Association between serum ferritin and the severity of drug eruptions

Abstract Background Early recognition and treatment for severe drug eruption are important in improving clinical outcomes. A few studies have reported laboratory parameters to evaluate the severity of drug eruptions. This study aimed to determine the association between serum ferritin and the severity of drug eruptions. Methods We retrospectively reviewed patients diagnosed with drug eruptions in our hospital from 2013 to 2018. Results We identified 85 patients (mean age 53.4 years), 20 in the severe cutaneous adverse drug reactions (SCADRs) group and 65 in the non‐SCADRs group. Serum ferritin level was higher in the SCADRs group compared with that in the CADRs group (P<.001). Serum ferritin was positively associated with peripheral white blood cell count, aspartate aminotransferase level, alanine aminotransferase level, blood glucose level, blood creatinine level, and body temperature. Receiver operating characteristic (ROC) analysis revealed a good diagnostic value of ferritin (area under the curve [AUC]:0.87, 95% confidence interval [CI]:0.78‐0.96) with a sensitivity of 80% and a specificity of 87.7% at a cutoff value of 416.15 ng/mL. Conclusions Serum ferritin is significantly associated with the severity of CADRs and hence might be potentially used to evaluate the severity of CADRs.

activation. [7][8][9] Increased ferritin level is also observed in alcoholism, cytolysis (including hepatic and muscle cytolysis), metabolic syndrome, juvenile idiopathic arthritis, and adult-onset Still's disease. [10][11][12] One study revealed that ferritin could be a useful early prognostic factor in DRESS. 13 Thus, we determine the association between serum ferritin and the severity of drug eruptions.

| MATERIAL S AND ME THODS
We conducted a retrospective study of in 166 inpatients who were diagnosed with drug eruptions in the Department of Dermatology from July 31, 2013, to July 31, 2018. Patients with iron deficiency, iron deficiency anemia, and iron overload were excluded. Other exclusion criteria were as follows: conditions that could potentially affect the body's iron storage or ferritin including any acute or chronic inflammatory conditions, hemorrhagic disorders, pregnancy, hemoglobinopathies, cancer, diabetes, and liver disease. Patients who met the inclusion criteria were divided into the following two groups: the SCADRs group and the non-SCADRs group. The SCADRs group included patients who were diagnosed with SJS/TEN, DRESS/HSS, and AGEP, while the non-SCADRs group was consisted of patients who were diagnosed with MPE, FDE, lichenoid eruptions, and urticaria. Diagnoses were established according to the respective diagnostic criteria. 14 This study protocol was approved by the local ethics committee of our hospital.
We searched data from the medical records, including clinical symptoms, lesion distribution and morphology, laboratory findings, comorbid diseases, and medications. Some clinical variables that were retrieved and recorded were as follows: pulse rate (PR), body temperature (BT), peripheral white blood cell (WBC) count, eosinophil count, serum ferritin level, serum albumin (ALB) level, aspartate aminotransferase (AST) level, alanine aminotransferase (ALT) level, blood glucose (GLU) level, blood creatinine (CR) level, and blood urea nitrogen (BUN) level. Chemiluminescence microparticle immuno assay was used to assess serum ferritin level. The modified systemic inflammatory response syndrome (mSIRS) was used to evaluate the inflammatory status. It included WBC count, PR, BT, and respiratory rate, with the lowest score of 0 and highest score of 3 in each parameter.

| Statistical analysis
Statistics analysis was performed using SPSS ver. 19 statistical software. Intergroup comparison of the difference was performed using Mann-Whitney U test. The association between ferritin and other variables was performed using Pearson's correlation analysis. The association between ferritin level and mSIRS score was executed by Spearman's rank correlation test. The diagnostic value of ferritin level was evaluated using the receiver operating characteristic (ROC) curve. The test sensitivity, specificity, and cutoff value were estimated according to the Youden criteria. P value <.05 was considered to be statistically significant.

| RE SULTS
Eighty-five patients aged was 53.4 ± 17.39 years (mean ± SD), were included in our study, with 81 patients excluded. Twenty patients were part of the SCADRs group, and 65 patients were part of the non-SCADRs group ( Figure 1). There was no significant difference in age and sex between the two groups. All laboratory tests including ferritin level were performed at initial presentation to our hospital. Based on our observation, higher level of serum ferritin was associated with a more severe outcome. We presumed that ferritin level could be potentially used to evaluate the severity of the disease condition. Thus, a ROC analysis was performed to assess serum ferritin level. According to the analysis, the AUC was 0.87 (95% CI:0.78-0.96). Serum ferritin level had a sensitivity of 80% and a specificity of 87.7% with a cutoff value of 416.15 ng/mL according to the Youden criteria ( Figure 4).  Early identification of CADRs and initiation of appropriate treatment are helpful in achieving better clinical outcomes and in reducing inpatient fee. Our study revealed that serum ferritin level has a predictive value for the severity of drug eruptions in the early stage. It was significantly higher in the SCADRs group compared with the non-SCADRs group, a result that is consistent with some case reports and studies showing the association between hyperferritinemia and drug eruptions in SJS and DRESS. [15][16][17] The association between ferritin and some variables, which may reflect the severity of drug eruptions and physical condition of patients, was observed. A ROC analysis revealed that serum ferritin had high sensitivity and specificity in predicting SCADRs.

| D ISCUSS I ON
The mean age of patients was 53.4 years, which is considered an advanced age and could be due to the fact that elderly people are more susceptible to various diseases and have higher chances of taking medicine; thus, drug eruption is more common in this age group than other age groups. It showed that BT of patients was higher in the SCADRs group than in the non-SCADRs group, and ALB level was lower in the SCADRs group compared with the non-SCADRs group. Patients with SCADRs had inflammation, hence increasing their BT. In this process, the SCADRs group consumed higher ALB level than the non-SCADRs group. Ferritin level was significantly associated with mSIRS score, which is one of the indicators of systemic inflammation. The SIRS score was used to assess the mortality rate of patients in the intensive care units. Our findings were also supported by several reports showing that ferritin was associated with inflammation.
Although the findings of our study were encouraging, the mechanism underlying the increase of serum ferritin level in drug eruptions was unclear. Drug eruptions are usually considered as immunologically mediated reactions, and the immunopathogenesis involves several inflammatory and cytotoxic mediators. 18,19 On the contrary, there are studies that showed that serum ferritin level increased in the inflammation. 20,21 Hence, it probably was caused by the different inflammatory mediators such as cytokines, which were associated with the drug eruptions, specifically the severe forms of drug eruptions. 18,22 The limitation of our study was that this is a retrospective analysis with limited number of patients. Thus, a prospective study including more patients to identify the sequence of ferritin level increase and SCADRs occur  In conclusion, our study showed that serum ferritin levels were positively associated with the severity of CADRs at the time of diagnosis.
Ferritin detection might be a screening tool for the severity of CADRs.

ACK N OWLED G M ENTS
This work was supported by the National Natural Science Foundation of China (grant number 81673045).

CO N FLI C T O F I NTE R E S T
The authors declared no conflicts of interest.