Antimicrobial susceptibility of gram-positive and gram-negative bacteria: a 5-year retrospective analysis at a multi-hospital healthcare system in Saudi Arabia

Studying time-related changes in susceptible pathogens causing healthcare-associated infections (HAIs) is vital in improving local antimicrobial and infection control practices. Describe susceptibility patterns to several antimicrobials in gram-positive and gram-negative pathogens isolated from patients causing HAIs at three private tertiary care hospitals in Saudi Arabia over a 5-year period. Data on trends of antimicrobial susceptibility among bacteria causing HAIs events in children and adults at three tertiary private hospitals located in Riyadh and Qassim, Saudi Arabia, were collected retrospectively between 2015 and 2019 using the surveillance data datasets. Over a 5-year period, 38,624 pathogens caused 17,539 HAI events in 17,566 patients. About 9450 (53.8%) of patients who suffered HAIs were females and the average age was 41.7 ± 14.3 years (78.1% were adults and 21.9% were children). Gram-negative pathogens were 2.3-times more likely to cause HAIs compared to gram-positive bacteria (71.9% vs. 28.1%). The ranking of causative pathogens in decreasing order was: Escherichia coli (38%), Klebsiella species (15.1%), and Staphylococcus aureus (12.6%). Gram-positive isolates were mostly susceptible to linezolid (91.8%) whereas they were resistant to ampicillin (52.6%), cefoxitin (54.2%), and doxycycline (55.9%). Gram-negative isolates were mostly sensitive to tigecycline (95%) whereas they were resistant to cefotaxime (49.5%) and cefixime (59.6%). During the 5 years, there were relatively stable susceptibility patterns to all tested antimicrobials, except for cefotaxime which shown a susceptibility reduction by 41.4%, among Escherichia coli and Klebsiella species. An increase in the susceptibility of Acinetobacter and Enterobacter and Citrobacter species to all studied antimicrobials was observed except for colistin that had a slight sensitivity reduction in 2019 by 4.3% against Acinetobacter species. However, we noted reduced sensitivity of MRSA, CoNS and Enterococcus species to gentamicin; and increased resistance of MRSA to linezolid and vancomycin. The observed increase in susceptibility of gram-positive and gram-negative bacteria to studied antimicrobials is important; however, reduced sensitivity of MRSA, CoNS and Enterococcus species to gentamicin; and increased resistance of MRSA to linezolid and vancomycin is a serious threat and calls for effective antimicrobial stewardship programs.

Background Antimicrobial resistance (AMR) is a major threat to public health imposing significant health and economic burdens on healthcare system and patients [1,2]. Unless proactive solutions are found to address AMR, global costs are estimated to reach USD 3 trillion annually by 2050 and an additional 10 million people could die each year; cumulated costs could reach over USD 100 trillion [3]. Decreasing private sector investment in the development of new antimicrobials to treat AMR infections threatens global efforts to fight this danger; and AMR requires international attention and collaboration, because bacteria do not recognize borders.
In Saudi Arabia, misuse of antimicrobials is high and complicated primarily because antibiotics are available to buy by anyone over-the-counter via the community pharmacies without a legal prescription [4]. Only two years ago, Saudi Ministry of Health has implemented a nationwide ban on the sale of antibiotics without a legal prescription; however, despite this law, dispensing antibiotics without prescription is still common [4]. Routine clinical microbiology laboratory data provide a profile of the susceptibilities of specific bacteria to antimicrobial agents for monitoring and responding to emerging antimicrobial issues. Data can be utilized to help in the selection of empirical therapy by selecting the most appropriate antibiotics before susceptibility results are available, but remains generally unexploited for purposes of epidemiological surveillance. Although Antimicrobial stewardship programs focus on antibiotic prescribing practice, it is supported by an understanding of local antibiotic susceptibility trends, which in turn depends on the availability of a reliable medical microbiology laboratory resource. The Medical Group has implemented antimicrobial stewardship (AMS) programs since January 2014 and employs various strategies to reduce inappropriate utilization of antimicrobials, minimize the emergence of AMR and lower incidence of health-care-associated infections (HAIs) and reduce cost [1,5]. Several local studies have estimated the rates of susceptibility among gram-positive and gram-negative bacteria in Saudi Arabia [6][7][8][9][10], but none was comprehensive, and comparisons are complicated by variable methods and study periods that influence the findings explanation and interpretation.

