Epidemiology and Virulence Determinants including Biofilm Profile of Candida Infections in an ICU in a Tertiary Hospital in India

The purpose of this prospective study was to isolate, speciate, and determine antifungal susceptibility and virulence patterns of Candida species recovered from the intensive care units (ICUs) in an Indian hospital. Study included 125 medical/postoperative patients admitted to ICU. Identification and speciation of yeast isolates were done by the biochemical methods. Antifungal susceptibility was done by broth microdilution method. Virulence testing of Candida species was done by phospholipase, proteinase, and adherence assay. A total of 103 Candida isolates were isolated; C. tropicalis was the predominant species (40.7%), followed by C. albicans (38.83 %), C. glabrata (11.65%), C. parapsilosis (3.88%), and 1.94% each of C. krusei, C. kefyr, and C. sphaerica. 60Candida isolates (58.25%) showed resistance to fluconazole, while 7 (6.7%) isolates showed resistance to amphotericin B. Phospholipase and proteinase activities were seen in 73.8% and 55.3% Candida isolates with different species showing a wide range of activities, while 68.9%Candida isolates showed 4+ adherence activity.The present study revealed that nonalbicans Candida species (NAC spp.) causedmost of the cases of Candidemia in the ICU patients.The isolation ofC. tropicalis from a large number of cases highlights the ability of this pathogen to cause bloodstream infections.The presence of azole resistance is a matter of concern.


Introduction
Fungal infections are emerging as a problem of significant magnitude in hospitals during the last decade, especially in the ICUs, which are epicenters for infections such as Candidemia and invasive Candida infection (ICI).The escalating problem in ICUs is probably due to an increasing population of immunocompromised patients.A survey of the epidemiology of sepsis conducted in USA revealed that the incidence of fungal sepsis increased threefold between 1979 and 2000 [1].Fungal infections account for nearly 8% of all nosocomial infections; Candida is the responsible agent in 80% of the cases [2].Candida species are approximately the fourth most common cause of nosocomial infections in ICUs, according to data from the National Nosocomial Infections Surveillance System and the European Prevalence of Infection in Intensive Care [3].
Critically ill patients who are treated in the ICUs are very susceptible to infections due to acquired defects in host defense mechanisms from the immunosuppressive effect of the underlying disease, recent surgery, trauma, and concurrent drug therapy.Infections occur in 15-40% of all ICU admissions and central venous catheters are considered as the most common risk factor for the development of Candidemia in these patients [4,5].The Candida biofilm results in antifungal drug resistance and protection of the fungus from host defenses [6], which carry important clinical repercussions.
Candida albicans has historically been the most frequent cause of Candidiasis.During recent decades, several

