Increased prescription rate of anti‐infective agents after diagnosis of myelodysplastic syndromes

Abstract The a priori risk for infections in patients with myelodysplastic syndromes (MDS) is unknown. This study examines prescription rates of anti‐infective agents in MDS patients before and after diagnosis, in both in‐ and outpatient settings, to provide information on infection management in clinical practice. We performed a population‐based study using the HemoBase registry, containing data of all MDS patients diagnosed since 2005 in Friesland, the Netherlands. Community and hospital pharmacies provided prescription data from 1995 to 2020. Data were obtained for 203 of 292 patients (70%). Patients received significantly more anti‐infective agents, predominantly antibacterials (70%), after diagnosis compared to before: 148.7 defined daily dose/1000 days (DID) (95% CI: 146.9–150.5) and 55.1 DID (95% CI: 54.5–55.8, p < 0.01), respectively, corresponding to median 23.5 and 7.6 treatment days/year. Higher‐risk (449.9 DID) and lower‐risk patients (129.1 DID) both received significantly more anti‐infective agents after diagnosis; comorbidities, neutropenia, and age did not show significant differences relative to prescription rates. Before diagnosis, 10% of patients had infection‐related hospital admissions versus 38% after diagnosis. In conclusion, MDS patients received significantly more anti‐infective agents compared to before diagnosis. This is the first study that has quantified the prescription rate of anti‐infective agents within and beyond the clinical setting in MDS.

infections is high, with infectious complications present in 15%-40% of patients in the studied populations [5][6][7][8]. Moreover, infections are associated with increased mortality and account for the death of 20%-38% of MDS patients [8][9][10]. In addition, infections have a negative impact on quality of life and health care costs [10].
Current estimates of infection rates are based on results from clinical trials or retrospective analyses that only reviewed hospital records [5-8, 11, 12]. Most studies that analyzed infectious complications focused on specific subgroups of the MDS population, such as specific treatment (e.g. hypomethylating agents [HMA] or a hematopoietic stem cell transplantation [HSCT]), or type of infection (e.g., invasive fungal infections) [13][14][15][16][17]. These studies have shown that certain MDS patients have a higher risk for infection; however, the prescription rate of anti-infective agents in different MDS risk groups has not been investigated. Consequently, detailed information about the risk for infections translating into the prescription of anti-infective agents for MDS in clinical practice is lacking. Furthermore, infections and prescriptions for anti-infective agents in first-line care have not been considered.
In contrast to the inpatient setting, outpatient care generally involves empirical treatment, as bacterial cultures are rarely available.
One method of studying infections in an outpatient setting is to investigate the prescription of anti-infective agents. This approach is particularly useful in examining the burden of possible infections [18,19]. The HemoBase registry offers the means to assess the prescription of anti-infective agents in MDS patients in Friesland, a province in the Netherlands with ±650,000 inhabitants. This registry includes all hemato-oncological patients diagnosed since 2005 and provides population-based data on diagnosis, treatment, and the day-to-day practice of MDS patients [2,20]. Combined with first-line care data, it offers a unique view on how these patients are treated.
This study investigates the prescription rate of anti-infective agents in MDS patients (lower-and higher-risk MDS) in an outpatient and inpatient setting; it assesses the prescription of anti-infective agents over time before and after diagnosis of MDS and differences in prescription rates according to MDS risk groups, comorbidities, neutropenia, and age.

METHODS
A retrospective, population-based study was performed using the HemoBase registry [2,20,21]. The local Medical Ethics Committee confirmed the execution of the study without the need for ethical review. All living patients provided informed consent in accordance with the Helsinki Declaration (2013 revision) and Dutch regulations.
All persons newly diagnosed with MDS between January 1, 2005 and December 31, 2017 were included in the study and followed through December 31, 2020 (also see Supplementary Data) [2].

