Author's Reply.

Sir,

Thank you for pointing out the limitations of assessing data on the rates of adverse events for various classes of gadolinium chelates. These limitations apply to both allergic reactions and nephrogenic systemic fibrosis (NSF). Since the publication of our NSF review article,[1] several studies have added additional scientific basis for the concept of a lower rate of immediate adverse events for linear nonionic gadolinium chelates.[2-7] In particular, Prince et al.[2] reviewing experience in 158,796 gadolinium-based contrast agent (GBCA) enhanced examinations from two major institutions found an order of magnitude with a lower rate of immediate adverse events for the linear, nonionic agents, compared to the linear ionic agents and gadoteridol, a macrocyclic agent. Prince also reviewed the deaths from allergic reactions reported to the US food and drug administration (FDA), capturing the experience from 51 million GBCA administrations over a six-year period. These data also show that the lowest death rate is with linear nonionic gadolinium chelates.

Lee et al.[6] reviewing immediate anapylaxis in 141,623 patients showed a lower adverse events rate for gadodiamide, a linear nonionic chelate, compared to gadobenate, an ionic chelate (0.00013 vs. 0.00221; P<0.001). Taketomi-Takahashi et al.'s[4] review of 13,252 gadolinium administrations showed an adverse events rate of 0.14% for gadodiamide compared to 0.55% for Gd:DTPA, a linear ionic chelate. At the latest Yale NSF symposium, Raisch et al.[5] from the RADAR group, using a completely different analysis of tapping into FDA and legal databases, in addition to the published literature, showed projected reaction rates for gadodiamide and gadoversetamide (two linear nonionic agents) of 0.16 and 2.248, compared to rates of 5.03 and 11.41 for Gd:DTPA and gadobenate, two linear ionic agents, and 16.16 for gadoterate, an ionic macrocyclic agent.

We agree that the retrospective analysis of adverse events is not as good as the randomized, prospective, double-blinded trials. However, for the <1 in a million incidence of death from immediate allergic-type reactions to gadolinium and now nearly non-existent NSF, the randomized, double-blinded, prospective trials are not practical and we are forced to rely on retrospective data. Currently published prospective, controlled, double-blinded clinical trials do not confirm that all marketed GBCAs are similar with respect to hypersensitivity reactions, because they are massively underpowered. Interpreting failure to observe a statistically significant difference as proof that there is no difference is a common statistical error. Note that the US FDA has also made some differentiation among the GBCAs, giving a contraindication for three linear ionic agents, gadobenate, gadoxetate, and gadofosveset, in patients with a known history of allergic reaction to GBCA, while the nonionic GBCAs have no such contraindication. The growing abundance of evidence that nonionic linear GBCAs have a lower immediate adverse events rate compared to the ionic agents is not surprising, given that the same relationship exists for iodinated-based contrast agents.