Materials and Methods
The present study was conducted from 2006 to 2008 at the Department of Microbiology, Maulana Azad Medical College and Associated Lok Nayak Hospitals, New Delhi, India, which is a 1500-bedded tertiary care hospital.One hundred and twenty-five medical/postoperative patients admitted to ICU for >48 hours undergoing therapy for one or more acute organ system failure or requiring intensive postoperative monitoring were studied.A total of 356 samples including blood, urine, tracheal aspirate, urinary catheter, endotracheal tube, and abdominal drains were collected and processed from the patients.Cases that had antifungal drug exposure during hospitalization were excluded from the study.All samples were inoculated on Sabouraud dextrose agar slants containing gentamicin (0.02 mg/mL) and cycloheximide (0.5%).One set of inoculated slants was incubated at 25 ∘ C and the other at 37 ∘ C, and they were examined on alternate days for the growth up to 4-6 weeks.
Identification and speciation of the isolates were conducted by colony morphology, gram staining, germ tube formation, corn meal agar with Tween80, HiCrome Candida agar, and an enzymatic triphenyltetrazolium chloride reduction; for further characterization, each isolate was subjected to carbohydrate assimilation and fermentation tests as per standard recommended procedures [7].
The in vitro minimal inhibitory concentrations (MICs) of the compounds were determined by the broth microdilution method as per Clinical and Laboratory Standards Institute (CLSI) M27-A3 document using RPMI medium and MOPS buffer.The concentration ranges tested were 0.125-128 g/mL for fluconazole and 0.016-16 g/mL for amphotericin B. The quality control was performed by testing C. parapsilosis (ATCC22019) and C. krusei (ATCC6258) from the American Type Culture Collection (ATCC) with each batch of clinical isolates [8].The MIC breakpoints recommended by CLSI guidelines were followed (CLSI, 2008).For fluconazole, MIC breakpoints were as follows: sensitive, MIC ≤ 8 g/mL; susceptible-dose dependent (SDD), MIC 16-32 g/mL; resistant, MIC ≥ 64 g/mL.For amphotericin B, isolates with MICs of ≥ 1 g/mL were categorized as resistant (CLSI, 2008).
Candida proteinase was detected by the Mohandas and Ballal method [9] using bovine serum albumin medium.The proteinase activity (Pz) was determined in terms of the ratio of the diameter of the well to the diameter of the proteolytic zone [9].The extracellular phospholipase activity of the isolates was screened by growing them on egg-yolk agar and measuring the size of the zone of precipitation by the Samaranayake et al. method [10].Biofilm-formation was determined spectrophotometrically by the Branchini et al. method [11].

Results
The patients belonged to a wide age group from 13 years to 90 years with a mean age of 33 years.Out of the 125 study cases, 53 (42.4%) belonged to 16-30 years age group.Males were predominant with a male to female ratio of 1.2 : 1.In our study 86.29% patients were admitted to ICU for less than 1 week followed by 9.68% for 1-2 weeks, while very few patients (6%) had a stay of more than 2-3 weeks.The most common organ system involved in our patients was gastrointestinal system (27.2%cases) followed by central nervous system (19.4% cases) and respiratory system (19% cases).Diabetes mellitus was seen in 14.5% cases, while fever, multiorgan system involvement, and infection were seen in 37.5%, 17.5%, and 16% cases, respectively.
A total of 356 patient samples were collected.They included blood in 125 cases (100%), urine in 112 cases (89.6%), tracheal aspirate in 88 cases (70.4%), urinary catheter in 18 cases (14.4%), endotracheal tube in 10 cases (8%), and abdominal drains in 3 cases (2.4%).The most common risk factor found in our patients was that of presence of indwelling device in 100 patients followed by prolonged antibiotic therapy in 78 patients and surgery in 73 patients.Prolonged antibiotic therapy was also seen in Candidemia cases as well as Candida colonization cases (Table 1).

