Abstract

Objectives

Multidrug-resistant Pseudomonas aeruginosa and Acinetobacter baumannii are becoming increasingly important nosocomial pathogens worldwide. To study the evolution of non-fermenters in a tertiary care hospital, we undertook a retrospective 10 year (1999–2008) trend analysis of antimicrobial consumption and resistance in non-fermenters causing bacteraemia.

Methods

Antibiotic consumption and resistance were analysed by linear regression. The Pearson correlation coefficient was used for assessing correlation between them.

Results

A total of 69 010 blood cultures were performed, which grew 15 465 isolates (22% positivity rate), of which 1525 isolates (771 isolates of P. aeruginosa and 754 isolates of A. baumannii) were non-fermenters. Overall antibiotic consumption showed an increasing trend, from 158 to 319 defined daily doses (DDDs)/100 bed-days (r2 = 0.62, P = 0.007). The largest relative increase in antibiotic consumption was seen for carbapenems (r2 = 0.68, P = 0.022), followed by β-lactam/inhibitor combinations (r2 = 0.45, P = 0.033), whereas third-generation cephalosporins, fluoroquinolones and aminoglycosides showed no significant changes. A significant increase in resistance in A. baumannii to fluoroquinolones (r2 = 0.63, P = 0.006), aminoglycosides (r2 = 0.63, P = 0.011) and carbapenems (r2 = 0.82, P = 0.013) and in P. aeruginosa to aminoglycosides (r2 = 0.59, P = 0.01) was observed. Carbapenem consumption was associated with the development of resistance in A. baumannii (r = 0.756, P = 0.049), whereas no such association was observed for other antimicrobials among non-fermenters.

Conclusions

Our study highlights the growing problem of high antimicrobial consumption. The increasing prevalence of non-fermenters and the emergence of multidrug-resistant A. baumannii are associated with the consumption of carbapenems. The data cannot prove cause and effect.

Introduction

Hospitals worldwide are continuing to face a crisis in the upsurge and dissemination of antimicrobial-resistant bacteria, particularly non-fermenters causing nosocomial infections.1 The increasing prevalence of multidrug-resistant Pseudomonas aeruginosa and Acinetobacter baumannii (MDRPA and MDRAB) organisms is alarming, as effective antimicrobial options are severely limited.2 Antimicrobial resistance among bloodstream isolates is considered a significant problem worldwide, and inappropriate treatment results in increased mortality rates.3 Although there is evidence of a causal link between antibiotic consumption and resistance,4 it remains a complex issue. Other factors, such as inter-hospital transfer of resistance due to patient transfers, community contribution to resistance, structures of healthcare systems and infection control policies and practices, may also play a role in determining the prevalence of resistance in a hospital. Recently there has been a focus on antimicrobial consumption and resistance patterns to understand the local epidemiology in formulating hospital antibiotic policy. Although regional surveillance data on antimicrobial consumption and resistance trends among bacteria are needed to assess their evolution, these data are lacking from the Indian region. Information and data on trends in antibiotic prescriptions and resistance, although available from other parts of the world,5 are urgently needed in this part of the world to fill this gap.

We present the first report on a 10 year trend analysis of antimicrobial consumption and the development of resistance among non-fermenters in a tertiary care hospital in Delhi, India.

Methods

Setting and timing

Sir Ganga Ram Hospital is a 650-bed, multi-specialty, tertiary care teaching hospital in Delhi, India. It is an active centre for liver, renal and bone marrow transplant surgeries. This hospital also serves as a referral hub for Delhi as well as the whole of India for super-specialty care. The study is a 10 year (1999–2008) retrospective trend analysis of antimicrobial consumption, resistance and their association in non-fermenters (P. aeruginosa and A. baumannii) causing bacteraemia in admitted patients [wards and intensive care units (ICUs)].

Antimicrobial testing

All the blood samples received were cultured by BacT/ALERT 3D (bioMérieux, Marcy l'Étoile, France). The isolates were identified by routine methods and antimicrobial susceptibility was determined by the Kirby Bauer disc diffusion method as per CLSI criteria.6 Manual identification and susceptibility testing were upgraded in 2002 to an automated system using a Vitek identification and MIC testing system (bioMérieux). Non-fermenters were studied for the following antimicrobial groups as per the ATC classification using ABC calc software (available free online):7 J01DD (third-generation cephalosporins; ceftazidime and ceftriaxone); J01CR03 and J01CR05 (β-lactam/inhibitor combinations; piperacillin/tazobactam and ticarcillin/clavulanate); J01MA (fluoroquinolones; ciprofloxacin and levofloxacin); J01GB (aminoglycosides; gentamicin and amikacin); and J01DH (carbapenems; imipenem and meropenem). Tigecycline and colistin were excluded from our trend study, as these were tested from 2008 onwards. Antibiotic resistance data were analysed using a FoxPro-based indigenously designed program until 2006; after that, customized software (Speedminer, Petaling Jaya, Malaysia) was used to extract the data from the hospital information system (Intersystems, Cambridge, MA, USA). Repeat isolates were excluded from the data if the same organism was grown from the same patient in a 15 day interval.

