If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
1 They both contributed equally to this article as co-corresponding authors.
Seoung Ju Park
Correspondence
Corresponding author. Department of Internal Medicine, Jeonbuk National University Medical School, Research Institute of Clinical Medicine of Jeonbuk National University-Biomedical Research Institute of Jeonbuk National University Hospital, 20, Geonji-ro, Deokjin-gu, Jeonju, 54907, Korea
1 They both contributed equally to this article as co-corresponding authors.
Sae-Hoon Kim
Correspondence
Corresponding author. Department of Internal Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173beon-gil, Bundang-gu, Seongnam, 13620, Korea
1 They both contributed equally to this article as co-corresponding authors.
Affiliations
Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, KoreaDepartment of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea
Although beta-lactams are 1 of the major causative agents of severe cutaneous adverse reactions (SCAR), their epidemiology and clinical aspects have been poorly studied. This study aimed to investigate the characteristics of SCAR caused by beta-lactams in the Korean SCAR registry.
Methods
We retrospectively analyzed beta-lactam-induced SCAR cases collected from 28 tertiary university hospitals in Korea between 2010 and 2015. The SCAR phenotypes included Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), SJS-TEN overlap, and drug reaction with eosinophilia and systemic symptoms (DRESS). Beta-lactams were classified according to their chemical structures: penicillins, cephalosporins, and carbapenems. The causative beta-lactams, clinical and laboratory features, treatments, and outcomes were evaluated.
Results
Among the 275 antibiotic-induced SCAR cases, 170 patients developed SCAR induced by beta-lactams. Beta-lactam antibiotic-induced SCAR showed more frequent SJS/TEN compared to SCAR induced by non-beta-lactam antibiotics (SJS/TEN/SJS-TEN overlap/DRESS: 36.5/11.2/5.9/46.5% vs. 23.8/10.5/2.9/62.9%, P = 0.049). Cephalosporin was the most common culprit drug. Particularly, 91 and 79 patients presented with SJS/TEN and DRESS, respectively. The odds ratio (OR) for poor prognosis, such as sequelae and death, was significantly increased in subjects with SJS-TEN overlap and TEN and carbapenem as culprit drug in the multivariate analysis (OR, 35.61; P = 0.016, OR, 28.07; P = 0.006, OR 30.46; P = 0.027).
Conclusion
Among antibiotic-induced SCAR, clinical features were different depending on whether the culprit drug was a beta-lactam antibiotic or SCAR type. The poor prognosis was related to SJS-TEN overlap, TEN type, and carbapenem as the culprit drug.
Drug hypersensitivity reactions are common obstacles that interfere with patient care. These reactions are caused by immunological mechanisms that occur in approximately 2.3–3.6 cases per 1000 patients.
Although most drug eruptions are mild maculopapular exanthema, they also include severe phenotypes, such as Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), and drug-induced hypersensitivity syndrome (DIHS)/drug reaction with eosinophilia and systemic symptoms (DRESS), termed severe cutaneous adverse drug reactions (SCAR).
The clinical features of SCAR are diverse. They may have various types of skin involvement with characteristic mucocutaneous involvement. In contrast, DIHS/DRESS has almost no skin detachment and is mainly accompanied by internal organ involvement and hematological abnormalities, such as fever, hepatitis, and eosinophilia. However, there can be overlap in some of the clinical features of SJS/TEN and DRESS, and there is no clear boundary.
In the case of SJS/TEN, there may be skin scarring and ophthalmic, genitourinary, and respiratory complications, and in the case of DIHS, end organ failure, and autoimmune disease may develop.
Efficacy of the HLA-B(∗)58:01 screening test in preventing allopurinol-induced severe cutaneous adverse reactions in patients with chronic renal insufficiency-A prospective study.
However, there are insufficient studies on other factors related to SCAR occurrence and prognosis.
Beta-lactam antibiotics, including penicillins and cephalosporins, are the most commonly used antibiotics for infectious diseases in clinical practice and the most common cause of drug allergy.
Nevertheless, few studies have been conducted on beta-lactam antibiotic-induced SCAR; therefore, we aimed to study the clinical characteristics and factors related to poor prognosis of SCAR caused by beta-lactam antibiotics.
Methods
Study subjects and resources
This study was based on data collected from the Korean SCAR registry, these data were described in previous studies.
The Korean SCAR registry is web-based and originated from the Regional Pharmacovigilance Center (RPVC) of the Korean Food and Drug Administration (KFDA). Currently, 36 tertiary hospitals are participating, and at least 1 allergist reviews cases at each institution participating in this registry.
Briefly, the Korean SCAR registry retrospectively collected drug allergy data from 36 tertiary hospitals nationwide from 2010 to 2015. Each case was registered after reviewing medical records by 2 physicians, including at least 1 allergist, and the definition of SCAR was based on the same criteria as in previous studies. The causal relationship between the suspected drug and SCAR was evaluated using the World Health Organization–Uppsala Monitoring Center (WHO-UMC) criteria.
The WHO-UMC criteria is based on 5 items: plausible time relationship between drug intake and events, cannot be explained by disease or other drugs, response to withdrawal plausible drug, event definitive pharmacologically or phenomenologically, and rechallenge satisfactory. Causalities were then classified into 4 stages: certain, probable, possible, and unlikely (Supplementary Table 1).
