Denise Van Hout

ANTIBIOTIC RESISTANCE

The human body is full of microorganisms, including viruses, fungi, and bacteria. Bacteria are single-celled microorganisms without a nucleus. Generally, a distinction is made between Gram-positive and Gram-negative bacteria based on their cell wall structure. They occur on the skin, in the nose, mouth, and in our intestines. This is called bacterial 'colonization' or 'carriage'. The bacteria in the intestines are collectively called the 'microbiota' (formerly known as 'intestinal flora'). Bacteria are usually harmless and are even important for our metabolism and immunity. However, under certain conditions, bacteria can multiply and cause diseases, such as skin or urinary tract infections. Infections caused by bacteria can usually be treated with antibiotics. However, bacteria can become resistant (insensitive) to certain types of antibiotics. In that case, a bacterium can survive and multiply in the presence of an antibiotic that would normally stop it from growing or kill it. This is called antibiotic resistance. We know that more antibiotic resistance occurs when many antibiotics are used. This is due to the phenomenon of selection pressure. Under the pressure of antibiotics, only the fittest bacteria survive and multiply. If antibiotic-resistant subpopulations exist within a group of bacteria and there is exposure to antibiotics, the resistant subpopulations will survive and be selected (Figure 1).

Bacteria can pass antibiotic resistance to each other, for example, through cell division. This is called vertical transmission. Another way is by exchanging plasmids. Plasmids are circular strands of DNA on which antibiotic resistance genes can be located. This is also called horizontal transmission. A well-known example of (mainly) plasmid-mediated antibiotic resistance in bacteria is the production of an enzyme that can break down antibiotics, called ESBL ('extended-spectrum beta-lactamase'). It is not directly dangerous if a bacterium we carry becomes resistant to certain antibiotics. However, if an infection subsequently develops with a resistant bacterium, it must be treated with a different type of antibiotic (often with a broader spectrum). Fortunately, in the Netherlands, there are usually enough other types of antibiotics available. We do our best to keep it that way by being cautious in prescribing antibiotics. Additionally, Dutch hospitals have many infection prevention measures to minimize the exchange of antibiotic-resistant bacteria between patients. An important way to counter the consequences of antibiotic resistance as much as possible is through surveillance. Surveillance is the systematic collection and analysis of data on bacterial resistance to antibiotics. Longitudinal surveillance can make shifts in resistance visible over time. These data are important for developing, evaluating, and optimizing infection prevention. In this thesis, we focus on various projects concerning hospital surveillance, particularly of antibiotic-resistant Gram-negative bacteria.

SUMMARY OF THE RESEARCH IN THIS THESIS

Patients admitted to intensive care (ICU) are generally seriously ill and therefore have an increased risk of infections. Bacteria and yeasts colonizing the gut and oropharynx play an important role and are the target of decontamination strategies, such as selective digestive decontamination (SDD) and selective oropharyngeal decontamination (SOD). The SDD strategy consists of antibiotic prophylaxis via mouth paste, a suspension given through a nasogastric tube, and intravenous antibiotics during the first four days of ICU admission. SOD consists only of the mouth paste and intravenous antibiotics. The largest and most recent study conducted in the Netherlands showed a survival benefit for ICU patients receiving SDD compared to SOD. However, the SDD strategy is more extensive in both antibiotic prophylaxis and microbiological surveillance. There was uncertainty about which strategy should be used in Dutch ICUs. In Chapter 2, we investigated this through a cost-effectiveness analysis (CEA). We pooled data from two cluster-randomized studies conducted in the Netherlands. The results showed that the SDD strategy was cost-effective compared to SOD: SDD reduced hospital mortality without a significant difference in costs. This was mainly because SDD patients were transferred faster from the ICU to a regular ward. Fear of antibiotic resistance development as a reason to withhold SDD is unfounded based on current evidence.

In Chapter 3, we investigated whether adding a selective plate, the SuperPolymyxin™ medium, to the laboratory process would improve the detection of colistin-resistant Gram-negative bacteria. Colistin is one of the SDD antibiotics, but resistance is difficult to detect. We found that using the SuperPolymyxin™ medium indeed identified more colistin-resistant bacteria. We calculated that if the costs of the plate did not exceed €5.09 per test, it would be cheaper to implement as a screening medium than existing processes, while reducing the number of strains needing definitive confirmation by 76%.

Chapter 4 evaluated the Dutch 6-point MDRO risk assessment questionnaire used upon hospital admission. We found that the screening identified previously unknown carriers in only 0.06% of all admissions. The number of questionnaires needed to detect one new carrier was 1,778 (or 10,420 questions). Furthermore, the majority of carriers (91.8%-99.6%) likely remained undetected because the risk factors asked do not predict carriage well or don't apply to most carriers. We concluded that the current mass screening strategy needs strong reconsideration.

Chapter 5 describes the EPIGENEC study protocol, focusing on molecular surveillance of Escherichia coli. Chapter 6 found that non-ESBL-producing and ESBL-producing E. coli strains causing bacteremia differed significantly in terms of serotypes, subtypes, and virulence/resistance genes. Non-ESBL strains were more heterogeneous. In Chapter 7, we explored using clinical samples as a proxy for surveillance of the community E. coli reservoir. We compared the genomes of 183 community carriage strains with clinical isolates from primary care, hospitals, and blood cultures. We found promising similarities, suggesting clinical samples can reliably estimate the clonal composition of E. coli circulating in the community.

Chapter 8 summarizes the main findings and discusses limitations and practical implications, concluding that hospital-based surveillance is essential but must be continuously evaluated for efficiency and effectiveness.