- General
Oxygen therapeutics are derived from hemoglobin- or fluorochemical-based compounds. The starting material for hemoglobin-based oxygen carriers may be a stroma-reduced hemoglobin, commonly referred to as stroma-free hemoglobin (SFH), chromatographically purified hemoglobin obtained from sources including outdated human blood or bovine blood, or recombinant hemoglobin. Stable and functional oxygen therapeutics are produced from these starting materials by various chemical and/or genetic manipulations. The resulting products include intra-tetrameric cross-linked hemoglobin, polymers of hemoglobin tetramers (intra- and inter-cross-linked), hemoglobin tetramers conjugated to non-protein macromolecules, or genetically stabilized tetramers.
Fluorochemical-based oxygen therapeutics are compounds in which hydrogen atoms on cyclic or straight chain hydrocarbons are replaced with fluorine. Some compounds also contain other atoms such as bromine. These compounds are emulsified in electrolyte solutions containing surface-active agents (surfactants) (perfluorochemical emulsions). Fluorochemical-based compounds are capable of dissolving large quantities of oxygen (without binding) at high concentrations of inspired oxygen.
- Safety Considerations
A number of new and largely unresolved safety-related problems have arisen during the preclinical and clinical development of the current generation of hemoglobin-based products and perfluorochemical emulsions. The mechanisms of various observed toxicities have not been elucidated fully. There is evidence to suggest that unmodified hemoglobin, which is filtered through the renal glomerulus, is nephrotoxic. This evidence suggests that unmodified hemoglobin (i.e., native tetrameric hemoglobin, mw - 64,000) should not be present in the final product and should be removed to the maximum extent possible. However, nephrotoxicity does not explain all the effects observed when hemoglobin-based oxygen carriers are infused. For hemoglobin-based oxygen carriers, the recommended testing scheme is based on the hypothesis that one cause of toxicity involves the interaction of oxygen radicals or iron with cellular metabolism. Thus, the modified oxygen carrying capacity of hemoglobin-based oxygen carriers could affect safety as a result of excess oxygen supply to cells and tissues, deficient supply, or oxidative effects of the hemoglobin moiety itself. Similarly, the toxicities observed in animals following administration of perfluorochemical emulsions may also be clinically significant. Activation of complement and procoagulant cascades may affect the safety of perfluorochemical emulsions and result in dose limiting toxicities in clinical practice. For both hemoglobin-based oxygen carriers and perfluorochemical emulsions, it is hypothesized that the effects seen are in part the result of activating any number of triggered enzyme or cellular systems.
Current oxygen therapeutic research efforts are aimed at gaining a better understanding of reactions in humans and developing safe and effective products. Following is a list of toxicities and laboratory findings known or thought to be associated with the use of hemoglobin-based oxygen therapeutics. The list is not intended to be all-inclusive, and it also includes some hypothetical toxicities that have not been demonstrated experimentally or clinically, but which, nevertheless, merit consideration.
- Vasoactivity
Almost all hemoglobin products, regardless of their molecular weight distribution, are associated with vasoactivity in humans and in animal models. The etiology and clinical consequences of the vasoactive property of hemoglobin are still controversial and not completely understood. One hypothesis is that the interactions between cell-free hemoglobin and nitric oxide (NO), the endothelium-derived relaxing factor, may be a primary event that contributes to a vascular inflammatory response progressing to multi-organ failure. Other hypotheses suggest that cell-free hemoglobins modulate adrenergic receptor sensitivity and stimulate endothelin-1, a peptide with vasoconstrictor activity. Even modified hemoglobin products with small effects on blood pressure may result in significant vasoconstriction and increased vascular resistance.
- Cardiac Toxicity
Administration of a hemoglobin-based oxygen carrier has been reported to be associated with the development of myocardial lesions characterized by mild to moderate focal and multifocal degeneration and/or necrosis of cardiac myocytes in certain animal models. The lesions have been reported to be seen 24-48 hours after a single infusion of product. Left ventricular myocardium near the base of the papillary muscle is most prominently affected, with other affected regions including the interventricular septum and the right ventricle. Animal species most commonly affected include rhesus monkey and pig. There are significant animal species differences in sensitivity to the effect of the oxygen therapeutic; dogs, sheep, and rats did not develop the lesion after single infusion of the hemoglobin-based oxygen carrier.
