Triple Combination Therapy with High-Dose Ampicillin/Sulbactam

Le Infezioni in Medicina, n. 1, 11-16, 2019 ORIGINAL ARTICLE 11

Triple combination therapy with high-dose ampicillin/sulbactam, high-dose tigecycline and colistin in the treatment of ventilator-associated pneumonia caused by pan-drug resistant Acinetobacter baumannii: a case series study

Stelios F. Assimakopoulos1, Vassilis Karamouzos2, Aikaterini Lefkaditi2, Christina Sklavou2, Fevronia Kolonitsiou3, Mirto Christofidou3, Fotini Fligou2, Charalambos Gogos1, Markos Marangos1 1Department of Internal Medicine, Division of Infectious Diseases, University of Patras Medical School, Patras, Greece; 2Department of Anaesthesiology and Intensive Care Medicine, University of Patras Medical School, Patras, Greece; 3Department of Microbiology, University of Patras Medical School, Patras, Greece

SUMMARY Acinetobacter baumannii has evolved in recent decades cline MICs > 2μg/mL. Nine patients (90%) exhibited a as a major problem in carbapenem-resistant gram-nega- successful clinical outcome, accompanied by microbio- tive nosocomial infections, associated with high mortal- logical eradication in seven of them. All clinically cured ity rates especially in intensive care units (ICUs). Recent patients survived at 14 and 28 days. Acute kidney injury reports highlight the increasing prevalence of resistance (AKI) was observed in one patient. In view of the in- to colistin, a last resort therapeutic option for carbapen- creasing prevalence of pan-drug resistant A. baumannii em-resistant A. baumannii. We retrospectively evaluated infections in ICUs, its associated high rates of mortality the potential efficacy, in terms of clinical and microbio- and the lack of effective treatment options, we feel that logical cure and mortality, of a combination of intrave- there is an emerging need for our results to be further nous colistin and high-dose ampicillin/sulbactam and validated in larger prospective studies. high-dose tigecycline, concurrently administered with inhaled colistin, in 10 ICU patients with ventilator-as- Keywords: Acinetobacter baumannii, ventilator-associat- sociated pneumonia (VAP) caused by carbapenem- and ed pneumonia, pan-drug resistant, colistin, tigecycline, colistin-resistant A. baumannii strains, with high tigecy- sulbactam

n INTRODUCTION resistant gram-negative nosocomial infections, especially in the intensive care unit (ICU). Its intrin- cinetobacter baumannii has evolved over the sic antimicrobial resistance together with its abili- Alast decades as a major problem in multidrug ty to easily adopt new resistance mechanisms has driven the evolution of extensively drug-resistant (XDR) and even pandrug-resistant (PDR) isolates Corresponding author [1]. The most frequent healthcare-associated in- Stelios F. Assimakopoulos fection caused by A. baumanii is ventilator-associ- E-mail: [email protected] ated pneumonia (VAP). For carbapenemase-pro- 12 S.F. Assimakopoulos, et al.

ducing multi drug-resistant (MDR) A. baumanii failures and mortality were observed either when strains, colistin represents the cornerstone of an- tigecycline was used as monotherapy or as part of timicrobial therapy, while tigecycline represents a combination therapy with colistin [4, 5]. From an additional option. However, the wide use of this point of view, the present study included pa- colistin for the treatment of MDR Gram negative tients with VAP caused by real PDR A. baumannii nosocomial infections have forced the increasing strains. Sulbactam alone has been found to have rates of resistance to this last resort antibiotic. In a intrinsic activity against Acinetobacter spp., but the recently published Southern Europe multicentre only available preparation of sulbactam in Greece study (Greece, Italy and Spain) on the molecular is within a fixed combination of ampicillin/sul- epidemiology, and antimicrobial susceptibility of bactam. However, it has been previously demon- A. baumannii isolates from respiratory tract sam- strated that the activity of ampicillin/sulbactam ples of patients with VAP, carbapenem resistance against Acinetobacter spp. is exclusively due to sul- was almost universal, while colistin resistance bactam [6]. raised to 47.7% [1]. Over the course of this study, Patients were treated with intravenous colistin 9 colistin resistance rates increased from 32% in million IU loading dose, followed after 12 h by 2012-13, to 62% in 2014-15. Of note, 34% of isolat- maintenance dose of 4.5 million IU every 12 hours ed strains were XDR and 31% PDR highlighting and high-dose intravenous tigecycline 200 mg the difficulty in treating these patients [1]. loading dose followed after 12 h by 100 mg every We present herein, our tertiary centre experience 12 hours and high dose intravenous ampicillin/ with ten patients hospitalized in the intensive sulbactam 9 g (6+3) every 8 hours, while inhaled care unit (ICU), suffering from VAP caused by colistin 2 million IU every 8 hours was concur- PDR A. baumannii strains, defined as carbapen- rently administered. In patients with a calculat- em- and colistin-resistant with high tigecycline ed creatinine clearance of less than 60 ml/min, MICs >2μg/mL. The reported patients received a according to the Cockcroft-Gault formula, the combination treatment consisting of intravenous maintenance dose of colistin was reduced based colistin and high-dose ampicillin/sulbactam and on a modified formula from Garonzik et al. (daily

