To analyse factors related to mortality and influence of antibiotic treatment on outcome in patients with nosocomial infection due to multidrug and carbapenem-resistant Acinetobacter baumannii (MDR-C AB).Patients and methods
Observational and prospective study of a cohort of adult patients with MDR-C AB infection. Data collection from clinical records was done according to a standard protocol (January 2007 through June 2008). Patients with MDR-C AB infection were identified by review of results of microbiology cultures from the hospital microbiology laboratory. Epidemiological and clinical variables and predictors of mortality were analysed.Results
24 out of 101 cases were considered colonizations and 77 infections (27 bacteraemia); global mortality in infected patients was 49% (18 cases with bacteraemia and 20 with no bacteraemia). In the multivariate analysis, including the 77 cases of infection, the prognosis factors associated with mortality were age (OR 1.09; 95% CI 1.02–1.2), McCabe 1 (OR 33.98; 95% CI 4.33–266.85), bacteraemia (OR 9.89; 95% CI 1.13–86.13), inadequate empiric treatment (OR 16.7; 95% CI 2.15–129.79), and inadequate definitive treatment (OR 26.29; 95% CI 1.45–478.19). In the multivariate analysis including the 57 cases of infection with adequate definitive treatment, the prognosis factors associated with mortality were McCabe 1 (OR 24.08; 95% CI 3.67–157.96) and monotherapy versus combined treatment (OR 7.11; 95% CI 1.63–30.99).Conclusions
Our cohort of patients with MDR-C AB infection is characterised by a very high mortality (49%); the severity of patients and inadequate treatment or monotherapy are statistically associated with mortality.
Analizar los factores asociados a mortalidad y la influencia del tratamiento antibiótico en la evolución de los pacientes con infección por Acinetobacter baumannii multiresistente y resistente a carbapenem (ABMDR-C).Pacientes y métodos
Estudio observacional y prospectivo de una cohorte de pacientes adultos con infección por ABMDR-C (Enero 2007 a junio 2008). Los casos de infección se identificaron a partir de los resultados de los cultivos microbiológicos. Analizamos las variables epidemiológicas y clínicas asociadas a mortalidad.Resultados
Veinticuatro de los 101 aislamientos de ABMDR-C se consideraron colonizaciones y 77 fueron definidos como infecciones (27 bacteriemias); la mortalidad global de los casos de infección fue del 49% (18 pacientes con bacteriemia y 20 sin bacteriemia). En el análisis multivariado que incluía a los 77 pacientes con infección, los factores predictores de mortalidad fueron: edad (OR 1,09; IC95% 1,02-1,2), McCabe 1 (OR 33,98; IC95% 4,33-266,85), bacteriemia (OR 9,89; IC95% 1,13-86,13), tratamiento empírico inadecuado (OR 16,7; IC95% 2,15-129,79), tratamiento definitivo inadecuado(OR 26,29; IC95% 1,45-478,19). En el análisis multivariado que incluía los 57 pacientes que recibieron tratamiento definitivo adecuado, los factores predictores de mortalidad fueron McCabe 1 (OR 24,08; IC95% 3,67-157,96) y tratamiento en monoterapia versus combinaciones (OR 7,11; IC95% 1,63-30,99).Conclusiones
En nuestra cohorte de pacientes con infección por ABMDR-C la mortalidad es elevada (49%); los pacientes graves y el tratamiento antibiótico inadecuado o el tratamiento definitivo adecuado en monoterapia se asocian de forma estadísticamente significativa a mayor mortalidad.
The clinical impact of severe infections due to multidrug and carbapenem-resistant Acinetobacter baumannii (MDR-C AB) is extensively discussed in the medical literature, since results and conclusions are contradictory, probably due to differences in methodology, pathogenicity and virulence of the isolated strains within different geographical areas.1–3
Faced with these discrepancies, we carried out a prospective analysis of a cohort of patients with MDR-C AB infection, to evaluate the prognostic factors associated with mortality and the influence of adequate treatment.Patients and method
We carried out a prospective and observational study of all documented nosocomial infections due to MDR-C A. baumannii corresponding to adult patients admitted to Hospital Universitario Virgen de la Arrixaca in the period comprised between January 2007 and June 2008. Information about patients was collected according to a previously established protocol. Patients admitted to either the paediatrics or the maternal units were excluded as they corresponded to polymicrobial isolates. This cohort has partially been analysed in a previous publication.6Microbiological study
The blood cultures were processed using the BACT-ALERT system (Becton-Dickinson, Sparks, MD). The samples from other locations were processed by the Service of Microbiology following the standard procedure and by means of the usual microbiological techniques.