Community pharmacies and (dispensing) general practitioners (GPs)
provided information about the prescription of anti-infective agents from 10 years before MDS diagnosis (i.e., from 1995 onwards, when available) to the end of follow-up. The information systems of community pharmacies contain detailed, up-to-date information on prescribed, and over-the-counter medication, which guaranteed a complete overview of the prescriptions for anti-infective agents without potential recall bias. Dutch pharmacies are by law obliged to keep records of all patient data for at least 20 years [22]. Many records date back even further, yielding rich data on prescriptions for anti-infective agents (see Supplementary Data) [19]. In addition, The Netherlands has a strict policy for prescribing anti-infective agents, and the use of these agents is much lower compared to other countries [23,24].
Therefore, pharmacy records provide a valid estimate of the use of anti-infective agents in MDS patients [19,23]. Patients served as their own control; we compared the prescription rate of anti-infective agents in patients with MDS to their prescription rate before diagnosis, with a maximum of 10 years prior to diagnosis. Complete case analysis was performed for DID calculations (i.e., only patients with data before and after diagnosis were considered).
In-and outpatient data were combined and calculated with the total outpatient follow-up before and after diagnosis to calculate the total DID. Descriptive statistics were used to describe patient characteristics. Day-to-day use of anti-infective agents in all patients before versus after MDS diagnosis was graphically depicted. In this analysis, each day of follow-up was assessed for the prescription of anti-infective agents.
More than one prescription of an anti-infective agent on the same day in an individual patient was counted as a single prescription. Univariate analyses were performed to assess differences in the delta of prescription rates for anti-infective agents before diagnosis compared to after diagnosis in patients according to different baseline parameters:   Table 2.
Both the lower-and higher-risk MDS groups showed a rela-

