Claus Østergaard

Treatment of bacterial vaginosis in pregnancy in order to reduce the risk of spontaneous preterm delivery: a clinical recommendation

Bacterial vaginosis (BV) is characterized by a dysbiosis of the vaginal microbiota with a depletion of Lactobacillus spp. In pregnancy, prevalences between 7 and 30% have been reported depending on the study population and the definition. BV may be associated with an increased risk of spontaneous preterm delivery (sPTD). However, it is controversial whether or not BV‐positive pregnant women will benefit from treatment to reduce the risk of sPTD. We could not identify any good‐quality guideline addressing this issue. Consequently we aimed to produce this clinical recommendation based on GRADE.

Rapid first-line discrimination of methicillin resistant Staphylococcus aureus strains using MALDI-TOF MS.

Fast and reliable discrimination of methicillin-resistant Staphylococcus aureus (MRSA) isolates is essential in identifying an outbreak. Molecular typing methods, such as S. aureus protein A (spa) typing, multi locus sequence typing (MLST) and pulse field gel electrophoresis (PFGE) are generally used for this purpose. These methods are all relatively time-consuming and not performed routinely in all laboratories. The aim of this study is to examine whether MALDI-TOF MS can be used as a fast, simple and easily implemented method for first-line discrimination of MRSA isolates. Mass spectra from 600 clinical MRSA isolates were included in the study, representing 89 spa types, associated with 16 different known clonal complexes. All spectra were obtained directly from colony material obtained from overnight cultures without prior protein extraction. We identified 43 useful discriminatory m/z-values (peaks) and used a concept of arranging these peaks into pairs or small clusters within a small mass range, allowing for quality control of the spectra obtained. Using this concept we could reproducibly characterise and arrange the isolates into 26 MALDI-TOF groups, which strongly correlated with spa types and clonal complexes. The results of this study clearly show that MALDI-TOF MS can be used for first-line discrimination of MRSA isolates, using a simple and fast method that is easy to implement as part of routine testing.

High rates of reduced susceptibility in the Bacteroides fragilis group isolated from blood cultures: The first national survey in Denmark

The authors acknowledge the critical reading of the manuscript by Dr Edward Chan as well as useful discussion with members from SC’s laboratory. Funding: This work was supported by the Chinese National Key Basic Research and Development (973) Programme [2013CB127200] and the Research Fund for the Control of Infectious Diseases from the Food and Health Bureau, the Government of Hong Kong SAR [ZJE4 to SC]. Competing interests: None declared. Ethical approval: Not required.

WGS-based surveillance of third-generation cephalosporin-resistant Escherichia coli from bloodstream infections in Denmark

Objectives To evaluate a genome-based surveillance of all Danish third-generation cephalosporin-resistant Escherichia coli (3GC-R Ec ) from bloodstream infections between 2014 and 2015, focusing on horizontally transferable resistance mechanisms. Methods A collection of 552 3GC-R Ec isolates were whole-genome sequenced and characterized by using the batch uploader from the Center for Genomic Epidemiology (CGE) and automatically analysed using the CGE tools according to resistance profile, MLST, serotype and fimH subtype. Additionally, the phylogenetic relationship of the isolates was analysed by SNP analysis. Results The majority of the 552 isolates were ESBL producers (89%), with bla CTX-M-15 being the most prevalent (50%) gene, followed by bla CTX-M-14 (14%), bla CTX-M-27 (11%) and bla CTX-M-101 (5%). ST131 was detected in 50% of the E. coli isolates, with the remaining isolates belonging to 73 other STs, including globally disseminated STs (e.g. ST10, ST38, ST58, ST69 and ST410). Five of the bloodstream isolates were carbapenemase producers, carrying bla OXA-181 (3) and bla OXA-48 (2). Phylogenetic analysis revealed 15 possible national outbreaks during the 2 year period, one caused by a novel ST131/ bla CTX-M-101 clone, here observed for the first time in Denmark. Additionally, the analysis revealed three individual cases with possible persistence of closely related clones collected more than 13 months apart. Conclusions Continuous WGS-based national surveillance of 3GC-R Ec , in combination with more detailed epidemiological information, can improve the ability to follow the population dynamics of 3GC-R Ec , thus allowing for the detection of potential outbreaks and the effects of changing treatment regimens in the future.

