Improving Outcomes In The Critically Ill Patient

At an industry-sponsored symposium held in conjunction with CHEST 2002, the annual meeting of the American College of Chest Physicians, four leaders in pulmonary medicine, surgery, gastroenterology and nutrition presented the latest information on the prevalence, impact and treatment of sepsis and Adult Respiratory Distress Syndrome (ARDS). Topics included emerging theories and investigational treatments, fluid resuscitation in patients with pulmonary failure, preventing the adverse effects of
gastrointestinal hypoperfusion and nutritional support in patients with pulmonary failure.

This program was supported by an unrestricted educational grant from Wyeth Pharmaceuticals.

Sepsis Syndrome: Emerging Theories and Investigational Treatments

Systematic inflammatory response syndrome (SIRS) is defined as system-wide microvascular injury with tissue edema and inflammation. Sepsis is a confirmed or suspected infection plus at least two of the SIRS criteria. Severe sepsis is diagnosed with the addition of at least one organ dysfunction. Septic shock occurs when there is fluid-refractory hypotension.

“One of the problems with these definitions is that they are descriptive,” said Neil R. MacIntyre, MD, FCCP, Duke University Medical Center, Durham, NC. “It ignores the heterogeneity of the mechanisms and the multiple mediators involved. This may cause us to miss important subgroup benefits to therapies in large clinical trials.”

It is important to remember that even perfect antibiotic use does not guarantee survival. A study by Pittet and others showed that appropriate antibiotic therapy lowers sepsis mortality up to 15%. However, mortality remained substantial (Pittet D, et al. Am J Respir Crit Care Med. 1996;153:684).

“While it is clearly important to treat the infection, we need to go beyond that and support the patient until their own defense mechanisms kick in,” said Dr. MacIntyre.

The ARDS Network Trial showed that reducing tidal volume (TV) to the 6 mL/kg range improved the proportion of patients alive and off the ventilator at day-28 when compared with a TV of 12mL/kg. They found a significant increase in organ failure-free days and reduced circulating inflammatory mediators in the reduced volume cohort (ARDS Network. NEJM. 2000;342: 1301).

There have been a number of studies on the ideal hemoglobin level. The Canadian Clinical Trials Group showed that in general ICU populations, hemoglobin in the 7-8 range had outcomes comparable to those transfused at higher triggers. There are some ongoing studies indicating weekly doses of erythropoietin might boost hemoglobin in long-term ICU populations. Whether that translates into clinically important outcomes remains unanswered (Hebert PC. NEJM. 1999;340: 409).

Using inotropes remains controversial. A study of trauma patients by Shoemaker and others looked at outcomes associated with pushing oxygen delivery to high levels using dobutamine. In young trauma patients there seemed to be a benefit (Shoemaker WC, et al. Chest. 1988;94:1176). However, larger trials of medical intensive care unit patients showed worse outcomes in the elderly (Hayes MA, et al. NEJM.1994; 330:1717).

A study by Rivers and others looked at patients presenting with florid sepsis who were randomized to a protocol calling for aggressive fluid resuscitation versus control. Those in the protocol arm used more dobutamine and three-quarters of the patients also received transfusions. Mortality was significantly better in the protocol group (Rivers E, at al. NEJM. 2001;345:1368).

The results suggested to Dr. MacIntyre “these modalities may have utility in the septic patient.”

High frequency oscillatory ventilation may help reduce lung injury. One study of 143 patients with ARDS showed a drop in mortality from 53% to 38%. Although the small number resulted in a P value of only .06, the results were promising enough that the Food and Drug Administration released the device to general use (Derdak S, et al. Am J Resp Crit Care Med.2002;166:801).

Recent studies using instillation of surfactant and important surfactant-related proteins has shown promising results (Walmrath D, et al. Am J Resp Crit Care Med. 1996;154:57). There are currently two additional trials underway that may show a benefit.

“There is an emerging concept of relative adrenal insufficiency in sepsis,” said Dr. MacIntyre. “Many organs get hit in sepsis so why shouldn’t the adrenal gland?”

Work by Annane and others show that physiologic doses of adrenal corticosteroids improved mortality (Annane D, et al. JAMA. 2000;283:1038). Dr. MacIntyre suggested a cortisone stimulation test. If the patient responds, consider stopping steroids. If not, steroids should probably be continued.

Another exciting area addresses the issue of inflammation and coagulation being linked to each other in sepsis. The monocyte stimulated by the endotoxin creates the inflammatory response that is key to the sepsis syndrome. The same cell also activates the coagulation system to form small fibrin clots. Stratgies to control this coagulopathy have shown promise (Bernard GR. NEJM. 2001;344: 699).