high-dose tigecycline, concurrently with inhaled maintenance colistin dose [IU] = CLCR/10 + 2) [7, colistin. 8]. Ampicillin/sulbactam dose was also reduced according to the creatinine clearance value (<60

mL/min); for CLCR 31-60 mL/min the dose was n PATIENTS AND METHODS reduced by 25% without changes in interval dos-

ing, while for CLCR 7-30 mL/min the dose was re- We retrospectively evaluated the potential effica- duced by 50% and administered twice daily [9]. cy, in terms of clinical and microbiological cure The diagnostic criteria of VAP were in accord- and mortality, of a combination of intravenous ance with current clinical practice guidelines set colistin and high-dose ampicillin/sulbactam and by the Infectious Diseases Society of America and high-dose tigecycline, concurrently administered the American Thoracic Society: (a) mechanical with inhaled colistin, in 10 ICU patients with VAP ventilation >48 h; (b) satisfied two of the follow- caused by carbapenem- and colistin-resistant A. ing: body temperature >38 °C or < 36 °C, leuko- baumannii strains, with high tigecycline MIC >2 penia or leukocytosis, or purulent secretions; (c) μg/mL. PDR phenotype was defined according new or progressive chest infiltrates, for patients to the international expert proposal for Interim with underlying pulmonary or cardiac disease, standards guidelines [2]. Accordingly, PDR A. bau- two serial chest radiographs were required for mannii isolates were non-susceptible to all agents assessment; (d) a quantitative culture of respira- in all antimicrobial categories, except for tigecy- tory sample taken non-invasively (endotracheal cline, where in the latest EUCAST clinical break- aspiration) yielding ≥106 CFU/mL [10]. A. bau- points (v. 8.1, May 15, 2018) no clinical breakpoint mannii strains isolated from respiratory cultures is set due to insufficient evidence [3]. However, deriving from ICU patients were identified using previous clinical studies with XDR A. bauman- Vitek 2 Advanced Expert System (bioMérieux, nii infections, have shown that when tigecycline Marcy l’Étoile, France). Antibiotic susceptibil- MIC is >2 μg/ml, significantly higher therapeutic ity was performed by the agar disk diffusion Triple combination therapy PDR A. baumannii VAP 13