Antimicrobial consumption data

Ten years of antimicrobial consumption data were evaluated from the antibiotic purchasing data from the hospital information system. Hospital purchasing data were used as a proxy for consumption, assuming that the same amount was dispensed to the patients.8 The amount of antimicrobial drug in grams was converted into the number of defined daily doses (DDDs)/100 bed-days using ABC calc.

Statistical analysis

Trend analysis of consumption and resistance for each antimicrobial was carried out using linear regression. The presence or absence of an association with resistance was tested using the Pearson correlation coefficient (r) using SPSS (version 17) (Chicago, IL, USA) software.

Results

Isolation

During the entire study period a total of 69 010 blood cultures were performed, with an increasing trend in the number of blood cultures performed; 4380 cultures per year in 2000 to 8608 cultures per year in 2008. A total of 15 465 isolates (22% positivity rate) grew, of which 1525 isolates (771 isolates of P. aeruginosa and 754 isolates of A. baumannii) were non-fermenters. A. baumannii incidence rose more rapidly (9-fold) as compared with P. aeruginosa (6-fold) (Table 1). There was also a remarkable emergence of non-fermenters causing bacteraemia, increasing from 4% (37 isolates) in 1999 to 28.6% (374 isolates) in 2008.

Table 1.

Trends of isolation rates, antimicrobial consumption and resistance in non-fermenters

 Year
 
   
  1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Rate of change (%) r2 P value 
Isolates/1000 patient-days 
P. aeruginosa 0.16 0.25 0.45 0.27 0.65 0.51 0.67 0.49 0.31 1.02 5.60 0.49 0.024 
A. baumannii 0.15 0.16 0.17 0.29 0.14 0.1 0.81 0.84 0.78 1.39 11.00 0.64 0.005 
Antibiotic consumption (DDDs/100 bed-days) 
 total antibiotics 158 201 211 209 207 223 307 263 212 319 2.40 0.62 0.007 
 3GC (J01DD) 40.4 66.7 43.3 37.5 29.4 31.7 42.8 43.6 41.5 35 −1.10 0.11 0.342 
 aminoglycosides (J01GB) 21.8 25 35.3 18.1 19.6 12.6 21.5 22.1 17.1 28.5 −0.70 0.03 0.654 
 fluoroquinolones (J01MA) 27.6 30 47 58.1 61.1 71.1 91.9 67.4 44.2 47.8 2.90 0.29 0.11 
 β-lactam/inhibitor combinations (J01CR03,05) 15 15.1 20.6 17.5 15.9 16.3 18 26.3 18.8 61.4 4.10 0.45 0.033 
 carbapenems (J01DH)    1.1 5.2 5.7 5.3 8.2 11.6 12.40 0.68 0.022 
P. aeruginosa (% resistance) 
 3GC 42 50 52 75 79 56 86 83 72 42 1.30 0.11 0.358 
 aminoglycosides 30 45 49 58 74 55 80 70 66 65 3.20 0.59 0.01 
 fluoroquinolones 49 54 58 49 32 56 71 61 68 65 1.60 0.24 0.153 
 β-lactam/inhibitor combinations    33 66 43 52 51 49 44 0.70 0.02 0.733 
 carbapenems    42 40 32 57 57 67 55 3.80 0.52 0.068 
A. baumannii (% resistance) 
 3GC 37 63 53 50 42 35 43 88 93 88 3.30 0.38 0.059 
 aminoglycosides 29 13 58 53 59 58 81 90 80 7.70 0.63 0.011 
 fluoroquinolones 32 30 35 42 58 80 94 86 11.80 0.63 0.006 
 β-lactam/inhibitor combinations    51 41 46 31 74 89.5 80 5.30 0.45 0.11 
 carbapenems    10 49 64 84 74 23.80 0.82 0.013 
 Year
 