Cases with possible or higher levels of causality were included.
If more than 2 possible causal drugs were used simultaneously, all suspected drugs assigned as culprit drugs based on expert judgment. Demographic information, hospitalization status, vital signs, laboratory tests, medical history, drug history, and clinical course were collected from a review of medical records.
Institutional Review Board (IRB) approval was obtained from each research participating institution (IRB approval number of the representative institution, Seoul National University Bundang Hospital is B-1802-450-401.).
Evaluation of clinical data on beta-lactam SCAR
The cases of SCAR caused by antibiotics were divided into cases of beta-lactam antibiotic-induced SCAR and cases of non-beta-lactam antibiotic-induced SCAR. Within beta-lactam antibiotic-induced SCAR cases, penicillins, cephalosporins, and carbapenems-were classified according to the culprit drug classifications. We evaluated whether there was a difference in the clinical characteristics and prognosis according to the beta-lactam or non-beta-lactam antibiotics and SCAR types.
Statistical analysis
Statistical analysis was performed using SPSS software (version 22.0; SPSS, IBM Inc., Chicago, IL, USA). Continuous variables were presented as means ± standard deviation or median interquartile range and analyzed using the t-test or Mann-Whitney U test. Categorical variables were presented as numbers or percentages and were analyzed using Pearson's χ2 test or Fisher's exact test. A P-value <0.05 indicated statistical significance. Binary logistic regression analysis was performed to find factors related to the prognosis of SCAR, and statistically significant factors in univariate analysis and factors judged to be clinically important, such as age and sex, were adjusted.
Results
Demographic and clinical characteristics
Data of 275 patients with SCAR due to antibiotics were extracted, of which 170 were classified as SCAR due to beta-lactam antibiotics and 105 as SCAR due to non-beta-lactam antibiotics (Table 1, Fig. 1A). Patients with beta-lactam antibiotic-induced SCAR had a mean age of 50.8 years, and 55.9% were male. Hypertension was the most common comorbidity, followed by diabetes mellitus, cancer, chronic kidney disease, and chronic liver disease (Table 1). On average, 72.4% of the body surface area (BSA) was affected. Mucosal involvement was detected in 61.6% of patients, fever in 62.3%, and lymphadenitis in 14.9%. The median disease duration was 18.0 (Interquartile range, IQR 13.0–27.0) days, the administration duration of culprit drugs was 8.0 (IQR 3.0–17.0) days, latent period which is the period form the antibiotics intake to the onset of reaction was 9.0 (IQR 1.0–22.0) days, and admission duration was 17.0 (IQR 10.0–35.5) days. In 14.6% of the patients, there was a history of exposure to the suspected drug. As treatment, 78.3% received systemic corticosteroids, 19.0% received intravenous immunoglobulin (IVIG), and 4.7% received intensive care. The mean SCORTEN score was 1.0 on day 1 and 1.7 on day 7. However, 85.7% of patients recovered without sequelae, sequelae remained in 7.7%, and 6.5% of patients died.
Table 1Demographic and clinical characteristics of antibiotics-induced SCAR patients.
Total
Non-beta-lactam
Beta-lactam
P-value
Number of cases
275
105 (38.2%)
170 (61.8%)
Age
53.3 ± 20.6
54.7 ± 21.5
50.8 ± 22.1
0.158
Male %
146 (53.1%)
51 (48.6%)
95 (55.9%)
0.264
BMI
22.6 ± 3.7
22.8 ± 4.0
21.6 ± 3.3
0.022
Smoking history (Non-/Ex-/Current smoker)
177/20/23 (80.5/9.1/10.5%)
77/6/7 (85.6/6.7/7.8%)
100/14/16 (76.9/10.8/12.3%)
0.298
33/238 (13.9%)
18/88 (20.5%)
15/150 (10.0%)
0.032
History of drug allergy
18/191 (9.4%)
11/76 (14.5%)
7/108 (6.1%)
0.075
Comorbidities
Diabetes mellitus
44/246 (17.9%)
15/93 (16.1%)
29/153 (19.0%)