method (Kirby-Bauer) for all antibiotics includ- No No No No No No No No No Yes

ing carbapenems, while MIC to tigecycline was of Tx determined by Etest (AB Biodisk) and MIC to AKI end colistin by broth microdilution method. Results No No No No No No No No No Yes were interpreted according to EUCAST guide- AKI 7d No No No No No No No No No Yes 48h lines (EUCAST, 2018). Blood cultures (two sets AKI from peripheral veins) were taken concomitant to tracheal aspirates at clinical suspicion of VAP No Yes Yes Yes Yes Yes Yes Yes Yes Yes 28 d and thereafter as clinically indicated (presence of survival fever ≥38.0 °C, clinical suspicion of bloodstream No Yes Yes Yes Yes Yes Yes Yes Yes Yes infection). Patients were considered immuno- 14 d survival compromised if they had received a transplant, had a splenectomy, a gamma globulin deficiency, No Yes Yes Yes Yes Yes Yes Yes Yes Yes success an HIV infection, were currently under anti-can- Clinical cer chemotherapy, or received at least 20 mg of prednisone/day for ≥30 days. The Charlson comorbidity index determined the extent of comor- No No Yes Yes Yes Yes Yes Yes Yes bid illnesses [11]. A successful clinical outcome success Microbiological Microbiological Superinfection was defined as the resolution of symptoms and KPC Klebsiella signs of infection and improvement of relevant 7 11 15 14 12 10 16 17 14 laboratory data at the end of therapy. Microbio- 14 duration logical success was defined as a follow up nega- Treatment tive culture for A. baumannii at days 7 or 14. Mor- tality was evaluated at days 14 and 28. Patients with culture-proven polymicrobial infections linezolid linezolid linezolid linezolid linezolid linezolid linezolid linezolid antibiotics ceftaroline were excluded. Patients receiving concurrent vancomycin empiric coverage for MRSA were not excluded Co- administered from our analysis. Therapeutic failure was con- 4 3 3 4 4 3 3 3 3 sidered if clinical findings worsened or persisted 12 Tigecycline without improvement after 4-7 days of treat- MIC (μg/ml) ment. VAP-related death was considered when persistent symptoms and signs were present at 27 23 25 28 20 23 20 22 18 25 the time of death without other definite causes. score (admission) Acute kidney injury (AKI) was defined according II APACHE to the RIFLE criteria; increase of serum creatinine VAP VAP VAP VAP VAP VAP VAP VAP of at least 50% from baseline (defined as Risk), VAP Type of Type Infection doubled serum creatinine level from the baseline + BSI VAP (defined as Injury), or three times increase in se- in 11 14 19 34 79 14 39 90 16 21 ICU rum creatinine (defined as Failure) [12]. Days 0 0 4 6 0 0 0 3 5 11 index n Charlson

PATIENTS AND RESULTS comorbidity

The present study included ten patients with PDR No No No No No No No No No A. baumannii VAP, while one patient additionally No Immuno- presented VAP-associated secondary bacterae- compromised 38 50 62 71 49 47 51 60 85 mia. Patients had a mean age of 56.4±13.5 years, 51 Age no one patient was immunocompromised, half of F F M M M M M M M M them had a Charlson comorbidity index ≥3 and Sex 1 2 3 5 4 6 7 8 were hospitalized in the ICU for a mean dura- 9 no 10 Patient VAP treated with high-dose ampicillin/sulbactam 1 - Clinical and microbiological outcome of PDR A. baumannii VAP Table and high-dose tigecycline colistin combination. infection; unit; BSI, blood stream ventilator associated pneumonia; ICU, intensive care VAP, resistant; female; PDR, pan-drug M, male; F, Health Evaluation; MIC, minimum inhibitory concentration; KPC, Klebsiella pneumoniae Acute Physiology and Chronic carbapenemase; A APACHE, tion of 33.7±28.3 days. The mean duration of the KI, acute kidney injury; Tx, treatment 14 S.F. Assimakopoulos, et al.