The identification and study of antimicrobial susceptibility of the isolated microorganisms were conducted by means of the automated Vitek-2 system (Biomerieux, Lyon, France), using the ID-GN and AST-N059 cards. The activity of tygecycline, tobramycin and amikacin was checked by means of the E-test (AB Biodisk, Solna, Sweeden). Multiresistance associated with carbapenem-resistance was defined as the presence of in vitro resistance to imipenem (CMI≥16μg/ml) and with more than four of the following antibiotics: piperacillin-tazobactam, aztreonam, ciprofloxacin, ceftazidime, ampicillin-sulbactam and tobramycin or amikacin. Molecular studies were not performed.Patients
Patients were assessed with regard to the prognosis of their underlying disease, according to the criteria of McCabe and Jackson.7 Their condition was then classified as “rapidly fatal” when death was expected to take place in days or weeks, “ultimately fatal” when death was likely to take place in some months or years, and “nonfatal” when death was not predictable. The severity of the patient's condition at presentation was assessed according to Winston et al.,8 as follows: “critical” when the patient's clinical condition was rapidly deteriorating and the probability of death during the first 24h was high; “poor” when the clinical condition was deteriorating and death was probable but not imminent; “fair” when the clinical condition was deteriorating but death was not probable; and “good” when the clinical situation did not change during the first 24h and death was not probable.Infection study
The epidemiological characteristics of patients with documented nosocomial infection were assessed. Only samples collected after 72h of hospital admission or within the first 72h after admission in the case of antecedent of admission in the previous month were included. The patients were divided into 2 groups: patients with colonization of MDR-C AB (group A) and patients with documented infection (group B), who were subdivided into two groups, patients with bacteraemia (group B1) and patients without bacteraemia (group B2). The sources and types of infection were determined in accordance with Atlanta's CDC criteria.9 The study of the prognostic factors related to mortality was carried out with the infected patients, excluding the cases of colonization.
Colonization was defined as the presence of A. baumannii in a non-sterile sample in a patient without symptoms or signs of infection.Prognostic factors
In order to identify prognostic factors, all clinical and epidemiological characteristics, all complications, and the type of antibiotic treatment were examined in relation to the final evolution of patients, with “recovery” understood to be the disappearance of all active signs and symptoms of infection. Patients with poor prognosis were divided in 3 subgroups based on the time of death; in the first subgroup, death occurred within the first 3 days after the documentation of infection; in the second subgroup death occurred between the 4th and 7th day; and in the third group death occurred after the 8th day following the documentation of infection. Patients were followed up until death or discharge.Type of treatment
Empiric treatment was defined as the antibiotics prescribed by the physician who assessed a patient with clinical symptoms of infection when the microbiological samples were obtained. Definitive treatment was defined as the antibiotics prescribed according to the microbiological information (A. baumannii isolation and susceptibility pattern).
The empirical treatment was considered to be inadequate when the patient did not receive any in vitro active antibiotic in the adequate doses. Definite treatment was considered to be inadequate when the patient did not receive in the adequate doses any of the in vitro active antimicrobials against A. baumannii, at least during a whole day.
Likewise, definite treatment was classified into monotherapy and combined treatment, considering as combinations those specified in Table 1.
|Type of antibiotic treatment||Patients (n=77) N (%)||Mortality (n=38) N (%)|
|Monotherapy||33 (43)||21 (55)|
|Tigecycline||11 (14)||9 (24)|
|Sulbactam||6 (8)||4 (10)|
|Amikacin||8 (10)||3 (8)|
|Tobramycin||4 (5)||3 (8)|
|Colistin||4 (5)||2 (5)|
|Combinations||37 (48)||10 (26)|
|Tigecycline+rifampicin±amikacin||9 (12)||3 (8)|
|Sulbactam+amikacin||8 (10)||2 (5)|
|Colistin+rifampicin||1 (1)||0 (0)|
|Tigecycline+colistin+rifampicin||19 (25)||5 (13)|
|Without treatment||7 (9)||7 (18)|
Data were analysed using the statistics program SPSS 15.0 (SPSS Software, Chicago, IL, USA). The relation or association between pairs of qualitative variables was determined through analyses of contingency tables by means of Pearson's χ2-test, complemented by an analysis of residues with the aim of determining the directional dependence. In the case of quantitative variables, means were compared using Student's t-test. The difference was considered significant at p<0.05. A multivariate analysis was carried out in order to detect the predictors of global mortality in patients with infection due to MDRAB-C; this analysis consisted of a non-conditional logistic regression in which death was considered as a dependent variable and all those that had been statistically associated with mortality in the bivariate analysis were included as independent variables.Results
Out of 101 patients with A. baumannii isolation, 24 were excluded because they were considered as cases of colonization (group A) and 77 were considered to have developed an infection (group B) (27 cases of bacteraemia – group B1 – and 50 non-bacteraemic infections – group B2); all cause mortality in both groups of patients (A and B) was of 41.6% (4 patients in group A and 38 in group B – 18 in group B1). From here on we will present a description and analysis of the cohort 77 infected patients.
