Secondary Logo

Journal Logo

Antimicrobial Reports

Spectrum and Antifungal Resistance of Candidemia in Neonates With Early- and Late-Onset Sepsis in Pakistan

Rattani, Salima MBBS*; Farooqi, Joveria FCPS*; Hussain, Ali Shabbir FCPS; Jabeen, Kauser FCPS*

Author Information
The Pediatric Infectious Disease Journal: September 2021 - Volume 40 - Issue 9 - p 814-820
doi: 10.1097/INF.0000000000003161
  • Free

Abstract

Candidemia is an emerging cause of increased morbidity, mortality and overall healthcare cost.1 It is commonly encountered in premature infants, patient receiving long-term broad-spectrum antibiotics, immunocompromised individuals or post-abdominal surgeries.2 Its prevalence in low-middle income countries (LMIC) like Pakistan, India and Bangladesh has been estimated to be high.3–6 Neonatal candidemia has also been on the rise globally.4,7,8 Studies from Pakistan reported non-albicans Candida species to be causative agents in 73.5%–78% cases of neonatal candidemia.2 Neonatal invasive fungal infections are divided into early-onset disease (EOD ≤7 days) and late-onset disease (LOD >7 days).9 Its reported incidences ranged from 1.6% to 9% in very low birth weight (VLBW) and 10%–16% in extremely low birth weight (ELBW) infants.7 Factors associated with neonatal candidemia include use of total parenteral nutrition, presence of central lines, prematurity, gastrointestinal surgery, artificial ventilation and prior fungal colonization.9–11

Few recent studies evaluating the etiologic spectrum and antifungal resistance in neonatal candidemia in developing countries like Pakistan are available.12 More specifically the difference in the spectrum and risk factors between early- and late- neonatal disease has not been studied. This knowledge could be used for appropriate empirical treatment of early- versus late-onset neonatal candidemia. Thus, we evaluated etiology, antifungal resistance and trends in isolation of non-albicans Candida species in blood cultures from neonates.

METHODS

This was a cross-sectional study of laboratory data, conducted at the Aga Khan University Hospital (AKUH), Karachi, Pakistan. This is a central laboratory receiving annually 112,500 blood culture samples from all over Pakistan through a widespread network of collection centers across the country.

After exemption from the AKUH ethics review committee (Reference # 2019-2031-5603), a laboratory database was used to extract information on all Candida positive blood cultures from 2014 to 2019 in neonates of age 0–28 days. For all patients, information regarding age in days at the time of sample collection and gender were available. For the year 2019, additional information about gestational age, birth weight, mode of delivery, the reason for hospital admission and use of antibiotics was collected as part of routine clinical reporting. Outcome data were not collected as follow-up phone calls were not made.

Neonatal candidemia was defined as blood culture positive for any Candida species in neonates at ≤7 days of life (EOD) and at >7 days of life (LOD), respectively.9 Duplicate cases were excluded if more than one blood cultures from a single patient were positive for the same Candida species.

Gestational age was classified as post-term or >42 weeks, term or 37–42 weeks, moderate to late preterm or 32 to ≤37 weeks, very preterm or 28 to ≤32 weeks, extremely preterm or <28 weeks13 and birth weight was classified as normal birth weight or 2500–3500 g, low birth weight or <2500 to ≥1500 g, VLBW or <1500 to ≥1000 g, ELBW or <1000 g.14 There were no infants with birth weight above 3500 g. Neonates with blood cultures positive for bacteria before current candidemia and those with blood cultures positive for both bacteria and Candida species at the same time were classified to have prior bacterial infection and coinfection with bacteria, respectively.

Species were identified based on conventional phenotypic characteristics which included production of a germ tube, morphology on BBL BiGGY Agar (BD Diagnostic Systems, Sparks, MD) and chrome agar, growth with cycloheximide, urease production and morphology on cornmeal/Tween 80 agar. For isolates that could not be identified by these methods (excluding C. albicans, Candida tropicalis and Candida parapsilosis), the identification profile was generated using API 20C AUX (BioMérieux, Durham, NC). The Candida species which could not be identified to species level with confidence on API 20C AUX were termed “rare Candida species.” “Other Candida species,” subsequently, will refer to Candida species other than the one mentioned before the current phrase and may include rare Candida species as well.

