Contrast Media Safety: What Do We Know and How Do We Know It?

Mark M. Mishkin, MD

Use of an intravenous injection of a water-soluble iodine compound (sodium iodide) for visualizing the urinary tract was first described by Osborne in 1923. However, sodium iodide was accompanied by a high incidence of side effects and the image quality was not good.

As newer, improved contrast agents were intro- duced, a number of investigators reported on the incidence of death and side effects incurred by their use. In 1975, the first large-scale, multicenter, pro- spective study on the incidence of fatal and nonfatal reactions was reported by Shehadi. In thls series, the overall incidence of adverse reactions was 5.65% of which 3.77% were mild, 1.77% were moderate, 0.02% were severe, and 0.0007% were fatal. Shehadl also noted that a history of allergy doubled the risk of adverse reactlon and a history of a previous adverse reactlon to contrast material trebled the risk; also, the risk of a reaction to an in- traarterial injection was half that of the risk associ- ated with an intravenous injection.

With the introduction of low osmolar nonionic agents, other large-scale studies were performed because these agents appeared to cause signifi- cantly fewer adverse reactiins. However, they were from 10 to 15 times more expensive than their predecessors.

In 1987, Lasser reported on the increased safe- ty of ionic agents when patients were pretreated with steroids. Palmer in Australia and Katayama in Japan collected 120,000 and 335,000 patlents, re- spectively, and compared the incidence of adverse reactions of the ionic versus the nonionic agents, Wolf recently reported a study that included a di- rect comparison of ionic agents with pretreatment with sterolds versus nonionic agents.

The Katayama study is the most powerful sta- tistically, and this, supported by both Wolf’s study and Palmer’s study, indicates that the single most important risk factor is the type of contrast agent used, with a nonionic agent having between 14 and 17% of the lncldence of adverse reactions, and be-

From Jefferson Medical College, Philadelphia, Pennsylvania. Address for reprints: Mark M. Mishkin, MD, Department of Radi-

ology, Suite 1009 Main Building, Thomas Jefferson University Hospi- tal, Philadelphia, Pennsylvania 19 107.

tween 4 and 17% of the incidence of severe ad- verse reactions.

These studies have resolved the issue of safety of the ionic agents versus the nonionic agents. The safety of ionic agents with sterold pretreatment re- mains to be established. Choice of a contrast agent now rests on considerations of cost, ethics and li- ability.

(Am1 Cardiol 1990;66:34F-36F)

I n 1923, Osborne et al1 of the Mayo Clinic reported on the first in vivo use of an intravenous injection of a water-soluble iodine compound (sodium iodide) for

visualizing the urinary tract. In 1927, Moniz2 used sodi- um iodide for carotid angiography. Sodium iodide pro- duced a large number of adverse reactions, many of which were severe. Over the next 2 decades, extensive industrial research yielded less toxic contrast agents with better visualization of the organs being studied. In the early 1950s the diatrizoates were introduced, and in the 1960s iothalamate was introduced. These latter 2 com- pounds were much better tolerated and were relatively safe and efficacious.

In 1942, Pendergrass et al3 reported the results of an informal survey, which showed an incidence of 39 deaths/million urograms. In 1955, a follow-up survey by Pendergrass et al4 yielded an incidence of 6.5 deaths/ million urograms. In 1973, Whitten et al5 reported on 32,964 outpatients undergoing excretory urography us- ing diatrizoate contrast media. The overall incidence of adverse reactions was 6.8%, with severe reactions repre- senting 0.09%. There was 1 death in this series.

In 1975, Shehadi6 reported on a prospective survey in 30 teaching hospitals in the United States, Canada, Eu- rope and Australia consisting of 112,003 patients. The patients had undergone intravenous and intraarterial in- jections of ionic contrast media. There was an overall incidence of adverse reactions of 5.65%, of which 0.02% were severe. Eight fatalities were reported.

In 1980, Shehadi and Toniolo7 reported on a prospec- tive survey of 302,083 patients examined with intravascu- lar contrast media. This was an international study in- volving the U.S., Canada, Australia and several Europe- an countries. The overall incidence of adverse reactions was 4.7% with 0.072% being severe, and there were 18 fatalities. The risk of adverse reactions doubled with a history of allergy. With a history of a previous adverse reaction to contrast material, the risk trebled. An intraar- terial injection resulted in half the number of adverse reactions as an intravenous injection.

34F THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 66

This modern era of ionic agents was characterized by a relatively low incidence of adverse reactions and the drugs were regarded as safe and effective. However, the persistence of occasional severe reactions (including death) resulting from strictly diagnostic procedures spurred the search for an even safer agent.

