Journal of Infectious Diseases & Therapy
Open Access

Nosocomial Candida colonization may lead to higher rates of infections and invasive infections, frequently occur in patients with prolonged hospital stay and high-risk groups including the immunocompromised patients who are steadily growing like patients with Hematological malignancy, solid organ transplant patients, immunosuppressed patients, patients on chemotherapeutic agents, on steroids and antibacterial therapies [1-6]. Candida may colonize the oral, esophageal, vaginal and cutaneous structures, and it may cause invasive infections like bloodstream infection, hepato-splenic candidiasis, and peritonitis carrying a poor prognosis [7-9]. The epidemiology of Candida species has been changing in the last decades with the decrease in the prevalence of C. albicans augmented by the wide scale use of fluconazole as prophylaxis [10], allowing the relative increase in the prevalence of non-albicans Candida species, leading to more difficult-to-treat infections [11]. The growing candida resistance to antifungal agents are alarming, in a "USA-SENTRY" based study, nosocomial Candida was more resistant than community-onset candida stains, especially to the triazoles and echinocandins reaching rates somewhat above 5%, and some strains of C. glabrata were resistant to both classes [12]. Furthermore, multidrug-resistant species emerged and being reported from different countries, a new resistant strain i.e. C. auris was reported from Japan in 2009, it demonstrates resistance to several antifungal classes, it was associated with patients who had diabetes mellitus, recent surgery, a central venous catheter, patients receiving systemic antifungal therapy, and a patient with immotile cilia syndrome (primary ciliary dyskinesia) and bronchiectasis. Later reports showed that C. auris may have been misidentified as C. haemulonii and C. famata [13-19]. A nosocomial outbreak in a cardiac surgery center in London reported a genotypically closely related C. auris   infection in the period 2015–2016, possible or proven C. auris constituted 44% (22/50) including Candidemia 18% (9/50) [20].

Mortality from bloodstream candidiasis is high, reaching up to 31.8%, especially if was associated with delayed empiric treatment, and failure to remove the causative central venous catheter [21,22]. The study aims to shed light on Candida in Jordan; its nosocomial prevalence, susceptibility, and the associated mortality (then after crude mortality) adding to the knowledge about this common yeast in the region.

Settings and study conduct

A multicenter prospective study focusing on the epidemiology, susceptibility, and in-hospital mortality of patients with candidiasis. The study was conducted in the period January 2018 to November 2019, it was conducted in Al Takhassusi, Al Khalidi and Jordan Hospitals, all located in Amman – Jordan, they encompass 650 beds including 60 ICU beds and serve as referral hospitals for patients coming from within and nearby countries, they contain operating rooms designed for cardiac surgery, kidney, and liver transplant surgery (Jordan hospital). The internal review and ethics board of each hospital approved the study. Laboratory logbooks and patients' records were reviewed for Candida growth, candida was considered as an isolate, except blood-growth candida or when isolated from a sterile body site then it was considered as invasive candidiasis. Patients' records were queried for other clinical data, including, the characteristics and demography of patients, diagnoses, prescribed antibacterial agents, comorbidities, antifungal agents prescribed as prophylaxis or treatment, duration of treatment or prophylaxis, the body site from which candida was isolated, white blood cell counts, and the site of surgery (Figure 1).

Figure 1: The prevalence of candida species in hospitalized patients in Jordan expressed as percent (the total number=307).

*Note that Candida albicans makes up 34.5% (106) Non-albicans Candida makes up 65.5% (201) of the isolated Candida species.

Outcome measures

To identify the epidemiology of nosocomial Candida infections, the prevalence of the regional candida species including the notorious C. auris   should it suspected or uncovered, their susceptibility patterns, and the mortality associated with isolating Candida or invasive candidiasis [23,24].

Candida identification and susceptibility

General-purpose media that are commonly used for fungal culture, Sabouraud dextrose agar is used here. The standard temperature for incubation of fungi is 300C and cultures are incubated in a humidified environment. Once colonies are visible, they were inspected carefully for their morphology and germ tube. An updated version 8.1 software VITEK-2 (BioMérieux Marcy l'Etoile. 376, Chemin de l'Orme. 69280 Marcy l'Etoile. France) is used for yeast identification and susceptibility. A reagent card for the identification of different organism classes is used. A Suspension is prepared, a sterile swab or an applicator stick is used to transfer a sufficient number of colonies of a pure culture and the microorganisms were suspend in 3.0 ml of sterile saline (aqueous 0.45% to 0.50% NaCl, pH 4.5 to 7.0) in a 12 × 75 mm clear plastic (polystyrene) test tube. The turbidity is adjusted for yeast at I.80-2.20, accordingly turbidity is measured using a turbidity meter: the DensiChekTM (BioMérieux).

