Invasive Candidiasis in Hospitalized Patients with Major Burns

Background: Invasive candidiasis (IC) is a major cause of morbidities and mortality in patients hospitalized with major burns. This study investigated the incidence of IC in this specific population and analyzed the possible risk factors. Materials and Methods: We retrospectively analyzed data from the National Health Insurance Research Database (NHIRD) of Taiwan. We identified 3582 patients hospitalized with major burns on over 20% of their total body surface area (TBSA) during 2000–2013; we further analyzed possible risk factors. Result: IC was diagnosed in 452 hospitalized patients (12.6%) with major burns. In the multivariate analysis, patients older than 50 years (adjusted odds ratio (OR) = 1.96, 95% confidence interval (CI) 1.36–2.82), those of female sex (adjusted OR = 1.33, 95% CI 1.03–1.72), those with burns on the head (adjusted OR = 1.33, 95% CI 1.02–1.73), and those with burns over a greater TBSA had higher risks of IC. Conclusion: Treating IC is crucial in healthcare for major burns. Our study suggests that several risk factors are associated with IC in patients hospitalized with major burns, providing reliable reference value for clinical decisions.


Introduction
Burns are estimated to result in 180,000 deaths globally every year [1]. In developed countries, the survival rate of patients with major burns has improved in recent years because of advances in surgical treatment and intensive care [2,3], but patients still experience prolonged hospitalization, disfigurement, and disability. Infection is a major challenge during hospitalization in patients with major burns because the skin barrier is extensively compromised.
The incidence of systemic candidiasis among patients with burns has increased from 2.19% to 4.3% in the 1990s [4,5], to 6.3% in a 2008 report [6], and to 11% in a 2015 report [3]. Some literature has suggested that the increasing trend of opportunistic fungal infections is associated with the increasing early use of broad-spectrum antibacterial agents [3,6]. In patients with burns, a fungal infection may be associated with several factors, including age, burn size, and chronic comorbidities. Candida species infections have outnumbered other species of fungus in literature [7,8]. In this study, we used the Taiwan National Health Insurance Research Database (NHIRD) for a retrospective analysis, focusing on patients with severe burns who required hospitalization. This study investigated the incidence of invasive candidiasis (IC) among these patients in Taiwan and related risk factors that severe burn patients might encounter.

Database
Taiwan's National Health Insurance program was established in 1995, and it currently covers 99.6% of Taiwan's population. The NHIRD is maintained by the National Health Research Institutes of Taiwan; the NHIRD is widely used for medical research and is validated as being of high quality and as being accurate in diagnoses [9]. The identification information of every individual is encrypted before the data are publicly available for research purposes. The diagnostic codes in the NHIRD are based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Patients with major burns, defined as a burn area equal to or more than 20% of the total body surface area (TBSA) or involving the eyes or head area, are included in the list of catastrophic illnesses in the NHIRD.

Study Design
A flowchart of patient inclusion in this study is presented in Figure 1. In this study, we retrospectively identified all patients hospitalized with burns in the NHIRD in 2000-2013 with ICD-9 CM codes 940-949. To define patients with major burns, we identified patients who had been registered as having catastrophic illness within 30 days after admission, with catastrophic illness codes 948.2-948.9 (TBSA 20-29% to >90%). Patients with no registration of catastrophic illness, those with no date of birth recorded, or those who were admitted for less than 3 days were excluded from this study. The burn site was defined according to ICD-9-CM codes. An inhalation injury was defined as ICD-9CM codes 947.0, 947.1, and 947.2 [10]. The season of a patient's hospitalization was defined as summer (May-October) or winter (November-April), with an average outdoor temperature cutoff of 24 • C.