0.611
Hypertension
81/245 (33.1%)
31/93 (33.3%)
50/152 (32.9%)
1.000
Chronic liver disease
12/236 (5.1%)
4/92 (4.3%)
8/144 (5.6%)
0.770
Chronic kidney disease
16/229 (7.0%)
6/89 (6.7%)
10/140 (7.1%)
1.000
Cancer
24/229 (10.5%)
6/87 (6.9%)
18/142 (12.7%)
0.189
Admission route
SCAR onset during hospitalization (%)
71/273 (26.0%)
25 (23.8%)
46 (27.4%)
0.572
Via OPD (%)
76/273 (27.8%)
34 (32.4%)
42 (25.0%)
0.212
Via ER (%)
126/273 (46.2%)
46 (43.8%)
80 (47.6%)
0.618
SCAR type (SJS/TEN/SJS-TEN overlap/DRESS or DHS)
87/30/13/145 (31.6/10.9/4.7/52.7%)
25/11/3/66 (23.8/10.5/2.9/62.9%)
62/19/10/79 (36.5/11.2/5.9/46.5%)
0.049
Presenting symptoms
Skin involvement, BSA (%)
74.2 ± 30.2
76.9 ± 29.7
72.4 ± 30.5
0.246
Mucosal involvement
115/204 (56.4%)
38/79 (48.1%)
77/125 (61.6%)
0.062
Fever
173/261 (66.3%)
74/102 (72.5%)
99/159 (62.3%)
0.107
Lymphadenitis
21/119 (17.6%)
10/45 (22.2%)
11/74 (14.9%)
0.330
Highest body temperature
39.0 ± 0.7
39.1 ± 0.7
39.0 ± 0.8
0.432
Duration of fever
6.2 ± 7.2
7.3 ± 8.8
5.3 ± 5.5
0.093
Highest WBC count
18653.0 ± 27901.0
24258.2 ± 42531.7
14847.0 ± 7506.7
0.110
Highest eosinophil count
1493.9 ± 2083.8
1612.3 ± 2023.0
1408.7 ± 2131.2
0.501
Highest creatinine level
2.0 ± 2.2
2.3 ± 2.6
1.7 ± 1.8
0.130
Highest ALT level
211.4 ± 286.1
241.7 ± 300.2
190.5 ± 275.1
0.205
Administration of culprit (days)
11.0 (IQR 4.0–21.0)
17.0 (IQR 7.0–27.0)
8.0 (IQR 3.0–17.0)
0.140
Duration of drug administration after symptom onset (days)
3.0 (IQR 2.0–5.0)
4.0 (IQR 2.0–9.0)
2.0 (IQR 1.0–4.0)
0.006
Latent period
11.0 (IQR 2.0–23.0)
13.0 (IQR 6.0–26.0)
9.0 (IQR 1.0–22.0)
0.314
Disease duration (days)
20.0 (IQR 14.0–30.0)
22.0 (IQR 14.0–34.0)
18.0 (IQR 130–27.0)
0.640
Admission duration (days)
18.0 (IQR 11.0–33.0)
18.0 (IQR 11.0–32.5)
17.0 (IQR 10.0–35.5)
0.673
SJS
13.0 (IQR 7.75–26.25)
14.0 (IQR 6.5–28.0)
13.0 (IQR 8.0–20.0)
0.904
SJS/TEN Overlap
26.0 (IQR 18.0–39.0)
27.0 (IQR 26.0–27.0)
25.0 (IQR 17.25–39.0)
0.346
TEN
32.0 (IQR 14.0–51.5)
26.0 (IQR 16.0–40.0)
34.5 (IQR 13.0–56.0)
0.715
DRESS or DHS
18.0 (IQR 11.0–37.0)
18.0 (IQR 11.75–32.25)
19.0 (IQR 11.0–39.0)
0.912
Past exposure to culprit drug
16/144 (11.1%)
4/62 (6.5%)
12/82 (14.6%)
0.180
Treatment
Systemic steroid
216/268 (80.6%)
86/102 (84.3%)
130/166 (78.3%)
0.267
IVIG
46/250 (18.4%)
16/92 (17.4%)
30/158 (19.0%)
0.866
Other immunosuppressant
5/248 (2.0%)
0/91 (0.0%)
5/157 (3.2%)
0.161
ICU care
13/275 (4.7%)
5/105 (4.8%)
8/170 (4.7%)
1.000
SCORTEN Day 1 (total 7)
1.0 ± 0.9
0.9 ± 1.0
1.0 ± 0.8
0.681
SCORTEN Day 7 (total 7)
1.8 ± 0.9
1.7 ± 1.0
0.843
Prognosis (%)
0.588
Improved
228/272 (83.8%)
84/104 (80.8%)
144/168 (85.7%)
With sequelae
24/272 (8.8%)
11/104 (10.6%)
13/168 (7.7%)
Death (%)
20/272 (7.4%)
9/104 (8.75)
11/168 (6.5%)
SCAR, severe cutaneous adverse reaction; BMI, body mass index; OPD, out-patient department; ER, emergency room; SJS, Stevens-Johnson syndrome; overlap, combined Stevens-Johnson syndrome and toxic epidermal necrolysis; TEN, toxic epidermal necrolysis; DRESS, drug reaction with eosinophilia and systemic symptoms; DHS, drug hypersensitivity syndrome; BSA, body surface area; WBC, white blood cells; Cr, creatinine; ALT, alkaline phosphatase; IQR, Interquartile range; IVIG, intravenous immunoglobulin; ICU, intensive care unit; SCORTEN, SCORe of Toxic Epidermal Necrosis
Compared with patients with non-beta-lactam antibiotic-induced SCAR, patients with beta-lactam antibiotic-induced SCAR had a lower body mass index (BMI) (21.6 ± 3.3 vs. 22.8 ± 4.0, P = 0.022), and fewer allergic diseases (10.0% vs. 20.5%, P = 0.032); SJS/TEN was more common, and DRESS was less in SCAR type (SJS/TEN/SJS-TEN overlap/DRESS 36.5/11.2/5.9/46.5% vs. 23.8/10.5/2.9/62.9%, P = 0.049). The duration of drug administration after symptom onset was shorter in beta-lactam antibiotic-induced SCAR (3.0 vs. 7.0 days, P = 0.006).