Antifungal susceptibility testing was performed by disc diffusion for fluconazole and voriconazole for C. albicans, C. tropicalis and C. parapsilosis, and interpreted according to CLSI M60. For all other non-albicans Candida species, antifungal susceptibility testing was performed by broth microdilution with fluconazole, itraconazole, voriconazole, posaconazole, anidulafungin, caspofungin, micafungin and amphotericin B as described by the Clinical and Laboratory Standards Institute using Sensititre YeastOne YO9 AST Plate. This algorithm was designed for cost-effectiveness. Candida glabrata with an MIC ≥64 µg/ml was considered resistant to fluconazole, MICs ≥0.5 µg/ml were considered resistant to caspofungin and anidulafungin and MIC ≥0.25 µg/ml was resistant to micafungin. Candida krusei was considered intrinsically resistant to fluconazole, MIC ≥2 µg/ml was resistant to voriconazole and MICs ≥1 µg/ml were considered to be resistant to caspofungin, micafungin and anidulafungin. Candida guilliermondii with an MICs ≥8 µg/ml were considered to be resistant to caspofungin, micafungin and anidulafungin.15 For other Candida species including Candida pelliculosa and Candida utilis, fluconazole was considered resistant at an MIC of ≥4 µg/ml in accordance with clinical breakpoints for fungi by EUCAST.16 In the absence of clinical breakpoints, epidemiologic cutoff values (ECVs) were used to interpret MICs as those conforming to wild type or non-wild type for a particular antifungal agent against a specific species. ECVs for fluconazole were 8 µg/ml for C. guilliermondii, 4 µg/ml for C. pelliculosa and 1 µg/ml for Candida lusitaniae. ECVs for voriconazole were 0.25 µg/ml for C. guilliermondii, C. glabrata and C. pelliculosa and 0.06 µg/ml for C. lusitaniae.16–18 Similarly, ECVs of anidulafungin were 4 µg/ml for C. guilliermondii, 1 µg/ml for C. lusitaniae, and those for caspofungin were 2 µg/ml for C. guilliermondii, 1 µg/ml for C. lusitaniae and 0.12 µg/ml for C. pelliculosa.17,18 ECVs for micafungin were 2 µg/ml for C. guilliermondii and 0.5 µg/ml for C. lusitaniae.17,18 The MICs of amphotericin and itraconazole were also interpreted according to ECVs.18,19 For amphotericin, the ECV for C. glabrata, C. krusei, C. guilliermondii and C. lusitaniae was 2 µg/ml. For itraconazole, ECVs were 1 µg/ml for C. lusitaniae and C. krusei, 2 µg/ml for C. guilliermondii and 4 µg/ml for C. glabrata.

Means and SD were used to compute continuous variables like age in days. Frequency and percentage were used to analyze qualitative variables like gender, birth weight category, gestational age category, antimicrobial therapy, abdominal distension and prior or coinfection with bacteria. Risk factors associated with early- and late-onset candidemia were determined. χ2 tests and Fisher exact tests were applied, as appropriate. Univariate analysis was performed for all variables and those with P < 0.2 were included in multivariable regression for which P value of <0.05 was taken as significant. The crude odds ratio (COR) and 95% confidence interval (CI) were calculated for variables found to have significant associations.

RESULTS

A total of 669 neonates with blood cultures positive for Candida species were identified from the year 2014 to 2019. The geographical origin of these isolates was from patients admitted to AKUH, Karachi (77), other private and public healthcare facilities in Karachi (435) and other cities of Pakistan including Hyderabad, Nawabshah, Larkana, Badin, Tando Mohammad Khan, Mithi, Mathelo from Sindh province (115); Lahore, Multan, Bahawalpur, Sialkot, Wah Cantt, Rahim Yar Khan and Sahiwal from Punjab province (77); Peshawar and Mardan from Khyber Pakhtunkhwa province (7); and 7 from Quetta, Balochistan province. Of these, 346 neonates had EOD while 323 had LOD. Pure growth of Candida species was isolated from 621 patients while multiple Candida species were identified in 48 neonates. Bacterial coinfection was present in 300 cases, which included coagulase-negative Staphylococcus species (37%), Burkholderia cepacia (26.3%), enterobacterales including Escherichia coli, Klebsiella pneumoniae, Enterobacter species, Serratia species (11.3%), Acinetobacter species (3.3%), Enterococcus species (2.7%), Staphylococcus aureus (1%), Pseudomonas aeruginosa (0.7%) and other Pseudomonas species (0.7%). Multiple bacteria were isolated in 17% of cases with bacterial coinfection. Table 1 shows the demographic and clinical characteristics of neonates with candidemia. EOD was found to be associated with non-albicans Candida species, respiratory distress and bacterial coinfection. LOD was associated with C. albicans, prior bacterial infection, other reasons for admission than necrotizing enterocolitis, respiratory distress syndrome, birth asphyxia and presumed sepsis.

TABLE 1. - Demographic and Clinical Characteristics of Neonates with Candidemia
Variable All cases, n (%) EOD, n (%) LOD, n (%) P
Total cases (%) 669 346 (51.7) 323 (48.3)
Mean age in days (range) 9.7 (0–28) 3.7 (0–7) 16.1 (8–28)
No. males (%) 425 (63.5) 220 (63.6) 205 (63.5) 0.975
 Candida species 0.000*
 C. albicans 89 (13.3) 30 (8.7) 59 (18.3)
Non-albicans Candida species 580 (86.7) 316 (91.3) 264 (81.7)
Coinfection (%) 300 (44.8) 170 (49.1) 130 (40.2) 0.021*
Prior bacterial infection (%) 28 (4.2) 7 (2) 21 (6.5) 0.004*
Gestational age n = 105 n = 43 n = 62
Full term 37–42 weeks 43 (41) 22 (50 21 (34.4) 0.109
Moderate to late preterm 32 to ≤37 weeks 39 (37.1) 16 (36.4) 23 (37.7) 0.888
Very preterm 28 to ≤32 weeks 178 (16.2) 4 (9.1) 13 (21.3) 0.093
Extremely preterm <28 weeks 2 (1.9) 1 (2.3) 1 (1.6) 0.815
Birth weight n = 91 n = 34 n = 57
Appropriate for gestational age 61 (67.0) 14 (41.2) 27 (47.4) 0.566
Low-birth weight <2500 to ≥1500 g 34 (37.4) 16 (47.1) 18 (31.6) 0.140
Very low-birth weight <1500 to ≥1000 g 14 (4.8) 3 (8.8) 11 (19.3) 0.180
Type of delivery n = 76 n = 33 n = 43
Vaginal 21 (27.6) 12 (36.4) 9 (20.9) 0.136
C-section 55 (72.4) 21 (63.6) 34 (79.1)
Reason for admission n = 124 n = 57 n = 67
Necrotizing enterocolitis/abdominal distension 23 (18.5) 7 (12.3) 16 (23.9) 0.098
Respiratory distress syndrome/pneumonia 67 (54) 39 (68.4) 28 (41.8) 0.003*
Birth asphyxia 8 (6.5) 5 (8.8) 3 (4.5) 0.332
Presumed sepsis 23 (18.5) 9 (15.8) 14 (20.9) 0.466
Others 20 (16.1) 4 (7) 16 (23.9) 0.011*
Antibiotic usage n = 107 n = 47 n = 60
Cephalosporin use 59 (55.1) 25 (53.2) 34 (56.7) 0.720
Carbapenem use 72 (67.3) 28 (59.6) 44 (73.3) 0.132
Colistin use 33 (30.8) 11 (23.4) 22 (36.7) 0.140
Vancomycin use 50 (46.7) 17 (36.2) 3355 0.053
Fluconazole prophylaxis use 4 (3.7) 0 (0) 4 (6.7) 0.068
*Represents significant association between risk factors and neonatal candidemia.
†N is the number of patients for each category of risk factors for which clinical detail was available.
Early-onset disease (EOD): neonatal candidemia at ≤7 days of life. Late-onset disease (LOD): neonatal candidemia at >7 days of life.