NONIONICS Alm&8,9 theorized and eventually developed the first

clinically used nonionic contrast agent, metrizamide, in the 1970s. This agent was primarily used for myelogra- phy, but the experience with intravascular use quickly demonstrated that it produced less discomfort and seemed to be safer than the ionic class of agents. Metriza- mide, because of its instability, was lyophilized and had to be mixed into solution before its use. The late 1970s saw the introduction of stable nonionic contrast agents for intravascular use. Schrott et alla in 1986 reported on 50,666 consecutive patients undergoing intravenous urography with a nonionic agent in a prospective multi- center study. The overall incidence of reactions was 2%, while the incidence of severe reactions was 0.01%. The patients were divided into 2 groups. Those without risk factors had an overall incidence of 1.3% and a severe reaction rate of 0.004%. In the patients with risk factors, the overall incidence was 2.7% and the rate for severe reactions was 0.02%. This series demonstrated that the high-risk patients were at 2 times the risk for reactions overall, and at 5 times the risk for severe reactions when using nonionic contrast agents.

COMPARISON OF IONIC AND NONIONIC AGENTS

There are 2 very large studies comparing the inci- dence of adverse reactions in patients receiving either ionic or nonionic contrast media. Palmer” in Australia and Katayamat2 in Japan collected 109,546 and 337,647 contrast studies in a prospective fashion, respectively. Palmer found the incidence of severe adverse reactions in the ionic group to be 0.09% and in the nonionic group, 0.02%. In a more detailed and controlled study, Ka- tayama12 reported the incidence of severe adverse reac- tions in the ionic group to be 0.25% while the severe adverse reactions in the nonionic group was 0.04%.

PRETREATMENT WITH STEROIDS In 1987, Lasser et alI3 reported on the use of pretreat-

ment with steroids in patients receiving ionic contrast media in a prospective, randomized, double-blind study. The incidence of overall reactions with steroid pretreat- ment was 6.4% and of severe reactions (which he called class III reactions) was 0.2%. He compared these results with those reported by Schrott et ali0 and concluded that pretreatment with steroids and the use of ionic agents offered the same order of safety as did the nonionic agents.

Because Lasser’s class III reactions did not match Katayama’s definition of severe reactions, Lasser reana- lyzed his data using Katayama’s definitions (personal

communication) and reported an incidence of severe re- actions in the steroid plus ionic category of 0.08% as compared to Katayama’s 0.04% in patients receiving non- ionic agents.

In 1989, Wolf et ali4 reported on a direct comparison of patients receiving ionic agents alone, pretreatment with steroids plus ionic agents and a nonionic agent. In a series of 15,668 patients, little difference was noted be- tween the group receiving the ionic agent alone and the group receiving pretreatment with steroids followed by the ionic agents. There was a sevenfold decrease in overall adverse reactions in the patients receiving the nonionic agent and a 25-fold decrease in severe adverse reactions in patients receiving the nonionic agent.

DISCUSSION Multiple studies from around the world have demon-

strated an improved safety profile of the nonionics as compared to the ionic contrast agents, with the safety factor for overall reactions ranging from 2: 1 to 7: 1 and the safety factor for severe reactions ranging from 4:l to greater than 25:1, all of these in favor of the nonionic agents.

None of the aforementioned studies is without weak- nesses. Only I was randomized and double-blinded (Lasser et aP3). However, Lasser et al did not include the use of nonionic contrast agents in their study and, because the study was relatively small, the confidence limits of the statistical data are very broad.

In the study of Wolf et al, i4 which directly compared an ionic agent alone, pretreatment with steroids plus an ionic agent, and a nonionic agent, the patients were not randomized and the study was not blinded. It was a crossover study and the individual hospitals participating in the study were blinded and randomized. Because of the difficulty with getting compliance for the patients sched- uled to receive pretreatment with steroids plus ionic con- trast agents, this group is small, again yielding broad confidence limits for the comparison of this group with the group receiving ionic agents alone and the group receiving nonionic agents.

The study of Palmer l1 lacked details necessary for analysis and comparison of the various subgroups of pa- tients within the 2 broad categories of ionic and nonionic agents. Further, there is no available information on the number of patients whose results were unreported or the reasons for their not being reported.

Because of the large size of the study of Katayama,12 it is statistically powerful. A retrospective audit of this study estimated that between 59 and 82% of the patients scheduled for the study were, in fact, enrolled. Thus, there is the potential for bias. However, an analysis of the details of the 2 arms of the trial, and specifically the distribution of the various characteristics analyzed in this study, fails to demonstrate any significant trend indica- tive of bias. Children (under 30 kg) were preferentially given the nonionic contrast media. Similarly, patients undergoing intravenous digital subtraction angiography were more frequently given nonionic contrast media.

THE AMERICAN JOURNAL OF CARDIOLOGY OCTOBER 26,199O 35

A SYMPOSIUM: CARDIOVASCULAR IMAGING IN THE 1990s

Both of these subsets represent very small components of the study and it is unlikely that either of these factors significantly influenced the results. In all of the other important categories, there was good balance between the group of patients receiving ionic contrast agents and those receiving nonionic contrast agents. Thus,there is no evi- dence of systematic bias in any of the other subgroups analyzed.