Statistical Analysis
We evaluated the patients' baseline characteristics, including age, sex, dwelling area, family income, and chronic diseases, from the NHIRD data. We assessed the risk factors for IC among the patients with major burns by using odds ratios (ORs) and accompanying 95% confidence intervals (CIs) with a chi-squared test.
Data management and statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). Statistical significance was indicated by a two-tailed p-value of <0.05.
This study was approved by the institutional review board of Taichung Veterans General Hospital (No. CE17178A-4). For privacy protection, the identities of the patients, physicians, and institutions were deidentified in accordance with the Personal Electronic Data Protection Law. IC was defined as a hospitalized patient with a diagnosis of ICD-9CM codes 112.5 (disseminated candidiasis), 112.81 (Candida endocarditis), or 112.83 (Candida meningitis) [11] and a prescription for systemic antifungal agents.

Results
From the NHIRD database for 2000-2013, we identified 3582 patients hospitalized with major burns of over 20% of TBSA, with an incidence of 0.9-1.8 in every 10,000 persons/year (Table 1). IC occurred in 452 patients, accounting for 12.6% of all patients hospitalized with major burns (annual incidence ranging from 6.58 to 17.31 per 100 patients), as presented in Table 2.
The characteristics of the patients included in our study are listed in Table 3. The distribution of age was notably different between patients with burns, and with and without IC (p < 0.001); those with IC were older. Family income distributions were different between the two groups, but sex and urbanization of dwelling areas were not. More patients with IC had inhalation injuries than did those without IC (9.1% vs. 6.5%, p = 0.048). The seasons of patient hospitalization were not significantly different between the two groups. The head and hands were the most common burn sites in the IC group, more common than in the non-IC group. More patients in the IC group had diabetes and hypertension than those in the non-IC group.
The 30-day mortality rates were 10.4% in the IC group and 10.2% in the non-IC group (p = 0.876). The 180-day mortality rates were significantly different between IC and non-IC patients (29.7% and 12.6%, p < 0.001).
In the multivariate analysis (Table 4), patients older than 50 years were associated with a higher risk of IC compared with those younger than 30 years (adjusted OR 1.96, 95% CI 1.36-2.82, p < 0.001). Male sex was associated with less IC than female sex was (adjusted OR 0.75, 95% CI 0.58-0.97, p = 0.029). Burns occurring on the head were also associated with IC (adjusted OR 1.33, 95% CI 1.02-1.73, p = 0.036). Patients with burns covering more than 30% of TBSA were also more likely to have IC compared with those with burns covering 20-29% of TBSA.