Hyperosmolar Solutions for Fluid Resuscitation During Critical Illness: Prevention of ARDS

“Traditionally, the clinical picture of ARDS consists of hypoxemia, reduced lung compliance, and fluffy infiltrates on X-ray,” said Ori D. Rotstein, MD, Toronto, Canada General Hospital. “Looking at large numbers of patients, about 80% of them have a pulmonary artery wedge pressure exceeding normal at sometime during hospitalization. This contributes to net water leakage.”

It seems appropriate to induce a negative fluid balance in patients by restricting fluids and some modest diuresis. A study by Mitchell and colleagues suggested that gentle fluid restriction and diuresis was associated with reduced lung fluids, decreased ventilator days and a shorter ICU stay (Mitchell L, et al. Am Rev Resp Dis.1992;145:990).

There is a very fine clinical line between keeping fluid out of the lungs and not reducing perfusion or inducing organ damage. Martin and colleagues looked at a series of patients with acute lung injury and hypoalbuminemia given 25% albumin plus diuresis using furosemide versus a double-placebo group. Body weight decreased, PAO2 to FIO2 ratio increased and there was a shorter duration of ventilation and ICU stay in the treatment arm. There was no difference in mortality. Organ injury-free days showed this protocol may have a potential overall benefit (Martin GS. Critical Care. 2000;4:S21).

With survival beyond seven days, multiple organ failure becomes the principal cause of death in civilian patients sustaining major trauma. The “two-hit hypothesis” suggests that ischemia reperfusion primes the immune system for excessive activation. A relatively trivial secondary inflammatory stimulus leads to neutrophil sequestration, excessive release of cytokines and the development of multiple organ dysfunction.

To see if early interventions might lower the responsiveness to a second “hit”, Dr. Rotstein’s group constructed a model using rats shocked to 40 mm of Hg for an hour and resuscitated with shed blood plus an equal amount of Ringer’s lactate (RL). Later, they were given an intratracheal dose of LPS to simulate the second hit.

The combination of LPS with shock produces a lung leak that is greater than shock or endotoxin alone. Similarly, that combination provokes a profound increase in neutrophils. They also found a profound increase in cyto-kine induced neutrophil chemoattractant (CINC), a rodent analog for the major chemokine interleukin-8 (Fan J et al. J Immunol. 1998; 161:440).

“Using strategies that reduce CINC expression, you can markedly reduce neutrophil sequestration,” noted Dr. Rotstein.

Using this protocol, Rizoli and others examined resuscitation using either RL or hypertonic saline. Animals resuscitated with hypertonic saline had albumin leak measurements near control levels. Neutrophil sequestration was also markedly reduced using hypertonic saline (Rizoli SB, et al. J Immunol. 1998;161; 6288).

Normally, L-selectin is maintained during shock resuscitation and CD11b is up-regulated. With hypertonic saline, there was a profound shedding of the former and no up-regulation of the latter. This effectively rendered the neutrophil devoid of mechanisms used to promote sequestration.

The effect does seem to be reversible. Eighteen hours after the hypertonic stress, the neutrophils begin to work perfectly. A second dose of hypertonic saline can recapitulate the anti-inflammatory effects (Rizoli SB, et al. J Surg Res. 1999;83:130) .

Careful consideration of fluid balance and resuscitation fluids may serve to optimize outcomes in critically ill trauma patients.

Gastrointestinal Hypoperfusion: Striving to Prevent Adverse Consequences

Of Intensive Care Unit patients put on mechanical ventilation, three out of four have endoscopically evident mucosal damage. Between 1% and 4% will have clinically important bleeding that leads to hypotension, tachycardia, anemia or blood transfusions. (Fennerty MD. Crit Care Med 2002;30:S351)

“What goes wrong is a very complex question with an equally complex answer,” said Neil H. Stollman, MD, University of California at San Francisco. “Clearly the central element is splanchnic hypoperfusion. All paths seem to converge at local gut ischemia.”

This is well illustrated by Figure 1.

One concern is that systemic measurements do not adequately reflect gut perfusion. There may be a time when systemic parameters of hemodynamics are still normal, yet visceral blood flow is already impaired (Maynard N, et al. JAMA. 1993;270:1203).

“This raises the idea that the gut is the canary in the coal mine,” said Dr. Stollman. “Does this give us an early warning that things still looking normal are about to change?”

Mohsenifar completed a double-blind cohort study involving 29 mechanically ventilated patients. They measured gastric perfusion via tonometry. The endpoint was spontaneous ventilation for at least 24 hours following extubation.

They found no differences in “traditional” hemodynamic, respiratory or gas exchange parameters between successful and failed weaning attempts. However, all that had been successfully weaned had a normal intramucosal pH (pHI) and maintained it after weaning. In the group that failed, the pHI started lower and dropped dramatically after weaning (Mohsenifar Z et al. Ann Intern Med. 1993;119:794).