C. albicans was the causative organism in 13.3% (n = 89) cases of neonatal candidemia while non-albicans Candida species was isolated in 86.7% (n = 580) patients. This was statistically significant with neonatal candidemia being predominantly caused by non-albicans Candida species as compared with C. albicans over the past 6 years from 2014 to 2019 (P-value 0.024). Year-wise statistical analysis revealed that as compared with the year 2014, the isolation of non-albicans Candida species causing neonatal candidemia was significantly higher than C. albicans during the years 2016 (COR 3.97, 95% CI: 1.73–9.12, P-value 0.001), 2017 (COR 2.9, 95% CI: 1.24–6.79, P-value 0.01) and 2018 (COR 3.35, 95% CI: 1.35–8.05, P-value 0.007) but not for the year 2019 (COR 2.06, 95% CI: 0.94–4.53, P-value 0.07). Figure 1 represents year-wise distribution of isolation of Candida species in neonates with candidemia. C. tropicalis followed by C. parapsilosis were most common Candida species isolated. Analysis of data from neonates admitted at AKUH also showed a predominance of non-albicans Candida species (74.4%) as compared with C. albicans (25.6%) with C. parapsilosis (28.2%) being the most common.

FIGURE 1.
FIGURE 1.:
The chart shows the trend of isolation of Candida species from 2014 to 2019 with Candida tropicalis being the most common Candida species. Isolation of Candida albicans has decreased. The isolation of Candida parapsilosis and other Candida species have increased in neonates with candidemia. Candida species include C. lusitaniae, C. guilliermondii, C. utilis, C. glabrata, C. rugosa, C. krusei, C. kefyr and rare Candida species (other than C. auris, not identified by API 20C AUX – BioMérieux, Durham, NC).

The distribution of different Candida species and their association with early- and late-onset candidemia are shown in Table 2. C. pelliculosa and C. guilliermondii were associated with EOD while C. albicans was associated with LOD.

TABLE 2. - Characterization of Different Candida Species Isolated From Neonates With Candidemia in Pakistan from 2014 to 2019
Organism isolated All cases, n (%) Early-onset disease (EOD), n (%) Candida species with EOD compared with LOD: OR (95% CI), P Late-onset disease (LOD), n (%) Candida species with LOD compared with EOD: OR (95% CI), P
C. albicans 89 (12.4) 30 (8.1) 0.43 (0.27–0.68), 0.000 59 (17) 2.35 (1.47–3.76), 0.000
C. tropicalis 280 (39) 138 (37.3) 0.85 (0.62–1.15), 0.29 142 (40.8) 1.18 (0.87–1.61), 0.29
C. parapsilosis 108 (15) 55 (14.9) 0.96 (0.64–1.45), 0.86 53 (15.2) 1.04 (0.69–1.57), 0.86
C. pelliculosa 84 (11.7) 55 (14.9) 1.99 (1.23–3.23), 0.005 29 (8.3) 0.50 (0.31–0.81), 0.005
C. lusitaniae 49 (6.8) 25 (6.8) 0.97 (0.54–1.74), 0.92 24 (6.9) 1.03 (0.58–1.84), 0.92
C. guilliermondii 36 (4.9) 25 (6.8) 2.21 (1.07–4.57), 0.032 11 (3.2) 0.45 (0.22–0.94), 0.032
Other Candida species 1.27 (0.77–2.07), 0.348 0.79 (0.48–1.29), 0.35
C. rugosa 3 (0.4) 1 (0.3) 2 (0.6)
C. glabrata 4 (0.6) 2 (0.5) 2 (0.6)
C. krusei 12 (1.7) 8 (2.2) 4 (1.1)
C. kefyr 1 (0.3) 1 (0.3) 0 (0)
C. utilis 23 (3.2) 12 (3.2) 11 (3.2)
Rare Candida species 29 (4.0) 18 (4.9) 11 (3.2)
Early-onset disease (EOD): neonatal candidemia at ≤7 days of life. Late-onset disease (LOD): neonatal candidemia at >7 days of life. Values in bold type are statistically significant. C. albicans was associated with EOD. C. guilliermondii was associated with LOD. Rare Candida species include Candida species (other than C. auris) not identified by API 20C AUX (BioMérieux, Durham, NC).

