The increasing number of resistant K. pneumoniae strains to multiple antibiotics is a major challenge in medical centers worldwide. Heidary et al. (2018), in a systematic review and meta-analysis article, showed that there is a relatively high prevalence of drug resistant K. pneumoniae isolates in Iran [19]. The highest resistance rate in the K. pneumoniae isolates was observed against ampicillin (82.2%), aztreonam (55.4%), and nitrofurantoin (54.5%), while, in the present study, 52%, 51%, 43%, and 43% of isolates were resistant to SXT, CTX, FEP, and CRO, respectively. Khamesipour et al. (2016) indicated widespread resistance to CRO (41.1%), SXT (36.7%), AN (32.2%), FEP (34.4%) and GM (26.7%) [20]. As well as, in the study of Moghadas et al. (2018), the antibiotic resistance rates were as follows: IPM (7.5%), CIP (16.1%), SXT (32.9%), FEP (34.1%), AN (36.4%), and CAZ (42.7%) [21]. In contrast to our study, 89.5% of isolates were MDR in Hou et al. (2015) study [22]. This percentage was far higher than those reported in our study. This discrepancy may be related to geographic distance, antimicrobial-prescribing patterns in hospitals and level of hygiene. According to the results, in DDST, of 35 IPM-resistant K. pneumoniae isolates, 74.3% were MBL-positive. The prevalence of blaVIM, blaIMP, blaNDM, and blaOXA−48 was 7%, 11%, 5%, and 28%, respectively. In contrast to the present study, Carroll et al. (2013) reported that all isolates were negative for blaVIM, blaIMP, and blaSPM genes [23]. An interesting point in this study was the presence of blaNDM gene. The presence of blaNDM gene in K. pneumoniae was first reported by Shahcheraghi et al. (2010) [24]. According to Fallah et al. (2014), 𝑏𝑙𝑎NDM is an MBL-encoding gene, which was newly recognized and described from New Delhi, India, for the first time followed by other areas such as Pakistan. Neighborhood of these countries with our country and many travels between the countries on the one hand and the ease of resistance transfer among microorganisms on the other led us to postulate that our isolates are likely have the same gene [25]. In the study of Seifi et al. (2016), of 94 K. pneumoniae isolates, 33% formed fully established biofilms, 52.1% were categorized as moderately biofilm-producing, 8.5% formed weak biofilms, and 6.4% were non-biofilm-producers [26]. Li et al. (2012) suggested that the expression of different adhesion, their cognate receptors, and exopolymeric components by individual cell types within a biofilm community can contribute to the general biofilm development. In particular, many bacteria are capable of using a quorum sensing mechanism to regulate biofilm formation and other social activities [27]. In this study, most of the biofilm producer strains were MDR. Our data revealed that 75% of K. pneumoniae were biofilm-producing isolates. These data are similar with the findings of Seifi et al, (2016) [26]. Zheng et al. (2018) found that biofilm formation was more pronounced among magA (K1), aero+, rmpA+, rmpA2+, allS+, wcaG+, and iutA + isolates than in isolates which were negative for these virulence factors [18]. Wu et al. (2011) concluded that treC and sugE affect biofilm formation by modulating capsular polysaccharide (CPS) production [13]. The importance of treC in gastrointestinal tract colonization suggests that biofilm formation contributes to the establishment and persistence of K. pneumoniae infection. In agreement with Seifi et al. (2015) and Boisvert et al. (2016), strong-biofilm producing phenotypes were higher in strains isolated from sputum samples compared to other specimens [26, 28]. This indicates the important role of biofilms in the survival and colonization of microbes in the lungs, causing bacterial resistance to pulmonary clearance. In addition, previous study showed that luxS was shown to be upregulated in biofilm-grown XDR K. pneumoniae strains [12]. Notably, in our study the luxS gene was detected in about 98% of the tested isolates. Using a rat model of middle ear challenge, Yadav et al. (2018) demonstrated that the functional defect in LuxS, leading to the reduced colonization capability of pneumococci in vivo [29]. However, in contrast to our results, Pakhshan et al. [30] concluded that the susceptible isolates to antibiotics tend to form stronger biofilms compared with the resistant strains.