Ten-year trends of antibiotic prescribing in surgery departments of two private sector hospitals in Central India: a prospective observational study

Antimicrobial resistance (AMR) is a global health threat that contributes to more than five million deaths every year worldwide [1]. AMR is primarily driven by the inappropriate use of antibiotics, which ultimately leads to a decrease in their effectiveness and an increased necessity for frequent use of broad-spectrum antibiotics to treat common infections [1, 2]. Despite rising awareness of the AMR problem, antibiotic consumption is on the rise globally. Especially in the low-and middle-income countries (LMICs), antibiotic consumption continues to increase, due to still high infectious disease burden, insufficient vaccination coverage, rising incomes, lower drug prices, increased access, unregulated prescription and sales, inadequate infection prevention and control measures, lack of diagnostic ability and surveillance, and lack of education and public awareness [2‐5]. India is one of the biggest antibiotic consumers in the world [2, 4, 6] and showed the overall greatest increase (103%) in antibiotic consumption between 2000 and 2015 among all LMICs [4]. In 2019, an analysis of aggregate levels of drug resistance showed that India had the worst drug resistance index among 41 countries, resulting in the lowest level of antibiotic effectiveness [7].

In 2017, the World Health Organization (WHO) developed the Access, Watch, and Reserve (AWaRe) Classification to promote rational use of antibiotics, global comparison and support local and national policy development and antibiotic stewardship efforts [1]. Antibiotics are classified into AWaRe groups, based on their AMR potential and preference for use. The WHO has set a target that by 2023 in every country, a minimum of 60% of all prescribed antibiotics should be from the Access group [1]. In India, total consumption of Access antibiotics decreased by 13% from 2011 to 2019, whereas consumption of Reserve antibiotics increased by 247% [2], which indicates rising levels of AMR [4].

Despite widespread antibiotic use and advancements in surgical techniques, surgical site infections (SSIs) remain a significant cause of morbidity, mortality, and increased healthcare costs globally [8‐10]. SSIs are the most common healthcare-associated infections that develop due to contaminated instruments or environment at the healthcare facility within 30 days after surgery or within 90 days if an implant is left in place; and can involve skin, tissues under the skin, organs, or implanted material [11‐13]. One of the recommended measures to minimize the risks and consequences of SSIs is the administration of systemic antibiotics before or during surgery, i.e., perioperative antibiotic prophylaxis (PAP) [14, 15]. It is estimated that 30-50% of all antimicrobials in hospitals are used for PAP [16], up to 60% of surgical patients receive postoperative antibiotics during hospital stay, and up to 50% are discharged with antibiotics [17]. A significant proportion of antibiotic prescribing in surgery departments is reported to be inappropriate, which consequently contributes to AMR in a vicious cycle [16‐18]. The most common drivers of inappropriate antibiotic prescribing are incorrect dose, frequency or duration of therapy, and use of broad-spectrum antibiotics when narrow-spectrum would suffice [19, 20]. Reported reasons for inappropriate antibiotic prescribing in surgery are uncertainty in diagnosis, complex comorbidities, prescriber's lack of experience or training, unfamiliarity with local resistance patterns, lack of laboratory capacity, and mistakes in microbiological results' interpretation [10].

The incidence of SSIs is considerably higher in LMICs than in high-income countries [8, 21]. In India, the risk of acquiring SSI varies largely across the country and between healthcare facilities and ranges between 1.6% and 38% [22]. SSI incidence in the study setting was estimated between 5 and 7.5% in different time periods and departments (general surgery and orthopaedic surgery) [23, 24]. Despite the high SSI risk, there is no national policy nor guidelines for PAP and antibiotic prescribing in surgery, and antibiotic choices are often empirical [25, 26].