Discussion
Patients in ICU are at a higher risk of acquiring nosocomial infections compared with patients in general wards due to the severity of the underlying illnesses and iatrogenic factors related to the high frequency of invasive procedures needed for the monitoring and treatment [12] which include insertion of intravascular catheters, endotracheal intubation, and positive pressure ventilation, urinary catheterization and surgical operations.
Studies from different parts of the world on nosocomial infections have shown that patient and treatment factors are risk factors for the development of nosocomial infections.Some identified risk factors related to treatment have been length of hospital or ICU stay, presence of indwelling devices, and intake of drugs like broad spectrum antibiotics and immunosuppressants [13] which were also seen in our cases as 86.29% patients were admitted to ICU for less than 1 week followed by 9.68% for 1-2 weeks.According to Sydnor and Perl, admission to ICU itself was a significant risk factor; roughly 25% of nosocomial infections occur in intensive care units (ICUs), which have been estimated to increase ICU length of stay by 4.3 to 15.6 days [14].Length of stay in the ICU is also associated with increased risk for Candida infections, which rises rapidly after 7-10 days [1].History of smoking was elicited in 26% of cases corroborated by a recent study from Mumbai also showing the association of nosocomial pneumonia and mortality with tobacco smoking [15].Workers from Delhi have reported a higher risk of nosocomial infections in ICUs attributed to the severity of the underlying illness and the development of iatrogenic factors related to the high frequency of invasive procedures required in these patients [12].
Peritonitis, multiple organ system involvement, duration of stay >1 week, and immunosuppressant were significantly associated with the group with Candida colonization ( ≤ 0.05) unlike the Candidemia cases.The number of cases of Candidemia was may be too low to get a significant association.Similar association with antibiotic and immunosuppressant therapy along with catheterization and surgery in causation of nosocomial Candidiasis has been quoted in other studies in North India [16,17].
In ICU patients, the most common types of Candida infections are seen to comprise bloodstream, catheter-related, intra-abdominal, and urinary tract infections [6].In our patients, culture isolations were obtained in 39.8% of urine samples followed by 31.06% of tracheal aspirate in addition to a varying percentage from other samples like urinary catheter (10.6%), blood (8.7%), endotracheal tube (6.7%), and abdominal drains samples (2.9%).Different studies reported the predominance of C. albicans as a leading cause of invasive candidiasis.However, recently, several countries around the world have witnessed a change in the epidemiology of Candida infections, characterized by a progressive shift from a predominance of C. albicans toward a predominance of NAC species [1].Although C. albicans remains the most commonly isolated species from cases of Candidemia in USA, Europe, and South America (Brazil), its dominance is slowly giving way to an increase in NAC species such as C. glabrata [18].
However, in India, C. tropicalis appears to be emerging as the most common cause of Candidemia [18].50% of our cases of Candidemia were due to C. albicans followed by C. tropicalis (40%) and C. glabrata (10%).Findings of our study quite similar to the epidemiological data from the Indian subcontinent showed that 50% of our nosocomial candidemia cases were due to NAC spp. of which C. tropicalis was the most dominant [19].Interestingly, in our study, C. tropicalis was most often isolated from urine (35.71%) followed by tracheal aspirate (26.19%), urinary catheter (14.28%), endotracheal tube (11.90%), blood (09.52%), abdominal drains (2.38%), and sputum (0%).C. albicans was more often isolated from urine (42.50%) and tracheal aspirate (32.50%).However, C. parapsilosis, C. krusei, C. kefyr, and C. sphaerica were isolated only from urine, urinary catheter, and tracheal aspirate.36.9% of isolates of C. tropicalis were from colonization sites (Table 2).Some of these emerging species have been correlated with increased virulence.An increasing role for NAC spp. was also noticed in studies performed among ICU patients, although the issue is somewhat controversial [20].The increased use of fluconazole has been seen to be associated with the predominance of many NAC spp., especially C. tropicalis species [21].Findings in our study have confirmed and corroborated the earlier reports of increased resistance to fluconazole among isolates of C. tropicalis from the Asia-Pacific region [22].Also C. parapsilosis, C. krusei, C. kefyr, and C. sphaerica have shown a predilection towards isolation from urinary catheters and tracheal aspirates in our cases, explaining their role in inherent biofilm formation leading to difficulty in eradication from the site and development of resistance in their long stay.Several investigators have postulated that the widespread use of fluconazole would have a big role in selecting out yeast species intrinsically resistant or less sensitive to fluconazole, such as C. krusei, C. glabrata, or C. tropicalis [20].In our study, a large proportion of the C. tropicalis (57.14%, 24/42) was resistant to fluconazole; 90.4% of C. tropicalis isolates were responsible for Candida colonization with many of them (50%) being involved in biofilm formation.Punithavathy et al. [23] from South India, however, reported that (22%, 11/50) isolates of C. tropicalis were resistant to fluconazole.
Blood stream isolates of C. albicans, C. tropicalis were 40% and 25% while the urinary isolates were 70.6% and 60% resistant to fluconazole respectively, in comparison to a higher resistance going up to 100% seen in isolates of endotracheal tubes, tracheal aspirate and abdominal drains.Our results of 25% resistance to fluconazole in blood stream isolates of C. tropicalis were consistent with a report from Maharashtra and Delhi, in which 37% and up to 83%, respectively, were resistant to fluconazole [24,25], while the isolates from urine and indwelling devices showed a higher resistance quite like a similar picture being seen in Chennai, South India, reflecting the high level of resistance in colonized isolates involved in biofilm formation predisposing to candidemia [26].