studied triple combination therapy for A. bau- tigecycline distribution in the lung, although its mannii VAP was 13±3 days. The characteristics levels in patients with pneumonia are expected of the studied patients and their clinical and mi- to be higher than healthy volunteers. Standard crobiological outcomes are presented in Table 1. doses are probably inadequate to reach maximal- Nine patients (90%) exhibited a successful clinical ly efficacy, especially against resistant pathogens outcome accompanied by microbiological erad- on the upper end of the MIC distribution (1 to 2 ication in 7 of them. This discordance between μg/mL) [16]. High dose tigecycline therapy, de- clinical success and microbiological eradication fined as 200 mg loading dose followed by 100 mg has been previously described in non-sterile sites, every 12 hours, attains a better PK-PD profile and especially when foreign bodies are involved, such is associated with significantly improved out- as in patients with VAP, and constitutes coloni- comes [16]. With regard to the comparison of an- zation [13]. All clinically cured patients survived timicrobial options usually administered to treat at 14 and 28d, while a patient died later from no- patients with nosocomial pneumonia caused by socomial sepsis owing to KPC-positive Klebsiella MDR/XDR A. baumannii, in a recent metanalysis pneumoniae bacteraemia. The only patient who of twenty-three studies including 15 antimicrobi- failed the combination treatment and died from al treatments, sulbactam (normal and high-dose) VAP was the second older patient in this series was the best treatment in terms of survival ben- (79 years old) with a high Charlson comorbidity efit, followed by fosfomycin and colistin, com- index of 6, with the highest APACHE II score (28) bined inhaled and intravenous colistin, high-dose on admission, who was complicated by KPC-pos- tigecycline and colistin monotherapy [17]. With itive Klebsiella pneumoniae superinfection. Colis- the emergence of colistin resistance in A. bauman- tin administration was proven safe in term of its nii, our therapeutic armamentarium to treat these potential nephrotoxicity and acute kidney injury infections becomes very limited. New antibiot- (AKI) was developed in only one patient, the one ics against Gram negative bacilli do not offer a with the accompanying bacteraemia from A. bau- solution to this problem, because ceftolozane-ta- mannii, who was receiving concomitant vancomy- zobactam has no reliable activity against A. bau- cin for empiric MRSA coverage. mannii and ceftazidime-avibactam has activity similar to ceftazidime alone against A. baumannii [18]. Therefore, in carbapenem- and colistin-re- n DISCUSSION sistant (PDR) A. baumannii severe infections, it is reasonable to use combination regimens to take VAP and bacteraemia caused by XDR A. bauman- advantage of antibiotic synergism. Diverse com- nii strains, defined as resistant to all antibiotics binations of antibiotics have demonstrated a syn- except colistin and tigecycline, result in >50% ergistic action against XDR A. baumannii, such as mortality rates [5, 14]. In infections by XDR A. colistin combined with rifampicin or tigecycline baumannii strains, a number of studies attempted or sulbactam or vancomycin and carbapenems to answer the question whether there is any ben- combined with colistin, or sulbactam or amino- efit of colistin-based combination therapy over glycosides or rifampicin [14]. The combination of monotherapy with colistin, but the results were high dose ampicillin/sulbactam with colistin has contradictory [14]. However, a recent randomised demonstrated promising results in in-vitro, exper- controlled trial including 406 patients with severe imental and clinical studies, but clinical data is infections caused by carbapenem-resistant A. bau- scarce, especially in patients with PDR A. bauman- mannii demonstrated that the addition of merope- nii severe infections [19-21]. nem to colistin did not improve clinical outcome Patients analysed in the present study suffered [15]. Regarding the use of tigecycline in VAP, as from VAP by PDR A. baumannii, resistant to car- previously stated significantly higher mortali- bapenems and colistin, with high tigecycline MIC ty was observed with this treatment in MICs >2 >2 μg/mL. This difficult to treat population re- μg/mL [4, 5]. However, in the above-mentioned ceived a salvage combination treatment regimen, studies the standard dose of tigecycline, 100 mg consisted of intravenous colistin and high dose loading followed by 50 mg every 12 hours, was ampicillin/sulbactam and high-dose tigecycline. used. Several concerns have been raised about Also, inhaled colistin as adjunctive to its intrave- Triple combination therapy PDR A. baumannii VAP 15