   
  1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Rate of change (%) r2 P value 
Isolates/1000 patient-days 
P. aeruginosa 0.16 0.25 0.45 0.27 0.65 0.51 0.67 0.49 0.31 1.02 5.60 0.49 0.024 
A. baumannii 0.15 0.16 0.17 0.29 0.14 0.1 0.81 0.84 0.78 1.39 11.00 0.64 0.005 
Antibiotic consumption (DDDs/100 bed-days) 
 total antibiotics 158 201 211 209 207 223 307 263 212 319 2.40 0.62 0.007 
 3GC (J01DD) 40.4 66.7 43.3 37.5 29.4 31.7 42.8 43.6 41.5 35 −1.10 0.11 0.342 
 aminoglycosides (J01GB) 21.8 25 35.3 18.1 19.6 12.6 21.5 22.1 17.1 28.5 −0.70 0.03 0.654 
 fluoroquinolones (J01MA) 27.6 30 47 58.1 61.1 71.1 91.9 67.4 44.2 47.8 2.90 0.29 0.11 
 β-lactam/inhibitor combinations (J01CR03,05) 15 15.1 20.6 17.5 15.9 16.3 18 26.3 18.8 61.4 4.10 0.45 0.033 
 carbapenems (J01DH)    1.1 5.2 5.7 5.3 8.2 11.6 12.40 0.68 0.022 
P. aeruginosa (% resistance) 
 3GC 42 50 52 75 79 56 86 83 72 42 1.30 0.11 0.358 
 aminoglycosides 30 45 49 58 74 55 80 70 66 65 3.20 0.59 0.01 
 fluoroquinolones 49 54 58 49 32 56 71 61 68 65 1.60 0.24 0.153 
 β-lactam/inhibitor combinations    33 66 43 52 51 49 44 0.70 0.02 0.733 
 carbapenems    42 40 32 57 57 67 55 3.80 0.52 0.068 
A. baumannii (% resistance) 
 3GC 37 63 53 50 42 35 43 88 93 88 3.30 0.38 0.059 
 aminoglycosides 29 13 58 53 59 58 81 90 80 7.70 0.63 0.011 
 fluoroquinolones 32 30 35 42 58 80 94 86 11.80 0.63 0.006 
 β-lactam/inhibitor combinations    51 41 46 31 74 89.5 80 5.30 0.45 0.11 
 carbapenems    10 49 64 84 74 23.80 0.82 0.013 

3GC, third-generation cephalosporins.

Antimicrobial consumption

There was a 2-fold increase in the rate of total antimicrobial consumption between 1999 and 2008. Table 1 shows that, in 2008, the β-lactam/inhibitor group was the major antimicrobial consumed of those included, constituting 19.2% of total use. This was followed by fluoroquinolones (15%), third-generation cephalosporins (11%), aminoglycosides (9%) and carbapenems (3.6%). Carbapenems and the antipseudomonal β-lactam/inhibitor combination piperacillin/tazobactam were introduced in our hospital formulary in 2002. Since then, the largest relative increase in consumption (10-fold) was observed for carbapenems, followed by a 4-fold increase in consumption of β-lactam/inhibitor combinations. Less substantial changes in the consumption of other antimicrobials were seen, as shown in Table 1. We observed a significant increasing trend in fluoroquinolone consumption from 1999 to 2005, peaking at 91.9 DDDs/100 bed-days (r2 = 0.954, P < 0.001), and subsequently a decreasing trend (r2 = 0.586, P = 0.445).

Antimicrobial resistance

Table 1 shows a high rate of resistance in A. baumannii in 2008 to third-generation cephalosporins (88%), fluoroquinolones (86%), aminoglycosides (80%), β-lactam/inhibitor combinations (80%) and carbapenems (74%). A significant and dramatic increase in resistance to aminoglycosides, fluoroquinolones and carbapenems in A. baumannii was seen over the 10 years of this study. A significant but less dramatic increasing resistance trend was seen in P. aeruginosa to aminoglycosides. Likewise the resistance rates in P. aeruginosa to third-generation cephalosporins, fluoroquinolones, aminoglycosides, β-lactam/inhibitor combinations and carbapenems were 42%, 65%, 65%, 44% and 55%, respectively, in 2008, which were lower compared with A. baumannii. Overall, Table 1 highlights that trends of increasing resistance were much stronger for A. baumannii compared with P. aeruginosa across all classes of antimicrobials. Colistin and tigecycline susceptibility tests were routinely performed from 2008 onwards. In 2008 colistin resistance was less than 1% in A. baumannii and P. aeruginosa, whereas tigecycline showed a high resistance rate of 70% in A. baumannii.