Candida glabrata has also become important because of its increasing incidence worldwide; it was seen in 11.6% of our study cases with 10 out 11 isolates from colonization sites exhibiting a high resistance to azoles.C. krusei, found in 2 cases again from colonization sites in our study, is another important Candida species which is gaining clinical significance because of its intrinsic resistance toward azoles and poor susceptibility to all other antifungals, including amphotericin B [27].C. sphaerica was isolated from urine and urinary catheter samples only and 50% of C. sphaerica isolates were resistant to fluconazole, while no resistance to amphotericin B was seen.On the other hand both the isolates of C. kefyr from urine and tracheal aspirate were sensitive to fluconazole and amphotericin B. However, in a study in Turkey, all the C. sphaerica strains (3/3) were sensitive to Fluconazole and resistant to amphotericin B and 25% (2/8) Candida kefyr strains were susceptible dose dependent (SDD) and 37.5% (3/8) sensitive to fluconazole and 37.5% (3/8) were SDD and 25% (2/8) resistant to amphotericin B as reported by Abaci and Haliki-Uztan [28] (Table 3).
Virulence activity of C. albicans and C. tropicalis isolates from blood samples were 40% and 75% for phospholipaseactivity, 60% and 25% for proteinase-activity and 100% each for adherent-activity, respectively.But the same was much more in the biofilm isolates from urinary catheter, endotracheal tube, and tracheal aspirate: 33.3% to 100% for phospholipase activity, 61.5% to 100% for proteinase activity, and 100% adherent activity in the urinary catheter, endotracheal tube and tracheal aspirate going up to 100% phospholipase activity, proteinase activity, and adherent activity in abdominal drains.
The proteinase-producing capacity of C. albicans (42.10%) was less than that of NAC spp.(57.89%) in our patients, whereas workers from Manipal showed NAC spp.(50.45%) to be lower proteinase producers than C. albicans (67.34%) [29].In our study among NAC spp.C. tropicalis (66.66%) was the highest proteinase producer, whereas C. kefyr showed no proteinase activity consistent with studies of Al-Nedaithy [30] which showed C. tropicalis (95%) to be proteinase producer and no activity with C. kefyr.50-100% of our Candida isolates showed phospholipase activity similar to the previous studies conducted by Price et al. [31] and Wu et al. [32] who reported phospholipase activity in 30-100% of their Candida isolates from various groups of patients and from various sites.Also a total of 22 (28.95%) of C. albicans isolates and 54 (85.7%) of NAC spp.isolates produced phospholipase in our cases, while Mohandas and Ballal [29] from Manipal and Ibrahim et al. [33] reported 23 (46.93%) of their C. albicans isolates and 26 (42%) of NAC spp.isolates producing phospholipase with greater extracellular phospholipase activity.In our study among NAC spp., C. tropicalis (64.81%) was the highest phospholipase producer concordant with studies of Negri et al. [34] and Júnior Jr. et al. [35] showing C. tropicalis to be the species with the highest number of positive isolates for phospholipase activity (up to 91.7%).
Biofilm production has been seen to promote persistence of infection as was evidenced in Candida isolates of OPC patient in a major medical center in India [36].Our study showed that 60.78% of isolates of NAC spp.produced biofilm in comparison to 39.21% of C. albicans producing lower slime (Figure 1).Shin et al. [5] and Pathak et al. [37]  and C. albicans [5].Kumar et al. [38] also reported C. tropicalis to be strongly positive for biofilm production.However, Hawser and Douglas [39] reported that isolates of C. parapsilosis and C. glabrata were significantly less likely to produce biofilms than the more pathogenic C. albicans.Biofilm may be a specific trait of a Candida that is not associated with differences in adherence but with slowness of growth; however, Candida biofilms contribute both to the pathogenesis of superficial and systemic Candidiasis as they are notoriously resistant to antifungal drugs [36].Advances in modern medicine and patient management have seen increased use of prosthetic biomaterials.This has led to the complication associated with biofilm formation on these materials, the constituent organisms of which are different in their growth rate and antifungal susceptibility as compared to isolates not associated with biofilms.The knowledge of the local epidemiological trends in the Candida species isolated is important to guide therapeutic choices.Our study demonstrates a shift of C. albicans predominance towards a NAC spp. in causation of candidiasis among ICU population with the isolation of C. tropicalis from a large number of cases highlighting the importance of this pathogen.The presence of high resistance to fluconazole by C. albicans followed by C. tropicalis is another important finding in our study and may have important clinical repercussions.It was seen that NAC spp.had greater biofilm-forming ability when compared to C. albicans species especially C. tropicalis showed the highest ability to produce phospholipase, proteinase, and biofilms.Thus in the light of virtual absence of such data from our area, this study would help in understanding the properties of the biofilms, with respect to growth characteristics and drug sensitivity, and should be a great help in understanding the present trends and predicting the overall long-term trends that will be helpful to physicians and to antibiotic use committees in establishing guidelines for the appropriate use of antifungal agents.This will help in reducing complications associated with biofilm formation and devise control strategies and consequently reducing morbidity, mortality, and treatment costs.The results obtained here open perspectives of new investigations that are aimed at establishing and broadening knowledge of understanding relationship between strain types and properties such as pathogenicity, commensality, and infectivity, an important interplay in host-pathogen relationship specially as far as Candida infections are concerned.

