TO THE EDITOR
Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are extremely rare life-threatening mucocutaneous skin reactions most commonly attributed to medications. Incidence rates of SJS/TEN reported in the general population range from 1–2 cases per million per year (Rzany et al., 1996; Schopf et al., 1991). A recent large observational study on the epidemiology of SJS/TEN published in the Journal of Investigative Dermatology demonstrated that active cancer may be associated with increased risk of SJS/TEN (Frey et al., 2017). This finding is consistent with the EuroSCAR-study that reported an increased risk (RR 2.7) of SJS/TEN in individuals with active malignancy (Mockenhaupt et al., 2008). To comprehensively asses this association, we conducted a retrospective cohort study to identify the incidence and triggers of SJS/TEN in a large series of cancer patients. We compared the incidence of SJS/TEN in our cancer patient cohort to that reported by in the general population.
Patients were identified through a search of electronic health records at Moffitt Cancer Center (MCC). All inpatient and outpatient patients treated for cancer between June 15, 2002 and June 15, 2015 who had International Classification of Disease (ICD-9) codes available in the electronic health records were eligible for inclusion. Patients entered the cohort in 2002 or at time of their first visit; follow-up occurred until time of event, death, or last visit. The ICD-9 codes 695.1, 695.12, 695.13, 695.14, and 695.15 were utilized to identify patients who may have experienced SJS/TEN. Manual retrospective chart reviews of identified patients were completed independently by two clinicians to confirm SJS/TEN. Patients were classified as (1) confirmed SJS/TEN (cases diagnosed at MCC), (2) possible SJS/TEN (cases diagnosed at an outside hospital while a patient at MCC or uncertain differential diagnoses), (3) historical SJS/TEN (cases occurring prior to cohort entry), or (4) unconfirmed SJS/TEN. Culprit drugs were abstracted from chart reviews and represent the hypothesized triggers according to the patients’ physicians.
To maintain comparability with published reports, the primary outcome was average annual incidence of SJS/TEN. Annual incidences were calculated as the number of unique patients with a SJS/TEN diagnosis divided by the total number of patients screened that year. Incidence rates were calculated as the total number of cases divided by the total number of person-years of observation. Incidences were calculated for confirmed SJS/TEN cases and confirmed plus possible SJS/TEN cases. Timing of historical cases relative to cancer diagnosis was unknown, so these cases were excluded from calculations.
A total of 104,062 unique cancer patients were screened for SJS/TEN ICD-9 codes, and 121 possible cases were identified. Chart reviews confirmed SJS/TEN diagnosis in 20 patients, possible SJS/TEN in 30 patients, and historical SJS/TEN diagnoses in 12 patients (Table 1). The average follow-up period for the cohort was 2.7±3.1 yr. The average annual incidence of confirmed cases of SJS/TEN was 5.7 (range 0–15.8) per 100,000 individuals; the average annual incidence of confirmed and possible cases SJS/TEN was 14.9 (range 0–28.4) per 100,000 individuals (Table S1). The incidence rates were 7.2 and 17.9 per 100,000 patient-years, for confirmed and confirmed plus possible cases, respectively. Confirmed cases of SJS/TEN were more common in hematologic diagnoses compared to solid malignancies (0.1% vs. 0.007%, p<0.0001). There was one possible case with a differential diagnosis of SJS/TEN vs. graft versus host disease; distinguishing between these diagnoses poses challenges (Macedo et al., 2014).
Table 1. Patient demographics.
Summary information is provided for entire cohort screened and all possible cases of Stevens-Johnson syndrome (SJS) and/or toxic epidermal necrolysis (TEN), sorted by classification based on manual chart review.
| Characteristic | Patients screened* | Confirmed cases | Possible cases | Historical cases | All cases |
|---|---|---|---|---|---|
| (n = 104,062) | (n = 20) | (n = 30) | (n = 12) | (n = 62) | |
|
| |||||
| Age # | 59 ± 14 | 53 (29–78) | 58 (23–76) | 60 (32–80) | 59 (23–80) |
|
| |||||
| Sex | |||||
| Female | 51,720 (50%) | 7 (35%) | 18 (60%) | 6 (50%) | 31 (50%) |
| Male | 52,342 (50%) | 13 (65%) | 12 (40%) | 6 (50%) | 31 (50%) |
|
| |||||
| Race | |||||
| White | 92,239 (89%) | 18 (90%) | 27 (90%) | 12 (100%) | 57 (92%) |
| Black | 5,050 (5%) | 1 (5%) | 2 (7%) | 0 | 3 (5%) |
| Asian | 1,099 (1%) | 1 (5%) | 0 | 0 | 1 (2%) |
| American Indian | 126 (0.1%) | 0 | 0 | 0 | 0 |
| Unknown | 5,548 (5%) | 0 | 1 (3%) | 0 | 1 (2%) |
|
| |||||
| Ethnicity | |||||
| Non-Hispanic | 77,186 (74%) | 17 (85%) | 25 (83%) | 12 (100%) | 54 (87%) |
| Hispanic | 5,510 (5%) | 2 (10%) | 4 (13%) | 0 | 6 (10%) |
| Unknown | 21,366 (21%) | 1 (5%) | 1 (3%) | 0 | 2 (3%) |
|
| |||||
| Cancer Type | |||||
| Solid | 88,455 (85%) | 6 (30%) | 18 (60%) | 9 (75%) | 33 (53%) |
| Hematologic | 13,633 (13%) | 14 (70%) | 10 (33%) | 3 (25%) | 27 (44%) |
| Unknown | 1,974 (2%) | 0 | 2 (7%) | 0 | 2 (3%) |
Cancer type represents primary site of first cancer diagnosis as reported in Florida Cancer Registry.
