Correlation between Antibiotic Consumption and Resistance of Invasive Streptococcus pneumoniae

There is a lack of long-term studies that correlate different metrics of antibiotic consumption and resistance of invasive S. pneumoniae. The present study aims to investigate the correlation between national outpatients total antibiotic, penicillin and broad spectrum penicillins consumption expressed in daily doses per 1000 inhabitants per day (DID) with the ATC/DDDs, WHO version of 2019 (new version) and 2018 (old version), number of prescriptions per 1000 inhabitants per year (RxIDs) and number of packages per 1000 inhabitant per day (PIDs) with the resistance of invasive S. pneumoniae in Slovenia in the period from 2000 to 2018. The prevalence of penicillin resistance of invasive S. pneumoniae decreased by 47.13%, from 19.1% to 10.1%. Decline of resistance showed the highest correlation (R = 0.86) between RxIDs followed by PID (R = 0.85) and resistance of S. pneumoniae. Higher correlation between total use of antibiotics expressed in DID WHO version 2019 (R = 0.80) than for WHO version 2018 (R = 0.78) was found. Very high (R = 0.84) correlation between use of β-lactams expressed in PID, and RxIDs (R = 0.82) and reasonable (R = 0.59) correlation expressed in DIDs version 2019 was shown as well. The consumption of broad -spectrum penicillins (J01CA and J01CR02) expressed in PID (R = 0.72) and RxIDs (0.57) correlated significantly with the resistance of S. pneumoniae as well. A new finding of this study is that RxIDs correlated better with the resistance of S. pneumoniae than total consumption of antibiotics expressed in DID and significant correlations exist between use of broad-spectrum penicillins expressed in PID and RxIDs.


Introduction
Streptococcus pneumoniae (also known as pneumococcus) is an important pathogen that causes severe infections in children and older adults worldwide [1]. Since 2017, S. pneumoniae has been listed as one of the twelve priority pathogens by the World Health Organization (WHO) [2]. Rising rates of resistance to penicillin and other antibiotics, as well as the continuing high burden of disease, have renewed interest in antibiotic therapy and prevention [3][4][5][6]. Antimicrobial resistance (AMR) is a threat to public health systems around the world [7,8]. Antibiotic use is a known risk factor for the emergence of antibiotic resistance, but demonstrating a causal relationship between antibiotic use and resistance is challenging. Several studies have found associations between total antibiotic and penicillin use with Streptococcus pneumoniae resistance to penicillin [9][10][11][12][13]. In contrast a study published by Olesen et al. showed weak or no correlations between total antibiotic use and the use of β-lactams with resistance to S. pneumoniae [14]. In all these studies, the only metric of antibiotic consumption was daily doses per 1000 inhabitants per day (DID and "old" DDDs were used [10,13,14]. The Collaborating Centre for Drug Statistics Methodology of the World Health organization (WHO-CC-DSM) recently changed some DDD definitions. From 1st January 2019, DDD measurement was revised for several Anatomical Therapeutic Chemical (ATC) classification codes for drugs commonly used in ambulatory care (e.g.,oral drugs, amoxicillin (J01CA04), amoxicillin/clavulanic acid (J01CR02)] and midecamycin (J01FA03) [15]. As a consequence of these changes there is a need to see the effect of antibiotic consumption expressed in new DIDs on the development of resistance. There is also a lack of long-term studies that correlate different metrics of antibiotic consumption and resistance of invasive S. pneumoniae. DDD is standardized tool for drug utilization research allowing presentation, comparison and benchmarking of drug consumption statistics at international and other levels. The feasibility and usefulness of this standardized metric strongly favours its complementation with another metric.
The present study aims to investigate the correlation between total antibiotic, penicillin and separately broad-spectrum penicillins consumption expressed in three metrics; DID the WHO version of 2019 (new version) and 2018 (old version), number of prescriptions per 1000 inhabitants per year (RxIDs) and number of packages per 1000 inhabitant per day (PIDs) with the resistance of invasive S. pneumoniae in Slovenia in the period from 2000 to 2018. Understanding the links between antibiotic use and S. pneumoniae resistance is critical for improving guidelines for therapy and vaccination strategy in Slovenia.

Results
Slovenia is a small country with a population of roughly 2.08 million inhabitants, according to the 2019 census [16]. In Slovenia, virtually all residents (>99%) have mandatory basic health insurance, and a prescription is required for any type of antibiotics. In addition, physicians prescribe antibiotics only for humans. Data on the number of antibiotic packages, the cost of antibiotics, the age and sex of patients and the identity number of the physicians and health care institutions prescribing antibiotics have been collected and published annually since 1976. In the period of 2000-2018, data on outpatient antibiotic use were collected using the ATC classification (ATC/defined daily doses (DDD) classification (WHO version 2018 and 2019) [15,17] The number of prescriptions for insured individuals and outof-pocket paid antibiotics ("white" prescriptions for uninsured persons and prescriptions before travel) were provided by the National Institute of Public Health of Slovenia (NIPH) and Health Insurance Institute of Slovenia (HIIS). National antimicrobial stewardship activities were recently published [18,19]. In Slovenia, the surveillance of invasive diseases in children caused by S. pneumoniae has been continuously monitored since 1993 and in adults since 1996 [12,13,20].

