Nosocomial invasive candidiasis (IC) has emerged as a major problem in neonatal intensive care units (NICUs). We investigated herein the temporal clustering of six cases of neonatal IC due to
Invasive candidiasis (IC) has substantially increased in neonatal intensive care units (NICUs) over the two past decades and is still associated with a high morbidity and mortality [
In order to type outbreak-related isolates and to assess their clonality and identify the source and the routes of their transmission, many molecular techniques have been used. They include electrophoretic karyotyping, southern blot hybridization, restriction fragment length polymorphism (RFLP) analysis, randomly amplified polymorphic DNA (RAPD) analysis, PCR-based fingerprinting, and multilocus sequence typing [
In the NICU of our hospital, six cases of IC due to
Six cases of
Molecular investigations were conducted retrospectively to assess clonality of the
A total of 20
The sequence of isolates, their anatomical origin, and the time of isolation are summarized in Table
Description of the 20
Neonate or HCW* | Isolate | Site of isolation | EK pulsotype | PFGE-Sfi I pattern | Date of sampling (day/mo/yr) | Time of hospitalisation |
---|---|---|---|---|---|---|
N1 | 1 | Blood | I | 1 | 11/09/2006 | |
2 | Endotracheal tube | I | 1 | 13/09/2006 | ||
3 | Postoperative wound | I | 1 | 26/09/2006 | Admission: 01/09/2006 | |
4 | Hepatic abscess | I | 1 | 27/09/2006 | Discharge: 07/11/2006 | |
5 | Drain | I | 1 | 04/10/2006 | ||
6 | Postoperative wound | I | 1 | 04/10/2006 | ||
N2 | 7 | Blood | II | 2 | 15/09/2006 | Admission: 06/09/2006 |
8 | Mediastinal fluid | III | 3 | 19/09/2006 | Discharge: 08/11/2006 | |
9 | Blood | III | 3 | 11/10/2006 | ||
N3 | 10 | Urine | IV | 4 | 26/09/2006 | |
11 | Blood | IV | 4 | 29/09/2006 | Admission: 16/09/2006 | |
12 | Umbilical catheter | IV | 4 | 05/10/2006 | Discharge: 08/11/2006 | |
N4 | 13 | Blood | III | 5 | 29/09/2006 | |
14 | Umbilical catheter | III | 6 | 05/10/2006 | Admission: 20/09/2006 | |
15 | Umbilical catheter | III | 7 | 16/10/2006 | Death: 16/10/2006 | |
16 | Blood | III | 7 | 16/10/2006 | ||
N5 | 17 | Blood | III | 5 | 13/10/2006 | Admission: 05/10/2006 |
N6 | 18 | Umbilical catheter | III | 5 | 04/10/2006 | Admission: 21/09/2006 |
HCW1 | H1 | Fingers’s nails | IV | 8 | 26/10/2006 | — |
HCW2 | H2 | Fingers’s nails | III | 9 | 27/10/2006 | — |
ATCC 90028 | A | V | 10 | — | — |
*Health care worker.
The isolates collected from neonates and HCWs were identified as
The preparation of DNA plugs of
Chromosomal DNA bands were separated on 1% agarose gels in 0.5x TBE buffer (0.045 mol L−1 Tris-HCl, 0.045 mol L−1 borate, 1 mol L−1 EDTA) in contourclamped homogeneous electric field (CHEF) electrophoresis system (CHEF-DRII, Bio-Rad). Electrophoresis was performed at 150 V for 24 h with a 180 s ramping switch interval and then at 110 V for 22 h with a 360 s ramping switch interval. The temperature of the running buffer was maintained at 14°C.
The plugs were placed in 200
In both molecular techniques, after electrophoresis, the gel was stained with ethidium bromide solution for 15 min and destained with distilled water. A ladder of
For both EK and PFGE-Sfi I data analysis, the bands were identified and their size evaluated by using Quantity One 1D analysis software (Bio-rad). The genetic relationships among isolates were established by cluster and ordination analysis performed on the matrix of genetic similaritics. Cluster analysis was performed on the genetic distance matrix with the unweighted paired group method using arithmetic average (UPGMA) and the Jaccard’s correlation coefficient calculated on the basis of ERIC-2 patterns by using the MVSP 3.1 software. The Jaccard’s coefficient ranges from 1.00 (the two patterns are identical) to 0.00 (no common bands between both patterns). We considered that (i) a Jaccard’s coefficient of 1.00 is indicative of strains of the same clone, (ii) a Jaccard’s coefficient ranging between 0.90 and 0.99 is indicative of highly similar but nonidentical strains, (iii) a Jaccard’s coefficient ranging from 0.80 and 0.89 represents less related isolates, and (iv) a Jaccard’s coefficient below 0.79 represents unrelated isolates [
The 20 clinical isolates generated four different karyotypes of six to eight bands each. The karyotypes generated by the HCWs’ isolates and nine isolates from neonates are shown in Figure
Examples of pulsotypes obtained in EK analysis. M: size marker (in base pairs). H1 and H2: isolates obtained from nurses HCW1 and HCW2, respectively. 6, 12, 17, and 18: isolates obtained from neonates N1, N3, N5, and N6 respectively. 7 and 8: isolates obtained from neonate N2. 13, 14, and 15: isolates obtained from neonate N4. A: ATCC 90028 reference strain.
Dendrogram of the genetic relatedness of 20
The clinical isolates generated nine different band patterns of 13 to 19 bands each. Figure
Examples of PFGE patterns of Sfi 1 digested genomic DNA. M: size marker (in base pairs). H1 and H2: isolates obtained from nurses HCW1 and HCW2, respectively. 6, 12, 17, 18: isolates obtained from neonates N1, N3, N5, and N6, respectively. 7 and 8: isolates obtained from neonate N2. 13, 14, and 15: isolates obtained from neonate N4. A: ATCC 90028 reference strain.