Age for screened individuals represents the mean and standard deviation of the age at first cancer diagnosis; age for cases represents the median and range at time of SJS/TEN ICD-9 code
There were 69 physician-reported culprits for the 62 cases of SJS/TEN (Table 2). The most common trigger in all identified cases of SJS/TEN was antibiotics (53.2%), with trimethoprim/sulfamethoxazole being the most frequent (22.6%). Anticonvulsants (e.g., phenytoin) and antineoplastic agents were the second most commonly stated trigger (9.7% each). Antineoplastics attributed to SJS/TEN included immunomodulatory agents (lenalidomide, interferon, and methotrexate).
Table 2. Triggers of Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN).
Triggers were identified from patient medical records and are reported according to SJS/TEN occurrence: confirmed cases (n = 20 patients), possible cases (n = 30) and historical cases (n = 12). For some patients, there was not a single definitive trigger, resulting in attribution to and discontinuation of multiple medications (Table S2).
| Trigger | SJS/TEN classification | Total | ||
|---|---|---|---|---|
| Confirmed | Possible | Historical | ||
|
| ||||
| Single attributed culprit | 10 | 24 | 11 | 45 |
|
| ||||
| sulfamethoxazole/trimethoprim | 4 | 4 | 3 | 11 |
| unknown | 0 | 3 | 3 | 6 |
| phenytoin | 1 | 3 | 0 | 4 |
| sulfonamides (unspecified) | 0 | 2 | 0 | 2 |
| piperacillin/tazobactam | 2 | 0 | 0 | 2 |
| cefepime | 1 | 0 | 0 | 1 |
| clindamycin | 0 | 0 | 1 | 1 |
| erythromycin | 0 | 0 | 1 | 1 |
| levofloxacin | 0 | 1 | 0 | 1 |
| minocycline | 0 | 1 | 0 | 1 |
| penicillin | 0 | 0 | 1 | 1 |
| fluconazole | 0 | 1 | 0 | 1 |
| voriconazole | 0 | 0 | 1 | 1 |
| allopurinol | 1 | 0 | 0 | 1 |
| colchicine | 0 | 1 | 0 | 1 |
| capecitabine | 0 | 1 | 0 | 1 |
| interferon | 0 | 1 | 0 | 1 |
| lenalidomide | 1 | 0 | 0 | 1 |
| pembrolizumab | 0 | 1 | 0 | 1 |
| temozolomide | 0 | 1 | 0 | 1 |
| bisphosphonate | 0 | 1 | 0 | 1 |
| clobetasol | 0 | 0 | 1 | 1 |
| epoetin alfa | 0 | 1 | 0 | 1 |
| saline | 0 | 1 | 0 | 1 |
| Herpes simplex virus | 0 | 1 | 0 | 1 |
|
| ||||
| Multiple uncertain culprits | 10 | 6 | 1 | 17 |
We utilized our institutional electronic health system biorepository of over 100,000 patients to determine the incidence and identify potential triggers of SJS/TEN in cancer patients. The average annual incidence of SJS/TEN observed was 5.7–14.9 cases per 100,000 per year. Most conservatively, this corresponds to a 28–57-fold higher incidence in cancer patients than reported in the general population; leniently, the incidence observed is 75–150-fold higher in cancer patients. Reported triggers of SJS/TEN were similar to those previously associated with the reaction; however, some cases of SJS/TEN were attributed to cancer-specific therapies.
Cancer patients have increased mortality rates from SJS/TEN compared to individuals without cancer (Wu et al., 2015); therefore, a thorough understanding of the factors that increase risk in cancer patients is especially critical. The immunocompromised state of many cancer patients, our observation of immunomodulators as a trigger, and the higher prevalence of SJS/TEN in HIV patients (Mittmann et al., 2012) suggest a possible role of the immune system in the increased risk of SJS/TEN. Other possible explanations include increased exposure to culprit medications, as alluded to by Frey and colleagues, the cancer disease process, or synergy between risk factors.