Table 4 :
Sample wise in vitro antifungal resistance profile of Candida isolates to fluconazole and amphotericin B. .FLU: fluconazole drug.AMB: amphotericin B drug.

Table 1 :
Analysis of potential risk factors found for Candida infection in ICU patients.
resistance to fluconazole.The results of Table2showed that 67.5% C. albicans were resistant to fluconazole followed by 57.14% C. tropicalis, 50% each of C. parapsilosis and C. sphaerica, and 45.4% C. glabrata, while a lower resistance to amphotericin B of 4.7%, 5%, and 9% was seen by C. tropicalis, C. albicans, and C. glabrata, respectively.The blood stream a : patients admitted to the same ICU with no Candidemia.Significant  values ( ≤ 0.05) are indicated in bold.NA: not applicable.

Table 2 :
Profile of Candida species isolated from different clinical samples.
isolates of C. albicans and C. tropicalis showed a resistance of 40% and 25%, respectively, to fluconazole in contrast to the urinary and the indwelling isolates which were much more resistant (Table4).Phospholipase activity was detected in 73.78% isolates.Out of 73.78% phospholipase positive isolates, 71.05% isolates were NAC spp.and 28.95% were C. albicans.Among NAC spp., C. tropicalis (64.81%) was the highest phospholipase producer followed by C. glabrata (20.37%), whereas C. kefyr (1.85%) was least producer.The proteinase activity was seen in 55.33% Candida isolates.Out of 55.33% proteinase positive isolates, 57.89% were NAC spp.and 42.10% were C. albicans.

Table 3 :
The in vitro antifungal susceptibilities of Candida isolates to fluconazole and amphotericin B.
also reported higher biofilm-forming ability of NAC spp.than C. albicans species.Our data also suggest the production of stronger biofilms by C. tropicalis (40.19%) followed by C. albicans (39.21%).Weak biofilm production was seen with C. kefyr, C. krusei, and C. sphaerica (1.96% each).Biofilm positivity has also been seen to occur most frequently in C. tropicalis isolates followed by C. parapsilosis, C. glabrata,