Among the 170 patients with beta-lactam antibiotic-induced SCAR, the most common suspected drug class was cephalosporins, especially third-generation cephalosporins, and the second most common suspected class was penicillins, especially aminopenicillins (Fig. 1B ). SJS/TEN was more frequent in penicillins and DRESS was more frequent in carbapenems, but the differences were not statistically significant (Fig. 1C).
Fig. 1The flowchart of case classification (A), distribution of subjects (B) and the proportions of SCAR types (C) according to culprit beta-lactam antibiotic class
In the analysis of SCAR types, patients with DRESS had the highest BMI (SJS/SJS-TEN overlap/TEN/DRESS 21.0/21.6/20.2/22.6, P = 0.021), and other allergic diseases were most frequently accompanied by TEN (26.7%, P = 0.042) (Table 2). Mucosal involvement was lowest in patients with DRESS (SJS/SJS-TEN overlap/TEN/DRESS 86.8, 88.0, 92.5, 21.7, P < 0.001), and BSA was highest in patients with TEN (48.3%, P < 0.001). As expected, subjects with DRESS more frequently showed fever, lymphadenitis and higher white blood cell (WBC) count, eosinophil count, and alanine aminotransferase (ALT) levels than those with other SCAR phenotypes. Conversely, the duration of fever was longest in patients with TEN, and serum creatinine (Cr) level was highest in patients with SJS-TEN overlap. The duration of drug administration even after symptom onset and latent period were longest in patients with DRESS as 14.6 days (P < 0.041) and 24.7 days (P < 0.001), respectively. The disease duration and hospitalization days were longer in patients with SJS-TEN overlap and TEN, at 40.4 days (P < 0.011) and 33.2 days (P < 0.001), respectively. Systemic corticosteroids were used more frequently in patients with SJS/TEN than in those with DRESS (SJS/SJS-TEN overlap/TEN/DRESS 93.2/100.0/94.9/77.1, P < 0.001), and IVIG was administered to 51.9% of patients with TEN (P < 0.001). Although the SCORTEN score on day 1 was highest in patients with SJS-TEN overlap (1.6 ± 0.7 [P = 0.010]), the SCORTEN score on day 7 was highest in patients with TEN (2.2 ± 1.2 [P = 0.017]), and the proportion of patients with a score of ≥2 on day 7 was also highest in patients with TEN (SJS/SJS-TEN overlap/TEN/DRESS 51.3, 68.6, 76.8, 62.0, P < 0.001). The most favorable prognosis was in patients with DRESS, with 92.4% of recovered patients and the worst prognosis was in patients with TEN, with 26.3% of the complications or death (P < 0.001). SCAR types and prognosis according to WHO-UMC causality evaluation was provided in Supplementary Table 2.
Table 2Demographics and clinical presentations of beta-lactam antibiotics-induced SCARs according to phenotypes.
SJS (n = 62)
Overlap (n = 10)
TEN (n = 19)
DRESS (n = 79)
P-value
Age (median, year)
46.1 ± 25.4
53.3 ± 24.7
52.1 ± 25.2
53.9 ± 17.7
0.479
Sex (% of males)
59.7
50.0
57.9
53.2
0.860
BMI (kg/m2)
21.0 ± 3.3
21.6 ± 2.6
20.2 ± 4.9
22.6 ± 2.7
0.021
Other allergic disease (%)
3.8
0.0
26.7
12.5
0.042
Other drug hypersensitivity (%)
4.8
0.0
9.1
7.5
0.856
Comorbidities
Chronic kidney disease
6.7
11.5
14.3
10.0
0.190
Chronic liver disease
5.9
0.0
2.7
6.3
0.476
Diabetes mellitus
14.1
20.7
15.4
19.3
0.370
Hypertension
35.1
46.9
38.7
36.6
0.613
Cancer
7.3
3.3
13.9
11.4
0.164
Presenting symptoms
Mucosal involvement (%)
86.8
88.0
92.5
21.7
< 0.001
Involved BSA (%)
11.0 ± 28.7
19.9 ± 12.5
48.3 ± 28.7
14.1 ± 32.6
< 0.001
Fever (%)
48.0
56.3
63.6
67.7
< 0.001
Lymphadenitis (%)
3.1
6.7
9.3
28.3
< 0.001
Highest body temperature
38.9 ± 0.7
38.9 ± 0.6
39.1 ± 0.8
39.0 ± 0.7
0.231
Duration of fever (days)
3.9 ± 4.1
5.1 ± 4.3
9.2 ± 10.6
5.4 ± 6.2
< 0.001
Highest WBC count
13494.2 ± 7079.5
12098.8 ± 6570.4
14645.7 ± 7224.6
20165.7 ± 24518.5
< 0.001
Highest eosinophil count
868.4 ± 1462.1
797.7 ± 1175.4
