Most dermatologic diseases have very low mortality rates and are generally managed on an outpatient basis. Never- theless, some dermatologic diseases can be life-threatening and require intensive care as a result of the disease itself or its complications. Analysis of dermatologic diseases and their associated risk factors could provide information to predict or prevent patient death. However, only a few studies have assessed the mortality rates among patients with derma- tologic diseases.

We have recorded 13 patient deaths in the dermatology ward from 1998 to 2014. The aim of this study is to in- vestigate dermatologic diseases and the associated factors that lead to death by analyzing these cases.

1. Data collection

This is a retrospective study conducted at a tertiary hospital in Korea (Department of Dermatology at Pusan National University Hospital, Republic of Korea). All data were collected from the medical records of dermatologic inpatients from 1998 to 2014. Patients who died from malignancy, including melanoma and non-melanoma skin cancer, were excluded. This study was approved by the Institutional Review Board of the Pusan National University Yangsan Hospital (IRB No. 05-2022-094).

Patient medical records and death certificate data were retrospectively reviewed. We analyzed the clinical findings according to the following: (1) patient demographic data, including age, sex, and comorbidities; (2) diagnosis of derma- tologic disorders resulting in admission; (3) treatments and progression including length of hospital stay (LoS); (4) cause of death; (5) and laboratory findings, including bacteriological profile and radiologic findings.

2. Statistical analysis

We compared the clinical parameters between expired and recovered patients with Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN). Statistical analyses were performed using SPSS ver. 21.0 (IBM Co., Armonk, NY, USA). Comparisons of clinical parameters between the groups were performed using Fisher's exact test for categorical variables and Mann-Whitney U test for numerical variables. p-values less than 0.05 were considered statistically significant.

A total of 740 patients were hospitalized in the derma- tologic ward from 1998 to 2014 (Table 1, admission for the surgical treatment of malignant or benign cutaneous neo- plasia, laser treatment, and aesthetic treatment was excluded). Thirteen patients who eventually died during hospitalization were enrolled in this study. The proportion of expired cases was 1.76% (13 deaths/740 dermatologic inpatients) during the study period. Table 2 shows the patient demographics, comorbidities, and dermatological disorders causing mortality.

1. Demographics

Fig. 1 shows the age distribution of the patients. The mean age was 63.8 years (range, 18~86). Eleven of out 13 patients (84.6%) were older than 55 years, and the highest mortality was recorded in patients older than 70 years (7 out of 13 patients, 53.8%). A slight female dominance among the patients was observed (male : female ratio, 1:1.6). Ten of 13 patients (76.9%) had more than one comorbidity, including hypertension (n = 8), diabetes mellitus (n = 6), angina pectoris (n = 2), and chronic kidney disease (n = 1). None presented with HIV infection.

Figure 1. Age distribution of expired patients

Modify Delete

2. Diagnosis of dermatologic disorders

Dermatologic complications responsible for admissions were drug reactions, including SJS/TEN and exanthematous drug eruption (n = 7), as well as bullous pemphigoid (BP, n = 2), exfoliative dermatitis (n = 2), generalized pustular psoriasis (GPP, n = 1), and dermatomyositis (n = 1).

The offending medications responsible for drug reac- tions were antibiotics in two patients, non-steroidal anti-inflammatory drugs in two patients, allopurinol in two patients, and "unknown" in one patient.

SJS/TEN was the most common dermatologic disorder in this study (n = 6). The mean detached body surface area (BSA) was 42.0 ± 19.2% (range, 20~70), and the mean SCORTEN was 2.8 ± 1.3 (range, 1~4) points upon admission. Compared with the recovered patients with SJS/TEN during the same period (n = 58), expired patients had significantly lower serum albumin levels (expired patients 3.0 ± 0.5 vs. recovered patients 3.9 ± 0.6, Mann-Whitney U test, p = 0.002) and higher blood urea nitrogen levels (expired patients 49.6± 39.7 vs. recovered patients 18.9 ± 9.6, Mann-Whitney U test, p = 0.016). Other laboratory findings, SCORTEN, and BSA were not significantly different between the two groups (data not shown).

3. Treatments and progression

Twelve patients (92.3%) were treated with immunosup- pressive agents, such as systemic steroids (n = 10, 76.9%), cyclosporine (n = 1, 7.7%), and methotrexate (n = 1, 7.7%). Antibiotics were administered in seven patients (53.8%) from admission. None of the expired patients with SJS/TEN were treated with intravenous immunoglobulin.

Two patients (15.4%) died in the dermatology ward, and 11 patients (84.6%) were transferred to another department or an intensive care unit. A total of 11 patients were trans- ferred because of suspected sepsis (n = 6), continuous fever (n = 2), renal dysfunction (n = 2), or abnormal liver function (n = 1). The average LoS in hospital (from admission to expiration) of all patients was 31.6 ± 24.7 days (range, 7~99). The mean LoS in dermatology wards (from admission to expiration or transfer) was 9.5 ± 8.5 days.

4. Cause of death

Seven of 13 patients (53.8%) developed septic shock and died as a result. Five patients (38.5%) died from acute respiratory failure caused by pneumonia (n = 4) or diffuse alveolar hemorrhage (n = 1). One patient (7.7%) died from acute myocardial infarction.

5. Laboratory findings

Abnormal laboratory findings upon admission included leukocytosis (n = 5), electrolyte imbalance (n = 3), throm- bocytopenia (n = 2), anemia (n = 2), abnormal kidney function (n = 2), and abnormal liver function (n = 2).