nous administration was concurrently adminis- infection and the concomitant administration of tered. The rationale of this combined regimen was colistin with glycopeptides are independently to take advantage of the best available antimicro- associated with high risk of AKI development in bial weapons to fight a very serious infectious critically ill patients [12]. Therefore, AKI in this enemy. The very high rates of clinical response patient was attributed to the aforementioned and 28d survival (90%) observed in the pres- co-factors and not solely to the combination treat- ent case series are promising and have not been ment regimen. previously attained with dual combination treat- A major limitation of our findings is that this is ments. Moreover, it is noteworthy that this posi- only a small case series study, retrospectively tive clinical response and survival was attained in looking at the results of our experience and clin- VAP cases caused by PDR A. baumannii strains, as ical practice, when dealing with such difficult to compared to most previously published studies treat infections. In view of the increasing preva- which refer to XDR A. baumannii infections. The lence of PDR A. baumannii infections in ICUs, its mean duration of the triple combination therapy associated high rates of mortality and the lack of in our study was 13±3 days, which is longer than effective treatment options, we feel that there is an the currently recommended 7-day course of anti- emerging need our results to be further validated microbial therapy in VAP [10]. However, in VAP in larger prospective studies. caused by non-glucose-fermenting Gram negative bacilli including Pseudomonas and Acine- Conflicts of interest tobacter, there is a concern that short courses of The authors declare no competing financial antibiotics are associated with recurrent infection, interests. as demonstrated in a systematic review including 6 randomised controlled trials [22]. Beyond the efficacy of the triple combination n REFERENCES antibiotic therapy used in the present study, the [1] Nowak J., Zander E., Stefanik D., et al. High inci- increased survival rate observed could be addi- dence of pandrug-resistant Acinetobacter baumannii iso- tionally attributed to host and pathogen factors. lates collected from patients with ventilator-associated Regarding host factors, studied patients were rel- pneumonia in Greece, Italy and Spain as part of the atively young (only two patients were >65 years MagicBullet clinical trial. J. Antimicrob. Chemother. 72, old), without underlying immunosuppression 3277-3282, 2017. (100%) or major comorbidities (Charlson comor- [2] Magiorakos A.P., Srinivasan A., Carey R.B., et al. bidity index was 0 in 5 patients). Also, although Multidrug-resistant, extensively drug-resistant and not analysed in the present study, differences in pandrug-resistant bacteria: an international expert pro- pathogenicity among carbapenem-resistant A. posal for interim standard definitions for acquired re- baumannii clones might explain the positive clini- sistance. Clin. Microbiol. Infect. 18, 268-281, 2012. [3] The European Committee on Antimicrobial Suscep- cal outcomes [23]. Mutations that lead to antimi- tibility Testing. Breakpoint tables for interpretation of crobial resistance might impair bacterial fitness MICs and zone diameters. Version 8.1, 2018. http:// and virulence potential. Previous studies have www.eucast.org. shown that emergence of colistin resistance in A. [4] Chuang Y.C., Cheng C.Y., Sheng W.H., et al. Effec- baumannii strains, especially in mutants that lose tiveness of tigecycline-based versus colistin- based lipopolysaccharide, cost a fitness and virulence therapy for treatment of pneumonia caused by multi- deficit [24, 25]. drug-resistant Acinetobacter baumannii in a critical set- Regarding the safety of the triple antibiotic com- ting: a matched cohort analysis. BMC Infect. Dis. 14, 102, bination, three patients developed diarrhoea, but 2014. none of these cases were associated with Clostrid- [5] Cheng A., Chuang Y.C., Sun H.Y., et al. Excess mortality associated with colistin-tigecycline compared ium difficile infection. Acute kidney injury (AKI) with colistin-carbapenem combination therapy for ex- developed in only one patient, the one with the tensively drug-resistant Acinetobacter baumannii bacte- accompanying bacteraemia from A. baumannii, remia: a multicenter prospective observational study. who was receiving concomitant vancomycin for Crit. Care Med. 43, 1194-1204, 2015. empiric MRSA coverage. Previous studies have [6] Corbella X., Ariza J., Ardanuy C., et al. Efficacy of sul- demonstrated that the presence of bloodstream bactam alone and in combination with ampicillin in no- 16 S.F. Assimakopoulos, et al.