All of the studies directly comparing adverse reactions and severe adverse reactions in the ionic and nonionic groups demonstrate that the single largest risk factor for developing an adverse reaction is the use of ionic contrast agents. Other important risk factors include a previous reaction to contrast agents, a history of allergy or asthma, and a history of cardiac disease. None of these is as great as the use of an ionic contrast agent.

IONIC AGENTS VERSUS NONIONIC AGENTS: WHICH TO CHOOSE?

This question leads to the dilemma we currently face. The newer agents are 10 to 15 times more expensive. The greatest risk to the patient is the use of the ionic agents. Yet, severe reactions and death from the ionic agents are uncommon. Before the introduction of nonionic agents, approximately 10 million contrast examinations were performed annually in the U.S. The cost of contrast ma- terial for these studies was $100 million/year. The cost for performing the same number of studies using only nonionic agents would have been approximately $1 bil- lion/year. t5 Are there more important ways that these additional dollars could be spent? What is the physician’s responsibility to the individual patient referred for a study and how does that compare with the physician’s responsi- bility to the cost of delivery of medical care in the U.S.? What are the ethical duties and legal liabilities associated with the use of an ionic contrast agent in a patient with no other identifiable risk factors? Certainly, if the cost were the same or if the new agents were only 2 or perhaps 3 times more expensive, virtually everyone agrees that there would be no reason to use ionic agents.

If the figures from the Katayama study are trans- ferred to the U.S., we would expect 22,000 severe adverse reactions each year if only ionic agents were used. If everyone received nonionic agents, there would be an estimated 4,000 severe adverse reactions per year. What is an appropriate expenditure of dollars to prevent 18,000

severe reactions per year? Although Katayama reported only 1 death in each arm of the study, it is stated that no causal relation to the contrast media could be established. Of the 18,000 severe reactions that are potentially avoid- able every year in the U.S., some percentage are going to have a fatal outcome. What is an appropriate expenditure to avoid these deaths?

The answers to these questions require input from physicians, hospital management, ethic&, attorneys, third-party payors and the government.

In conclusion, despite the lack of a “perfect study,” the data overwhelmingly demonstrate the relative safety of the nonionic agents when compared to the ionic agents. The safety of premeditation with steroids plus ionic agents is still unanswered, The issue comes down to cost, ethics and liability.

REFERENCES 1. Osborne ED, Sutherland CG, Scholl AJ, Rowntree LG. Roentgenography of urinary tract during excretion of sodium iodide. JAMA 1923;80:368-373. 2. Moniz E. L’Angiographe Cerebale. Paris: Massons et tie, 1934. 3. Pendergrass EP, Chamberlin GW, Godfrey EW, Burdick ED. A survey of deaths and unfavorable sequelae following the administration of contrast media. Am J Radio1 1942;48:741-762. 4. Pendergrass BP, Hodes PJ, Tondreau MD, Powell CC, Burdick ED. Further considerations of deaths and unfavorable sequelae following theadministration of contrast media in urography in the United States. Am J Radio1 1955;74:262-287. 5. Whitten DM, Hirsh FD, Hartman G W. Acute reactions to urographic contrast medium. Am J RoentgenoI 1973;119:832-840. 6. Shehadi WH. Adverse reactions to intravascularlv administered contrast me- dia: a comprehensive study based on prospective survey. Am J Radial 1975:124:145-152. 7. Shehadi WH, Toniolo G. Adverse reactions to contrast media. Radio& 1980;137:299-302. 8. Alm6n T. Development of nonionic contrast media. Invest Radial 1985;2O(suppl):S2-59. 9. Almen T. Contrast agent design. Someaspects on thesynthesis of water-soluble contrast agents of low osmolality. J Theor 5iol 1969,24:216-226. 10. Schrott KM, Behrends B, Clauss W, Kaufmann J, Lehnert J. lohexol in excretory urography: results of the drug monitoring program. Forsch Med 1986;104:153-156. 11. Palmer FJ. The RACR survey of intravenous contrast media reactions: a preliminary report. Australm Radio1 1988;32:8-11. 12. Katayama H. Report of the Japanese committee on the safety of contrast media. Presented at the Radiological Society of North America Meeting, Novem- ber 1988. 13. Lasser EC, Berry CC, Tainer LB, Santini LC, Lang EK, Gerger FH, Stol- berg HO, and other Contrast Material Reaction Study Participants. Pretreatment with corticosteroids to alleviate reactions to intravenous contrast materials. N Engl J Med 1987;3 17~845~895. 14. Wolf G, Mishkin M, Roux S, Halpern E. Adverse drug reactions to contrast media in routine clinical practice (abstr). Chicago, Illinois: Proceedings, Annual Meeting RSNA, 1989. 15. Evens RG. Economic impact of low-osmolarity contrast agents on radiology procedures and departments. Radiology 1987;162:267-268.

36F THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 66