This study was conducted between January 1st, 2013 and December 31st, 2014 in a rural hospital with 30 NICU beds which caters to both inborn and out-born babies. The antibiotic policy for neonatal sepsis was initiated on January 1st, 2014. A protocol was developed for empirical therapy of neonatal sepsis based on the review of blood culture susceptibility data obtained from NICU between January and December 2013. Ampicillin and gentamicin were considered as first line antibiotics for community acquired infections (CAIs). Combination of amikacin and ciprofloxacin was considered if there was deterioration on first line antibiotics for CAIs and as first line for healthcare associated infections (HAIs). Meropenem was recommended for empiric therapy only in very severe cases of HAIs. Third generation cephalosporins were recommended only when an intracranial infection was suspected. Empirical treatment choices were adjusted subsequently to narrow spectrum antibiotics whenever possible based on culture and antibiotic susceptibility results.

Standard definitions for "possible sepsis", "probable sepsis" and "confirmed sepsis" were used. The duration of therapy for "possible sepsis" was 48-72 hours; 5-7 days for "probable sepsis"; 10-14 days for "confirmed sepsis" and 21 days for "meningitis" [4].

The primary outcome was to compare overall antibiotic consumption in the year before and after the initiation of antibiotic policy using WHO’s Anatomical Therapeutic Chemical (ATC)/Defined Daily Dose (DDD) per 100 patient-days [5]. DDD is the average maintenance dose per day for a drug used for its main indication in adults [6]. Secondary outcomes included proportion of admitted newborns on any antibiotic, the choice of antibiotics, overall mortality, and sepsis-related mortality. The clinical and laboratory data were collected from a computerized clinical database, case sheets and laboratory records.

Chi-square test and Fisher’s exact test were used for comparing categorical variables and Student t-test was used to compare continuous variables. Interrupted time-series (ITS) analysis was utilized to examine the impact of antibiotic policy change on overall antibiotic consumption. A P-value of <0.05 was considered to be statistically significant. Statistical analysis was performed using STATA 13. The institutional ethics committee at the RDT hospital, Bathalapalli approved the study.

Overall, 1176 newborns were admitted to NICU in the year prior and 1279 newborns in the year after antibiotic policy initiation. There were 7,409 patient-days in the year prior to and 7,748 patient-days in the year after antibiotic policy initiation. There was a significant decrease in the number of VLBW babies in the year after the antibiotic policy initiation (9% vs. 6% P<0.001) (Table I).

Table I	Patient Characteristics and Invasive Procedures in the Year Prior to and after Initiation of Antibiotic Policy

Total antibiotic consumption decreased from 12.47 to 11.47 DDD/100 patient-days before and after the implementation of antibiotic policy (Table II). However, in ITS analysis adjusted by average weight per month there was no statistically significant change neither in the level (P=0.57) nor the slope (P=0.17) of antibiotic consumption (Supplementary material). There was a significant decrease in the consumption of third generation cephalosporins (1.45 vs. 0.45 DDD/100 patient-days) that was confirmed in the ITS analysis (level change -7.6 DDD/100 patient-days, P=0.002).

Table II 	Antibiotic Consumption and Proportion of Babies on Antibiotics Prior to and After Initiation of Antibiotic Policy 

Overall, the proportion of babies on antibiotics decreased significantly (58% (n=681) vs. 46% (n=584); P<0.001) and the proportion of babies on first line antibiotics (ampicillin/gentamicin) increased signi-ficantly (66% (n=449) vs. 84% (n=491); P<0.001) (Table II). There was a significant decrease in the proportion of babies on 3GCs whereas, the proportion of babies on colistin and ciprofloxacin increased signi-ficantly. There were no significant differences in overall mortality (4% (n=50) vs. 3% (n=38); P=0.10) and sepsis-related mortality (3% (n=35) vs. 2% (n=28); P=0.28) for the two years.

With antibiotic policy, the proportion of babies on antibiotics decreased, the proportion of babies on first line antibiotics (ampicillin and gentamicin) increased and the consumption of third generation cephalosporins decreased significantly without affecting NICU mortality. However, there was no significant change in the overall consumption of antibiotics. We observed an increase in the consumption of ciprofloxacin and colistin in the year after the initiation of antibiotic policy. Although, gentamicin was used empirically more often, subsequent change to ciprofloxacin was done based on antibiotic susceptibility results, and thus its consumption increased. The increase in colistin consumption was due to a temporary outbreak of carbapenem-resistant Klebsiella species sepsis in our NICU.

Our study has limitations. Firstly, we do not know if the empiric choice and duration of antibiotics for various sepsis categories have been followed as we lacked individual patient level data. Secondly, we only looked at one year data; it will be important to monitor further to know if the results remain sustained.

In this study, there was no significant change in overall consumption of antibiotics with antibiotic policy, the increased use of first line agents and reduction in use of third generation cephalosporins is encouraging because the quality of antibiotic use determines the prevalence of multi-drug resistance infections. Decreased consumption of broad spectrum antibiotics is associated with reduction in infections due to multi-drug resistant pathogens in NICUs [6]. Thus, our findings indicate the utility of antibiotic policy in the era of high antimicrobial resistance. However, the success of antibiotic policy could be compromised by poor infection control practices. Thus, implementation of strict infection control practices is essential in order to maximize benefits of an antibiotic policy.

Contributors: DRJ: conception or design of the work; DRJ, SG, GA-U: Analysis and Interpretation of Data: All authors are accountable for drafting the work or revising it critically for important intellectual content and final approval of the version to be published.

Funding: None; Competing interest: None stated.