Emanuel Kanal, MD, FACR Director, Magnetic Resonance Services, Professor, Department of Radiology, University of Pittsburgh Medical Center

This presentation will briefly review the safety data for these four agents and then will present our experience at the University of Pittsburgh Medical Center (UPMC) with all of the agents, focusing on our decision-making process and the research we conducted in choosing a gadolium-based MR contrast agent.

Adverse Events Profiles

Among the four gadolinium-based contrast agents, there are many similarities in the adverse reaction profile, but there are some differences as well. Table 1 summarizes the adverse reaction profiles for each of these agents. In some areas, the numbers are similar between all the agents, but in other areas, in my opinion, the numbers are quite dissimilar.

Dosing and Administration

A second area of potential distinction is in dosing, administration, and injection rates. Although the agents are more or less similar in this regard, there are some variations in FDA approval to date (table 2). Physiochemically, there are potentially clinically important differences in osmolality and viscosity. There are differences in dosages and indications as well, although these discrepancies may have less clinical relevance. Despite these differences, once a drug is approved by the FDA, it is generally used by the physicians in this country as they believe is appropriate and defendable for the benefit of their patients.

Osmolality and Viscosity

Figure 1 shows the osmolalities (the relative concentration of the solute to the solvent) published in the package inserts for each of the gadolinium-based contrast agents. The osmolality for Magnevist is 1,960, or approximately 7 times that of the plasma (285) into which it is being introduced. The osmolality of Omniscan is 789, or about 2.5 to 3 times that of plasma. It is 1,110 for OptiMARK, roughly 4 times plasma, and 630, or 2.2 times plasma, for ProHance.

A second clinical concern is the relative viscosity of these agents (figure 2). Viscosity is very important when the agent is being hand injected or when it might serve to limit the injection rates during power injection for such studies as perfusion imaging or MR angiographic imaging. With Magnevist being about 5 on a relative scale, Omniscan is 2, OptiMARK is 3, and ProHance is 2.

The UPMC Experience

As noted earlier, we have used all four of the gadolinum-based agents at UPMC. The remainder of this presentation will focus on our experience and the research we conducted when choosing an agent.

At UPMC, we began using Magnevist when it was first approved in 1988 and then switched to ProHance in November 1992. The reasons for this switch included financial considerations, the availability of FDA approval for high-dose use of ProHance, and the lower osmolality and viscosity with the latter agent. In fact, I had been one of the proponents who had specifically requested the switch to ProHance, mainly due to the lower osmolality and viscosity.

Within about a month of switching, the chief nurse for MR in our institution walked into my office and insisted that we stop using the new ProHance. She said that the new agent was causing a tremendous increase in adverse reactions and the nursing staff wanted its use discontinued.

This was a concern to me since I had specifically requested ProHance because I wanted to give larger, more rapid doses. However, I agreed to recommend pulling this drug from the UPMC pharmacy based on a single condition: that we do a prospective study. I requested that we pursue the data in a scientific fashion so that when I had to go back to the administration with the request to terminate our usage of the drug that I requested we start to use, I would have solid scientific data defending that requested change. Therefore, we designed a prospective study to objectively assess the true incidence of adverse events associated with ProHance administration at UPMC.

Study Methods

Although initially intended to run for 3 months, the study actually included 6 months of continuous data collection. All UPMC patients who underwent MR imaging with a contrast agent, in this case, ProHance, were included in the study. There were no further selection criteria.

It is important to note that this study was not supported in any way by any pharmaceutical firm or MR manufacturer. In fact, we specifically did not inform anyone outside of the UPMC setting that we were performing this study, so that no political or other potential bias could be introduced on that basis.

For data collection, we created a drug utilization evaluation (DUE) form. These blue sheets were prospectively completed for every single patient. The study population included all patients who received ProHance at UMPC from May 1, 1993 through January 31, 1994. All of the doses were administered as rapid-bolus. The administration of the agent was performed by a radiology nurse who interrogated the patient for possible adverse reactions both before and after receiving the drug. The results were recorded and prospectively filled out, with the nurse initialing or signing the form to let us know who did the data evaluation on every single patient.