Characterization of Carbapenem Nonsusceptible Pseudomonas aeruginosa in Denmark: A Nationwide, Prospective Study

From January 1st 2011 through June 30th 2011, 116 nonreplicate, noncystic fibrosis-related Pseudomonas aeruginosa isolates with reduced carbapenem susceptibility were collected from 12 out of 13 Danish departments of clinical microbiology. The presence of acquired β-lactamases was assessed with combination tablet-diffusion methodology and polymerase chain reaction. In addition, antimicrobial susceptibility testing, an efflux pump inhibitor assay, and pulsed-field gel electrophoresis (PFGE) were performed. Isolates producing acquired β-lactamases were further investigated by serotyping and multi locus sequence typing. Eight isolates produced the metallo-β-lactamase (MBL) VIM-2, and one isolate produced OXA-10 and VEB-1-like extended-spectrum beta-lactamase (ESBL). Phenotypic indications of derepressed AmpC and efflux pump were seen in 56 and 43 isolates, respectively. Overall, the results indicate that mutational factors related to permeability--often combined with derepressed, chromosomal AmpC--is the main factor behind carbapenem nonsusceptibility in Danish P. aeruginosa isolates. The ESBL producer and all the VIM producers belonged to international clones. PFGE revealed that most of the isolates were unrelated, but clonal spread was seen; the 116 isolates distributed in 97 PFGE types, with the largest cluster consisting of 4 isolates (including three isolates from the same hospital with 100% similarity). Thirty-two isolates were pair-wise related, while the remaining isolates were clonally unrelated, as were all nine ESBL/MBL producers.

Emergence of vanA Enterococcus faecium in Denmark, 2005-15

Objectives To describe the changing epidemiology of vancomycin-resistant Enterococcus faecium and Enterococcus faecalis in clinical samples in Denmark 2005-15 according to species and van type, and, furthermore, to investigate the genetic relatedness of the clinical E. faecium isolates from 2015. Methods During 2005-14, all clinical VRE isolates were tested for the presence of vanA/B/C genes by PCR. In 2015, all clinical VRE isolates were whole-genome sequenced. From the WGS data, the presence of van genes and MLST STs were extracted in silico . Core-genome MLST (cgMLST) analysis was performed for the vancomycin-resistant E. faecium isolates. Results During 2005-15, 1043 vanA E. faecium , 25 vanB E. faecium , 4 vanA E. faecalis and 28 vanB E. faecalis were detected. The number of VRE was <50 isolates/year until 2012 to > 200 isolates/year in 2013-15. In 2015, 368 vanA E. faecium and 1 vanB E. faecium were detected along with 1 vanA E. faecalis and 1 vanB E. faecalis . cgMLST subdivided the 368 vanA E. faecium isolates into 33 cluster types (CTs), whereas the vanB E. faecium isolate belonged to a different CT. ST203-CT859 was most prevalent (51%), followed by ST80-CT14 (22%), ST117-CT24 (6%), ST80-CT866 (4%) and ST80-CT860 (2%). Comparison with the cgMLST.org database, previous studies and personal communications with neighbouring countries revealed that the novel cluster ST203-CT859 emerged in December 2014 and spread to the south of Sweden and the Faroe Islands during 2015. Conclusions VRE increased in Denmark during 2005-15 due to the emergence of several vanA E. faecium clones.

Subdivision of MRSA CC398 isolates using MALDI-TOF MS

Outbreak investigations demand a fast and discriminative typing method. MALDI-TOF MS has been shown to be a rapid, easy and inexpensive method of subtyping MRSA.The aim of the present study is to explore whether it is possible to subdivide isolates of MRSA CC398, commonly livestock associated, using an enhanced version of the MALDI-TOF MS typing method that we previously described (Østergaard et al, 2015). We included MALDI-TOF spectra from 378 isolates of MRSA belonging to CC398, of which 322 were epidemiologically independent. We identified 17 peaks as discriminatorily useful and could therefore reliably subdivide the isolates into 23 subtypes, including a distinct type corresponding to a strain from an unusual and initially undiscovered hospital outbreak. Whole genome sequencing was carried out for 193 of the isolates and compared with both the spa type and an antibiogram of these strains. The proposed MALDI-TOF subdivision method for MRSA CC398 was found to be more discriminative than both spa typing and resistotyping, and had a high negative predictive value for ruling out a close genetic relationship between pairs of strains with different MALDI-TOF types. We conclude that the MALDI-TOF-based typing method can be used for rapid and inexpensive routine subdivision of MRSA belonging to CC398.