Ninety-three percent of all hospitals in India belong to the private sector [27] and it is estimated that up to 90% of total drug consumption occurs in the private sector health facilities [2]. Yet the research on antibiotic prescribing has mostly been conducted in public hospitals. Consequently, leading to a knowledge gap in antibiotic prescribing patterns in the major healthcare sector of India [28‐30]. Thus, it is crucial to estimate the actual use of antibiotics before planning and implementing antibiotic stewardship programs in the private hospitals. Previous research conducted in private hospitals in Madhya Pradesh suggested that the most commonly prescribed antibiotics were broad-spectrum and often prescribed empirically without a clear indication [28‐31]. The WHO emphasizes on the importance of long-term surveillance data for better reliability in developing and implementing antibiotic prescribing policies and antibiotic stewardship programs. However, there was no research done on long-term antibiotic prescribing trends in surgery departments in private-sector hospitals in Central India. Therefore, the present study aims to analyse and present 10-year antibiotic prescribing patterns, trends, and their appropriateness at group level in surgery departments of two private sector hospitals in Madhya Pradesh.

In total, the study population consisted of 29,515 surgical patients (TH, 15,016; NTH, 14,499, Table 1). In both hospitals, most patients were male (TH, 78%; NTH, 71%) and below 46 years of age (TH, 53%; NTH, 61%). Median length of hospital stay was longer in the TH (12 days) than in the NTH (4 days). In both hospitals, the majority of the patients were prescribed one antibiotic for five days. Culture and susceptibility tests were rarely performed (TH, 2%; NTH, 1%). More patients were prescribed antibiotics at discharge in the NTH (50%) compared to the TH (14%, Table 1).

Table 2 shows that in total 257,205 (TH- 178,712; NTH- 78,493) antibiotic prescriptions were prescribed during the study period. Fluoroquinolones (20%) in the TH and 3^rd generation cephalosporins (41%) in the NTH were the most prescribed antibiotic classes, whereas oral metronidazole (10%) in the TH and ceftriaxone (16%) in the NTH were the mostly prescribed antibiotics. Adherence to the NLEMI (TH, 80%; NTH, 69%) was higher than adherence to the WHOMLEM (TH, 77%; NTH, 66%) in both hospitals. Fixed-dose combinations (FDCs) of antibiotics were more frequently prescribed in the NTH (31%) compared to the TH (17%). Most of the prescribed antibiotics belonged to the Access group in the TH (57%) and the Watch group in the NTH (66%, Table 2).

The most common diagnoses for admission and operation were inguinal hernia (12%) in the TH and calculus of kidney and ureter (13%) in the NTH, followed by hyperplasia of prostate (TH, 9%; NTH, 10%, Table 3). In total, 5,733 patients (38%) were operated in the TH and 7,825 patients (54%) in the NTH. In both hospitals, all operated patients were prescribed antibiotic prophylaxis. In the TH, the most prescribed antibiotic prophylaxis was norfloxacin (19%), followed by lincomycin (15%); while in the NTH, the most prescribed antibiotic prophylaxis was ceftriaxone (24%), followed by amikacin (11%, Table 3).

This was the first study to our knowledge to assess the long-term antibiotic prescribing trends at surgery departments of private-sector hospitals in Central India. Results of this study show that the most common diagnoses were inguinal hernia (TH) and calculus of the kidney and ureter (NTH). Most prescribed antibiotic classes were fluoroquinolones (TH) and 3^rd generation cephalosporins (NTH). All operated patients were prescribed antibiotic prophylaxis, mostly norfloxacin (TH) and ceftriaxone (NTH). In both hospitals, antibiotic prescriptions were rarely accompanied by culture and susceptibility tests, and adherence to the NLEMI was higher than adherence to the WHOMLEM. The majority of prescribed antibiotics belonged to the Access group in the TH, and the Watch group in the NTH. Total antibiotic use was two times higher in the NTH compared to the TH; showed an increasing trend until 2012 in the TH, and until 2014 in the NTH, and after that started decreasing in both hospitals. Nevertheless, the proportion of Watch antibiotic prescriptions significantly increased in both hospitals over 10 years.