Aim
This study aimed to examine patterns of antimicrobial susceptibility of gram-positive and gram-negative pathogens isolated from inpatients and outpatients causing HAIs using the surveillance data datasets collected from three HMG hospitals (Altakhassusi, Arryan and Qassim) over a 5-year period, in Saudi Arabia.

Settings
The private tertiary medical group is considered as one of the largest private healthcare providers in the Middle Eastern region. Currently, the medical group operates 14 medical facilities across Saudi Arabia, UAE and Bahrain, including 7 hospitals and 6 medical centers.
Study was conducted at three tertiary and specialized health facilities with adequate medical professional resources with 237-bed capacity, 365-bed and 150-bed capacity, respectively located in two different cities in Saudi Arabia.
These facilities provide healthcare services to a wide range of patients in various specialties and subspecialties. Yearly, the three healthcare facilities encounter over 127,364 surgical cases, nearly 1,742,144 visits to emergency departments, and over 360,587 admissions.

Study design
Data of trends in antimicrobial susceptibility among of all reports of four types of gram-positive isolates [Staphylococcus aureus, Methicillin-resistant Staphylococcus aureus (MRSA), Coagulase-negative staphylococci (CoNS) and Enterococcus species] and six types of gram-negative isolates [Escherichia coli, Klebsiella species, Pseudomonas species, Acinetobacter species, Proteus species, and Enterobacter and Citrobacter species] causing HAIs, collected from the infection control and prevention surveillance data between January 2015 and December 2019 from adult and pediatric patients in three tertiary private hospitals in Saudi Arabia, were extracted using standard customized Excel data collection sheets (Microsoft Corp, Redmond, WA, USA). The antimicrobial susceptibility patterns for selected antimicrobials were analyzed and reported.
We extracted the following patient data from the patient records meeting the inclusion criteria: age, gender, patient location (wards, intensive care units, and

Antimicrobial susceptibility testing
Species identification of isolates and their antimicrobial susceptibility profiles were obtained with different automated systems at every single laboratory of the three facilities using (VITEK ® 2 system, BioMariex, France), BD Phoenix system (BD Biosciences, NJ, USA), MicroScan plus (Beckman Coulter, CA, USA), and BD BACTEC system (BD Biosciences) according to manufacturers' specifications, between 2015 and 2019, with susceptibility interpretations based on the Clinical and Laboratory Standards Institute (CLSI) broth microdilution and breakpoint criteria [11]. To ensure data compatibility, quality control was performed using control strains from the following American Type Culture Collection (

Statistical analysis
Variables that were continuous were presented as means and categorical variables were presented as frequencies and percentages. Susceptibility patterns of pathogens were presented over time. The difference in sensitivity trends between 2015 and 2019 was examined using the multivariate analysis of variance (MANOVA) and a two-sided P-values < 0.05 were considered to be statistically significant. The proportion of susceptible isolates was calculated as the sum of susceptible organisms (neither intermediately susceptible nor resistant) relative to the total number of organisms tested. SPSS (Version 25.0. Armonk, NY: IBM Corp) and Microsoft Excel Professional Plus 2019 (Microsoft Corp., Redmond, WA, USA) were used for all statistical analyses.
Our study was performed in accordance with the ethical standards of the Declaration of Helsinki and its later amendments or comparable ethical standards. Ethics approval (RC20.10.95-2) was obtained by the Ethics Committee of the coordinator center (IRB Committee of Dr. Sulaiman Al Habib Medical Group, Riyadh, Saudi Arabia).  (Fig. 1).