Statistical analysis

The total prevalence of Candida species isolates and invasive strains was calculated as means, medians and trimmed mean per the number of admissions. Also, calculating the numbers and the rates of Candida and their susceptibility, and most common morbidities among the isolates and the invasive strains (descriptive statistics), Chi-square (Χ2) and 2-sided Fisher's Exact Test is used to assess the differences in the categorical and continuous variables and ANOVA for the significant differences among means. Data were collected by upload to a Google cloud form, an excel sheet is generated and imported to an SPSS version 22 (IBM corporation) for analysis.

There were 307 non-duplicate Candida isolates, the prevalence of Candida species calculated per admission was 15.2% and the invasive species was 2.48%. The overall Candida mean length of hospital stay was 19 days (95% CI, 15.87 – 22.18, trimmed mean 15.18 days), for C. albicans the mean was 18.53 days (95% CI 17.4–19.65) and for Non-albicans Candida the mean was 20.46 days (95% CI 19.77 ‒ 21.15).

C. albicans accounted for 34.5% (106) and Non-albicans Candida 65.5% (201) (p=0.000). No significant gender difference was found between C. albicans and non-albicans Candida (p=0.148), and no significant difference in age (p=0.305). Comorbidities; haemato-malignancy, solid tumors, diabetes mellitus, central venous catheter use, the used antimicrobials, and the different immunodeficient states, were not significantly different between C. albicans and non-albicans Candida (p=0.194), neither were the rates of the surgical intervention (p=0.166). Low white blood cell count did not differ significantly between the two Candida groups (p=0.595). Seventy-nine patients died; 31 in C. albicans group and 48 in the non-albicans Candia group, but there was no significant difference (p=0.337) (Table 1).

Table 1: The features and characteristics of patients with 307 Candida species.

Candida was collected from several sources; numerically C. albicans was more than each other Candida species, but as a group Non-albicans Candida was numerically higher than C. albicans (Table 2 and Figure 1). The non-albicans candida were; C. glabrata 55 (17.9%), C. tropicalis 39 (12.7%), C. ciferii 20 (6.5%) and C. famata 14 (4.6%). The other were C. krusei 3.3%, C. parapsilosis 2.9%, C. kefyr 2.6% and C. lusitanae 2%. The unidentified Candida species were collected before the update of VITEK 2 database to version 8.01, the database update may have identified some of the "Others" including C. auris   (Figure 2). The invasive strains; blood 6 C. albicans and 19 Non-albicans, the peritoneum recovered 6 invasive C. albicans and 19 Non-albicans Candida (Table 2). There was no significant proportion difference within both Candida groups between invasive strains and isolates (p=0.104). Associated mortality among patients was significantly more in all patients with invasive Candidiasis rather than isolates (p=0.000). Nested analysis demonstrated mortality to significantly differ within non-albicans Candida between isolates and invasive infection (p=0.000) but not for C. albicans (p=0.150), (Table 3).

Figure 2: Candida species isolated from various sources, C. albicans was considered separate versus all other species.

Table 2: The distribution of all Candida species isolated from different sources.

Table 3: The mortality rates among invasive C. albicans and non-albicans infections compared with colonizers.

Among 73 tested C. albicans species 94.5% were susceptible to Fluconazole with MIC distribution for all strains ≤ 1 µg/mL but one isolate (5.5%) was resistant (16 µg/mL). Among 35 C. tropicalis isolates, one (2.8%) was fluconazole-resistant. C. glabrata were 16 isolates with 68.8% susceptibility, and among 7 isolates of C. parapsilosis 85.7% were susceptible to fluconazole. Among 68 C. albicans species susceptibility to voriconazole was 100%, and the same 100% susceptibility for C. tropicalis (n=34), C. krusei (n=8) and C. parapsilosis (n=7). For echinocandins, C. albicans was 100% susceptible (n=72 for micafungin and 73 for Caspofungin). Also, 100% susceptibility for both echinocandins; C. tropicalis (n=34 and 33), C. krusei (n=9 and 9), and C. parapsilosis (n=7 and 7) respectively, but susceptibility in C. glabrata was 97.8% for micafungin, and 95.6% for caspofungin (n=46 and 46) respectively (Figure 3).

Figure 3: The distribution of the minimum inhibitory concentration (MIC) of Candida species to some antifungal agents.

Numbers next to Candia names denote the CLSI breakpoint; those without numbers next to them have no established CLSI breakpoints.

CLSI: clinical laboratory institute

There are no Amphotericin-B CLSI breakpoints for Candida species (25), however, the CLSI most frequent low MIC for Amphotericin-B is (0.5-2) µg/mL that includes 92% - 99% of species. Our isolates showed that Amphotericin-B MIC of <2 µg/mL for C. albicans was 93.6% (3/47) and for Non-albicans Candida 92.9% (7/99), the 7 out of range (unsusceptible) species were: 3/16 C. Ciferii, 2/25 C. glabrata , 2/10 C. famata (Figure 4). No multidrug-resistant Candida was observed when tested by several antifungal agents.

Figure 4: The distribution of minimum inhibitory concentration (MIC)* of Amphotericin B for C. albicans and Non-albicans Candida.