After the MR study was performed and the patient was removed, he or she was once again interrogated as to possible adverse reactions. It is important to note that the study was designed intentionally so that the individuals who were responsible for acquiring and collecting this data and completing the forms were the same exact nursing staff who requested that we stop using ProHance. Therefore, if there was a potential bias built into the study, which admittedly there was, the bias was geared toward ensuring accurate reporting and, in fact, potentially even over-reporting.

The DUE forms were then submitted to a central location where the department of pharmacy actually ran the study. The recorded data included: the administered dose; the indication for the study; patient tolerance to the agent, both subjectively and objectively; and any adverse reactions reported or observed. Neither I nor anyone else from the department of radiology had direct access to the data as it was being collected. At the completion of the study, the data were unlocked and evaluated. The results of this study were published shortly thereafter. ¹ The results are summarized below, but the reader is directed to the original study for full data.

Study Findings

Of the 8,377 patients who underwent MR imaging study during that time period, 3,558 patients received ProHance. ¹ This more-than-40% rate of contrast use is much higher than the then national average estimated to be roughly 25% to 35%. We observed 101 adverse drug reactions in a total of 75 patients (2.1%) with no single adverse reaction having an incidence greater than 1%. The most common occurrence was nausea at 0.98%; emesis, at 0.34%, was the next most common (table 3). Five patients required medical intervention, although all five adverse events resolved after either oral or intravenous diphenhydramine was administered. None required either long-term follow-up or long-term care.

We then compared the results of this study with a retrospective study that we had conducted on 5,280 patients who had received Magnevist between June 1988 and March 1990. This retrospective study found a spontaneously recorded adverse reaction rate of 1.3%.

It is important to note, however, that the rate of spontaneously reported adverse reactions is generally significantly lower than that of reports that are specifically sought, such as in a prospective study. In fact, a study published in Radiology in 1995 demonstrated that the rate of adverse events found in a retrospective study is typically half that found in a prospective study. ² Specifically, this study of 15,496 patients who received Magnevist found a 2.4% rate of adverse reactions. In the same group, spontaneously reported adverse reactions were only 1.2%. This means that if you wait for the patient to say "I'm having a problem," total adverse events were 1.2%. Whereas, if you seek out adverse events and ask each patient, "How are you doing?" the rate doubles to 2.4%. The results of this study seem to suggest what I believe is confirmed elsewhere in the literature: that retrospective studies pick up approximately half of all the adverse events that occur.

Therefore, using the 1.3% adverse rate found in our retrospective study of 5,280 bolus administrations of Magnevist, it is possible to assume that a prospective study would have found an adverse event rate of approximately 2.6%. This is not significantly different from the 2.1% rate found in our prospective study of ProHance.

A Second UPMC Study

In the spring of 2000, UPMC, which had previously switched from ProHance to Omniscan for financial considerations, switched back again to ProHance. The decision to return to use of ProHance was made for multiple reasons including financial considerations and lower osmolality and viscosities. The decision to switch was made by a joint collection of all the physician representatives and administrative representatives from the various hospitals in the UPMC group.

One of the member hospitals, Shadyside University Hospital, decided that before they were willing to switch to ProHance they wanted to perform a prospective trial, run entirely by their personnel at their facility. They looked at the data that they had for Omniscan, which they had been using, and then began using, and prospectively evaluating, ProHance. The study design was virtually identical to the one described earlier. All patients were prospectively and repeatedly asked about any adverse reactions. After approximately 100 patients were studied, I personally received a phone call from the physician at that hospital who was running the study. She inquired of me if it was acceptable to us if they prematurely terminated the study. When I asked why, she replied that they had found no adverse events in their first 100 administrations and that they felt that "that was good enough for us." They terminated the study and switched to ProHance, which they have been using ever since. In fact, on March 27, 2001, I contacted the physician in charge of the study at Shadyside University Hospital and asked her what their experience was to date, to ensure that nothing had changed in the interim. Essentially, to summarize her words, she said that the agent was performing exactly as stated on the package insert and that they are happy to continue to use it.

Conclusion

In conclusion, there are many areas in which these agents are in fact relatively equivalent. I believe that, in general, all four of the gadolinium-based contrast agents are among the safest drugs that we as diagnostic radiologists use. In my opinion, the safety profiles are similar. Areas in which they differ, including osmolality and viscosity, helped to tip the scales in favor of our decision to switch to ProHance, which we have done and with which we are quite satisfied to date.