Trends of susceptibility among gram-positive bacteria
A total of 79,280 gram-positive pathogen sensitivity events against 14 clinically important antimicrobials occurred at HMG Hospital in Altakhassusi (38.2%), HMG Hospital in Arryan (36.6%), and HMG Hospital in Qassim (25.2%). Gram-positive bacteria showed an overall susceptibility of ≥ 52.6%. Antimicrobial susceptibility patterns in gram-positive pathogens over time are presented in Table 1.
Over the 5 years, sensitivity of nitrofurantoin to overall gram-positive bacteria was the only antimicrobial to increase significantly (30.2% increase, p-value = 0.032). Prominent insignificant increase in the susceptibility of specific gram-positive bacteria to some antimicrobials occurred in 2019 compared to 2015 by: 30 Overall, among the studied antibiotics the gram-positive isolates were mostly sensitive to linezolid (91.8%) whereas they were resistant to ampicillin (52.6%), cefoxitin (54.2%), and doxycycline (55.9%) ( Table 1).

Trends of susceptibility among gram-negative bacteria
A total of 314,624 gram-negative pathogen sensitivity events against 21 clinically important antimicrobials occurred at HMG Hospital in Altakhassusi (35.9%), HMG Hospital in Arryan (39.3%), and HMG Hospital in Qassim (24.8%). Gram-negative bacteria showed an overall susceptibility of ≥ 49.5%. Antimicrobial susceptibility patterns in gram-negative pathogens over time are presented in Table 2.
Generally, the highest susceptibilities of gram-negative pathogens to antimicrobials were seen towards: tigecycline, meropenem, imipenem and amikacin by  were sensitive to amikacin (82.5%) and meropenem (80.3%); and Enterobacter and Citrobacter species were sensitive by ≥ 70% to most of the tested antimicrobials over the 5-year period. In contrary, lowest susceptibilities of gram-negative pathogens to antimicrobials were seen to cefotaxime and cefixime by Acinetobacter species, 35.2% and 37.6%, respectively. Acinetobacter species shown low sensitivity of ≥ 40% almost to all antimicrobials; and both Klebsiella and Pseudomonas species were slightly sensitive to cefotaxime (45% and 46%, respectively).
Over the 5 years, sensitivity of cefazolin and ofloxacin to Pseudomonas species increased significantly (21.5% and 31.5% increase, p-values = 0.046 and 0.030; respectively). The small sensitivity increase of Acinetobacter species towards tigecycline was found to be significant (0.4% increase, p-value = 0.022). In addition, large difference in susceptibility were found for both ampicillin and cefuroxime towards Enterobacter and Citrobacter species (41.2% and 35.1% increase, p-values = 0.023 and 0.003; respectively).
For a 5-year difference, sensitivity of Acinetobacter species to antimicrobials shown many insignificant increases: (rate of sensitivity increase: for cefepime, 40%; for cefotaxime, 38 Overall, among the studied antibiotics the gram-negative isolates were mostly sensitive to tigecycline (95%) whereas they were resistant to cefotaxime (49.5%) and cefixime (59.6%) ( Table 2).