Correlation

We observed a significant correlation between total antimicrobial use and the isolation rate of P. aeruginosa (r = 0.838, P = 0.003) and A. baumannii (r = 0.838, P = 0.02) (Figure 1). A significant but weak association was observed for carbapenem consumption and resistance in A. baumannii, whereas no association was observed for carbapenem consumption and resistance in P. aeruginosa (Table 2). Similarly no association was seen between antimicrobial consumption and the development of resistance for other classes of drugs in our study.

Table 2.

Pearson correlation coefficient (r) between antibiotic consumption and resistance

 P. aeruginosa
 
A. baumannii
 
Antibiotic correlation coefficient (rP value correlation coefficient (rP value 
Third-generation cephalosporins −0.182 0.615 0.213 0.554 
Aminoglycosides −0.197 0.585 −0.397 0.256 
Fluoroquinolones 0.241 0.502 0.291 0.414 
β-Lactam/inhibitor combinations −0.183 0.694 0.516 0.236 
Carbapenems 0.458 0.301 0.756 0.049 
 P. aeruginosa
 
A. baumannii
 
Antibiotic correlation coefficient (rP value correlation coefficient (rP value 
Third-generation cephalosporins −0.182 0.615 0.213 0.554 
Aminoglycosides −0.197 0.585 −0.397 0.256 
Fluoroquinolones 0.241 0.502 0.291 0.414 
β-Lactam/inhibitor combinations −0.183 0.694 0.516 0.236 
Carbapenems 0.458 0.301 0.756 0.049 
Figure 1.

Association between antimicrobial consumption and isolation rate of non-fermenters.

Figure 1.

Association between antimicrobial consumption and isolation rate of non-fermenters.

Discussion

Non-fermenters are emerging as important opportunistic nosocomial pathogens in our institute. The increase in the incidence of non-fermenters was significantly associated with total antimicrobial consumption, indicating that there is selecting out of MDRAB and MDRPA under antimicrobial pressure. The increase may also be explained by the increasing number of blood cultures being performed over the years and thus an increase in the isolation rate. The increasing number of blood cultures drawn may also reflect periods of increasing antibiotic consumption, as shown by Lamoth et al.9 Similar trends of increasing prevalence of nosocomial non-fermenters have been noted in other surveillance studies as well.1,10

Despite its importance, there is very little available information about hospital antimicrobial consumption in India. This may be because of a lack of systems that allow assessment of antibiotic consumption. During our 10 year analysis we observed a significant increase in the rate of drug consumption, which doubled from 157.8 DDDs/100 bed-days to 318.5 DDDs/100 bed-days. These figures are much higher than the 108.5 DDDs/100 bed-days reported by Shankar et al.11 in neighbouring Nepal. In another large survey of 25 hospitals in the Mediterranean region by Borg et al.,12 antimicrobial consumption varied greatly, from 84 to 428 DDDs/100 bed-days, with a median consumption of 112 DDDs/100 bed-days. One limitation in the measurement of antibiotic consumption in DDDs/100 bed-days is its inability to adjust antibiotic use according to variations in the case mix over time. The high antibiotic consumption rate in our hospital is possibly due to a higher case mix index (CMI). CMI is an economic parameter that is calculated using diagnosis-related groups, a measure that is today routinely used in various countries as a basis for hospital reimbursement.13 Kuster et al.14 found a significant correlation between CMI and antibiotic use when calculated in DDDs/100 bed-days. They also showed that the antibiotic use varied from 21 in the rheumatology unit, with a lower CMI, to 323 DDD/100 patient-days in the transplantation unit, with a higher CMI.14 Since Sir Ganga Ram Hospital is an active centre for liver, renal and bone narrow transplant surgeries, it is possible that the increase in antibiotic consumption is related more to the higher CMI in our hospital compared with other hospitals. The other reason for high consumption of antimicrobials in our setting may be because of high antimicrobial resistance and thus the tendency to treat infection with multiple antimicrobials. Also, Sir Ganga Ram Hospital, being a tertiary care referral centre, performs a large number of complicated referred surgeries. A total of 23 874 total surgeries performed in 2008 is a case in point. We have many post-surgical cases where there is a general tendency to prescribe multiple antimicrobials, often in the absence of microbiological data. No significant trends in antibiotic consumption were seen for aminoglycosides, third-generation cephalosporins and fluoroquinolones. This reflects the emergence of MDR strains, and consequently fewer prescriptions of these first-line antibiotics. Since combination therapy of β-lactams and fluoroquinolones/aminoglycosides is recommended for treatment of nosocomial non-fermenters,15 the increasing prevalence of MDR in non-fermenters has led to a switch to second-line antibiotics, such as β-lactam/inhibitor combinations and carbapenems, in our hospital, which is also reflected in the significant increase in the consumption of these antimicrobials. Similar trends of third-generation cephalosporins, β-lactam/inhibitor combinations and carbapenems were also seen in a 10 year (1997–2007) MYSTIC surveillance programme.16 Although increasing consumption trends for β-lactam/inhibitor combinations and carbapenems were not reported in surveillance studies by other authors,17,18 this was probably due to their short study period or restrictions on antibiotic prescriptions during the study period, which was not the case in our study.