The resistance to fluconazole and voriconazole in Candida species in neonates with EOD and LOD from 2014 to 2019 is shown in Table 3. Most isolates were susceptible to fluconazole (96%) and voriconazole (99%). Resistance to fluconazole was seen in C. glabrata (25%), C. tropicalis (0.7%), C. pelliculosa (13.1%), C. utilis (13.0%), C. guilliermondii (11.1%) and C. lusitaniae (8.2%) and rare Candida species (15.2%). The isolation of fluconazole nonsusceptible non-albicans Candida species was significantly higher in patients with early-onset (5.9%) versus late-onset (2%) neonatal candidemia (P-value 0.005; COR 2.73, 95% CI: 1.34–5.53). The MIC50 and MIC90 values of Candida species other than C. albicans, C. tropicalis and C. parapsilosis isolated from neonatal candidemia patients in the year 2019 are shown in Table 4. Two C. pelliculosa, 3 C. guilliermondii and 5 C. lusitaniae strains showed non-wild-type behavior in respect to fluconazole with an MIC of 8 µg/ml, 16–32 µg/ml and 2 µg/ml, respectively. Non-wild-type behavior with regards to voriconazole was seen in single C. lusitaniae and C. guilliermondii strain with MIC of 0.12 µg/ml and 0.5 µg/ml. Two strains each of C. pelliculosa and C. glabrata strains showed non-wild-type behavior with an MIC of 0.5 µg/ml and 0.5–1 µg/ml, respectively. All isolates displayed amphotericin B MICs of <1.0 µg/ml, well below the ECV of 2.0 µg/ml.

TABLE 3. - Resistance to Fluconazole and Voriconazole by Candida Species in Neonates With Early-onset Disease and Late-onset Candidemia in Pakistan from 2014 to 2019
Species Neonates EOD (age 0–3 d), n (%) Neonates LOD (age 3–28 d), n (%)
FLC-NS
22 (5.9)
VOR-NS
5 (1.4)
All isolates
370
FLC-NS
7 (2.0)
VOR-NS
3 (0.9)
All isolates
348
C. albicans 0 0 30 0 0 59
C. tropicalis 2 (1.4) 1 (0.7) 138 0 0 142
C. parapsilosis 0 0 55 0 0 53
C. glabrata 1 (50) 1 (50) 02 0 1 (50) 02
C. krusei NA 1 (12.5) 08 NA 0 04
C. pelliculosa* 10 (18.2) 1 (1.8) 55 1 (3.5) 1 (3.5) 29
C. utilis* 2 (16.7) 12 1 (9.1) 11
C. lusitaniae 1 (4) 0 25 3 (12.5) 1 (4.2) 24
C. guilliermondii 3 (12) 1 (4) 25 1 (9.1) 0 11
Other Candida species 3 (15) 20 1 (7.7) 13
EOD, early-onset disease; LOD, late-onset disease; FLC-NS, fluconazole nonsusceptible; VOR-NS, voriconazole nonsusceptible.
EOD: neonatal candidemia at ≤7 days of life. LOD: neonatal candidemia at >7 days of life. Other Candida species include C. rugosa, C. kefyr, C. famata and rare Candida species (other than C. auris) not identified by API 20C AUX (BioMérieux, Durham, NC).
*Fluconazole was considered nonsusceptible at an MIC of ≥4 µg ml−1 in accordance with clinical breakpoints for fungi by EUCAST, voriconazole breakpoints are not available.
†Fluconazole and voriconazole clinical breakpoints are not available, interpreted according to ECV as given by CLSI.
‡Fluconazole nonsusceptible strains were significantly associated with EOD (P value 0.005; COR 2.73, 95% CI: 1.34–5.53).