Discussion
This retrospective study describes the distribution of pathogens causing HAIs and susceptibility patterns for a very high number of samples collected from both the ward and clinics in Saudi Arabia from 2015 to 2019, with an emphasis on the antibiotic classes frequently utilized to treat common infections given by a huge national surveillance program. The most commonly encountered organisms were Escherichia coli, Klebsiella species, and Staphylococcus aureus. Though various studies have previously described susceptibility rates in several infectious isolates, Saudi data are limited either to single-center studies [7,[12][13][14][15][16][17][18][19][20] or to research concentrating on the susceptibility to single or double antimicrobial classes [21][22][23][24][25][26].
One of the vital findings of the data analysis of this study was the significant increase of sensitivity for overall gram-positive bacteria to nitrofurantoin over the 5 years (30.2% increase, p-value = 0.032). Interestingly, the susceptibility of Staphylococcus aureus to tetracycline; MRSA to gentamicin and clindamycin; CoNS to cefoxitin, tetracycline, clindamycin, trimethoprim-sulfamethoxazole and nitrofurantoin; and Enterococcus species to trimethoprim-sulfamethoxazole, nitrofurantoin and vancomycin; increased insignificantly over time by ≥ 30% although this was likely due to the change of followed guidelines used for antimicrobial susceptibility testing at the Medical Group facilities, a shift from the Clinical and Laboratory Standards Institute (CLSI) to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [11,27].
A comparison of the current results with findings from previous studies can offer some validation of the findings of this present study and identify methodological distinctions in their approaches. As expected, HAI events were more frequent in the ICUs (37.2%) compared with non-ICU locations [HAI events in wards and outpatients were 32.9% and 29.9%, respectively], a finding which was previously described in local studies [10,28] and may reflect the epicenter role of ICU in both infections and antimicrobial resistance. The predominant isolates to cause HAIs were gram-negative organisms (71.9% vs. 28.1%); this finding was similar with many Saudi studies made in different cities in Saudi Arabia including Riyadh [6,28,29], Makkah [30,31], Dhahran [32], Bisha [33], and Aljouf [10]; with the majority being Escherichia coli (38%) accounting approximately for 52.9% of the gram-negative bacterial growth in line with previous national studies [7,29,30,32,34]. The second predominant isolates of the gram-negative organisms were the Klebsiella species (15.1%), this finding was similar to the bacterial isolates prevalence studies from Dhahran [7], and Riyadh  [6]. The proportions of Klebsiella were 17.2% in Dhahran [7], and 14.7% in Riyadh [6]. The culture rate in our study for Proteus species (2.3%) was comparable to previously reported rates in two different studies in Riyadh (1.2% and 1.8%) [6,35]. Also, the incidence of Acinetobacter species in our study was very close to the rate reported before (4.2% vs 5.5%) [6]; in contrast to the much higher rates found in two separate studies in Riyadh (31.7% and 25.3%) [35,36]. Our prevalence of Staphylococcus aureus was similar to the rate described in a previous report done in Riyadh (12.6% vs 13.9%) [6]. We report a lower rate of MRSA (15.9%) compared to two previous studies made in Riyadh (24.4% and 30.3%, respectively) [28] but similar to the rate reported before in another study in Riyadh (17.5%) [29]. We report a higher susceptibility of Enterococcus species to vancomycin (85.9% vs 79.7%) compared to one study in Riyadh [6]. In our study, proportion of Pseudomonas species that caused HAIs is less than what was reported in Riyadh (10.1% vs 15.4%) [6]; however, our prevalence was in agreement to the bacterial frequency in a study from Dhahran (12.8%) [7]. Frequency of CoNS in causing HAIs in this study is in line with a study from Riyadh (5.2% vs 6.5%) [6] but much lower than the rate reported previously in a study in Riyadh (28.4%) [29]. In our study, incidence of Enterococcus species as causative pathogens for HAIs is almost half of the reported rate by a study in Riyadh (4.5% vs 8.6%) [6]; however, rate was in parallel to the prevalence reported in other study in Riyadh (5.9%) [28] but contradicts with the rate reported in another study in Riyadh (15.8%) [29]. Our data analysis regarding the susceptibility patterns of antimicrobials confirm or contradict the findings of previous local studies. For example, Pseudomonas and Acinetobacter were most susceptible to colistin and amikacin in a study in Riyadh [29], whereas in our study, tigecycline and colistin had higher susceptibility rates. On the other hand, Escherichia coli, Klebsiella pneumonia, Enterobacter and Citrobacter species were most sensitive to amikacin, imipenem and meropenem [29], whereas in our study, Escherichia coli and Enterobacter and Citrobacter species were most susceptible to tigecycline, and Klebsiella species was most susceptible to amikacin. However, our study support the finding that CoNS were most susceptible to vancomycin and linezolid [29] and we found the susceptibility of Staphylococcus aureus to clindamycin and trimethoprim-sulfamethoxazole were almost identical to the results of the aforementioned study (92% vs 94% and 87.7 vs 87%, respectively). This is might be due the fact that the sample was drawn from tertiary private hospitals in Saudi Arabia where the level of environmental hygiene is higher and staff are highly restricted to infection control practices.
