This retrospective study was conducted at JIPMER, a 1591 bedded tertiary care teaching hospital in Pondicherry, South India. The total number of beds included medical services, surgical beds, ICU, obstetrics, ophthalmology, urology and ENT. The data were collected retrospectively from Pharmacy Stores, Medical Records Department (MRD) and Microbiology Department. The total number of inpatients in the months June and December of 2010-13 were obtained from MRD for calculating the DDD per 1000 patients.

The antibiotics issued to the hospital were recorded in the pharmacy stores on a monthly basis. Deliveries to the hospital wards were assumed to reflect usage [10]. In the present study, the consumption data of parenteral antimicrobials in the Anatomic Therapeutic Chemical (ATC) Class J01 were obtained for the months June and December of 2010 to 2013 (Dec 2010 to Jun 2013) and expressed in “Daily Defined Doses” (DDD) measuring unit [11]. The gram amounts of antimicrobials were converted to DDD per 1000 in-patients for the top six frequently issued antibiotics i.e. gentamycin, amikacin, ceftazidime, ceftriaxone, ciprofloxacin and meropenem [10]. The number of DDD was calculated by multiplying the quantity issued by the DDD conversion factor [12].

Antibiotic resistance was determined by Kirby Bauer’s disc diffusion method according to Clinical Laboratory Standards Institute (CLSI) guidelines. Escherichia coli ATCC 25922 was routinely used for the quality control of Kirby Bauer’s disc diffusion method [13].

The antimicrobial resistance data during 2010-2013 were collected for the months for which consumption was noted and their consecutive months, i.e. June-July and December-January of the consecutive year [10].

Correlations between antibiotic consumption and resistance rates or burden were analysed with the Spearman rank correlation using IBM SPSS Statistics 21. All p-value of less than 0.05 were considered statistically significant.

The antimicrobial consumption for the selected parenteral antimicrobials during Dec 2010 - Jun 2013 on a six-month interval was expressed as DDD per 1000 patient as given in [Table/Fig-1]. The data in [Table/Fig-1], show that the consumption of ceftriaxone was highest among all parenteral antimicrobials having a DDD of 113.62 in Dec’10, which decreased drastically to 92.53 with further decrease to 86.41 in Jun’12 and again increased to 125.11. The DDD of gentamycin increased from 19.50 in Dec’10 to 24.29 in Jun’11 and decreased to 12.41 in Dec’12. In the case of amikacin, the consumption decreased from 41.65 to 24.50 in Jun’12 and increased to 45.41 in Jun’13. The consumption of ceftazidime decreased slightly from 17.15 to 12.10 and gradually increased to 15.72 in Dec’12. There was a drastic reduction in consumption of ceftriaxone from 113.62 to 86.41, compared to the mild change in ciprofloxacin consumption from 8.44 in Dec’10 to 13.54 DDD in Dec’12. Meropenem did not show much change in first three intervals and decreased suddenly in June ’12 to 9.64 followed by drastic increase in Dec’12 to 17.96 [Table/Fig-2].

[Table/Fig-2]:

Consumption of drugs in group J01 (Anti-infectives for parenteral use) at JIPMER from 2010 to 2013

[Table/Fig-3] provides the number of isolates processed in the corresponding months and percentage resistance for selected antimicrobial agents against gram-negative pathogens, Escherichia coli and Klebsiella pneumoniae. From the data, it was noted that, the resistance of Escherichia coli for ceftriaxone was highest. The resistance patterns of Escherichia coli and Klebsiella pneumoniae were almost same for most of the antimicrobial drugs, i.e., the resistance decreases in Jun’ 12 and again increases in Dec’ 12. The resistance pattern of Escherichia coli for amikacin showed a steady increase and that of meropenem was fluctuating.

