporeal support. And these tests were repeated at regular intervals to monitor for progress and safety. All I had to do was to closely observe these tests, keeping in mind that I was trying to discern the oxygen pressure field (if it existed) and detect the lethal corner (a portent of death) if it developed. But where was I to start? Could these labs be interpreted using a new perspective? New perspectives are difficult to achieve unless someone else shows you how.
I explained all this to the surgeon I was working with then and asked him for guidance. He looked as if I had stunned him with a ball peen hammer blow to the forehead. He said, “Don’t ask me! Go ask a REAL doctor!” So I went to ask Stanley Hellerstein MD. Stan was a nephrologist who, in the absence of a hematologist at the hospital, had been caring for all the sickle cell (SC) children. He was the smartest doctor I ever knew. I came to know him years earlier when he had asked me to help him with exchange transfusions on SC children who had strokes due to their illness.
I asked Stan what number or parameter I could look at to quantify how sick a cardiopulmonary (CP) patient was, i.e. detect the lethal corner. He told me, in his experience, if the CP patient could maintain a normal anion gap (AG), they would most likely survive. So he thought I should study the patient’s AG before and during ECMO.
As I reviewed the labs on the first 120 ECMO patients, I found that patients with normal AGs during ECMO had the lowest mortality and patients with high AGs had the highest mortality. Of all the expired patients, half of them had high AGs while half still had normal AGs. So an elevated AG had good positive predictive value for death. But a normal AG had poor negative predictive value for survival. Nonetheless, this was an important first step in discerning the oxygen pressure field and searching for the lethal corner.
I had concluded that the elevated AG was a reflection of the conversion of a portion of the tissues (those in the lethal corner) to anaerobic oxygenation resulting in the production of organic acids like lactic acid. I would later find that for each milliequivalent per liter (mEq/L) that the AG was above the normal limit, mortality would increase by an additional 10%. For example, patients with an average AG of 10 mEq/L during ECMO had a10% mortality. Patients with an AG of 15 mEq/L had about 50% mortality. And the unfortunate patients with an average AG of 20 mEq/L had 100% mortality. In essence, the magnitude of the AG told me how big the lethal corner was. I would later refer to this as an ‘anoxic lethal corner’ for reasons described below.
Despite finding this lethal corner using the AG, there was something important missing. I still had dead patients with normal AG values. My focus had been on studying the effects of anaerobic metabolism due to the lack of oxygen at the cellular level. But after being placed on ECMO, patients had all the oxygen they needed. Unexpectedly, the second piece of the puzzle had nothing to do with oxygenation.
I came across a 1991 article by Johnson and Weil describing how the CO2 gradient between the arterial and venous blood gases could be used to predict mortality in cardiac patients (9). I was taught to use venous blood gases early on in my career. I had always insisted on drawing both arterial and venous gases on ECMO patients and my CPB patients. But I had not considered that an elevated pvCO2 was an indication of intercellular CO2 retention which caused a detrimental pH change in the cells, stopping their normal metabolic processes despite adequate oxygenation.
In reviewing those expired ECMO patients with normal AG values, I found that they had elevated pvCO2 values. This was the other puzzle piece I needed! I noticed that for every 1 mmHg that the CO2 gradient was above normal, the mortality increased by 10%. For example, patients with an average CO2 gradient of 7 mmHg had about 10% mortality. Those with a gradient of 12 mmHg had about 50% mortality. And those with a 17 mmHg gradient had 100% mortality. This, then, was another type of lethal corner, not caused by oxygen deprivation, but by intracellular CO2 accumulation; a hypercapnic lethal corner. (See figure below).
Since the AG scale for mortality and the CO2 gradient scale for mortality were about the same (10% increase for each unit increase above normal), I determined that the AG and CO2 gradient scores could be added together as a ‘viability index’ score (11). I found this to be true in older patients as well, not just neonates. After reviewing 294 ECMO patients, those with an average viability index score of 17 (AG of 10 + CO2 gradient of 7) had 5% mortality. Those with a viability index score of 27 (AG of 15 + CO2 gradient of 12) had about 50% mortality. Those with an index score of 37 ((AG of 20 + CO2 gradient of 17) had virtually 100% mortality. The one thing that seemed to violate this rule was if a patient had a lethal anatomy such as certain congenital diaphragmatic hernia or cardiac patients. In that case, maintaining a low score on ECMO could not save the patient.
I concluded that when patients with a lethal corner were placed on ECMO and suddenly reperfused, they suffered a reperfusion injury that led to the most common lethal complications. I believe that when the viability index is high the enzymes that protect the cell from reperfusion injury are deactivated by the change in intracellular pH. If reperfusion occurs before the intracellular pH can be restored to normal the cell is susceptible to reperfusion damage. Some of my data was published in 2009 and 2010, 19 years after discovering August Krogh’s Oxygen Pressure Field Theory (1,10).
So, should hope be abandoned for ECMO patients with a lethal corner? No! Rather, we should consider a ‘reperfusion strategy’ for the most vulnerable patients before and during ECMO that counter-acts the ravages of reperfusion injury. How that is done is a discussion for another time.
This article does not have the scope to list all the many of the details about what I found in my search for the lethal corner. Those can be seen on my web site; <perfusiontheory.com>.