Results
Characteristics of the study subjects and index HSRs
We assessed the results of 482 skin test panels performed on 431 patients with a history of LOCM-related immediate HSR (including 186 patients with severe index HSR) (Table 1). In 17.6% of the patients, the initial HSR occurred on their first exposure to LOCM. The mean age of these patients was 57.0 ±13.2 years, and 269 (62.4%) patients were female.
According to the severity of the index HSR, there were 216 moderate cases (44.8%), 186 severe cases (38.6%), and 80 mild cases (16.6%). The most common culprit LOCM was iopromide (23.1%), followed by iohexol (22.8%), iobtridol (18.7%), iopamidol (16.5%), iomeprol (8.2%), ioversol (8.0%), and iodixanol (3.8%).
The outcome of skin test with LOCM and time interval
Skin test positivity to any LOCM was only 3.1% (15/482) in the skin prick test but increased to 51.8% (250/482) when the IDT was performed. The mean number of reactive LOCMs per test was 2.29 ± 1.34 in cases showing positivity in skin tests. Of the LOCMs, iohexol showed the highest positive rate (70.0%), while iopamidol showed the lowest positive rate (44.1%, Figure 1A). Positive rate was not different between monomers and dimer (56.6% vs. 58.3%, respectively).
The occurrence of previous severe HSRs was associated with a higher likelihood of a positive skin test (Figure 1B). Among the mild HSR group, 38.7% (31/80) of patients had positive results. Furthermore, 45.8% (99/216) of patients with previous moderate HSR demonstrated positivity, while 64.0% of those with severe HSR (119/186) showed positive results (P = 0.002 among three groups, Figure 1B). Skin test positivity was dependent upon the time interval between the index HSR and the skin test; 67.2% (39/58) < 4 weeks, 60.7% (34/56) in 4-8 weeks and 46.0% (160/348) > 8 weeks, respectively. (P = 0.007, Figure 1C).
Cross-reactivity classified by common N-(2,3-dihydroxypropyl) carbamoyl side chain
To evaluate cross-reactivity among LOCMs, we calculated the co-positivity between different LOCM pairs. Among the 250 patients with skin test positivity, 157 cases (62.8%) showed co-positivity to at least two different LOCMs (Supplement B). The most common co-positivity was observed in the iohexol-iomeprol pair (36.3%). The co-positivity proportion in the LOCM pairs sharing common side chain was 21.5%, which was significantly higher than the 13.3% that was observed in the LOCM pairs without common side chain (P = 0.008) (Figure 2). This significant difference was observed in the severe index HSR group (20.7% vs 11.5%, P = 0.003), but not in the non-severe cases.
Outcomes of re-exposure to LOCM
Of 431 patients of LOCM skin tests, 355 cases were subsequently re-exposed to LOCM and the consequences of LOCM re-exposure were assessed based on the skin test results. The overall BTR rate was 12.3% in the 244 cases in which the LOCM was changed, which is significantly lower than the 46.6% BTR rate observed upon re-exposure to the culprit LOCM (Figure 3-A, P-value = 0.004). The subgroup analyses by severity showed that patients who had a severe index HSR exhibited a significantly lower rate of recurrent HSR when changing the LOCM compared to those who had re-exposure to the same culprit LOCM (11.3% vs. 100%, P-value < 0.001). However, reduction of recurrence due to the changing of the contrast media was not significant in patients who had non-severe HSR to LOCM.
The recurrence of HSR was evaluated based on the presence of the N-(2,3-dihydroxypropyl) carbamoyl side chain. Among cases that used an alternative LOCM on re-exposure, the reaction rate was 15.1% when the alternative LOCM had an identical side chain to the culprit LOCM. This was slightly higher than the 11.8% rate observed when the alternative LOCM had a different side chain than the culprit LOCM (Figure 3-B). However, this did not reach statistical significance (P -value = 0.428). When cases were divided into two groups according to the severity of their index HSR, analysis revealed that, in patients that had severe index HSRs, the use of an alternative LOCM with a different side chain significantly reduced the BTR rate from 24.0% to 7.8% (P = 0.049). However, this difference was not observed in those who had a non-severe HSR to LOCM. In addition, changing between a monomer and dimer did not demonstrate an advantage in reducing the risk of HSR (Data not shown).