TABLE 4. - Candida Species Causing Neonatal Candidemia From Pakistan (2019); Frequencies and Antifungal Susceptibility Profiles Tested by Commercial Colorimetric Broth Microdilution
Species n Antifungal MIC range
(µg/ml)
MIC50
(µg/ml)
MIC90
n (µg/ml)
Resistant/
non-wild type (n)
C. lusitaniae 09 Caspofungin 0.015–0.25 0.06 0.25 0
Fluconazole 0.25–2.0 1.0 2.0 5
Voriconazole 0.008–0.12 0.06 0.06 1
Itraconazole 0.06–0.12 0.12 0.25 0
Amphotericin B 0.12–0.5 0.12 0.25 0
C. guilliermondii 08 Caspofungin 0.03–1.0 0.5 1.0 0
Fluconazole 0.25–32.0 8.0 16.0 3
Voriconazole 0.015–0.5 0.12 0.25 1
Itraconazole 0.25–1.0 0.5 1.0 0
Amphotericin B 0.12 0.12 0.12 0
C. pelliculosa 15 Caspofungin 0.008–0.06 0.015 0.06 0
Fluconazole 0.12–8.0 2.0 4.0 2
Voriconazole 0.008–0.5 0.12 0.25 2
Itraconazole 0.015–0.25 0.12 0.25
Amphotericin B 0.12–0.5 0.12 0.25
C. utilis 14 Caspofungin 0.008–0.12 0.008 0.06
Fluconazole 0.5–4.0 2.0 4.0
Voriconazole 0.015–0.25 0.12 0.12
Itraconazole 0.03–0.25 0.12 0.25
Amphotericin B 0.12–0.5 0.12 0.5
C. glabrata 02 Caspofungin 0.06–0.12
Fluconazole 16.0–32.0
Voriconazole 0.5–1.0 2
Itraconazole 0.5 0
Amphotericin B 0.25–0.5 0
C. krusei 04 Caspofungin 0.06–0.5
Fluconazole 64.0–128.0
Voriconazole 0.5–1.0 0
Itraconazole 0.5 0
Amphotericin B 0.25–0.5 0
Other Candida species 05 Caspofungin 0.008–2.0
Fluconazole 1.0–128
Voriconazole 0.015–8.0
Itraconazole 0.06–0.5
Amphotericin B 0.12–1.0
Minimal inhibitory concentrations (MIC) not available for C. albicans, C. tropicalis and C. parapsilosis as susceptibility testing was done by disc diffusion and interpreted according to CLSI guidelines. Susceptibility results of these organisms are shown in Table 4.
Other Candida species include C. rugosa, C. kefyr, rare Candida species (other than C. auris) not identified by API 20C AUX (bioMérieux).

Univariate analysis showed that LOD was negatively associated with non-albicans Candida species (COR 0.43, 95% CI: 0.27–0.68) as causative agents of candidemia, but multivariate analysis found it to have no significant association. However, late-onset candidemia in neonates was more likely to occur with use of antibiotics like vancomycin (COR 3.89, 95% CI: 1.39–10.89), less likely in full-term neonates (COR 0.27, 95% CI: 0.09–0.49) with gestational age of 37–42 weeks and when respiratory distress (COR 0.16, 95% CI: 0.05–0.78) was a reason for intensive care admission compared with EOD. Early-onset neonatal candidemia was more likely in patients with vaginal delivery (COR 4.16, 95% CI: 1.42–12.23) and neonates with respiratory distress as a cause for ICU admission (COR 3.31, 95% CI: 1.05–10.42), after adjusting for extremely preterm neonates with gestational age of <28 weeks (Table 5).

TABLE 5. - Association of Risk Factors With Neonates With Candidemia
Variables Univariate OR (95% CI) Multivariate OR (95% CI)
Late-onset neonatal candidemia
° Non-albicans Candida species 0.43 (0.27–0.68)* 0.95 (0.25–3.65)
° Respiratory distress 0.34 (0.16–0.71)* 0.16 (0.05–0.49)*
° Vancomycin use 2.16 (0.99–4.72) 3.89 (1.39–10.89)*
° Full term 37–42 weeks 0.53 (0.24–1.16) 0.27 (0.09–0.78)*
Early-onset neonatal candidemia
° Vaginal delivery 0.46 (0.17–1.29) 4.16 (1.42–12.23)*
° Respiratory distress 0.34 (0.16–0.71)* 3.31 (1.05–10.42)*
° Extremely preterm <28 weeks 1.4 (0.09–22.93) 0.59 (0.03–11.8)
*Represents significant association between risk factors and neonatal candidemia.
Early-onset disease (EOD): neonatal candidemia at ≤7 days of life. Late-onset disease (LOD): neonatal candidemia at >7 days of life.

DISCUSSION

Our study shows non-albicans Candida species to be the predominant causative agent in neonates (age 0–28 days) with candidemia compared with C. albicans. C. tropicalis was the most common Candida species isolated in neonatal candidemia patients from Pakistan (39% of all cases). C. pelliculosa and C. guilliermondii were associated with EOD while C. albicans was with LOD. The isolation of fluconazole nonsusceptible non-albicans Candida species was significantly higher in patients with early- versus late-onset neonatal candidemia. We also found non-albicans Candida species to be negatively associated with late-onset candidemia and that EOD was more likely in patients with vaginal delivery.

Samples were received from all over the country including many large public sector hospitals and certain specialized neonatal centers from main Pakistani cities predominating. Thus, the data can be applied to the basic neonatal population of Pakistan. It is true that distribution is skewed to certain locations, but in the absence of a national surveillance network, our data are a very close representation of neonatal candidemia in Pakistan.

There was a high incidence of bacterial coinfection in our study which may raise the question about Candida spp. as blood culture contaminants. Most of the organisms isolated with Candida spp. in our study are common causes of neonatal sepsis globally and in Pakistan.20,21

The incidence of EOD was comparatively high in our study (51.7%) as compared with LOD (48.3%). Other studies from developed countries show a higher incidence of LOD as compared with EOD.9 However, there are no prior studies from LMICs which provide comparative data on neonatal candidemia. This may be related to the difference in the characteristics of the babies born in developed countries versus LMICs. The higher number of term babies compared with higher-income countries in our study population is reflective of the higher proportion of term births in LMICs reported previously from Pakistan and other developing countries.22