- Gastrointestinal Toxicity
- Discomfort.
Symptoms reported in clinical trials include nausea, vomiting, dysphagia, or generalized abdominal pain. These symptoms are believed to be related to NO scavenging by hemoglobin, causing localized spasm throughout the gastrointestinal tract.
- Bacterial Translocation.
Animal experiments indicate changes in the architecture of the intestinal microvilli within minutes of infusion of some hemoglobin-based oxygen carriers. These experiments suggest the possibility that the incidence or the severity of bacterial translocation across the gut epithelium may be increased.
- Pro-inflammatory Activity
Infusion of hemoglobin in small animals stimulates monocyte procoagulant activity resulting in disseminated intravascular coagulation. Early in vivo studies in rabbits, but not in other animal species, demonstrated pulmonary arteritis and thrombotic lesions possibly related to procoagulant activity. In vitro studies on the effect of hemoglobin on leukocytes in whole blood demonstrate that in the presence of other blood components, leukocytes release pro-inflammatory cytokines such as tumor necrosis factor (TNF) and interleukin-8 (IL-8).
- Oxidative Stress
Numerous in vitro and some in vivo reports suggest that hemoglobin solutions may induce an oxidative stress. Oxidative stress may, in part, be explained by the ability of hemoglobin to serve as a source of toxic oxygen species and/or the ability of hemoglobin to remove nitric oxide, an important component of normal antioxidant mechanisms. Indirect evidence in humans of such effects is seen with reports of increased enzyme activity including creatine phosphokinase (CK), lactic dehydrogenase (LDH), and pancreatic enzymes, lipase and amylase (see below).
- Pancreatic and Liver Enzyme Elevation
Reports of elevated pancreatic and liver enzymes in a number of animal models of exchange transfusion suggest possible free-radical mediated injury. Increased levels of amylase and lipase after hemoglobin infusions in animals and humans have been frequently observed. The pattern and magnitude of lactate, CK, and LDH elevations seen in animal studies are similar to those reported in humans and also suggest possible free-radical mediated injury.
- Endotoxin Synergy with Hemoglobin
Endotoxin, a bacterial product, and hemoglobin have been shown to exert synergistic toxicity when hemoglobin is given in a clinically relevant dose as a resuscitative fluid. Direct biochemical interactions between hemoglobin and endotoxin have been shown to reduce clearance of endotoxin from the circulation and to enhance lethality in some animal models of sepsis. This is a worrisome feature since these products may be given in some instances to patients with ongoing infectious processes or to trauma victims with contaminated wounds.
- Neurotoxicity
Although experimental data suggest that acellular hemoglobins scavenge NO and may inhibit NO-related neurotoxicity, numerous published reports implicate free hemoglobin in the degenerative changes in the brain in experimental models of subarachnoid hemorrhage. Direct neurocytotoxicity of modified hemoglobin is suggested by pre-clinical studies in which neurons in culture were killed by the addition of hemoglobin, leaving glial cells intact. In one clinical report, use of an acellular hemoglobin product was an independent predictor of an unfavorable outcome at three months following acute ischemic stroke. The mechanism of this clinical toxicity has been hypothesized to be, in part, due to the potent vasoconstrictor effect of endothelin-1 which was increased in a dose-dependent manner by the administration of the acellular hemoglobin product.
Toxicities known or thought to be associated with one or more of the current perfluorochemical emulsions include the following:
- Thrombocytopenia
The basis for this side-effect in animals and humans is not fully understood, but may be related to metabolism and normal clearance process of these compounds. Fluorochemical compounds may lead to some changes in the surface characteristics of platelets and subsequent uptake of the altered platelets by the liver and the spleen. Platelets may appear "functional" in aggregation and bleeding time measurements in vitro, but may potentially have a short half-life in circulation even though platelet production is normal.
- Complement Activation and Cytokine Release
At present, although a hypothetical toxicity, there are no experimental data to suggest that complement activation occurs with the use of perfluorochemical emulsions.
- Reticuloendothelial Blockade
After intravenous administration, the droplets of perfluorochemical emulsion are taken up by the reticulo-endothelial system. It has been hypothesized that administration of perfluorochemical emulsions may affect ability to clear circulating bacteria.