All patients in this study were prescribed antibiotics even though antibiotics were largely not indicated for the observed most common diagnoses. In addition, not all patients admitted to surgery departments were operated (TH, 38%; NTH, 54%), so, the main reason for the prescription of antibiotics was not PAP. Furthermore, the most common diagnoses in both hospitals- inguinal hernia, calculus of kidney and ureter, hyperplasia of the prostate- do not classify as infectious nor high-risk for infection. PAP is not recommended for elective inguinal hernia repairs [36]. For benign hyperplasia of the prostate, the most common surgical procedure is transurethral resection of the prostate (TURP), for which PAP is indicated in all cases [37]. However, before 2015, open prostatectomy was commonly performed in the TH for the treatment of benign hyperplasia of the prostate, which is a more invasive procedure associated with higher perioperative morbidity, longer hospitalization, and prolonged recovery [38]. Thus, open prostatectomy might warrant longer postoperative antibiotic treatment, though evidence suggests that only one day of postoperative antibiotic regimen at the time of the catheter removal is sufficient [39]. In the treatment of calculus of the kidney and ureter, ureteroscopy is a common, minimally invasive procedure for which risk-adapted minimal antibiotic usage is recommended, usually in the form of a single-dose PAP [40]. For removing larger kidney and ureter stones, percutaneous nephrolithotomy is a commonly performed procedure; however, it might cause postoperative infections such as fever and urosepsis. In order to prevent postoperative infections, for patients with negative urine culture- a single dose of PAP is recommended; for patients with positive urine culture- a preoperative antibiotic treatment based on the susceptibility pattern of urine culture in duration of seven days is recommended; and for patients with positive stone culture- postoperative antibiotics based on the stone culture susceptibility pattern are recommended [41]. However, in our study, very few culture and susceptibility tests were performed, despite the functional laboratories in both hospitals. Research shows that surgical patients are more likely than other patients to receive antibiotics during their hospital stay [18]. Antibiotic prescribing without indication was reported as a frequent medical error in surgery departments [10]. Surgeons often see antibiotic prescribing as a necessary intervention, even without any evidence of infection, mainly driven by the fear of SSIs and to avoid the blame in case of complications [18, 42]. This practice was reported as the main driver of inappropriate antibiotic use in the postoperative phase [18].

In this study, all operated patients were prescribed antibiotic prophylaxis. However, PAP is not indicated for all surgeries, i.e., it is not indicated for clean, minor, and non-prosthetic-associated procedures. PAP should not be used as a substitute for adequate infection prevention and control measures and skin preparation techniques [20]. Nevertheless, surgeons in our study setting often report poor hygienic conditions of patients, as well as severe and delayed clinical presentations as the reasons to always give PAP (personal communication). The selection of PAP is based on the anatomic region to be operated, bactericidal activity, pharmacokinetic and safety profile, ease of administration of the antibiotic agent, local resistance patterns, and cost [15]. Therefore, a 1^st generation cephalosporin cefazolin is the recommended choice for PAP for the majority of surgical procedures unless contraindicated, as it covers the most common causative pathogens of SSIs, has good pharmacokinetic and safety profile and is cost-effective [1, 13, 15, 20]. In our study, however, a fluoroquinolone-norfloxacin (TH) and a 3^rd generation cephalosporin- ceftriaxone (NTH) were mostly used as antibiotic prophylaxis, and cefazolin was only prescribed in 2% of cases in both hospitals. Surgeons in the two study hospitals preferred ceftriaxone as PAP because of the good coverage of gram-positive and gram-negative bacteria in skin and soft-tissue infections and in combination with metronidazole for intra-abdominal infections [43]. Nevertheless, ceftriaxone is not recommended for all skin and soft-tissue infections, but mainly for moderate non-purulent skin and soft-tissue infections [44]. Norfloxacin was mainly given as antibiotic prophylaxis at the time of admission for urinary tract infections, which were commonly observed by the surgeons in the study setting (personal communication). As norfloxacin comes in oral formulation, it is likely that it was not given as PAP, but as antibiotic prophylaxis for urinary tract infections on the day of surgery. In addition, for a majority of surgical procedures, a single dose of intravenous PAP is adequate. Postoperative administration of antibiotics is only required in special cases, e.g. lower limb amputations and some cardiac and vascular surgeries, for a maximum duration of 24 hours [15, 20]. In our study, the median length of antibiotic therapy in both hospitals was five days and, therefore, was beyond a single dose of PAP. Extended duration of PAP is frequently reported as one of the main drivers of inappropriate antibiotic use in surgery worldwide, which increases the likelihood of adverse reactions and development of AMR [20, 26, 42].