Before 2006, C. albicans had been the predominantly reported Candida species associated with neonatal candidemia in Pakistan.7 After 2006, several studies have indicated that non-albicans Candida species particularly C. tropicalis became the most common Candida species identified in neonates with candidemia.2,23 In contrast, a study from India shows C. parapsilosis and C. glabrata to be the most common Candida species as a cause of neonatal candidemia.24C. parapsilosis was found to be the most common Candida species in neonates admitted at our institute, similar to another study from the same center in neonates with device-associated infections25 where fluconazole prophylaxis is also in use for preterm babies. Since 2015, there has been greater awareness of sepsis due to educational sessions and with better medical care, consumption of carbapenems and improved utilization of blood culture, the incidence of neonatal candidiasis increased as shown in Figure 1. However, there is no way to prove this as the cause of the increasing number of candidemia cases. Higher rates of C. tropicalis in babies outside of our institute may represent vaginal flora as the source of acquisition, as fluconazole prophylaxis is not practiced in most other centers. Our study shows that among all Candida species, C. tropicalis followed by C. parapsilosis and C. pelliculosa were the most predominant in neonates with EOD and LOD candidemia. C. tropicalis has been the most common species over the past years from 2014 to 2019, but the isolation of other Candida species like C. parapsilosis increased in the year 2019. Most of these were from other institutes in Karachi and other cities of Pakistan with only 9 isolates from our center. No particular reason for this change in spectrum could be identified and this supports a need for standardized national surveillance for candidemia. Other common Candida species that were isolated include C. lusitaniae, C. guilliermondii and C. utilis. The true spectrum of neonatal candidiasis could not be assessed as only phenotypic methods were used for the identification of Candida species.

Sensitivity testing revealed fluconazole nonsusceptibility to be significantly higher in neonates with early-onset candidemia in our study. In a cohort of 110 infants with candidemia 3% were resistant to fluconazole similar to our study.26 Twenty-nine (4.03%) isolates were resistant to fluconazole which was slightly higher than previously reported. Fluconazole resistance among clinical C. parapsilosis isolates is an emerging problem but no fluconazole-resistant C. parapsilosis was isolated in our study.27 However, we could not accurately determine fluconazole resistance due to unknown and rare Candida species for which clinical breakpoints or ECVs are not available, supporting the empiric use of amphotericin for neonates with candidemia.28

Candida species are the third most common cause of neonatal sepsis worldwide,29 mostly associated with LOD.30–32 As discussed earlier, studies in ELBW neonates have shown C. albicans to be the most common Candida species causing neonatal candidemia and EOD and non-albicans Candida species as common causative agents of late-onset candidemia.7,9,33–35 Our collection did not have any ELBW babies, but it did show that non-albicans Candida species (86.7%) are more common in neonates with both early and late onset of candidemia. This is similar to studies from Pakistan, India and Italy which showed non-albicans Candida species as a cause of sepsis in neonates.2,36,37 However, we found C. albicans to be associated with LOD along with the use of antibiotics as a significant risk factor, similar to other studies from India.10,33,34

Despite high usage of carbapenems and cephalosporins in 2019 (60% and 53% in EOD and 73% and 57% in LOD, respectively) as compared with vancomycin (36% in EOD and 55% in LOD), the association of vancomycin with LOD was the only statistically significant finding. Studies show Gram-negative bacilli to be the most common cause of neonatal sepsis in our part of the world warranting empiric usage of carbapenems and third-generation cephalosporins, therefore their numbers are similar in EOD and LOD. Gram-positive coverage with vancomycin is added when there is high probability of central line bloodstream infections, necrotizing enterocolitis (NEC) and purulent infections.38 We found prior infection with bacteria more commonly in patients with LOD while coinfection occur more commonly in EOD. Studies show that NEC leading to late-onset neonatal sepsis is more commonly caused by bacteria.39 This might explain why C. albicans is seen with LOD, as the origin of candidemia is the gut rather than exogenous40 while non-albicans Candida species, except C. tropicalis, are more common skin flora.41 Other factors that were not assessed but can affect the spectrum of late-onset neonatal candidemia include the type of facility, that is, secondary or tertiary care, public or private hospital setting, intensive care setup or not and, infection control, antibiotic prescribing, use of fluconazole prophylaxis and stewardship practices followed by the healthcare professionals.

Secondary analysis revealed a positive association between non-albicans Candida species and EOD specifically with C. guilliermondii and C. pelliculosa. These children are very vulnerable to organisms spread by contact and hence organisms like C. guilliermondii and C. pelliculosa as well as C. parapsilosis have been reported to cause outbreaks in neonatal intensive care units.34,42 It also showed that early-onset candidemia was more likely to occur in neonates born via vaginal delivery or those with respiratory distress leading to the need for mechanical ventilation. This is similar to a study on ELBW infants from Canada which also found placental or endometrial candidiasis to be associated with EOD.9 Studies from different geographical regions show different Candida species as the cause of vaginal candidiasis.43,44 Intrauterine neonatal infections due to these maternal infections might occur during delivery through colonized vagina.45

A limitation of our study was that incomplete perinatal data collection prevented exploration of many important risk factors and outcomes as cases were identified from laboratory database for the years 2014 to 2018. History was not available for most of these patients other than those admitted to our hospital. For the year 2019, data were obtained via telephonic communication with patients’ physicians or guardians leading to gaps in information and different numbers of neonates in gestational age and birth weight groups. The study participants did not include any neonates with ELBW. The effect of this limited data may be seen with the wide CIs. Another limitation of our study was that all the Candida species were identified by phenotypic methods which might have led to improper identification of rare and unknown Candida species. Unfortunately, due to limited resources, we could not perform MICs for all antifungals on all isolates, which would have lent greater strength to our study.