In addition, pHI is a powerful prognosticator of mortality. In at least two different studies, those with normal initial intramucosal pH readings had mortality rates much lower than those with low pHI on admission. If low pHI results had not returned to normal within 12 hours, the mortality rate approached 87% (Doglio GR, et al. Crit Care Med. 1991;19:1037. Maynard N, et al. JAMA. 1993;270:1203).

There has been some major investigative work on ways to prevent stress-related GI bleeds. One candidate is H2 blockers.

A study by Martin and others looked at cimetidine compared with placebo. Overt bleeding (not necessarily clinically significant) was significantly reduced in the treatment group. Cimetidine also controlled gastric acid better than placebo (Martin LF, et al. Crit Care Med. 1993;21:19).

“It is pretty clear from this data that H2 blockers work and don’t seem to cause the nosocomial pneumonia problems we used to worry about,” said Dr. Stollman. “However, they are not terribly potent agents and there is concern about tolerance.”

Another group of medications getting a serious look are the proton pump inhibitors (PPIs). Merki and Wilder-Smith studied ICU patients who were receiving omeprazole, ranitidine or placebo. They reported bleeding in 31% of ranitidine group and only 6% in the omeprazole cohort. Omeprazole was also associated with better pH control (Merki HS, Wilder-Smith CH. Gastroenterology. 1994;106:60).

Levy and others compared the efficacy of sucralfate, ranitidine and omeprazole head-to-head in bleeding prophylaxis. Overt bleeding in the PPI arm was zero, compared to 10% among those taking ranitidine and 10% for sucralfate (Levy MJ, et al. Dig Dis Sci. 1997;42:1255).

“One thing becoming clear is that we are not stress prophylaxing very well,” said Dr. Stollman. “Institutions should develop treatment algorithms considering that H2 blockers are effective. Also keep in mind that PPIs may be superior in the inhibition of final pathways for acid secretion, have more
potent acid inhibition and lack tolerance.”

Nutritional Support of Patients With Pulmonary Failure

Nutrition decisions can have a major impact on respiratory status and ventilation, according to Charles M. Mueller, PhD, RD, Weill Medical College of Cornell University in New York City. One is on pulmonary physiology. The most important is probably compromise of muscles used in respiration. There may also be impaired immunocompetence, surfactant deficits and altered antioxidant defense with malnutrition.

The respiratory quotient (RQ) is a measure of patient metabolism. The RQ helps identify the types of substrate metabolism ongoing in a patient and how well they will use the substrates that are given.

Askanazi, et al. noted high parenteral carbohydrate (CHO) levels lead to an increase in PCO2. Parenteral protein caused an increase in O₂ consumption and changes in respiratory drive (Askanazi, et al.Anesthesiol. 1984;60:106).

“They came to the conclusion that hospitalized patients should be fed mixed substrate rather than the CHO and protein admixtures in use at the time,” said Dr. Mueller. “They also emphasized avoiding overfeeding.”

Others have suggested that the type of feeding impacts on immunity. The Moore brothers completed a meta-analysis concluding that early enteral feeding reduced septic complications and pulmonary infections leading to sepsis (Moore, FA, et al. Ann Surg. 1992;216:172). Studies in a mouse model by Kudsk and others indicated a loss of upper respiratory tract immunity with parenteral feeding (Kudsk KA, et al. Ann Surg. 1996;223:629).

Timing of the feeding may also prove to be important. There is a large 5-year study in press comparing burn patients given enteral nutrition within the first 24 hours and after 72. In their data there was a disturbing trend toward GI ischemia, infarction and death in those fed early (Gottschlich, M et al. J Burn Care Rehabil. 2002;23:401).

“Dr. Frank Cera years ago called metabolism in the ICU ‘auto-catabolism’,” said Dr. Mueller. “He means that no matter what you feed a person, it is their endogenous mechanisms that really drive substrate oxidation.”

Feeding decisions should be based on energy requirements, which can be as high as 25% of resting energy expenditure (REE) for those in impending pulmonary failure. Reduce this requirement by 50% during weaning to avoid overfeeding. Although requirements in pulmonary patients do not differ from others with chronic illnesses, protein can increase respiratory drive. CHO should be monitored to avoid overfeeding, a contributor to CO2 accumulation that may make weaning difficult.

Dr. Mueller also stressed the need for physicians to be aware of refeeding syndrome, often seen in anabolic patients. When a patient is making tissue the requirements for potassium, phosphorous and other mineral anions increases.

“These are the patients that get through their ICU stay, go to the floor and don’t eat,” said Dr. Mueller. “We sometimes forget about them until they come back to the ICU. Look for this at the end of the ICU stay or after transfer and try to prevent it by feeding them appropriately.”