The increasing trend in resistance to fluoroquinolones, β-lactam/inhibitor combinations and aminoglycosides among non-fermenters may have been the result of high fluoroquinolone consumption until 2005, as this is known to be a risk factor in the emergence of MDR bacteria.17 Although fluoroquinolone use declined after 2005, the increased use of broad-spectrum antibiotics, β-lactam/inhibitor combinations and carbapenems contributed to the emergence of MDR strains.

Our results show there was rapid emergence of carbapenem resistance in A. baumannii compared with P. aeruginosa. Similar trends of resistance in non-fermenters causing bacteraemia among non-fermenters were observed by Livermore et al.19 in the UK in a 6 year survey from 2001 to 2006. Falgas et al.,2 in Greece, also showed a significant increase in resistance in A. baumannii (44.4% in 2002 to 100% in 2005; P < 0.001) and no significant change in P. aeruginosa (62.5% in 2002 to 55.5% in 2005; P = 0.75), but the number of isolates was relatively small in this study. Resistance rates of 3%–82% to aminoglycosides, 19%–74% to fluoroquinolones, 8% to 34% to β-lactam/inhibitor combinations and 17%–74% to carbapenems among non-fermenters have been reported worldwide,4,10,17–20 indicating wide variation worldwide. Scarce published literature from Southeast Asia also show a high prevalence of carbapenem resistance (36%–90%) among non-fermenters.21,22 Additionally, to analyse the cause of carbapenem resistance among non-fermenters in 2008, we performed phenotypic testing for metallo-β-lactamase (MBL) by Etest MBL strips23 on 50 consecutive blood isolates each of imipenem-resistant P. aeruginosa and imipenem-resistant A. baumannii. Although MBLs in P. aeruginosa are the common cause of carbapenem hydrolysis,24 OXA carbapenemases are considered more common in A. baumannii.19 In contrast, we observed that MBLs were the main mechanism of carbapenem resistance in both P. aeruginosa (92.8%) and A. baumannii (88%). Similar emergence of MBLs as a prominent cause of carbapenem resistance (24.5%) was seen in Enterobacteriaceae in our hospital, which was 35.8% in 2008. Recently there have been reports of the emergence and rapid spread of a novel MBL designated as New Delhi MBL-1 (NDM-1) in Indian hospitals.25 This increase in MBLs in Enterobacteriaceae can also be explained by the rapid spread of NDM-1 in our hospital, as few of our MBL-producing isolates screened for NDM-1 were positive by PCR. As A. baumannii isolates in Indian hospitals have also been found to carry NDM-1 carbapenemase genes giving very broad-spectrum antibiotic resistance profiles,26 we fear that the emergence of MBLs in A. baumannii is due to NDM-1, although this needs to be confirmed by a genotypic method. As more than 70% of non-fermenters in our hospital are resistant to all antimicrobials, carbapenems and colistin are often administered presumptively in the treatment of suspected bacteraemia among critically ill patients. Even after the availability of susceptibility reports, colistin remains the only therapeutic option in the majority of our patients, although high-quality pharmacokinetic data and clinical outcome studies are lacking for polymyxin therapy.27 Tigecycline, introduced in 2006 in our hospital for MDR bacteria, is already showing resistance of up to 70% among A. baumannii isolates in ICUs. These trends worldwide indicate that carbapenem-resistant A. baumannii is emerging as a major nosocomial pathogen. Although MDR Acinetobacter infections are of low virulence, they are associated with high crude mortality rates (26%–68%)28–30 since they usually occur in critically ill patients in ICUs with few therapeutic options.