Discussion
The study shows a very high correlation between total use of antibiotics for systemic use (J01) expressed in RxIDs (R = 0.86), PID (R = 0.85) and DIDs version 2019 (R = 0.80) and resistance of invasive S. pneumoniae isolates. Total consumption expressed in DDDs version 2018 showed lower correlation (R = 0.78). The results are in agreement with previous studies that reported a significant correlation between total antibiotic use, expressed as DID, and resistance of S. pneumoniae [12,13].
A new finding of our study is that the consumption expressed in RxIDs had a higher correlation with the resistance of S. pneumoniae than the consumption of total use of antibiotics expressed in both DIDs. A very high correlation was also found between the use of penicillins, expressed as PID (R = 0.84) and RxIDs (R = 0.82) and lower (R = 0.59) correlation with DDDs version 2019. Nonsignificant correlation was found with DDDs version 2018 (R = 0.45, p = 0.05). In the current study, a lower correlation was found (R = 0.59 and R = 0.45 respectively) between use of penicillins expressed in DIDs and resistance of S.pneumoniae. Goossens et al. reported a Spearman's coefficient of 0.84 and van de Sande Bruinsma a coefficient of 0.78 [12,13]. Our data are in agreement with the Olesen et al. study [14]. The highest correlation (R = 0.84) between use of β-lactams expressed in PID and resistance of S. pneumoniae is in accordance with the Bruyndonckx study [21] If we have no substantial changes in the amount of DDDs per package, PID is a good metric for antibiotic consumption in outpatients [22]. DID and PID, are used for a better understanding and interpretation of outpatient antibiotic use and its relation to resistance. RxIDs counts number of treatments independently of the prescribed doses and is used in many countries. It is useful in countries where OTC antibiotic use is not available [23]. A significant correlation (R = 0.72), (p < 0.001) and (R = 0.57), (p < 0.05) between consumption of extended spectrum penicillin (J01CA) and penicillins with β-lactamase inhibitors (CR02), expressed in PID and RxIDs and resistance of S.pneumoniae was found as well. Previous studies analysed resistance of S. pneumoniae among different countries, however this study analyses the long-term correlation between the use of antibiotics and resistance of S. pneumoniae in one country [12,13]. The selective pressure of antibiotics is the most important driver of antimicrobial resistance, although the consumption of antibiotics does not always correlate with the prevalence of resistance. EARS-Net data from 2018 and antibiotic consumption in ESAC-Net data from 2017 show that there are always exceptions when comparing antibiotic consumption and resistance [24,25]. Antibiotic selective pressure, clonality of resistant clones, co-selection and the fitness cost of resistance are important factors [23,25]. The spread of resistant clones is found in many studies [26][27][28][29][30]. In this study, a large number of different serotypes were found among penicillin-resistant isolates, suggesting the absence of clonal spread. In Slovenia, there is a decrease in the selective pressure of penicillin (12.38%) and co-selection with decreased use of other classes of antibiotics. The consumption of co-selective antibiotics, used in outpatients, co-trimoxazole (45.4%), macrolides (55.9%) and tetracyclines (77.6%) substantially decreased.

Limitations of the Study
In Slovenia the universal vaccination in the national vaccine recommendations programme was introduced in 2015 and might have an influence on the prevalence of resistance of S. pneumoniae. Moderate coverage of vaccination (49-55%), low percentage of isolates (22%) in children and nonsignificant influence on the incidence of S. pneumoniae invasive infections diminish the influence of vaccination on the prevalence and resistance of S. pneumoniae [31].

Collection of Isolates and Serotyping
Isolates of S. pneumoniae obtained from sterile body sites from patients with suspected invasive infection were collected from all Slovenian microbiological laboratories in the Department for Public Health Microbiology, National Laboratory of Health, Environment and Food in Ljubljana. The isolates were identified by classical colony morphology and hemolysis on blood agar and further tested for optochin susceptibility (Optochin Disc, Oxford, UK) and bile solubility. A total of 4223 isolates of S. pneumoniae were collected from 2000 to 2018. Of these, 930 (22%) were obtained from children (0-14 years) and 3293 (78%) from adults. Resistance was determined in 4208 isolates, as 15 isolates died off. The number of isolates per year varied from 93, in 2002, to 330, in 2015. Isolates were serotyped by Neufeld Quellung reaction with antisera (Statens Serum Institut, Copenhagen, Denmark).

Data Analysis
Statistical analyses were performed using the free software R (R Core Team 2019, Vienna, Austria) [34]. Spearman's rho2 rank correlation coefficients at a 95% confidence level were calculated to infer possible monotonic associations between S.pneumoniae resistance and total antibiotic and penicillin consumption expressed with DDDs (WHO versions of 2018 and 2019), RxIDs and PIDs.

Conclusions
Data from the national long-term survey shows very high correlation between total antibiotic use expressed in RxIDs, PIDs and DIDs, WHO version 2019, and high correlation with DIDs WHO version 2018 with resistance of S.pneumoniae to penicillin. A new finding of this study is that the consumption expressed in RxIDs has a higher correlation with the resistance of S. pneumoniae than consumption of total use of antibiotics expressed in both DIDs. A significant correlation was also found between the use of penicillins expressed in PID and RxIDs and use of broad spectrum penicillins expressed in PID and RxIDs and resistance of S.pneumoniae respectively. Based on the stated results, RxIDs should be considered to be added to DID to monitor outpatient's antibiotic consumption. Recently, the number of prescriptions per defined population was included in consensually validated metrics to assess the quantity of antibiotic use in the outpatient setting, enabling (inter)national comparisons [23]. To reduce the prevalence of resistance of S. pneumoniae to penicillin reduced number of prescriptions of all antibiotics and reduction of penicillins and broad-spectrum penicillins use is needed.