- "Flu-like" Symptoms
Administration of perfluorochemical emulsions in humans has been associated with the development of flu-like symptoms and transient elevations in proinflammatory cytokines. The etiology of this phenomenon is not known.
- Central Nervous System Effects
Cerebrovascular accident has occurred in the context of administration of perfluorochemical emulsion. The pathophysiologic basis for the association has not been elucidated.
- Efficacy Considerations
In addition to the safety considerations described above, oxygen therapeutic products present several issues relating to efficacy. Efficacy endpoints may be direct measures of clinical benefit (improved survival, alleviation of symptoms) or they may be laboratory measurements or physical signs expected to correlate meaningfully with clinical benefit. The latter are referred to as surrogate endpoints and, once validated, are especially important in the case of oxygen therapeutics since direct demonstration of efficacy is likely to be very difficult (as it has been for red blood cells, per se). Validation of a surrogate endpoint for a therapy includes generation of clinical data demonstrating that effects of the therapy on the surrogate endpoint are reasonably likely to predict clinical benefit. See 21 CFR 601.41. Factors of importance when considering acceptability of surrogate endpoints include feasibility of using direct clinical measurements, risk/benefit assessments, and perhaps most importantly, knowledge and understanding of the disease and of the agent.
There has been extensive clinical experience with red cell transfusion, resulting in a practical appreciation of relevant indications, benefits, and risks. There is also an extensive collection of data on red blood cells, the anemic state, and their interaction, resulting from years of basic and applied research. Thus, although it is not possible to document the clinical benefit of all red cell transfusions with specific endpoints, the available knowledge relevant to such transfusions support use of surrogate endpoints such as the P50, the oxygen content and the hematocrit as suitable endpoints to demonstrate efficacy of red cell transfusions in clinical practice and in some clinical trials. Currently, we do not consider these surrogate endpoints to be acceptable as measures of the effects of hemoglobin- and perfluorochemical-based red cell substitutes, because knowledge of the effects of hemoglobin- and perfluorocarbon-based red cell substitutes and of the interaction of these agents with various clinical states is rudimentary. Further, no oxygen carrier presently approved by FDA has all the properties of the human red cell, nor are any two products identical. We recommend that the endpoints used in clinical studies of these agents be selected with these caveats in mind.
Under certain circumstances, oxygen therapeutic agents may qualify for Fast Track designation. Criteria and considerations for Fast Track designation may be found in "Guidance for Industry: Fast Track Drug Development Programs. Designation, Development, and Applicataion Review" dated September 1998. The guidance document may be found at http://www.fda.gov/cber/guidelines.htm.
There are several potential indications for oxygen therapeutics. Below, we discuss the following three such uses for these products: 1) local effects/regional perfusion, 2) perioperative indications and 3) trauma. As noted earlier, this discussion is not intended to represent the only approaches to evaluating clinical use of oxygen therapeutics, nor are investigators required to accept these categories.
- Local Effects/Regional Perfusion
This category might best be defined by considering two examples: perfusion during coronary angioplasty and enhancement of tumor radiosensitivity. Perfusion, via the central lumen of a catheter used for percutaneous transluminal coronary angioplasty (PTCA), is an FDA-approved indication for a perfluorocarbon preparation (Fluosol). The data that supported this approval included clinical studies utilizing surrogate endpoints of left ventricular function that had been validated as clinically relevant by recognized cardiologic investigations. Future studies for this indication could conceivably utilize similar clinical trial design, with specific endpoints appropriately updated. The rationale for use of oxygen carriers (systemically administered) in therapy of neoplasms is based on the observation that increased tumor tissue oxygen tension will increase the sensitivity of tumors to radiation or to chemotherapy more than that of normal tissue. Demonstration of increased oxygen tension in the target tumor can function as an important supporting argument for efficacy, but would not be likely to serve alone as the primary endpoint. Ultimately, we recommend that the endpoint used to establish efficacy be similar to that used in evaluation of cytotoxic agents for the stage and type of cancer under investigation.