The strength of our study is that this is data on culture-positive neonatal candidemia cases for which the etiologic spectrum and antifungal resistance data have been compared.

The trends for isolation of Candida species in neonates are changing with C. albicans being the most common agent before the year 2010, and C. tropicalis being the most common Candida species during the recent years.37 There is evidence of change in epidemiology and increasing rates of azole resistance in developed countries, which advocates continuing surveillance for species and the antifungal susceptibility spectrum.46 Despite infrequent isolation of C. glabrata and C. krusei, fluconazole resistance rates among other Candida species were found to be 4.9% in this study. Although preterm fluconazole prophylaxis has shown good outcomes in neonates in intensive care, fluconazole may not be considered a good empiric treatment choice for neonates with suspected fungal sepsis specifically those with EOD. Amphotericin thus remains the first-line option for neonatal candidemia in our setting, despite 7% of isolates being C. lusitaniae which may develop resistance over the course of therapy. Resistance may also emerge as the clinical use of antifungals increases creating a need for accurate identification to the species level using more sensitive methods as well as routine antifungal susceptibility testing for all Candida species.

ACKNOWLEDGEMENTS

We would like to acknowledge the clinical microbiology laboratory at the Aga Khan University Hospital for allowing us to conduct this research.

REFERENCES

1. Kaur H, Chakrabarti A. Strategies to reduce mortality in adult and neonatal candidemia in developing countries. J Fungi. 2017;3:41.
2. Farooqi JQ, Jabeen K, Saeed N, et al. Invasive candidiasis in Pakistan: clinical characteristics, species distribution and antifungal susceptibility. J Med Microbiol. 2013;62:259–268.
3. Jabeen K, Farooqi J, Mirza S, et al. Serious fungal infections in Pakistan. Eur J Clin Microbiol Infect Dis. 2017;36:949–956.
4. Bongomin F, Gago S, Oladele RO, Denning DW. Global and multi-national prevalence of fungal diseases—estimate precision. J fungi. 2017;3:57.
5. Gugnani HC, Denning DW, Rahim R, et al. Burden of serious fungal infections in Bangladesh. Eur J Clin Microbiol Infect Dis. 2017;36:993–997.
6. Chakrabarti A, Sood P, Rudramurthy SM, et al. Incidence, characteristics and outcome of ICU-acquired candidemia in India. Intensive Care Med. 2015;41:285–295.
7. Ariff S, Saleem AF, Soofi SB, et al. Clinical spectrum and outcomes of neonatal candidiasis in a tertiary care hospital in Karachi, Pakistan. J Infect Dev Ctries. 2011;5:216–223.
8. Thatrimontrichai A, Janjindamai W, Dissaneevate S, Maneenil G. The risk of mortality in neonatal invasive fungal infection over 29 years. Southeast Asian J Tropical Med Public Health. 2019;50:893–899.
9. Barton M, Shen A, O’Brien K, et al. Early-onset invasive candidiasis in extremely low birth weight infants: perinatal acquisition predicts poor outcome. Clin Infect Dis. 2017;64:921–927.
10. Juyal D, Sharma M, Pal S, et al. Emergence of non-albicans Candida species in neonatal candidemia. N Am J Med Sci. 2013;5:541–545.
11. Ali GY, Algohary EH, Rashed KA, et al. Prevalence of Candida colonization in preterm newborns and VLBW in neonatal intensive care unit: role of maternal colonization as a risk factor in transmission of disease. J Matern Fetal Neonatal Med. 2012;25:789–795.
12. Hussain AS, Ahmed AM, Arbab S, et al. CLABSI reduction using evidence based interventions and nurse empowerment: a quality improvement initiative from a tertiary care NICU in Pakistan. Arch Dis Child. 2021;106:394–400.
13. WHO. Preterm birth 2018. Available at: https://www.who.int/news-room/fact-sheets/detail/preterm-birth. Accessed February 7, 2020.
14. Gill SV, May-Benson TA, Teasdale A, et al. Birth and developmental correlates of birth weight in a sample of children with potential sensory processing disorder. BMC Pediatr. 2013;13:29.
15. CLSI. Performance Standards for Antifungal Susceptibility Testing of Yeasts. 2nd ed. 2017 CLSI supplement M60. Clinical and Laboratory Standards Institute; 2017.
16. European Committee on Antimicrobial Susceptibility Testing, Data from the EUCAST MIC distribution website. Available at: http://www.eucast.org. last accessed 02/07/2020.
17. Pfaller MA, Castanheira M, Diekema DJ, et al. Triazole and echinocandin MIC distributions with epidemiological cutoff values for differentiation of wild-type strains from non-wild-type strains of six uncommon species of Candida. J Clin Microbiol. 2011;49:3800–3804.
18. CLSI. Epidemiological Cutoff Values for Antifungal Susceptibility Testing. 3rd ed. CLSI supplement M59. Clinical and Laboratory Standards Institute; 2020.
19. Pfaller MA, Espinel-Ingroff A, Canton E, et al. Wild-type MIC distributions and epidemiological cutoff values for amphotericin B, flucytosine, and itraconazole and Candida spp. as determined by CLSI broth microdilution. J Clin Microbiol. 2012;50:2040–2046.