Infection developed in 11 patients (84.6%). The most com- mon infection site was the respiratory tract (pneumonia), followed by the kidneys (acute pyelonephritis) and skin (soft tissue infection). Table 3 summarizes the results of the blood cultures. Twelve microorganisms in 7 patients were identified using bacterial cultures from blood. The most common patho- gens were well-known microorganisms causing hospital-acquired infections (HAI), such as Pseudomonas aeruginosa, Staphylococcus aureus, Klebsiella pneumoniae, and Acineto- bacter baumannii. In most cases, the results of the blood cultures were reported after the patient's death. Among the patients who died from infection, a reversal of C-reactive proteins (CRP) from decreasing to increasing levels with positive microbial growth in the blood cultures was observed in three patients (Fig. 2). Sudden-onset neutropenia (Fig. 3A) and sudden-onset fever (Fig. 3B) with positive microbial growth in the blood cultures were observed in one patient.

Figure 2. CRP reversal from a decrease to an increase with positive blood cultures (X: patient death)

Modify Delete

Figure 3. Sudden-onset neutropenia (A) and sudden-onset fever (B) (VRE: vancomycin-resistant Enterococcus faecalis, X: patient death)

Modify Delete

The frequency of dermatologic diseases causing death is different for each country. The most common diseases were autoimmune blistering diseases in Brazil¹ and India², chronic ulcers and cellulitis in the United States3, and SJS/TEN in the United Kingdom4. In the present study, SJS/TEN was the most common disorder resulting in death, followed by BP, exfoliative dermatitis, GPP, and dermatomyositis in derma- tologic inpatients.

SJS/TEN is a well-known, life-threatening disorder. The overall hospital mortality rate of SJS/TEN is 22~25%⁵. In our study, the mortality rate of SJS/TEN was 9.4% (6 out of 64 patients with SJS/TEN). The prognosis score of SJS/TEN (SCORTEN) is useful, but the SCORTEN upon admission was not significantly different between the expired and recovered patients (expired patients 2.8 vs. recovered patients 2.0, Mann-Whitney U test, p = 0.224) in the present study. BP is the most common autoimmune blistering disease. Old age, poor general health, and the presence of anti-BP180 anti- bodies have been associated with a poor prognosis5. Exfoliative dermatitis is potentially life-threatening because of compli- cations, including fluid and electrolyte imbalance, thermo- regulatory disturbance, fever, tachycardia, high-output failure, and hypoalbuminemia. GPP can potentially result in life-threatening complications, such as hypocalcemia, bacterial superinfection, sepsis, and dehydration. Death occurred in 2% of the GPP cases⁶. The overall mortality rate of dermato- myositis and polymyositis was 22%. Variables associated with poor outcomes of dermatomyositis and polymyositis were older age, pulmonary and esophageal involvement, and cancer⁷.

It is well-known that old age contributes to increased mortality in diseases. In a population-based study examining death from non-neoplastic skin disease in the United States, most deaths occurred in patients aged 65 years and older3. Similarly, in the present study, 85% of deaths occurred in patients aged 55 years and older, and the highest mortality was reported in patients aged 70~79.

Infection developed in 11 patients (84.6%) and 7 patients (53.8%) died from septic shock. Inflammation, fissuring, and excoriation of the diseased skin increase susceptibility to bacterial colonization and infection in dermatologic diseases, such as SJS/TEN, exfoliative dermatitis, and blistering diseases5. Immunosuppressive agents are the mainstay of treatment in severe inflammatory skin diseases, but they can also predis- pose patients to infections⁸. Most patients (92.3%) were treated with immunosuppressive agents in the present study.

The most common pathogens were P. aeruginosa, S. aureus (including MRSA), K. pneumoniae, and E. coli, which are important microorganisms in HAI. Tao et al.⁹ reported that P. aeruginosa, K. pneumoniae, and E. coli are common pathogens in HAI. Vijayamohan et al.¹⁰ reported a high pre- valence of methicillin-resistant S. aureus (MRSA) infection among dermatology inpatients. Prolonged duration of illness, prolonged antibiotic and steroid therapy, and diabetes were relative risk factors for acquiring MRSA. In another study, prolonged LoS was an important factor in hospital-acquired MRSA¹¹. Therefore, precautions against HAI during hospital- ization, such as hand decontamination, personal hygiene, disinfection of patient equipment, and safe injection prac- tices¹², could decrease mortality among patients with severe dermatologic disease.

Among the patients who died from sepsis, sudden-onset fever, neutropenia, and CRP elevation were observed. Bor et al.¹³  reported a fourfold increase in the mortality rate of patients with fever compared with those without fever. CRP is an acute-phase reactant, suggesting that its levels rise in response to inflammation. In sepsis, IL-6, TNF-a, and IL-1b stimulate the release of CRP during the inflammatory cascade^14,15. Therefore, the reversal of CRP from decreasing to increasing levels could be a marker of infection, and serial CRP measurement could be a valuable tool in patients with severe dermatologic disorders. Leukopenia (defined as a white blood cell count of <4,000/μL) can be seen in sepsis as one of the criteria of systemic inflammatory response syndrome. Leukopenia indicates anomalies in the host's inflammatory response and is associated with poor outcomes16. Therefore, although fever, CRP elevation, and neutropenia can be observed without infection, such as those with connective tissue disease, malignancy, and cardiovascular disease^17,18, efforts should be made to identify any infection sites in these situations.

The limitations of this study include its retrospective design, small cohort size, and the single center nature of the study.

In conclusion, our study revealed that drug reactions, such as SJS/TEN, were the most common dermatologic disease leading to mortality that is associated with infection and sepsis. Factors including old age, epithelial loss by derma- tologic diseases, and immunosuppression could predispose infection in these patients. Infection, especially during hos- pitalization, was the most important factor contributing to mortality. Therefore, efforts to prevent HAI could contribute to decreasing mortality. This study aims to improve the under- standing of severe dermatologic diseases and contributes to the prevention of deaths in patients with dermatologic disorders.