There are varying reports in the literature both supporting and refuting the relationships between antibiotic consumption and resistance,4,17,18 suggesting local factors play an important role in such relationships. In our study we did not observe any association between antimicrobial consumption and development of resistance except for a weak association for carbapenems in A. baumannii (Table 2). The reason for the lack of correlation observed may be explained by the fact that resistance to one antibiotic is associated with cross-resistance to other antimicrobials.17 Lepper et al.18 demonstrated in their study that imipenem therapy was associated with resistance to piperacillin/tazobactam and ceftazidime in P. aeruginosa, whereas Messadi et al.17 showed a significant correlation between ciprofloxacin consumption and resistance to imipenem (r = 0.89, P = 0.043) in P. aeruginosa. Further imipenem resistance being integron borne imparts co-resistance to other antibiotics resulting in MDR organisms containing MDR gene cassettes.20 Therefore, in institutions where there is high carbapenem consumption and resistance, it predisposes to cross-resistance to all classes of drugs and thus masks the effects of antibiotic use and resistance for other classes of drugs, as also seen in our study.

We observed that the development of carbapenem resistance in A. baumannii was associated with carbapenem use (Figure 2), which is in accordance with other studies.4,17 The weak association (P = 0.049) may be explained by the fact that, other than for antimicrobial use, horizontal spread of A. baumannii because of its extended survival time has been shown to result in persistence and spread of endemic MDRAB infections in healthcare centres.27 These two reasons may also explain the preferential emergence of MDRAB over MDRPA during the 10 years of our study period. In addition, importation of de novo MDR bacteria from other institutes and their maintenance in susceptible hosts may also be contributing to the emergence of carbapenem-resistant non-fermenters in our hospital. To that end, we often receive critical patients from other hospitals in whom carbapenem-resistant non-fermenters are isolated within 48 h of admission.

Figure 2.

Association between carbapenem consumption and resistance in P. aeruginosa and A. baumannii.

Figure 2.

Association between carbapenem consumption and resistance in P. aeruginosa and A. baumannii.

Since the results of our study show alarming trends of increasing resistance and antimicrobial use, which greatly threaten the utility of our existing antimicrobial armamentarium, immediate remedial actions need to be implemented to salvage the situation. Although we have an active hospital infection committee and an antibiotic policy in place, more innovative measures need to be explored. Towards that aim, we are undertaking an intervention initiative where doctors get monthly feedback of their qualitative and quantitative antimicrobial prescriptions compared with their peer group and other specialties, as well as a monthly update on hospital antibiogram and antibiotic policy, to observe if this continuous feedback results in increased compliance to hospital antibiotic policy and more prudent use of antimicrobials. Restricted use of last-resort antibiotics, like imipenem, by written approval from an infectious disease specialist or by a computer approval system have also shown to result in an improved rate of appropriate empirical antibiotic treatment and reduction in antibiotic treatment costs.18,31

Enhanced environmental cleaning and improved hand hygiene are practical and effective measures to prevent the horizontal spread of non-fermenters.27 Special dedicated staffing and physical separation of high antibiotic use areas, such as oncology and liver transplant units from cardiology or medicine units, can be explored in referral hospitals like ours. This policy is currently being implemented for a paediatric ICU in our hospital with encouraging preliminary results.

We would like to address a few limitations of our study. Establishing an aetiological relationship between the rate of antimicrobial consumption and the prevalence of resistance may require other types of complex methodology with more timepoints.32 Our study, with yearly aggregation of timepoints, may not be sensitive enough to reflect subtle changes in the complex interaction between antimicrobial drug prescribing and resistance. Furthermore, as this study was retrospective and uncontrolled in nature, potential confounders such as changes in the length of stay and case matching, staffing level and hand-hygiene compliance could not be factored while observing the trends. We also could not perform genotyping on the isolates to ascertain the clonality to help in interventional policies.33

In conclusion, the results of this trend analysis are important, as this is the first of its kind from India documenting the trends of non-fermenters in bacteraemia over 10 years on a large number of isolates. Our study highlights the problems of high rates of antimicrobial consumption and the emergence of MDRAB due to carbapenem consumption as a prominent cause of bacteraemia. More studies are needed from other parts of India to determine the magnitude of the problem and formulate infection control guidelines relevant to our region.

Funding

No specific funding.

Transparency declarations

None to declare.

Acknowledgements

We thank Dr Karanvir Singh for assisting in the software development for obtaining antimicrobial consumption and resistance data and Parul Takkar for assisting in statistical evaluation.