The most prescribed antibiotic classes were fluoroquinolones in the TH and 3^rd generation cephalosporins in the NTH, both belonging to the Watch group of antibiotics. The predominant use of 3^rd generation cephalosporins and the increase in consumption of fluoroquinolones have been noted in all LMICs [4]. In India between 2011 and 2019, the amount of consumed 3^rd generation cephalosporins was higher than the combined amounts of 1^st and 2^nd generation cephalosporins; which resulted in 83% of bacterial isolates being resistant to 3^rd generation cephalosporins on the national level in 2019 [2]. Against the WHO recommendation [1], broad-spectrum 3^rd generation cephalosporins have long been used in India as a first choice in empiric antibiotic treatment for the respiratory tract, skin and soft tissue, and gonococcal infections, as well as for enteric fever; and have consequently become largely ineffective against infections of extended-spectrum beta-lactamase-producing (ESBL) Enterobacteriaceae [3, 4].

Further assessment of appropriateness of antibiotic prescribing at group level shows somewhat more appropriate antibiotic prescribing patterns and trends in the TH than in the NTH. In the TH, the proportion of prescribed Watch antibiotics and FDCs was lower compared to the NTH, whereas the adherences to the WHOMLEM and NLEMI were higher. Additionally, more patients were operated on and prescribed antibiotics at discharge in the NTH compared to the TH, despite the fact the hospital stay was significantly shorter in the NTH (4 days) compared to the TH (12 days). A possible explanation for this is that the TH provided services free of charge; whereas in the NTH, patients had to pay for services and medicines out-of-pocket; therefore, the expectations on surgery and antibiotic prescription both from the provider and patient side might be higher. Nevertheless, in both surgical departments, the antibiotic treatment was prolonged (5 days) and mostly empirical (≤2% culture and susceptibility tests performed), which may suggest inappropriate antibiotic use. However, in the absence of detailed information about patient comorbidities and postoperative complications of each patient, it is not possible to fully conclude on the appropriateness of antibiotic prescribing. Nonetheless, the misuse of antibiotics is a known driving force for the development of AMR [1].

In our study, total antibiotic use showed an increasing trend for the first four (TH) to six years (NTH), after which it started decreasing in both hospitals. This finding is somewhat different from the previous study conducted at the orthopaedic departments in the TH and the NTH during the same time period, which demonstrated an overall increasing trend in antibiotic use over 10 years [24]. In the study conducted by Fazaludeen Koya et al., which estimated the private-sector antibiotic consumption in India, the total antibiotic consumption, expressed by DDD/1000 inhabitants-day, decreased by 3.6% between 2011 and 2019 [2]. Furthermore, a study from Western China, also showed a decreasing trend in overall antibiotic use in hospitals between 2013 to 2015, possibly due to the National Antibiotic Stewardship Action Initiative campaign. However, the antibiotic consumption in hospitals in Western China was higher compared to the results of our study [45], keeping in mind that our study focused only on antibiotic use in surgery departments. In addition, when compared to the total antibiotic use in surgery departments in a private hospital in Indonesia (144.2 DDD/100 bed-days) [42], the results of our study showed lower mean antibiotic use in surgery departments in both hospitals. Additionally, adherence to NLEMI was higher in both hospitals (TH, 80%; NTH, 69%) compared to the reported NLEMI adherence in India (43.8%), and compared to the reported NLEMI adherence in Madhya Pradesh (44%) [2]. Furthermore, the proportion of prescribed FDCs was lower in both hospitals (TH, 17%; NTH, 31%) compared to the relatively high proportion of FDC prescriptions in Madhya Pradesh (41.2%), which is among the highest in India [2].