Discussion
Patients with burns are a high-risk group for invasive fungal infections [12], and fungal wound infection is independently associated with mortality in patients with burns [13]. Fungal wound infection also delays wound healing, causes prolonged hospitalization, and may promote scarring and contracture [7]. Indwelling medical devices, such as endotracheal tubes, mechanical devices, central venous catheters, urinary catheters, and possibly parenteral nutrition, are necessary for life support in patients with major burns, but they may be sources of IC. Antibiotics are usually prescribed for prolonged prophylaxis or treatment against bacteria and represent another risk factor for IC as an opportunistic infection [14].
In our study, the rate of IC among patients with major burns ranged from 6.58 to 17.31 per 100 hospitalized patients. According to recent reports, the incidence of IC has been increasing [15], but the rate has varied between different studies. In the study by Escrig et al. [14], the incidence of IC was between 0 and 46.3 per 1000 admissions. Zhou et al. re-ported an IC rate ranging from 6.06% to 17.54% in a single intensive care unit for burns [16]. Devrim et al. conducted a study of pediatric patients with burns and found an incidence of 11.5% for IC [17]. Additionally, exposure to spores at the burn scene when the patients tried to extinguish the fire increases the possibility of IC [18]. The improvement of laboratory techniques in the identification of Candida species is also related to this increasing trend [2]. The awareness and vigilance of clinicians to fungal infections could also explain a higher diagnostic rate, though this is rather difficult to measure.
The diagnosis and treatment of IC are challenging. Currently, culturing with subsequent susceptibility testing is the only diagnostic approach, but the sensitivity of this method ranges from only 21% to 71% [8,19]. Fungal infections may manifest only nonspecific symptoms, and fungal cultures are time-consuming; these are some of the reasons why documenting fungal infections with cultures is difficult even when suspected [6]. Delayed antifungal administration can lead to higher mortality in hospitalization. However, prophylactic antifungal agents may lead to a higher proportion of species other than Candida albicans being present and a higher rate of reduced susceptibility [15]. The distributions of Candida species can vary with the duration of prophylaxis and the antifungal agent used for prophylaxis [19]. Other treatments including aggressive surgical debridement and catheter changes are also imperative to controlling the infection [18,20].
The risk factors vary for IC in patients with burns among different studies. Fochtmann et al. reported that female sex, gastrointestinal complications requiring surgery, non-gastrointestinal thromboembolic complications, and inhalation trauma were factors associated with proven IC [3]. According to the study by Sharma et al. [21], the incidence of fungal infection increased with age and larger TBSA involvement. In the study by Zhou et al. [16], risk factors included inhalation injury, renal dysfunction with replacement therapy, severe gastrointestinal complications, T-cell lymphopenia, and prior Candida colonization but not the percentage of TBSA. In our study, some risk factors were consistent with those of previous studies, such as female sex and larger TBSA involvement. Further studies with more patients may be necessary for more conclusive answers.
Female sex was found to be one of the risk factors for developing IC among patients with major burns in this study. Fochtmann and colleagues reported being female as one of the factors associated with candidemia (OR 5.03, 95% CI 1. 25-24.9, p = 0.031) [3]. However, Zhou and colleagues reported that sex was not a significant risk factor of candidemia using multivariate logistic regression [16]. In an earlier review of patients with burns by Ha and colleagues, male sex was considered a risk factor for candidiasis (but not systemic candidiasis), with an OR of 2.4 [2]. In a report of autopsied burn patients conducted by Murray and colleagues, they found that the sex distribution was not different between patients with and without fungal infection [22].
Approximately 10-15% of asymptomatic women are colonized with Candida in their vulvovaginal areas [23]. During stress, skin and mucosa damage, urinary catheter dwelling, dysbiosis due to prolonged antibiotics use, and localized candidiasis might result in invasion candidiasis. Evidence suggested that estrogen may have an important role in Candida invasion and proliferation. Estrogens may inhibit Candida-specific human peripheral blood lymphocyte responses [24]. Estrogens also have direct effects on Candida cells with estrogenbinding proteins [25]. We propose that these could be part of the explanation for why the female sex is a risk factor for IC among patients with major burns. Nevertheless, further investigations are necessary to distinguish the role of sex differences in IC of patients with major burns.
The site of the burn injury was rarely reported or discussed in previous literature. We propose that burns involving the head indicate a more severe injury condition, which may not be reflected in TBSA. We excluded the effect of inhalation injury in the multivariate analysis; however, intubation and mechanical ventilation may be necessary in burn patients involving the head. Preexisting Candida colonization of the mucosal membranes of the nose and throat could represent a latent source of infection [26].
The mortality rate within 30 days was not significantly different between patients with burns and with and without IC; however, the mortality rate was significantly different between the two groups in 180 days. As we mentioned earlier, candidiasis acts mostly as an opportunistic infection and may not occur or be life-threatening in the early days of a burn injury. However, in long-term care, candidiasis can greatly affect patient survival. The mortality rate reported in previous literatures varied from 29% to 76% [3,11,14], but the variety could be partially explained by the different study designs (such as different definitions of major burns and exclusion criteria).
This study has some limitations. This was a retrospective database analysis study, and the patients' physical examinations, laboratory data, Candida species, and doses of antifungal agents were unavailable in the database. We could not distinguish candidemia-related mortality from non-candidemia-related via the retrospect database. The population in this study might not have been large enough to reach significance in some of the statistical analyses.

Conclusions
In conclusion, our study suggests that IC in patients hospitalized with major burns is associated with risk factors including an age older than 50 years, female sex, burns on the head, and larger TBSA involvement. Identifying high-risk patients with proper antifungal prophylaxis is paramount, and our study may provide more information for clinical decision-making.