20. Cortese F, Scicchitano P, Gesualdo M, et al. Early and late infections in newborns: where do we stand? A review. Pediatr Neonatol. 2016;57:265–273.
21. Piqueras AI, Rubio T, Lopez-Medina EM, et al. Recent changes in candidemia trends in a tertiary hospital (2011–2018). Revista Iberoamericana de Micología. 2020;37:87–93.
22. Pusdekar YV, Patel AB, Kurhe KG, et al. Rates and risk factors for preterm birth and low birthweight in the global network sites in six low- and low middle-income countries. Reprod Health. 2020;17(suppl 3):187.
23. Farooqi J, Jabeen K, Naqvi F, Malik F, Zafar A, editors. Spectrum of Neonatal Candidemia in Pakistan: review of laboratory data 2014-2016., Hoboken 07030-5774, NJ USA; Mycoses; 2017: Wiley 111 River St;
24. Sardana V, Pandey A, Madan M, et al. Neonatal candidemia: a changing trend. Indian J Pathol Microbiol. 2012;55:132–133.
25. Hussain AS, Ahmed AM, Arbab S, et al. CLABSI reduction using evidence based interventions and nurse empowerment: a quality improvement initiative from a tertiary care NICU in Pakistan. Arch Dis Child. 2021;106:394–400.
26. Autmizguine J, Tan S, Cohen-Wolkowiez M, et al.; NICHD Neonatal Research Network. Antifungal susceptibility and clinical outcome in neonatal candidiasis. Pediatr Infect Dis J. 2018;37:923–929.
27. Asadzadeh M, Ahmad S, Al-Sweih N, et al. Epidemiology and molecular basis of resistance to fluconazole among clinical Candida parapsilosis isolates in Kuwait. Microb Drug Resist. 2017;23:966–972.
28. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guideline for the management of candidiasis: 2016 update by the Infectious Diseases Society of America. Clin Infect Dis. 2016;62:e1–50.
29. Pammi M, Holland L, Butler G, et al. Candida parapsilosis is a significant neonatal pathogen: a systematic review and meta-analysis. Pediatr Infect Dis J. 2013;32:e206–e216.
30. Mehar V, Yadav D, Somani P, et al. Neonatal sepsis in a tertiary care center in central India: microbiological profile, antimicrobial sensitivity pattern and outcome. J Neonatal Perinatal Med. 2013;6:165–172.
31. Vergnano S, Menson E, Kennea N, et al. Neonatal infections in England: the NeonIN surveillance network. Arch Dis Child Fetal Neonatal Ed. 2011;96:F9–F14.
32. Wu JH, Chen CY, Tsao PN, et al. Neonatal sepsis: a 6-year analysis in a neonatal care unit in Taiwan. Pediatr Neonatol. 2009;50:88–95.
33. Fu J, Ding Y, Wei B, et al. Epidemiology of Candida albicans and non-C. albicans of neonatal candidemia at a tertiary care hospital in western China. BMC Infect Dis. 2017;17:329.
34. Fu J, Wang X, Wei B, et al. Risk factors and clinical analysis of candidemia in very-low-birth-weight neonates. Am J Infect Control. 2016;44:1321–1325.
35. Chen J, Jiang Y, Wei B, et al. Epidemiology of and risk factors for neonatal candidemia at a tertiary care hospital in western China. BMC Infect Dis. 2016;16:700.
36. Koppad B, Prakash KP. Neonatal candidiasis: clinical spectrum and epidemiology in a tertiary care centre. Int J Contemporary Pediatrics. 2017;4:438.
37. Caggiano G, Lovero G, De Giglio O, et al. Candidemia in the neonatal intensive care unit: a retrospective, observational survey and analysis of literature data. Biomed Res Int. 2017;2017:7901763.
38. Chaurasia S, Sivanandan S, Agarwal R, et al. Neonatal sepsis in South Asia: huge burden and spiralling antimicrobial resistance. BMJ. 2019;364:k5314.
39. Wu IH, Tsai MH, Lai MY, et al. Incidence, clinical features, and implications on outcomes of neonatal late-onset sepsis with concurrent infectious focus. BMC Infect Dis. 2017;17:465.
40. Vakiliamini M, Babaei H, Mohammadi M, Habibi R, Motamed H. Intestinal colonization rate of Candida albicans among low birth weight neonates after using oral synbiotic supplementation: a randomized placebo-controlled trial. Iranian J Neonatology IJN. 2020;11:51–56.
41. Zuza-Alves DL, Silva-Rocha WP, Chaves GM. An update on Candida tropicalis based on basic and clinical approaches. Front Microbiol. 2017;8:1927.
42. Shane AL, Sánchez PJ, Stoll BJ. Neonatal sepsis. Lancet. 2017;390:1770–1780.
43. Bitew A, Abebaw Y. Vulvovaginal candidiasis: species distribution of Candida and their antifungal susceptibility pattern. BMC Womens Health. 2018;18:94.
44. Gonçalves B, Ferreira C, Alves CT, et al. Vulvovaginal candidiasis: epidemiology, microbiology and risk factors. Crit Rev Microbiol. 2016;42:905–927.
45. Heerema-McKenney A. Defense and infection of the human placenta. APMIS. 2018;126:570–588.
46. Pfaller MA, Moet GJ, Messer SA, et al. Candida bloodstream infections: comparison of species distributions and antifungal resistance patterns in community-onset and nosocomial isolates in the SENTRY antimicrobial surveillance program, 2008-2009. Antimicrob Agents Chemother. 2011;55:561–566.
Keywords:

candidemia; neonates; early-onset disease; late-onset disease; Pakistan

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.