The WHO target of a minimum of 60% Access antibiotics of all prescribed antibiotics [1] was not met in our study. Overall, the TH had 57% of Access antibiotic prescriptions compared to only 34% in the NTH. The proportion of Access antibiotics is higher in the TH and lower in the NTH compared to the reported Access/Watch ratio (0.5) in the private sector in Madhya Pradesh between 2011 and 2019 [2]. In our study, the proportion of Access antibiotics increased in the NTH between 2008 and 2017, while in the TH it did not change significantly. The proportion of Watch antibiotics increased in both hospitals; whereas the proportion of Reserve antibiotics increased slightly in the TH, while in the NTH it did not change significantly. In comparison, when looking into the private sector consumption in India between 2011 and 2019, the reported consumption of Access antibiotics decreased by 13%, of Watch antibiotics increased by 4%, and of Reserve antibiotics increased by 247% [2]. Although direct comparison of levels of change is difficult between the two studies due to the different methods of analysis used, we can observe that the trend of increase in consumption of Watch antibiotics in our study is consistent with the overall trend in India. Furthermore, the proportion of consumed Reserve antibiotics between 2011 and 2019 in India remained approximately 1% [2], which is similar to the proportion of consumed Reserve antibiotics in our study (<1%). It has to be noted, however, that the WHO target of 60% of Access antibiotics refers to the overall antibiotic use at a country level [1], whereas our study setting is limited to two tertiary-care hospitals, and the use of Watch antibiotics is expected to be higher in the tertiary-care settings compared to the primary-care settings.

One of the biggest strengths of the present study is the long duration and a large sample size, which can be considered representative for the surgical patients in Central India. The data collection process was prospective, and the same data collection tool was used in both hospitals, training and monitoring was done by the same person to maintain the uniformity and enable the comparison. The study used the ATC/DDD system, which is the recommended method for the comparison of antibiotic use between healthcare settings. However, this method gives only one DDD per drug, while the recommended dose can differ based on age, type, and severity of diagnosis [32]. Additionally, DDDs can change over time, which can complicate the comparison of trends in antibiotic use over longer periods of time. Furthermore, although the latest versions of essential medicines lists (2022 NLEMI, 2023 WHOMLEM) were not applicable at the time of the study, they were used alongside AWaRe classification for the analysis of antibiotic prescribing to make it more relevant to present-day and comparable with the current recommendations.

One of the limitations of the manual data collection was the possibility of missing and incorrectly entered data. However, the utilization of nurses working in the department, for data collection is a novel concept which can be replicated at other similar settings that do not have computerized data entry system. The data collection form was not designed to collect information about PAP exclusively, therefore, antibiotic prescribed on the day of surgery was considered as PAP for analysis, that can be an underestimation. In addition, information about the culture and susceptibility tests started to be collected from 2011, approximately three years after the beginning of the study, therefore, it is possibly lowballed. Additionally, the diagnoses in this study were analysed at a category level, whereas analysis of the individual diagnoses, as well as more detailed information about the type of wound and surgery, associated patient and surgical risk factors, skin preparation, timing and duration of PAP, and postoperative complications would have been necessary to fully assess the appropriateness of antibiotic prescribing. Therefore, the appropriateness of antibiotic prescribing could only be partially assessed in this study.

Total antibiotic prescribing increased until 2012 in the TH and 2014 in the NTH, after which it started decreasing in both hospitals until 2017. Consumption of Watch antibiotics increased over 10 years in both hospitals. Antibiotic prophylaxis was prescribed to all operated patients, although it is not indicated for every surgery, and the observed most common diagnoses were not high-risk nor infectious. Most prescribed antibiotic prophylaxes were norfloxacin and ceftriaxone, contrary to the internationally recommended PAP standard cefazolin. Antibiotic prescriptions were largely empirical. The results of this study suggest the need to perform culture and susceptibility tests whenever appropriate in order to establish local resistance patterns. There is a pressing need for the development and implementation of antibiotic prescribing and PAP guidelines in surgery based on local resistance patterns, as well as antimicrobial stewardship programs in both hospitals.