894.7 ± 1906.9
2376.7 ± 3513.1
< 0.001
Highest serum Cr (mg/mL)
1.4 ± 1.2
2.4 ± 2.0
2.2 ± 2.0
2.2 ± 2.3
0.001
Highest serum ALT (IU/L)
172.7 ± 272.6
110.7 ± 116.3
197.1 ± 324.5
320.1 ± 506.8
< 0.001
Culprit drugs
0.600
Penicillins
27 (42.9%)
4 (6.3%)
8 (12.7%)
24 (38.1%)
Cephalosporins
31 (33.3%)
6 (5.9%)
10 (9.8%)
52 (51.0%)
Carbapenems
1 (20.0%)
0 (0.0%)
1 (20.0%)
3 (60.0%)
Administration of culprit (days)
19.0 ± 30.1
4.9 ± 3.6
4.0 ± 4.2
18.6 ± 14.3
0.268
Duration of drug administration after symptom
6.7 ± 9.1
3.3 ± 2.4
8.0 ± 12.3
14.6 ± 80.0
0.041
Latent period (days)
17.8 ± 25.8
18.7 ± 28.7
18.8 ± 28.2
24.7 ± 23.8
< 0.001
Disease duration (days)
18.9 ± 11.0
40.4 ± 38.7
27.9 ± 12.9
27.6 ± 28.9
0.011
Admission duration (days)
15.7 ± 12.6
25.5 ± 17.5
33.2 ± 25.4
24.5 ± 26.7
< 0.001
Previous exposure to culprit (%)
9.9
31.3
13.0
9.5
0.075
Treatment
Systemic steroid (%)
93.2
100.0
94.9
77.1
< 0.001
IVIG (%)
15.7
38.2
51.9
8.0
< 0.001
Other immunosuppressant
5.7
2.9
6.8
2.1
0.054
ICU care
4.6
17.6
26.3
6.3
< 0.001
SCORTEN Day 1 (total 7)
0.9 ± 0.8
1.6 ± 0.7
1.4 ± 0.9
0.9 ± 0.8
0.010
SCORTEN Day 7 (total 7)
1.6 ± 1.0
1.2 ± 0.6
2.2 ± 1.2
1.8 ± 0.9
0.017
SCORTEN Day 1 (% of ≥ 2)
23.2
28.6
36.6
24.1
0.092
SCORTEN Day 7 (% of ≥ 2)
51.3
68.6
76.8
62.0
< 0.001
Prognosis (%)
< 0.001
Improved
90.2
77.8
47.4
92.4
With sequelae
6.6
22.2
26.3
2.5
Death
3.3
0.0
26.3
5.1
SCAR, severe cutaneous adverse reaction; BMI, body mass index; OPD, out-patient department; ER, emergency room; SJS, Stevens-Johnson syndrome; overlap, combined Stevens-Johnson syndrome and toxic epidermal necrolysis; TEN, toxic epidermal necrolysis; DRESS, drug reaction with eosinophilia and systemic symptoms; DHS, drug hypersensitivity syndrome; BSA, body surface area; WBC, white blood cells; Cr, creatinine; ALT, alkaline phosphatase; IVIG, intravenous immunoglobulin; ICU, intensive care unit; SCORTEN, SCORe of Toxic Epidermal Necrosis
Regarding the factors related to prognosis, 28.7% of fully recovered patients developed SCAR during hospitalization, whereas no patient with sequelae developed SCAR during hospitalization (P = 0.021). DRESS was more frequent in patients with full recovery; TEN and SJS-TEN overlap were more frequent in patients with sequelae (SJS/TEN/SJS-TEN overlap/DRESS 38.2,6.3,4.9,50.7%vs.30.8,38.5,15.4,15.4%, P = 0.001) (Table 3). Moreover, the latent period was shorter (16.4 ± 23.6vs.4.9 ± 7.3, P = 0.016), and the rates of IVIG use and ICU care were higher in patients with sequelae (12.7%vs.41.7%, P = 0.019, 1.5%vs.33.3%, P < 0.001).
Table 3Prognosis factors of full recovery, sequelae, and mortality.
Survival
Mortality
P-value
Full recovery
Sequelae
P-value
Number of cases
144 (91.7%)
13 (8.3%)
157 (93.5%)
11 (6.5%)
Age
50.6 ± 22.3
45.7 ± 19.0
0.447
50.2 ± 22.0
59.6 ± 24.3
0.173
Male %
80/144 (55.6%)
6/13 (46.2%)
0.570
86/157 (54.8%)
8/11 (72.7%)
0.350
21.9 ± 3.3
20.5 ± 2.8
0.177
21.8 ± 3.3
19.7 ± 3.6
0.053
Smoking history (Non-/Ex-/Current smoker)
15/12/83 (13.6/10.9/75.5%)
1/0/10 (9.1/0.0/90.9%)
0.740
16/12/93 (13.2/9.9/76.9%)
0/2/6 (0.0/25.0/75.0%)
0.195
Smoking history (pack-year)
4.0 ± 10.2
1.8 ± 6.0
0.483
3.8 ± 9.9
0.0 ± 0.0
<0.001
Allergic disease
12/126 (9.5%)
0/12 (0.0%)
0.600
12/138 (8.7%)
3/10 (30.0%)
0.066
History of drug allergy
6/97 (6.2%)
0/9 (0.0%)
1.000
6/106 (5.7%)
1/7 (14.3%)
0.369
Comorbidities
Diabetes mellitus
24/130 (18.5%)
1/12 (8.3%)
0.692
25/142 (17.6%)
4/10 (40.0%)
0.098
Hypertension
42/129 (32.6%)
3/12 (25.0%)
0.752
45/141 (31.9%)
5/10 (50.0%)
0.300
Chronic liver disease
8/121 (6.6%)
0/12 (0.0%)
1.000
8/133 (6.0%)
0/10 (0.0%)
0.650
Chronic kidney disease
6/117 (5.1%)
1/12 (8.3%)
0.504
7/129 (5.4%)
2/9 (22.2%)
0.107
Cancer
14/119 (11.8%)
1/12 (8.3%)
1.000
15/131 (11.5%)
3/10 (30.0%)
0.118
Admission route