DISCUSSION
This study showed a significantly increased prescription rate of antiinfective agents in patients after diagnosis of MDS as compared to before diagnosis. To our knowledge, this is the first descriptive population-based study that provides an overall insight on prescribed anti-infective agents within and beyond the clinical setting in a welldefined representative MDS patient cohort. MDS patients had a 2.5 times higher chance of receiving anti-infective agents in an outpatient setting after diagnosis. Higher-risk MDS patients were particularly at risk for receiving anti-infective agents; however, lower-risk patients also showed a significant increase, as the prescription rate doubled In particular, higher-risk MDS patients showed an increased risk for receiving outpatient anti-infective agents after diagnosis. This may be explained by the severity of the disease, as this group more often has severe neutropenia, impaired function of (regulatory) T cells, a higher transfusion burden, and more interventions compared to lower-risk patients and are therefore prone to infections [27][28][29]. Of note is the relatively short follow-up of higher-risk patients as shown in Figure 3, which can be explained by their poor prognosis, censoring for transplantation, and overall survival [2]. Although higher-risk patients are predominantly treated with HMA or chemotherapy to restore bone marrow function, especially after the start of treatment, they are at increased risk for infections [5,27,28]. Indeed, there was an increased prescription rate for higher-risk MDS patients who received treatment.
However, lower-risk patients also showed a significant increase in prescription rate: the DID doubled following MDS diagnosis. The effect of treatment should be further investigated in future studies, as this topic was beyond the scope of this study.
This study found that patients received 55.1 DID before MDS diagnosis, which is higher than prescription rates presented in other studies [23,[30][31][32][33][34]. The average total rate in 2019 was 9.5 DID in the Netherlands and 19.4 DID in Europe [23]. However, These data only included anti-infective agents from ATC-group J01, which may explain the lower figures compared to our data [23]. As 70% of our prescriptions consisted of agents from ATC-group J01, our study still shows an increased rate (42.4 DID) compared to the literature. Our results suggest that MDS patients have an increased prescription rate before diagnosis compared to other patient groups, and the prescription rate further increases after diagnosis of MDS. Interestingly, the prescription rate was relatively stable over the 10 years leading up to diagnosis.
Two studies have described the prescription of anti-infective agents in nursing homes: one reported a DID of 44.8, the other described rates varying from 20 to 120 DID across European countries [32,35]. Our results are in accordance with these findings, which included elderly populations such as ours. Studies describing prescription rates in other haemato-oncological patients were, to our knowledge, not available.
Hence, it remains difficult to assess whether MDS patients have an increased prescription rate of anti-infective agents a priori, and further research is needed.
There are several possible explanations for the increased prescription rate of anti-infective agents after diagnosis of MDS and the increased susceptibility to infections in this patient population. The first explanation is a functional impairment on a cellular level [36][37][38].
Bento et al. studied the monocytes of MDS patients and found that abnormalities may be associated with an impaired immune system [36].
Secondly, several studies have identified neutropenia or neutrophil dysfunction as risk factors for infection [4,15,27,28,39,40]. Neutropenia was a common finding (42%) in this population and is not limited to a specific subtype of MDS, which supports our finding that the entire MDS population showed an increased prescription rate of antiinfective agents [4,28]. Finally, comorbidities may also contribute to infections in MDS [15,28]. Two-thirds of our study population had at least one comorbidity, which further supports the increased prescription rate of anti-infective agents in these patients. Although not statistically significant, patients with a CCI score ≥2 indeed had a greater increase in prescription rate than patients with a CCI < 2. In larger study populations, the effect of comorbidities might be more visible.
Antibacterial agents were most commonly prescribed, which agrees with earlier findings that MDS patients predominantly suffer from bacterial infections [28,40,41].
Regarding the study's strengths and weaknesses, one notable strength of this study is the large amount of detailed information we collected in an unselected MDS population. We had the opportunity to compare prescription rates of anti-infective agents before and after diagnosis of MDS based on comprehensive prescription data from Dutch pharmacies. The detailed, up-to-date pharmacy information systems provided a complete overview without potential recall bias. Moreover, Dutch policies on prescribing anti-infective agents are strict and therefore may act as an accurate reflection of infection rates [23,24]. In addition, we received the data of 70% of the MDS population, and our study population consisted of a representative cohort of MDS patients [3,[42][43][44]. This investigation is the first populationbased study able to quantify the prescription rate of anti-infective agents in MDS patients in outpatient and inpatient care. It can provide guidance for future investigations and the management of risk for infections in this patient population.
A limitation of this study is the age difference in the pre-post MDS diagnosis comparison. Due to this design, the estimates of prescriptions of anti-infective agents after MDS diagnosis were based on an older population. To limit this effect, a maximum follow-up was set to 10 years before diagnosis. Furthermore, age at the time of diagnosis was not associated with increased prescription rates. Because of this design, characteristics such as gender and medical history were equal before and after diagnosis. Therefore, we feel this study shows an accurate comparison of the real-world prescription rate of antiinfective agents before and after diagnosis of MDS. Another limitation was that not all information could be retrieved for all patients, as this was a retrospective study, and it is possible that some inpatient data might have been missed. In addition, it was not feasible to study the effect of factors like neutropenia or comorbidities over time.
These parameters should ideally be collected prospectively to avoid selection bias and confounding by indication. These data were unavailable in our dataset and should be further investigated. Still, this study encompassed many detailed data on MDS patients and their prescription rates over time, and we feel it provides important information on the management of MDS in clinical practice.
In conclusion, this study showed a significantly increased prescription rate of anti-infective agents after diagnosis of MDS compared to before diagnosis. Higher-risk patients were particularly at risk for receiving anti-infective agents; however, lower-risk patients also showed a significantly increased risk, as the prescription rate doubled after MDS diagnosis. To our knowledge, this is the first populationbased study that has quantified the prescription rate of anti-infective agents within and beyond the clinical setting in a well-defined representative MDS patient cohort. Further research on infectious complications in this patient group and the rational use of anti-infective agents in a real-world setting is warranted.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of the article at the publisher's website.