References

1
Friedland
I
Stinson
L
Ikaiddi
M
, et al.  . 
Phenotypic antimicrobial resistance patterns in Pseudomonas aeruginosa and Acinetobacter: results of a multicenter intensive care unit surveillance study, 1995–2000
Diagn Microbiol Infect Dis
 , 
2003
, vol. 
45
 (pg. 
245
-
50
)
2
Falagas
ME
Kasiakou
SK
Nikita
D
, et al.  . 
Secular trends of antimicrobial resistance of blood isolates in a newly founded Greek hospital
BMC Infect Dis
 , 
2006
, vol. 
6
 pg. 
99
 
3
MacGowan
AP
Clinical implications of antimicrobial resistance for therapy
J Antimicrob Chemother
 , 
2008
, vol. 
62
 
Suppl 2
(pg. 
ii105
-
14
)
4
Iosifidis
E
Antachopoulos
C
Tsivitanidou
M
, et al.  . 
Differential correlation between rates of antimicrobial drug consumption and prevalence of antimicrobial resistance in a tertiary care hospital in Greece
Infect Control Hosp Epidemiol
 , 
2008
, vol. 
29
 (pg. 
615
-
22
)
5
White
AR
The British Society for Antimicrobial Chemotherapy Resistance Surveillance Project: a successful collaborative model
J Antimicrob Chemother
 , 
2008
, vol. 
62
 
Suppl 2
(pg. 
ii3
-
14
)
6
Clinical and Laboratory Standards Institute
Performance Standards for Antimicrobial Susceptibility Testing: Fifteenth Informational Supplement M100-S15
 , 
2005
Wayne, PA, USA
CLSI
7
European Society of Clinical Microbiology and Infectious Diseases (ESCMID)
 
8
Wattal
C
Raveendran
R
Kotwani
A
, et al.  . 
Establishing a new methodology for monitoring of antimicrobial resistance and use in the community in a resource poor setting
J Appl Ther Res
 , 
2009
, vol. 
7
 (pg. 
37
-
45
)
9
Lamoth
F
Francioli
P
Zanetti
G
Blood samples drawn for culture can serve as a surrogate marker for case-mix adjustment of hospital antibiotic use
Clin Microbiol Infect
 , 
2007
, vol. 
13
 (pg. 
454
-
6
)
10
Karlowsky
JA
Draghi
DC
Jones
ME
, et al.  . 
Surveillance for antimicrobial susceptibility among clinical isolates of Pseudomonas aeruginosa and Acinetobacter baumannii from hospitalized patients in the United States, 1998 to 2001
Antimicrob Agents Chemother
 , 
2003
, vol. 
47
 (pg. 
1681
-
8
)
11
Shankar
RP
Partha
P
Shenoy
NK
, et al.  . 
Prescribing patterns of antibiotics and sensitivity patterns of common microorganisms in the Internal Medicine ward of a teaching hospital in Western Nepal: a prospective study
Ann Clin Microbiol Antimicrob
 , 
2003
, vol. 
2
 pg. 
7
 
12
Borg
MA
Zarb
P
Ferech
M
, et al.  . 
Antibiotic consumption in southern and eastern Mediterranean hospitals: results from the ARMed project
J Antimicrob Chemother
 , 
2008
, vol. 
62
 (pg. 
830
-
6
)
13
Roger
FH
Case mix use in 25 countries: a migration success but international comparisons failure
Int J Med Inform
 , 
2003
, vol. 
70
 (pg. 
215
-
9
)
14
Kuster
SP
Ruef
C
Bollinger
AK
, et al.  . 
Correlation between case mix index and antibiotic use in hospitals
J Antimicrob Chemother
 , 
2008
, vol. 
62
 (pg. 
837
-
42
)
15
Lister
PD
Wolter
DJ
Hanson
ND
Antibacterial-resistant Pseudomonas aeruginosa: clinical impact and complex regulation of chromosomally encoded resistance mechanisms
Clin Microbiol Rev
 , 
2009
, vol. 
22
 (pg. 
582
-
610
)
16
Patzer
JA
Dzierzanowska
D
Turner
PJ
Trends in antimicrobial susceptibility of Gram-negative isolates from a paediatric intensive care unit in Warsaw: results from the MYSTIC programme (1997–2007)
J Antimicrob Chemother
 , 
2008
, vol. 
62
 (pg. 
369
-
75
)
17
Messadi
AA
Lamia
T
Kamel
B
, et al.  . 
Association between antibiotic use and changes in susceptibility patterns of Pseudomonas aeruginosa in an intensive care burn unit: a 5-year study, 2000–2004
Burns
 , 
2008
, vol. 
34
 (pg. 
1098
-
102
)
18
Lepper
PM
Grusa
E
Reichl
H
, et al.  . 
Consumption of imipenem correlates with β-lactam resistance in Pseudomonas aeruginosa
Antimicrob Agents Chemother
 , 
2002
, vol. 
46
 (pg. 
2920
-
5
)
19
Livermore
DM
Hope
R
Brick
G
, et al.  . 
Non-susceptibility trends among Pseudomonas aeruginosa and other non-fermentative Gram-negative bacteria from bacteraemias in the UK and Ireland, 2001–06
J Antimicrob Chemother
 , 
2008
, vol. 
62
 