SCAR onset during hospitalization (%)
41/143 (28.7%)
0/13 (0.0%)
0.021
41/156 (26.3%)
5/11 (45.5%)
0.177
OPD (%)
34/143 (23.8%)
5/13 (38.5%)
0.313
39/156 (25.0%)
3/11 (27.3%)
1.000
ER (%)
68/143 (47.6%)
8/13 (61.5%)
0.394
76/156 (48.7%)
3/11 (27.3%)
0.219
SCAR type (SJS/TEN/SJS-TEN overlap/DRESS or DHS)
55/9/7/73 (38.2/6.3/4.9/50.7%)
4/5/2/2 (30.8/38.5/15.4/15.4%)
0.001
59/14/9/75 (37.6/8.9/5.7/47.8%)
2/5/0/4 (18.2/45.5/0.0/36.4%)
0.015
Presenting symptoms
Skin involvement
70.4 ± 31.3
81.8 ± 26.7
0.206
71.4 ± 31.0
87.5 ± 19.9
0.036
Mucosal involvement
62/105 (59.0%)
10/12 (83.3%)
0.126
72/117 (61.5%)
4/7 (57.1%)
1.000
Fever
81/135 (60.0%)
7/12 (58.3%)
1.000
88/147 (59.9%)
10/11 (90.9%)
0.053
Lymphadenitis
9/61 (14.8%)
1/8 (12.5%)
1.000
10/69 (14.5%)
1/4 (25.0%)
0.487
SCORTEN
1.0 ± 0.8
1.2 ± 0.7
0.396
1.0 ± 0.8
1.4 ± 0.9
0.127
Highest body temperature
38.9 ± 0.7
39.5 ± 0.8
0.061
39.0 ± 0.7
39.0 ± 1.1
0.958
Duration of fever
5.0 ± 5.0
4.0 ± 2.2
0.651
5.0 ± 4.9
9.0 ± 10.6
0.355
Highest WBC count
14426.2 ± 7051.0
10865.0 ± 3997.8
0.322
18872.4 ± 27816.1
21957.1 ± 23487.1
0.775
Highest eosinophil count
1456.6 ± 2208.1
709.5 ± 711.7
0.503
1428.9 ± 2174.3
1044.0 ± 1144.9
0.669
Highest creatinine level
1.6 ± 1.5
2.3 ± 3.0
0.573
1.6 ± 1.6
2.9 ± 2.9
0.244
Highest ALT level
190.7 ± 283.0
157.6 ± 199.5
0.761
188.7 ± 278.2
219.6 ± 235.0
0.775
Administration of culprit (days)
18.9 ± 21.9
7.0 ± 7.5
0.133
17.8 ± 21.2
5.9 ± 5.0
0.042
Duration of drug administration after symptom onset (days)
3.17 ± 2.6
2.6 ± 2.1
0.641
3.1 ± 2.3
2.3 ± 1.4
0.407
Latent period (days)
16.4 ± 23.6
4.9 ± 7.3
0.016
15.4 ± 22.9
14.3 ± 19.8
0.873
Disease duration (days)
24.4 ± 22.1
30.6 ± 28.2
0.348
25.0 ± 22.6
40.2 ± 49.5
0.334
Admission duration (days)
26.8 ± 30.3
27.2 ± 29.1
0.961
26.8 ± 30.1
48.3 ± 50.6
0.193
Past exposure to culprit drug
8/68 (11.8%)
1/7 (14.3%)
1.000
9/75 (12.0%)
3/7 (42.9%)
0.061
Treatment
Systemic steroid
107/142 (75.4%)
12/12 (100%)
0.069
119/154 (77.3%)
9/10 (90.0%)
0.693
Duration of steroid use (days)
Total dose of steroid (mg)
IVIG
17/134 (12.7%)
5/12 (41.7%)
0.019
22/146 (15.1%)
8/10 (80.0%)
<0.001
Other immunosuppressant
4/134 (3.0%)
1/12 (8.3%)
0.353
5/146 (3.4%)
0/9 (0.0%)
1.000
ICU care
2/134 (1.5%)
4/12 (33.3%)
<0.001
6/146 (4.1%)
5/11 (45.5%)
<0.001
SCORTEN Day 1 (total 7)
0.9 ± 0.8
1.2 ± 0.7
0.396
1.0 ± 0.8
1.4 ± 0.9
0.127
SCORTEN Day 7 (total 7)
1.7 ± 0.8
1.5 ± 1.5
0.771
1.7 ± 0.9
2.9 ± 1.2
<0.001
SCORTEN Day 1 (% of ≥2)
32 (22.2%)
4 (30.8%)
0.497
36 (22.9%)
5 (45.5%)
0.139
SCORTEN Day 7 (% of ≥2)
86 (59.7%)
5 (38.5%)
0.153
91 (58.0%)
9 (81.8%)
0.202
SCAR, severe cutaneous adverse reaction; BMI, body mass index; OPD, out-patient department; ER, emergency room; SJS, Stevens-Johnson syndrome; overlap, combined Stevens-Johnson syndrome and toxic epidermal necrolysis; TEN, toxic epidermal necrolysis; DRESS, drug reaction with eosinophilia and systemic symptoms; DHS, drug hypersensitivity syndrome; BSA, body surface area; WBC, white blood cells; Cr, creatinine; ALT, alkaline phosphatase; IVIG, intravenous immunoglobulin; ICU, intensive care unit; SCORTEN, SCORe of Toxic Epidermal Necrosis
In patients with mortality, the amount of cigarette consumption was lower (3.8 ± 9.9% vs. 0.0 ± 0.0, P < 0.001), TEN was more frequent SCAR types (SJS/TEN/SJS-TEN overlap/DRESS 37.6/8.9/5.7/47.8% vs. 18.2/45.5/0.0/36.4%, P = 0.015), and the involved skin area was broader than in patients who survived (71.4 ± 31.0 vs. 87.5 ± 19.9, P = 0.036). Moreover, administration days of culprit drug were shorter (17.8 ± 21.2 vs. 5.9 ± 5.0, P = 0.042), and the rates of IVIG use and ICU care were higher (15.1% vs. 80.0%, P < 0.001, 4.1% vs. 45.5%, P < 0.001). Finally, a higher SCORTEN score on day 7 was associated with increased mortality (1.7 ± 0.9 vs. 2.9 ± 1.2, P < 0.001) (Table 3).