Dendrogram of the genetic relatedness of 20
EK analysis showed that (i) the isolates H1 and H2, obtained from both HCWs one day apart, were unrelated to each other (Jaccard’s coefficient of 0.78). (ii) The isolate H1 collected from the HCW1 had an EK pattern identical to three isolates (isolates 10 to 12) collected from the neonate N3. (iii) The isolate H2 collected from the HCW2 yielded an EK pattern identical to eight isolates collected from four neonates (N2, N4, N5, and N6) and was less related (Jaccard’s coefficient of 0.86) to the isolates from neonate N1. These findings suggest the occurrence of two outbreaks (karyotypes III and IV) and of cross-contamination between the HCWs and five neonates.
PFGE-Sfi I analysis revealed different epidemiologic features. It showed that (i) the isolate H1 was unrelated to the all isolates collected from neonates (Jaccard’s coefficient ≤0.67). (ii) The isolate H2 was less related to the six isolates collected from the neonate N1 (Jaccard’s coefficient of 0.88) and unrelated to the other neonates’ isolates. (iii) Three isolates (isolates 13, 17, and 18) obtained from neonates N4, N5, and N6 showed the same banding pattern, suggesting that these isolates, collected within a two week-period from the three neonates, were clonal in origin. (iv) Three neonates (N1 to N3) had specific PFGE-Sfi I patterns suggesting coincidental emergence of unrelated strains.
Thus, the PFGE-Sfi I analysis broke the largest EK clone (represented by the karyotype III) into three clones: two clones including each the sequential isolates from the same neonate (N2 and N4) and one clone including isolates from different neonates (N4 to N6).
Of the six neonates included in our study, four had two or more isolates. In two neonates (N1 and N3), sequential isolates from each patient showed the same karyotype and the same PFGE-Sfi I pattern, suggesting that in both N1 and N3 neonates, isolates were clonally related.
In the third neonate (N2), two isolates were collected from blood (isolates 7 and 9) and one isolate from mediastinal fluid (isolate 8). The second blood isolate (isolate 9) and the mediastinal isolate obtained 22 days apart were identical, showing the same karyotype and PFGE-Sfi I pattern, but these two isolates were unrelated to the first blood isolate in both molecular analysis (Table
In the fourth neonate (N4), two isolates were collected from blood (isolates 13 and 16) and two umbilical catheter tips (isolates 14 and 15). Interestingly, PFGE-Sfi I analysis showed that the first blood isolate (isolate 13) and the first catheter isolate (isolate 14), collected one week later, were highly similar to each other, exhibiting minor band differences (Jaccard’s coefficient of 0.92). The second catheter isolate (isolate 15) and the second blood isolate (isolate16) collected 11 days later were identical to each other and highly similar to the isolate 14 (Jaccard’s coefficient of 0.95). These findings suggest that the four isolates derived from the same strain and this strain underwent two successive microevolutions.
Microevolutionary changes were not detected by EK analysis where the four isolates showed the same karyotype.
It is important to note that, according to the PFGE-Sfi I analysis, the first blood isolate (isolate 13) of the neonate N3 was identical to the isolates 14 and 15, collected from the neonates N4 and N5, suggesting that the isolates from the three neonates derived from the same strain which underwent microevolution in only neonate N4.
Nosocomial candidiasis has been increasing over the last decades and is still associated with a high morbidity and mortality rates despite the availability of novel antifungal agents [
At first insight, the temporal clustering in our NICU of six cases of neonatal IC caused by
The remaining N4 to N6 three neonates were infected by the same strain, as showed by identical PFGE-Sfi I patterns. As the three neonates occupied the same NICU at the same time, there is evidence supporting the horizontal transmission of
Many previous studies showed that
In our study, the sequential isolates were identical in both neonates N1 and N3. In neonate N1, isolates collected from three different body sites and from two medical devices over the 24-day period were identical and did not showe any genetic variation in both used molecular analyses. In neonate N3, three isolates were sequentially collected from urine, blood, and umbilical catheter, two and five days apart respectively. In this case, because the catheter was removed as soon as the blood culture was shown to be positive and because catheter colonization frequently results in catheter-related blood stream infection [
In neonate N4, the isolates recovered from the catheters were identical or highly similar to the blood isolates. This suggests that the portal of entry of the yeast was the catheter and that the four isolates derived from the same strain which underwent microevolution. It is worth mentioning that two successive microevolutionary changes occurred in the neonate N4. The same strain infected two additional neonates (N5 and N6) in whom only a single isolate was collected, which precludes checking the stability of the strain in both infants. These findings confirm the occurrence of microevolutionary changes in catheter-related candidemia but do not allow estimating its frequency because of the small number of cases investigated in our study. Shin et al. showed that microevolution occurred in 36% of catheter strains when compared to blood strains [
In neonate N4, microevolution could only be detected by PFGE-Sfi I, whereas, in EK analysis, all isolates were identical to each other and to isolates collected from the neonates N5 and N6. This finding confirms that PFGE-Sfi I method is a powerful means of detecting microevolution among sequential isolates of
In neonate N2, both blood isolates, collected approximately one month apart, showed different karyotypes and different PFGE-Sfi I patterns and were widely separated in both dendrograms. This finding suggests strain replacement in blood as change in the karyotype of serial clinical isolates of
In conclusion, our results showed that only three out of the six investigated
As a limited number of isolates were investigated, larger studies on