*Note that are no CLSI breakpoints for Amphotericin B for Candida species.

Nosocomial Candida infections are increasing in hospitalized patients. With the worldwide growth in the immunocompromised population, they account for 7% – 10% of patients. In our patients colonization accounts for 15.2% of admitted patients and the invasive infections accounted for 2.48% (considering blood and peritoneum as a proven invasive infection), the lower rates in our population may be due to exclusion of other sources, as the study design is not powered to identify invasive strains from all sources, and importantly due to the nature of the patients population where sicker patients and those who need longer hospital stay are being treated in public hospitals, and a referral cancer center in Jordan [26].

There were no significant distribution differences between C. albicans and the non-albicans Candida in gender, age, or for the other comorbidities (p>0.05). Though surgery is a known risk for candida infections [27] the type of surgery did not significantly differ in the two Candida groups (p>0.05), as well as the white blood cells count (Table 1). Like elsewhere the non-albicans Candida was significantly (p=0.000) dominant accounting for 65.5%, and C. glabrata and C. tropicalis were commonly isolated from most sources, C. glabrata vaginal isolates were 12.5% (1/8) (Table 2), this is (Figure 2) due to the low vaginal isolates (n=8) and the female population at two study hospitals were dominated by a healthy parturient women with a routine vaginal swabs before delivery [28]. A study from Mexico quoted higher rate for C. glabrata (19%), but the study included women with symptoms of vaginal candidiasis [29].

The susceptibility of Candia species (Figure 3) against the tested antifungal agents are all positively skewed, where the bulk of the MIC values are skewed to lower values. Fluconazole has a descent activity against C. albicans and C. tropicalis, but it was poor for C. glabrata; resistance rates in C. albicans was 5.5% (n=73), similar to an earlier report [12], and higher than other reports of 2%. C. tropicalis resistance rate was 2.8% (n=35) versus other reports of 9%. Resistance in C. parapsilosis was 14.3 % (n=7) versus up to 6%, and C. glabrata resistance here was 30.2% (n=16) versus up to 13%. Evidently, our resistance patterns are higher than what was reported, possibly driven by the wide-scale use of fluconazole and the oral antifungal gel/solution in the country in addition to the small sample size [30]. Other Candida species with no CLSI breakpoints have variable susceptibilities to fluconazole, and no comment can be done at the moment.

The susceptibility to voriconazole was 100% in C. albicans, C.tropicalis, C. Krusei and C. parapsilosis similar to what was reported from our region [31]. Candida susceptibility to the tested two echinocandins was 100% except for C. glabrata; for micafungin it was 97.8% and caspofungin 95.8% [32]. Though there are no clear CLSI breakpoints used for Amphotericin B, again the MIC distribution for Candia species is positively skewed, where the majority of the tested isolates were in the anticipated susceptibility region; C. albicans was 93.6% susceptible and non-albicans was 92.9% Susceptible (Figure 4).

No single isolate of C. auris was identified or suspected by Vitek 2 with the new updated software (version 8.1 YST-AST card) though it has its limitations in identifying C.auris   [(https://www.cdc.gov/fungal/candida-auris/recommendations.html), (https://www.cdc.gov/fungal/diseases/candidiasis/pdf/Testing-algorithm-by-Method-temp.pdf). However, there was no single isolate that was a multidrug-resistant among the few resistant cases to a single agent, including C. duobushaemulonii, C. haemulonii, or C. famata [33]

The associated mortality rate with Candida isolation among our patients was 25.7%; with C. albicans was 29.2% and non-albicans Candida was 23.3% (p=0.337). The associated mortality with the invasive candidiasis was (25/50) 50%, this high rate may be due to the selection of blood and peritoneum as invasive, while in a review it was 40% possibly due to the consideration of other sources [34]. The mortality associated with the non-invasive isolates of candida was (54/257) 21.0%, but there was no significant mortality difference between isolates and invasive strains for C. albicans (p=0.105), though there was a significant difference for the non-albicans Candida associated mortality (p=0.000). Also, there was no significant mortality difference for the proportions of the invasive strains between C. albicans (6/31=19.35%) and non-albicans Candida (19/48=39.58%) (p>0.05).

Non-albicans Candida was more in prevalence than C. albicans; C. galbrata, C. tropicalis and C. ciferii are the most predominant, followed by C. famata, C. krusei and C. parapsilosis. Our antifungal resistance rates are higher than what was reported, and the mortality from Non-albicans was not significantly higher than C. albicans. No single case of C. auris was documented or suspected in our Candida species.

We thank the help provided by: Deema Al Jammal, Nisreen Al Radaydeh, Marwa Abu Saad, Rasmiyyeh Abu Kwaik, and Walid Jamal from the microbiology laboratories.

JW receives honoraria for his lectures from Pfizer, MSD, Gilead and Hikma pharmaceuticals. All other authors have nothing to disclose.