Linezolid and vancomycin had the best susceptibility profile to Staphylococcus aureus, CoNS, and Enterococcus species while gentamicin shown low sensitivity towards MRSA, CoNS and enterococcus species. In the context of emergence of resistance of malicious grampositive bacteria to gentamicin, linezolid and vancomycin have become effective alternatives to gentamicin treatment frequently associated with nephrotoxicity [37]. Linezolid and vancomycin are active against the most serious gram-positive bacteria, including streptococci, vancomycin-resistant enterococci (VRE) and MRSA [38]. Nevertheless, we noted a low rate of susceptibility of linezolid and vancomycin against MRSA (54.9% and 50.9%, respectively) likely because of antibiotic selection pressure and possibly a reflection of selective reporting of susceptibility testing; this finding contradicts those of a recent study in Riyadh, which identified a 100%-sensitivity of both antimicrobial agents for MRSA [29]. The relatively lower susceptibility in gram-positive bacteria in the current study may be reflecting Saudi prescription trends in recent years that overuse fluoroquinolones [1,39] and carbapenems [40,41] at the expenses of other broad-spectrum such as linezolid and vancomycin due to increased availability and reduced cost of these drugs. However, nitrofurantoin maintained the greatest efficacy against MRSA in our study (82.6%); supporting the finding of a recent study in Aseer that shown 100% susceptibility of MRSA to nitrofurantoin [42].
Over the 5-year period, it is interesting to note imipenem and meropenem either retained its activity or shown susceptibility increase patterns towards all the studied gram-negative pathogens except for imipenem which was less sensitive in 2019 by 1.6% against Escherichia coli and for meropenem that shown a minor sensitivity reduction by 8.6% to Klebsiella species. Previous studies from Saudi found high susceptibility of Pseudomonas to carbapenems [7,41]; however, in other local studies, the susceptibility of Pseudomonas to meropenem declined over a five-year period [29] and nonsusceptibility of Acinetobacter and Pseudomonas aeruginosa to carbapenems was very high (68.3% and 76%) [6,41]. Furthermore, there were relatively stable susceptibility patterns to all tested antimicrobials, except for cefotaxime which shown a susceptibility reduction by 41.4%, among Escherichia coli and Klebsiella species; in opposite to the finding of a local study in Dhahran that shown a reduction trend in the susceptibility of antibiotics to Escherichia coli and Klebsiella species [7]. Moreover, we observed an increase in the susceptibility of Acinetobacter and Enterobacter and Citrobacter species to all studied antimicrobials except for colistin that had a slight sensitivity reduction by 4.3% against Acinetobacter species. This can be considered as a part of the success of the combating strategies implemented since January 2014 at the medical settings to reduce further emergence and spread of AMR, lower the percentage of HAIs and MDR organisms, and save on needless healthcare expenses [1]. Significant differences in antibiogram findings between different healthcare facilities and regions may suggest differences in populations of the served patients, patterns of antimicrobial use, or deficiencies in hospital infection control and hygiene practices that could be further explored.

Limitations
This study had a few limitations. Firstly, the retrospective design and the risk of misclassification and selection bias. For instance, even though the laboratories follow the highest standards, there may be a possibility that some isolates had some contaminants. Furthermore, since all three hospitals in this study are tertiary care hospitals, they receive more complicated cases that may be caused by resistant pathogens which may not indicate the antibiotic susceptibility trend and microbiology of the general population. Nevertheless, our study's findings will add to local and global data on antimicrobial susceptibility, especially with highly threatening infections.

Conclusion
Systematic collection and analysis of routine clinical laboratory data is important in assessing the antimicrobial resistance burden. Nationwide surveillance is urgently needed to provide policy makers, antimicrobial stewardship committees, infection preventionists, microbiologists, and epidemiologists with essential information to guide proper action plans. The observed increase in susceptibility of gram-positive and gram-negative bacteria to studied antimicrobials is important; however, reduced sensitivity of MRSA, CoNS and Enterococcus species to gentamicin; and increased resistance of MRSA to linezolid and vancomycin is a serious threat and calls for effective antimicrobial stewardship programs.