socomial infections caused by multiresistant Acinetobac- dose tigecycline in critically ill patients with severe in- ter baumannii. J. Antimicrob. Chemother. 42, 793-802, 1998. fections due to multidrug-resistant bacteria. Crit. Care [7] Garonzik S.M., Li J., Thamlikitkul V., et al. Popu- 18, R90, 2014. lation pharmacokinetics of colistin methanesulfonate [17] Jung S.Y., Lee S.H., Lee S.Y., et al. Antimicrobials and formed colistin in critically ill patients from a mul- for the treatment of drug-resistant Acinetobacter bau- ticenter study provide dosing suggestions for various mannii pneumonia in critically ill patients: a systemic categories of patients. Antimicrob. Agents Chemother. 55, review and Bayesian network meta-analysis. Crit. Care 3284-3294, 2011. 21, 319, 2017. [8] Karaiskos I., Giamarellou H. Multidrug-resistant [18] van Duin D., Bonomo R.A. Ceftazidime/avibac- and extensively drug-resistant Gram-negative path- tam and ceftolozane/tazobactam: second-generation ogens: current and emerging therapeutic approaches. beta-lactam/beta-lactamase inhibitor combinations. Expert Opin. Pharmacother. 15, 1351-1370, 2014. Clin. Infect. Dis. 63, 234-241, 2016. [9] Betrosian A.P., Frantzeskaki .F, Xanthaki A., Dou- [19] Lenhard J.R., Smith N.M., Bulman Z.P., et al. High- zinas E.E. Efficacy and safety of high-dose ampicillin/ dose ampicillin-sulbactam combinations combat poly- sulbactam vs. colistin as monotherapy for the treatment myxin-resistant Acinetobacter baumannii in a hollow-fib- of multidrug resistant Acinetobacter baumannii ventila- er infection model. Antimicrob. Agents Chemother. 61, tor-associated pneumonia. J. Infect. 56, 432-436, 2008. e01268-e01216, 2017. [10] Kalil A.C., Metersky M.L., Klompas M., et al. Man- [20] Leelasupasri S., Santimaleeworagun W., Jitwa- agement of adults with hospital-acquired and ven- sinkul T. Antimicrobial susceptibility among colistin, tilator-associated pneumonia: 2016 Clinical Practice sulbactam, and fosfomycin and a synergism study of Guidelines by the Infectious Diseases Society of Amer- colistin in combination with sulbactam or fosfomycin ica and the American Thoracic Society. Clin. Infect. Dis. against clinical isolates of carbapenem-resistant Acine- 63, e61-e111, 2016. tobacter baumannii. J. Pathog. 2018, 3893492, 2018. [11] Charlson M., Szatrowski T.P., Peterson J., Gold J. [21] Betrosian A.P., Frantzeskaki F., Xanthaki A., Geor- Validation of a combined comorbidity index. J. Clin. Ep- giadis G. High-dose ampicillin-sulbactam as an alter- idemiol. 47, 1245-1251, 1994. native treatment of late-onset VAP from multidrug-re- [12] Rocco M., Montini L., Alessandri E., et al. Risk sistant Acinetobacter baumannii. Scand. J. Infect. Dis. 39, factors for acute kidney injury in critically ill patients 38-43, 2007. receiving high intravenous doses of colistin meth- [22] Pugh R., Grant C., Cooke R.P., Dempsey G. Short- anesulfonate and/or other nephrotoxic antibiotics: a course versus prolonged-course antibiotic therapy for retrospective cohort study. Crit. Care 17, R174, 2013. hospital-acquired pneumonia in critically ill adults. [13] Kiem S., Schentag J.J. Correlations between microbi- Cochrane Database Syst Rev. CD007577, 2015. ological outcomes and clinical responses in patients with [23] Nishida S., Ono Y. Comparative analysis of the severe pneumonia. Infect. Chemother. 45, 283-291, 2013. pathogenicity between multidrug-resistant Acinetobac- [14] Poulikakos P., Tansarli G.S., Falagas M.E. Com- ter baumannii clinical isolates: isolation of highly patho- bination antibiotic treatment versus monotherapy for genic multidrug-resistant A. baumannii and experimen- multidrug-resistant, extensively drug-resistant, and tal therapeutics with fourth-generation cephalosporin pandrug-resistant Acinetobacter infections: a system- cefozopran. Infect. Drug Resist. 11, 1715-1722, 2018. atic review. Eur. J. Clin. Microbiol. Infect. Dis. 33, 1675- [24] Da Silva G.J., Domingues S. Interplay between 1685, 2014. colistin resistance, virulence and fitness in Acinetobacter [15] Paul M., Daikos G.L., Durante-Mangoni E., et al. baumannii. Antibiotics (Basel). 6, 2017. Colistin alone versus colistin plus meropenem for treat- [25] Carretero-Ledesma M., Garcia-Quintanilla M., ment of severe infections caused by carbapenem-resist- Martin-Pena R., Pulido M.R., Pachon J., McConnell M.J. ant Gram-negative bacteria: an open-label, randomised Phenotypic changes associated with colistin resistance controlled trial. Lancet Infect. Dis. 18, 391-400, 2018. due to lipopolysaccharide loss in Acinetobacter bauman- [16] De Pascale G., Montini L., Pennisi M., et al. High nii. Virulence 9, 930-942, 2018.