Discussion

In this retrospective observational study, we tested whether premedication with dexamethasone is effective in preventing trastuzumab-associated IRR in patients with HER2-positive breast cancer based on electronic medical record data. To the best of our knowledge, this study is the first to show that dexamethasone premedication is effective in preventing IRRs caused by trastuzumab.
In our study, premedication with dexamethasone reduced the incidence of IRR (final model OR=0.616; 95%CI, 0.440-0.861; p =0.005) (Figure 3). Hemophagocytic lymphohistiocytosis occurring in rheumatic disease (macrophage activation syndrome) is treated with glucocorticoids, IL-1 blockade, or cyclosporine A.48,49 Glucocorticoids remain the first-line drugs for adult-onset Still’s disease characterized by excessive cytokine production.50Thus, the results of this study showing that dexamethasone was effective in preventing IRRs caused by trastuzumab support the hypothesis that IRRs are cytokine-dependent. In a previous observational study, IRRs were prevented when all patients received an intravenous histamine H1 receptor antagonist (5 mg of dichlorpheniramine) as a premedication and 6.6 mg premedication with dexamethasone as an antiemetic agent for cetuximab therapy.51 The dexamethasone premedication used in the patients in our study was also intended as an antiemetic for other chemotherapy regimens administered in combination with trastuzumab, at doses of 6.6 mg or 8.25 mg. These suggest that doses such as those used as antiemetics are effective in reducing IRR risk with cetuximab and trastuzumab. There is little information on the addition of premedication to regimens that do not traditionally require premedication, such as trastuzumab monotherapy.39 Because unnecessary premedication may result in dexamethasone-induced adverse events, it is necessary to identify patients at high IRR risk and determine the need for dexamethasone premedication. For example, preoperative status and high-dose trastuzumab (i.e., first course patients) were identified as risk factors for trastuzumab-induced IRR, and these patients may require premedication with dexamethasone. A previous observational study revealed no differences in IRR prophylaxis for cetuximab between dexamethasone doses of 6.6 mg and 13.2 mg.51 Our results show that the OR was 0.616 per 1 mg increase in dexamethasone, which indicates that if 8 mg were premedicated, the OR would be 0.02, or a 1 in 50 risk of IRR occurrence. Further studies are needed to determine the optimal dose of dexamethasone to avoid unnecessary exposure.
In the current study, the risk of IRR in patients with breast cancer was higher in preoperative patients and in patients receiving high trastuzumab doses (Figure 3). The frequency of IRRs after rituximab administration is clearly higher in patients with tumors than in patients with rheumatoid arthritis.30,31Administration of rituximab to patients with a large number of tumor cells in the blood may increase the likelihood of a severe initial IRR.33,34 In a study investigating IRRs in response to rituximab in patients with B-cell lymphoma, multivariate logistic regression analysis confirmed that low-grade lymphomas (OR=2.81;p =0.017) and bulky disease (OR=2.52; p =0.037) were independent risk factors.52 The IRR risk factor identified in the present study, preoperative status, reflects a high tumor count, and a higher dose of trastuzumab indicates more tumors to be destroyed. In other words, these reports are consistent with the results of the present study, as they indicate patients with high cytokine release. Conversely, patients who received postoperative adjuvant chemotherapy have fewer tumor cells than preoperative patients and therefore do not release as many cytokines; postoperative immune dysfunction may be the reason for their low IRR risk.
Of all factors evaluated in a study by Thompson et al., high BMI, stage IV, and no prior medication use (diphenhydramine, meperidine, or hydrocortisone) were significantly associated with higher risks of trastuzumab-induced IRRs in breast cancer patients.39In contrast, in our results, BMI was not associated with IRR risk. We attribute the differences between these studies to differences in BMI distribution in our cohort. The numbers of patients with BMI (kg/m2) <18.5, 18.5–24.9, 25.0–29.9, and 30≤ were 0, 60, 55, and 82, respectively. In the previous study, they were 17, 113, 36, and 10, respectively, at baseline. The lower distribution of BMI in the present study may account for the disparities in results. Cochran-Armitage trend testing in our study also showed a statistically significant trend towards increased IRR incidence with advancing stage (Table 2). Furthermore, stage IV (i.e. metastasis) was associated with higher risk of IRR in univariate analysis (Figure 2) but not statistically significant in multivariate analysis. There are several possible reasons for this. First, metastasis was not detected because of the higher IRR risk of preoperative or trastuzumab dosage used as variables in this study. Second, the variables used in the multivariate analysis were different. Third, the patient background, such as BMI, was different. However, the details underlying this finding are not known, and the IRR risk in patients with stage IV breast cancer needs further investigation.