A notable strength of our study was the utilization of electronic medical records to confirm ICD-9 code diagnoses. It has been demonstrated that use of claims databases to investigate incidences of SJS results in significant overestimates; in fact, one study found that only 14.8% of individuals with SJS claims had a final diagnosis of SJS (Strom et al., 1991). Likewise, in our study we found that only 16.5% of individuals with SJS/TEN ICD-9 codes had confirmed diagnoses. The use of claims data alone would have resulted in our reporting a 600–1200-fold higher incidence of SJS/TEN in cancer patients.
Limitations of our study are inherent to the challenges of treating SJS. First, there are no definitive diagnostic criteria for SJS/TEN, so it is possible that cases were missed during the search process querying ICD-9 codes. Additionally, physicians likely preferentially attribute SJS/TEN to drugs with reported associations and patients will never be re-challenged, so culprits cannot be confirmed and causality is not definitive. Variable follow-up times may complicate the true interpretation of risk, further contributing to the conservative nature of estimates reported. Future studies are warranted to determine the role of the cancer disease process, immune system, and culprit drugs on SJS/TEN pathogenesis.
Supplementary Material
ACKNOWLEDGMENTS
This work has been supported in part by the Shared Resources at the H. Lee Moffitt Cancer Center & Research Institute, an NCI-designated Comprehensive Cancer Center (P30-CA076292). Research support: The DeBartolo Family Personalized Medicine Institute, the State of Florida Cancer Research Endowed Chair, the Collins Charitable Foundation, the American Foundation for Pharmaceutical Education, Khalid Ishaq Graduate Travel Stipend, and the Eshelman Foundation.
Abbreviations:
- ICD
International Classification of Disease
- MCC
Moffitt Cancer Center
- SJS
Stevens-Johnson syndrome
- TEN
toxic epidermal necrolysis
Footnotes
CONFLICT OF INTEREST
The authors report no conflicts of interest relevant to this work.
REFERENCES
- Frey N, Jossi J, Bodmer M, Bircher A, Jick SS, Meier CR, et al. The Epidemiology of Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis in the UK. J Invest Dermatol. 2017;In press. [DOI] [PubMed] [Google Scholar]
- Gravante G, Delogu D, Marianetti M, Esposito G, Montone A. Toxic epidermal necrolysis and Steven-Johnson syndrome in oncologic patients. Eur Rev Med Pharmacol Sci. 2007;11(4):269–74. [PubMed] [Google Scholar]
- Macedo FI, Faris J, Lum LG, Gabali A, Uberti JP, Ratanatharathorn V, et al. Extensive toxic epidermal necrolysis versus acute graft versus host disease after allogenic hematopoietic stem-cell transplantation: challenges in diagnosis and management. J Burn Care Res. 2014;35(6):e431–5. [DOI] [PubMed] [Google Scholar]
- Mittmann N, Knowles SR, Koo M, Shear NH, Rachlis A, Rourke SB. Incidence of toxic epidermal necrolysis and Stevens-Johnson Syndrome in an HIV cohort: an observational, retrospective case series study. Am J Clin Dermatol. 2012;13(1):49–54. [DOI] [PubMed] [Google Scholar]
- Mockenhaupt M, Viboud C, Dunant A, Naldi L, Halevy S, Bouwes Bavinck JN, et al. Stevens-Johnson syndrome and toxic epidermal necrolysis: assessment of medication risks with emphasis on recently marketed drugs. The EuroSCAR-study. J Invest Dermatol. 2008;128(1):35–44. [DOI] [PubMed] [Google Scholar]
- Rosen AC, Balagula Y, Raisch DW, Garg V, Nardone B, Larsen N, et al. Life-threatening dermatologic adverse events in oncology. Anticancer Drugs. 2014;25(2):225–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rzany B, Mockenhaupt M, Baur S, Schroder W, Stocker U, Mueller J, et al. Epidemiology of erythema exsudativum multiforme majus, Stevens-Johnson syndrome, and toxic epidermal necrolysis in Germany (1990–1992): structure and results of a population-based registry. J Clin Epidemiol. 1996;49(7):769–73. [DOI] [PubMed] [Google Scholar]
- Schopf E, Stuhmer A, Rzany B, Victor N, Zentgraf R, Kapp JF. Toxic epidermal necrolysis and Stevens-Johnson syndrome. An epidemiologic study from West Germany. Arch Dermatol. 1991;127(6):839–42. [DOI] [PubMed] [Google Scholar]
- Strom BL, Carson JL, Halpern AC, Schinnar R, Snyder ES, Stolley PD, et al. Using a claims database to investigate drug-induced Stevens-Johnson syndrome. Stat Med. 1991;10(4):565–76. [DOI] [PubMed] [Google Scholar]
- Wu J, Lee YY, Su SC, Wu TS, Kao KC, Huang CC, et al. Stevens-Johnson syndrome and toxic epidermal necrolysis in patients with malignancies. Br J Dermatol. 2015;173(5):1224–31. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.