Suppl 2
(pg. 
ii55
-
63
)
20
Perez
F
Hujer
AM
Hujer
KH
, et al.  . 
Global challenge of multidrug-resistant Acinetobacter baumannii
Antimicrob Agents Chemother
 , 
2007
, vol. 
51
 (pg. 
3471
-
84
)
21
Taneja
N
Maharwal
S
Sharma
M
Imipenem resistance in non-fermenters causing nosocomial urinary tract infections
Indian J Med Sci
 , 
2003
, vol. 
57
 (pg. 
294
-
9
)
22
Irfan
S
Zafar
A
Guhar
D
, et al.  . 
Metallo-β-lactamase-producing clinical isolates of Acinetobacter species and Pseudomonas aeruginosa from intensive care unit patients of a tertiary care hospital
Indian J Med Microbiol
 , 
2008
, vol. 
26
 (pg. 
243
-
5
)
23
Walsh
TR
Bolmstrom
A
Qwarnstrom
A
, et al.  . 
Evaluation of a new Etest for detecting metallo-β-lactamases in routine clinical testing
J Clin Microbiol
 , 
2002
, vol. 
40
 (pg. 
2755
-
9
)
24
Poirel
L
Nordmann
P
Carbapenem resistance in Acinetobacter baumannii: mechanisms and epidemiology
Clin Microbiol Infect
 , 
2006
, vol. 
12
 (pg. 
826
-
36
)
25
Kumarasamy
KK
Toleman
MA
Walsh
TR
, et al.  . 
Emergence of a new antibiotic resistance mechanism in India, Pakistan, and the UK: a molecular, biological, and epidemiological study
Lancet Infect Dis
 , 
2010
, vol. 
10
 (pg. 
597
-
602
)
26
Kumarasamy
KK
Thirunarayan
MA
Krishnan
P
Coexistence of blaOXA-23 with blaNDM-1 and armA in clinical isolates of Acinetobacter baumannii from India
J Antimicrob Chemother
 , 
2010
, vol. 
65
 (pg. 
2253
-
70
)
27
Maragakis
LL
Perl
TM
Acinetobacter baumannii: epidemiology, antimicrobial resistance, and treatment options
Clin Infect Dis
 , 
2008
, vol. 
46
 (pg. 
1254
-
63
)
28
Falagas
ME
Kopterides
P
Siempos
II
Attributable mortality of Acinetobacter baumannii infection among critically ill patients
Clin Infect Dis
 , 
2006
, vol. 
43
 (pg. 
389
-
90
)
29
Lee
NY
Lee
HC
Ko
NY
, et al.  . 
Clinical and economic impact of multi-drug resistance in nosocomial Acinetobacter baumannii bacteraemia
Infect Control Hosp Epidemiol
 , 
2007
, vol. 
28
 (pg. 
713
-
9
)
30
Abbo
A
Carmeli
Y
Navon-Venezia
S
, et al.  . 
Impact of multi-drug-resistant Acinetobacter baumannii on clinical outcomes
Eur J Clin Microbiol Infect Dis
 , 
2007
, vol. 
26
 (pg. 
793
-
800
)
31
Paul
M
Andreassen
S
Tacconelli
E
, et al.  . 
Improving empirical antibiotic treatment using TREAT, a computerized decision support system: cluster randomized trial
J Antimicrob Chemother
 , 
2006
, vol. 
58
 (pg. 
1238
-
45
)
32
Shardell
M
Harris
AD
El-Kamary
SS
, et al.  . 
Statistical analysis and application of quasi experiments to antimicrobial resistance intervention studies
Clin Infect Dis
 , 
2007
, vol. 
45
 (pg. 
901
-
7
)
33
Jonas
D
Meyer
E
Schwab
F
, et al.  . 
Genodiversity of resistant Pseudomonas aeruginosa isolates in relation to antimicrobial usage density and resistance rates in intensive care units
Infect Control Hosp Epidemiol
 , 
2008
, vol. 
29
 (pg. 
350
-
7
)