In the present study, high eosinophil levels were a risk factor for IRR occurrence in univariate analysis (Figure 2), but no statistically significant association was found when adjusted for other factors by multivariate analysis (Figure 3). In a previous study, low eosinophil levels were associated with trastuzumab-induced IRR in breast cancer patients.53 High eosinophil counts were a risk factor for cetuximab-induced IRR in patients with squamous cell carcinoma of the head and neck.42 Therefore, the influence of blood eosinophils on the development of IRRs caused by monoclonal antibody preparations shows conflicting results. In addition, eosinophil levels being used as a baseline characteristic in our study may have been a factor in the lack of associations with increased IRR risk. Because eosinophil data was not available for each infusion in our study, we were unable to obtain eosinophil levels at a consistent time point prior to each trastuzumab infusion, rather than at baseline, to assess the role of eosinophils in IRR development as a micro-level variable.
Although the risk of IRR is clearly higher with monoclonal antibody preparations at the first dose,37,39 course was not associated with the development of IRRs in during multivariate analysis (Figure 3). Since the dose of trastuzumab treatment was higher at the first dose (8 mg/kg) than at the maintenance dose (6 mg/kg), the same patient is at higher risk for IRRs with the first dose.
In this retrospective study of breast cancer patients treated with trastuzumab, the overall incidence of IRR was 3.4% of all infusions. Most IRRs occurred during the first dose (53 of 58 patients, 91.4%). The incidence of IRRs was higher with the first course and the first dose of 8 mg/kg, which is consistent with previous reports.3 Previous studies have reported a relatively high incidence of IRRs with trastuzumab (40%,23.4%,24 or 5.9%).25 Although the incidence of IRR in our study was on the low side, the wide range in IRR incidence between different reports seems to result from differences in comorbidities or premedication use.37 The use of trastuzumab has expanded to salivary gland and colorectal cancers, and the differences in IRR occurrence by disease state should be examined in the future.
There are several limitations to our study. First, we cannot distinguish whether the IRRs that occurred were due to trastuzumab or other chemotherapy. To minimize this effect, we qualified IRRs only as those symptoms that developed within 120 minutes of the start of trastuzumab infusion, based on reports that most trastuzumab IRRs occur during trastuzumab administration or within 2 hours after initiation.2 Second, it is not known from our study whether dexamethasone premedication affects the efficacy with trastuzumab. While previous studies have shown that pretreatment with glucocorticoids does not affect the efficacy of rituximab at 24 weeks in the treatment of rheumatoid arthritis,54 there is no similar evidence for trastuzumab for breast cancer. However, studies using BT-474 breast cancer cells suggest that dexamethasone at least partially inhibits the growth-suppressing effects of trastuzumab;55 therefore, clinical evaluations are needed.
In conclusion, to investigate how to prevent trastuzumab-induced IRR in breast cancer, we analyzed a model adjusted for patient background and found that premedication with dexamethasone is effective in preventing trastuzumab-induced IRR. However, the current lack of information on the risks of dexamethasone-induced adverse events and its effect on trastuzumab efficacy makes it impractical to unreservedly recommend premedication with dexamethasone based solely on this study. It is essential to select patients at high risk for IRR for dexamethasone premedication, such as those of preoperative status and who are receiving high trastuzumab doses based on the results of this study. In addition, dexamethasone should be limited to a minimal dose to lower IRR risk and avoid prolonged infusion time, and dexamethasone premedication should not be used in patients at high risk for dexamethasone-associated adverse events. Future studies are needed to determine the optimal dose of dexamethasone to prevent IRRs and the impacts of dexamethasone on the efficacy of trastuzumab in breast cancer.