Description and validation of a new automated surveillance system for Clostridium difficile in Denmark

The surveillance of Clostridium difficile (CD) in Denmark consists of laboratory based data from Departments of Clinical Microbiology (DCMs) sent to the National Registry of Enteric Pathogens (NREP). We validated a new surveillance system for CD based on the Danish Microbiology Database (MiBa). MiBa automatically collects microbiological test results from all Danish DCMs. We built an algorithm to identify positive test results for CD recorded in MiBa. A CD case was defined as a person with a positive culture for CD or PCR detection of toxin A and/or B and/or binary toxin. We compared CD cases identified through the MiBa-based surveillance with those reported to NREP and locally in five DCMs representing different Danish regions. During 2010-2014, NREP reported 13 896 CD cases, and the MiBa-based surveillance 21 252 CD cases. There was a 99·9% concordance between the local datasets and the MiBa-based surveillance. Surveillance based on MiBa was superior to the current surveillance system, and the findings show that the number of CD cases in Denmark hitherto has been under-reported. There were only minor differences between local data and the MiBa-based surveillance, showing the completeness and validity of CD data in MiBa. This nationwide electronic system can greatly strengthen surveillance and research in various applications.

Appreciable uncertainty regarding benefits and risks in the treatment of bacterial vaginosis to prevent preterm birth

Sir, As part of his speeches, the Roman senator Cato the Elder frequently uttered “Carthago delenda est” (Carthage must be destroyed). Similarly, Lamont and colleagues frequently utter that the “wrong patients with the wrong diagnosis were given wrong antibiotics at the wrong gestational age.” We respect that Lamont and colleagues find it difficult to accept meta-analyses, which “cut a toe and a bit of a heel” to merge evidence from more publications, and we certainly cannot exclude the possibility that future research will change clinical practice. However, in contrast to Lamont and colleagues, we emphasize that not only statistical significance but also other aspects should be taken into consideration in the development of clinical recommendations (1). Using GRADE, we found that appreciable uncertainty exists regarding the magnitude of benefits from clindamycin treatment compared with the risks. Given that the relative risk reported in Figure 5 represents the truth (2), the number needed to treat was 45 bacterial vaginosis (BV)-positive pregnant women to avoid one spontaneous preterm delivery (sPTD) before gestational week 37. This relatively minor effect should especially be measured against the microbiological concerns for iatrogenic damage when clindamycin is administered to approximately 10–15% of Danish pregnant women. These concerns include resistance development of BV-associated bacteria, Clostridium difficile colitis, and yet unknown adverse effects to the microbiome (3). The major concern raised by Lamont and colleagues seems to relate to whether the intervention with clindamycin before 22 weeks’ gestation would significantly reduce the incidence of sPTD (4). We agree that had the cut-off been gestational week 22, then the difference in incidence of sPTD would have been statistically significant with respect to clindamycin. Regardless, according to GRADE, the overall level of evidence for the outcome sPTD was rated low for reasons given in Supporting Information Table 2 (2). Regarding the PREMEVA 1 study, we agree that the risk of bias may have increased since October 2014 when we conducted the literature search. At this time, the PREMEVA1 abstract had only been published for 10 months (5). We sent out official emails to the authors; however, none responded. We agree that it is worrisome that the PREMEVA1 study is not yet published and we, as Lamont and colleagues, can only speculate why. Could publication bias due to negative findings be disregarded? Finally, it should be emphasized that we cannot confirm from our analyses that treatment of BV reduces the rate of early preterm birth and low birthweight infants, nor could we find the evidence in the review and meta-analysis by Lamont et al. from 2011 (6). In conclusion, we do not agree that the concerns mentioned by Lamont and colleagues would change the clinical recommendations in the recent AOGS publication (2). The weak recommendation against clindamycin reflects that some experts might treat BV in pregnancy whereas the majority would not, as “weak recommendations for or against intervention are made when guideline authors believe that most informed people would choose the recommended course of action, but a substantial number would not.”