When the risk factors for poor prognosis leading to sequelae or death were analyzed, a higher SJS-TEN overlap and TEN, serum Cr level, and SCORTEN score on day 7 and, if carbapenems were the causative agents, a higher odds ratio for poor prognosis was observed in the univariate analysis (Table 4). However, in the multivariate analysis, the statistical significance of other factors disappeared, and the odds ratio of poor prognosis remained significant with SJS-TEN overlap/and TEN and was also significant when carbapenems were the causative agents.
Table 4Factors associated with poor prognosis (sequelae or death) of beta-lactam-induced SCARs using binary logistic regression (univariate and multivariable analysis).
In this study, there was a clinical difference between beta-lactam antibiotic-induced SCAR and non-beta-lactam antibiotic-induced SCAR. SJS/TEN was more common than DRESS in patients with beta-lactam antibiotic-induced SCAR, and the number of culprit drug administration days was shorter. Moreover, among patients with beta-lactam antibiotic-induced SCAR, different clinical features were observed according to the SCAR type. Patients with DRESS showed a higher BMI, higher frequency of fever and lymphadenitis as presenting symptoms, more elevated WBC count, eosinophil count, and ALT level, and longer duration of drug administration after symptoms and latent periods. Patients with TEN were more commonly accompanied by other allergic diseases and showed larger involved BSA, longer duration of fever and admission duration, higher proportion of IVIG use, higher SCORTEN score on day 7, and a higher proportion of patients with poor prognosis (with sequelae or death). The factors associated with poor prognosis were SJS-TEN overlap, TEN as the SCAR phenotype. and carbapenems as the causative drugs.
To the best of our knowledge, this is the first large-scale study that has been conducted on beta-lactam antibiotic-induced SCAR. In 2014, a retrospective study on SCAR related to systemic antibiotics in 74 cases was published.
In this study, penicillins and cephalosporins were the most common causative agents of SJS, TEN, and acute generalized exanthematous pustulosis, and glycopeptides were the most common causative agents of DRESS. Although the ranking of penicillins and cephalosporins was reversed compared to that in our study, the most common causative antibiotic class, beta-lactam antibiotics, was identical. In that study, the mortality rate was the highest in the SJS and TEN groups, and it was associated with old age and underlying sepsis. The average latent period was 6.52 ± 4.59, 5 ± 3.52, and 11.3 ± 8.20 days for SJS, TEN, and DRESS, respectively, while it was 17.8 ± 25.8, 18.8 ± 28.2, and 24.7 ± 23.8, respectively, in our study. The latent period was longer in our study, but the longest latent period was observed in the DRESS subtype in both studies. The proportion of corticosteroid use was similar at 70.3% and 80.6%, respectively, but IVIG was used more frequently in our study (1.3% and 18.4%, respectively). The overall mortality was 20%, which was higher than 7.4% in our study, and mortality by SCAR type was TEN (66.7%), SJS or SJS-TEN overlap (20%), and DRESS (12%), which was also lower in our study (26.3% for TEN, 3.3% for SJS, 0% for SJS-TEN overlap, and 5.1% for DRESS. The average SCORTEN score was 1.25 in survivors and 2.77 in the dead; similarly, the SCORTEN score on day 7 was 1.7 and 2.9, respectively.
SCARs caused by all drugs were targeted, and antibiotics were the most common agents (88.1%). Among them, cephalosporins (23.7%) and penicillin (16.9%) were the most common culprit drugs, as shown in our study. In that study, the latent period was 6.2 days for SJS/TEN and 14.0 days for DRESS, which were shorter than those in our study. Systemic corticosteroids were used in 71.4% of the patients, and IVIG was not administered. There was female preponderance in SJS and TEN, and male preponderance in DRESS, while there were no sex differences in our study.