- Perioperative Indications
Sponsors must monitor their investigations and evaluate pertinent risks to research subjects. (21 CFR 312.56.) In the case of investigational oxygen therapeutic products for proposed perioperative indications, such risks include the risks to subjects receiving the oxygen therapeutic in lieu of allogeneic transfusion (e.g., inferior perfusion, undesirable hemodynamic responses, and other adverse drug reactions). In situations where red blood cells are used, reduction in the use of allogeneic blood is a surrogate endpoint for the avoidance of the risks associated with the use of allogeneic blood including, but not limited to, viral transmission, incompatibility, etc. In the past, we have accepted reduction in the use of allogeneic blood (for instance, in the indications for erythropoietin). However, use of both allogeneic transfusion and oxygen therapeutics entail risks. Therefore, safety is a critical element in any evaluation of oxygen therapeutic products for elective surgical use. In the setting of elective perioperative use, a mere delay in the requirement for allogeneic transfusion without reduction in the use of allogeneic red blood cells would probably not be considered of benefit to patients. The category of peri- and post-operative use of oxygen therapeutics includes situations such as hemodilution (with or without autologous predonation) and intra- and post-operative replacement. We recommend the investigators be aware of the present lack of objective criteria to define a broadly applicable transfusion trigger and strive to develop and validate physiologic markers of efficacy for individual oxygen carriers.
A trial to obtain an elective surgical indication alone, without evaluation of the product in unstable patients in a trauma setting, is unlikely to assure the safety of an oxygen therapeutic in elective surgical patients who become unstable or in trauma patients. While an oxygen therapeutic may appear to be safe in patients who are euvolemic and anemic, it cannot be assumed that such an oxygen therapeutic would be safe in hypovolemic or unstable patients, either surgical or trauma.
- Trauma
In trauma, mortality is an unambiguous endpoint that many consider to be the most meaningful of the potential indications related to clinical benefit of oxygen therapeutics. Although short-term (e.g., 24-48 hours) survival is helpful in assessing the physiologic activity of an oxygen therapeutic, long-term survival is the primary clinical benefit of interest to the patient and the patient's family. The benefit of short-term survival is limited if it does not lead to long-term survival. There is insufficient information at present to correlate short-term survival with long-term survival for oxygen therapeutics.
In designing clinical trials for a potential trauma indication in the hospital setting, we recommend that you consider designs where patients are able to provide consent for themselves or where consent can be obtained from a legally authorized representative. In the hospital setting, the use of an oxygen therapeutic would not be expected to result in a survival advantage over the use of red blood cells. Rather, oxygen therapeutics may be interchangeable with red blood cells in the short-term. Many, if not all, recipients of the oxygen therapeutic will need administration of other transfusion blood components. A clinical trial to demonstrate noninferiority, as opposed to superiority, of an oxygen therapeutic to blood in terms of mortality might not qualify for exception from informed consent under 21 CFR 50.24.
Historically, many researchers have considered field use as the most likely situation in which oxygen therapeutics could improve survival. However, in major urban areas with rapid transit to definitive care, there may only be a small percentage of patients for whom use of an oxygen therapeutic might provide a survival benefit. Some of the most seriously injured patients will die in spite of rapid availability of optimum definitive care and other, less seriously injured, patients will survive with current resuscitation measures. It is difficult to select deliberately and prospectively the small population of trauma patients for whom use of an oxygen therapeutic may provide a survival advantage. If a clinical trial is not designed to select deliberately and prospectively for this small subset of patients who are likely to benefit from treatment, a large number of patients would likely need to be tested for a survival advantage. Accordingly, many more subjects would receive an oxygen therapeutic agent than are likely to benefit from such use; therefore a careful overall safety assessment would be appropriate. In addition, for trauma patients who have also sustained head injuries, the heterogeneity in the severity of head injury may lead to mortality outcomes independent of the effect of blood loss or the use of an oxygen therapeutic agent.
In rural areas or other situations where there may be prolonged delay to definitive care, there may be greater potential for an oxygen therapeutic to provide clinical benefit than in urban settings. Such studies are difficult to control and may pose complications of trial design and data analysis, due, for example, to practical considerations such as differences in the length of time required to transport a patient to the hospital. Nevertheless, in the rural setting, a temporary treatment that sustains adequate tissue oxygenation and aerobic metabolism prior to control of bleeding and/or prior to obtaining cross-matched blood may provide a clinical benefit.