[Table/Fig-4–6] provide correlation between antimicrobial consumption for selected antibiotics and resistance rate for Escherichia coli and Klebsiella pneumoniae during the three-year period. Among Escherichia coli isolates, high correlations were discovered between the use of gentamycin (r= 0.943, p= 0.005) and ciprofloxacin (r= 0.943, p= 0.005) and the resistance for these agents, followed by meropenem (r=0.886, p= 0.019), ceftriaxone (r= 0.841, p= 0.036) and ceftazidime (r=0.771, p= 0.072). The positive slope of the line in [Table/Fig-5] demonstrated that with an increase in antibiotic consumption, the antibiotic resistance for Escherichia coli increases except for amikacin which showed a negative slope ( r= -0.132, p= 0.803).

[Table/Fig-5]:

Correlation of antimicrobial use and incidence of resistance for Escherichia coli the corresponding and next month of antibiotic consumption.

[Table/Fig-6]:

Correlation of antimicrobial use and incidence of resistance for Klebsiella pneumonia the corresponding and next month of antibiotic consumption

Among the Klebsiella pneumoniae isolates, the meropenem (r=0.943, p= 0.005) and gentamycin (r=0.886, p= 0.019) showed high correlations followed by ceftazidime (r= 0.829, p= 0.042), amikacin (r=0.783, p= 0.066), ceftriaxone (r=0.771, p= 0.072) and ciprofloxacin (r=0.543, p= 0.266). [Table/Fig-6] shows a positive slope for all the selected antibiotics, indicating that there was an increase in antibiotic resistance with an increase in antibiotic consumption.

Antimicrobial consumption analysis is a critical task which is based on the parameter used for quantifying [14]. The parameter for drug consumption in our study is DDD, defined by the WHO Collaborating Centre for Drug Statistics Methodology. DDD is most widely used method for measuring the consumption in drug utilization research. However, it is emphasized by WHO that the Defined Daily Dose is a statistical measurement of average dose of drug utilisation per day and does not necessarily represent the recommended therapeutic dose or actual Prescribed Daily Dose. Doses for individual patients and patient groups will often differ from the DDD and will necessarily have to be based on individual characteristics (e.g. age and weight) and pharmacokinetic considerations [11]. A study concluded that, DDD methods are useful for benchmarking purposes and do not imply the number of Days of Therapy (DOTs) or relative use for many antibacterial drugs [14]. Also, many studies demonstrated the use of DDD for measuring the Antibiotic consumption [12,15–18].