In another retrospective observational comparative study between antibiotic- and non-antibiotic-associated delayed cutaneous adverse drug reactions, 48% of the 84 patients were antibiotic-associated cases.
When compared with antibiotic-related SCAR in our study, age and sex were similar, while the latency period was longer in our study (11.0 vs. 6.0 days). As in our study, beta-lactam antibiotics (61.8% vs. 45.0%), especially cephalosporins, were the most common implicated drug. Mortality were lower in our study (7.4% vs. 10–20%).
Although studies on risk factors for drug allergy are still lacking, previous studies have suggested that female sex, age, systemic lupus erythematosus, and human immunodeficiency virus infection are risk factors.
Risk factors for the development of SCAR or the poor prognosis of SCAR have been rarely studied, and risk factors for the occurrence of beta-lactam antibiotic-induced SCAR could not be analyzed in our study. However, analysis of risk factors related to poor prognosis showed that there was no statistical significance with sex, age, comorbidities, or other allergic diseases; only TEN subtype and carbapenem use were significant risk factors in our study.
Although antibiotics and anticonvulsants are known to be the most common causative drugs of SCAR,
few studies have been conducted on antibiotic-induced SCAR. The clinical features of antibiotic-induced SCAR have not been studied; thus, we have no clinical information regarding this. Moreover, although cotrimoxazole, allopurinol, carbamazepine, phenytoin, phenobarbital, and oxicam-NSAIDs are known as “high risk” for the development of SCAR,
For the first time, carbapenem-induced SCAR was found to be associated with poor prognosis in our study. However, carbapenem is expected to be used in clinically severe and antibiotic-resistant patients. Therefore, it is likely that the prognosis was poor because patients who used carbapenems had severe baseline medical conditions, rather than because of a difference in the type of antibiotics. Although we considered the effect of comorbidities as a contributing factor in the univariate analysis, detailed patient conditions were not evaluated. Moreover, the number of patients with carbapenem-induced SCAR was small and larger-scale studies are required to confirm our result.
There are some limitations to our study. First, we only collected clinical data, and we did not study possible mechanisms. Therefore, this aspect needs to be further investigated. Secondly, the causative drugs were not certain. Since not all cases were confirmed by drug skin test or in vitro test, the culprit drug had to be presumed. In particular, it was more difficult to select a suspected drug when multiple drugs were administered simultaneously. In these cases, all suspected drugs used simultaneously were assumed to be causative agents based on expert judgment. Thirdly, we did not assess the effects of type or severity of infection for the prognosis of SCAR as unified and valid evaluation of infection severity is complicating. The association between specific beta-lactam antibiotics and outcome of SCAR needs to be clarified controlling these factors in the further study. Fourth, SCAR may be related to viral infection, but testing for this has not been performed. In the case of SJS/TEN, they may be related or mimic to herpes simplex virus (HSV) or Mycoplasma pneumonia infection. In addition, reactivation of human herpesvirus (HHV) is often seen in DRESS patients. Since the relationship between viral infection and SCAR cannot be ruled out, further investigation is needed in the next study.
Finally, the retrospective study design had limitations in predicting causal relationships.
In conclusion, we analyzed the clinical characteristics, common causative drugs, and risk factors of poor prognosis of beta-lactam antibiotic-induced SCARs, using large scale, nationwide data. SJS/TEN is the most common type of beta-lactam antibiotic-induced SCAR. Among the beta-lactam antibiotic-induced SCAR, prognosis was the most favorable in patients with DRESS and the worst in patients with TEN. The risk factors for poor prognosis leading to sequelae or death were SJS-TEN overlap and TEN, and carbapenems as the causative agents.
SCAR is a serious and lethal disease, and beta-lactam antibiotic-related SCAR accounts for a large proportion of SCAR cases. In particular, SJS/TEN which is very severe SCARs with poor prognosis were common type in beta-lactam antibiotic-related SCARs. Therefore, more attention should be paid to monitoring skin reactions while using beta-lactam antibiotics. In addition, it was observed in this study that the prognosis may be worse when SCAR caused by carbapenem occurs. Careful observation is also necessary in patients taking carbapenem, and early discontinuation and treatment would be needed when symptoms develop. Moreover, in-depth research is required to understand the mechanism in order to prevent SCAR from happening in the future.
Abbreviations
BMI, body mass index; BSA, body surface area; DIHS, drug-induced hypersensitivity syndrome; DRESS, drug reaction with eosinophilia and systemic symptoms; IVIG, intravenous immune globulin; OR, odds ratio; SCAR, severe cutaneous adverse reactions; SJS, Stevens-Johnson syndrome; TEN, toxic epidermal necrolysis; WBC, white blood cell.
Funding
There is no funding source of this study.
Availability of data and materials
The data used in this paper is not public data and therefore cannot be disclosed.
Ethics approval
IRB approval number of the representative institution, Seoul National University Bundang Hospital is B-1802-450-401.
Authors’ consent for publication
All authors consented to the publication of this paper.
Declaration of competing interest
Authors have no conflicts to disclose.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Efficacy of the HLA-B(∗)58:01 screening test in preventing allopurinol-induced severe cutaneous adverse reactions in patients with chronic renal insufficiency-A prospective study.
00114-4&id=gr1.jpg){kind=link}