According to WHO, “Antimicrobial resistance (AMR) is resistance developed in microorganism, for one or multiple antimicrobial agents, to which it originally was sensitive. Resistant microorganisms are able to combat effect of antimicrobial medicines, making standard treatments ineffective and long persistence of infections, increasing their risk of spread to others.” Infections caused by resistant microorganisms often fail to respond to conventional treatment, resulting in prolonged illness, greater risk of death and higher costs [19]. Many studies have evaluated the relationship between antimicrobial use and resistance in clinical isolates, and different results have been reported [10,20–25]. A study was conducted to determine the correlation between consumption of imipenem and β-Lactam resistance in Pseudomonas aeruginosa and it was found that imipenem utilisation was the major component associated with carbapenem and β-lactam resistance in endemic P. aeruginosa. Also, extensive use of imipenem was linked with a significant increase in their resistance [10]. In the present study also, we observed a significant correlation between the use of meropenem and its resistance rates. Another study was conducted to investigate the relationships between rates of antimicrobial consumption and the prevalence of antimicrobial resistance in Staphylococcus aureus and Pseudomonas aeruginosa isolates. From the data obtained from 47 French hospitals, it was concluded that a statistically significant relationship existed between the rate of fluoroquinolone consumption and the resistance rates among S. aureus and P. aeruginosa isolates [20]. The MYSTIC programme in North America evaluated the antimicrobial usage and resistance relationships for Enterobacteriaceae and P. aeruginosa over a period of three years (1999–2001) in 10–15 medical centres and many significant trends were identified in the prevalence of resistance among P. aeruginosa and Enterobacteriaceae. Firstly, increased use of ciprofloxacin was accompanied by a higher resistance among Enterobacteriaceae, and secondly a correlation existed between resistance rates of ciprofloxacin classes and levels of resistance to other antimicrobial classes in P. aeruginosa [21]. We also observed that the resistance rates of ciprofloxacin showed a significant correlation with its consumption. Similarly, in a recent study from India, a decreasing trend in the resistance rates was seen in P. aeruginosa to ciprofloxacin, ceftazidime, meropenem and imipenem. The authors have suggested that the decline in the use of ciprofloxacin could be the probable reason for the decline in the resistance rate [22]. Consumption of ciprofloxacin is known to upregulate MexEF-OprN efflux system and decrease the levels of outer membrane porin protein OprD, thereby mediating resistance to fluoroquinolones, as well as carbapenems [26,27]. Another study conducted at a tertiary hospital in India stresses the startling increase in resistance and antibiotic use. Further, they suggested that, in Klebsiella pneumoniae, evolution of pan-resistance could be due to the production of carbapenemases and the mechanism of resistance in Escherichia coli is by virtue of ESBL production [28]. Researchers in Finland examined the association between consumption of macrolide and its regional resistance rates in Streptococcus pyogenes during 1997–2001. In order to explain their association, a linear mixed model for repeated measures was used and it was found that a statistically significant association existed between regional resistance rates of erythromycin in S. pyogenes and the consumption of macrolides; however their association with azithromycin use alone was not found [23]. In Spain, researchers assessed the evolution of Streptococcus pneumoniae resistance to penicillin and erythromycin in relation to β-lactam and macrolide consumption over a 19-year period (1979-1997). In this study, a causal relationship could not be established between antibiotic consumption and development of resistance, but the study proposes that overuse of certain specific antibiotics is related to an increase in drug-resistant strains strains of S. pneumonia [24]. Another study was conducted in 20 hospital districts, in Finland, to explore the relationship between the antimicrobial consumption of Out-patients and their resistance in Escherichia coli during 1997-2005. Interestingly, only a few associations were found between antimicrobial consumption and its resistance. Including the association between consumption of fluoroquinolone and its resistance, the majority of the associations studied were not significant [25]. Most of the research discussed above showed a linear relationship between the rate of antimicrobial use and drug resistance.

In our study, only parenteral antibiotic used was analysed and oral ciprofloxacin, which was prescribed at out-patient clinics were not considered which may also account to antibacterial resistance.

Researchers from India reviewed the need to adopt and reinforce an antimicrobial policy to restrain antimicrobial resistance at community and hospital level in India [29]. An antibacterial policy was developed in JIPMER for containment of antimicrobial resistance. The first version of the antibiotic policy was introduced in January 2012 following which there was a decrease in the antibiotic resistance in June 2012. Similarly, the revised version was introduced in January 2013, following which the resistance rates decreased again. Antibiotic prescription policy of the hospital will have a significant impact on bacterial resistance rates. There are studies, supporting the concept that an antibiotic policy which optimises the consumption of various antibiotic classes may influence the microbial sensitivity patterns of hospitals [10,30].

In conclusion, we found that a statistically significant association prevail between consumption of the studied antimicrobials and resistance of Escherichia coli isolates except amikacin and ceftazidime. In the case of Klebsiella pneumonia isolates, gentamycin, ceftazidime and meropenem show statistically significant association between their consumption and resistance. The data from the above study showed a linear trend in the relationship between antimicrobial consumption and resistance exhibited by Escherichia coli and Klebsiella pneumoniae isolates, except Escherichia coli resistance for amikacin. The present study provides an important input in controlling the development of resistant strains of Escherichia coli and Klebsiella pneumoniae to ensure effective treatment and a better perception can be obtained to make necessary changes in antibiotic rotation.