Early mortality (first 72h after the microbiological isolation) was 10.39% in group B (14.8% in group B1 and 8% in group B2) (Table 2).
Comparative study of bacteremic and non-bacteremic infections.
|Characteristics||Total cohortn=77N (%)||Bacteremic infectionn=27N (%)||Non-bacteremic infectionn=50N (%)||p|
|Men||56 (73)||19 (70)||37 (74)||ns|
|Women||21 (27)||8 (30)||13 (26)|
|Yes||59 (77)||19 (70)||40 (80)||ns|
|Diabetes mellitus||24 (31)||3 (11)||21 (42)||<0.05|
|Arterial hypertension||37 (48)||9 (33)||28 (56)||ns|
|Alcoholism||10 (13)||5 (18)||5 (10)||ns|
|Smoking||18 (23)||6 (22)||12 (24)||ns|
|Dyslipidemia||10 (13)||2 (7)||8 (16)||ns|
|Lung disease||16 (21)||3 (11)||13 (26)||ns|
|Kidney failure||11 (14)||2 (7)||9 (18)||ns|
|Liver disease||9 (12)||4 (15)||5 (10)||ns|
|Heart disease||30 (39)||10 (37)||20 (40)||ns|
|Vasculopathy||9 (12)||3 (11)||6 (12)||ns|
|Immunosuppression||14 (18)||7 (26)||7 (14)||ns|
|Trasplantation||11 (14)||6 (22)||5 (10)||ns|
|Systemic disease||4 (5)||1 (4)||3 (6)||ns|
|Cancer||10 (13)||2 (7)||8 (16)||ns|
|1||28 (36)||14 (52)||14 (28)||<0.05|
|2||24 (31)||9 (33)||15 (30)|
|3||25 (32)||4 (15)||21 (42)|
|Medical||12 (16)||1 (4)||11 (22)||<0.05|
|Surgical||65 (84)||26 (96)||39 (78)|
|ICU admission||64 (83)||27 (100)||37 (74)||<0.05|
|Invasive procedures||69 (90)||27 (100)||42 (84)||ns|
|Mechanical ventilation||59 (77)||25 (93)||34 (68)||<0.05|
|Tracheotomy||45 (58)||18 (67)||27 (54)||ns|
|Central venous catheter||64 (83)||27 (100)||37 (74)||<0.05|
|Nasogastric tube||52 (67)||20 (74)||32 (64)||ns|
|Enteral nutrition||37 (48)||14 (52)||23 (46)||ns|
|Parenteral nutrition||44 (57)||20 (74)||24 (48)||<0.05|
|Red blood cells transfusions||64 (83)||26 (96)||38 (76)||<0.05|
|Urinary catheter||66 (86)||26 (96)||40 (80)||ns|
|Surgery||65 (84)||27 (100)||38 (76)||<0.05|
|Previous antibiotics||16 (21)||5 (18)||11 (22)||ns|
|Winston||61 (79)||22 (81)||39 (78)|
|≤2||52 (67)||22 (81)||30 (60)||ns|
|>2||25 (32)||5 (18)||20 (40)|
|<12,000cells/mm3||33 (43)||14 (52)||19 (38)||ns|
|>12,000cells/mm3||44 (57)||13 (48)||31 (62)|
|<10g/dl||63 (82)||25 (93)||38 (76)||ns|
|>10g/dl||14 (18)||2 (7)||12 (24)|
|Complications||36 (47)||17 (63)||19 (38)||ns|
|Kidney failure||21 (27)||10 (37)||11 (22)||ns|
|Acute respiratory distress syndrome||23 (30)||12 (44)||11(22)||<0.05|
|Hepatic failure||9 (12)||6 (22)||3 (6)||<0.05|
|Shock||25 (32)||12 (44)||13 (26)||ns|
|Exitus vitae||38 (49)||18 (67)||20 (40)||<0.05|
|Within the first 3 days||8 (10)||4 (15)||4 (8)|
|Between 4th and 7th day||4 (5)||1 (4)||3 (6)|
|After the 7th day||26 (34)||13 (48)||13 (26)|
When comparing group B1 and B2 statistically significant differences were found in the mortality rate, the proportion of patients with rapidly fatal underlying diseases, diabetes mellitus incidence, the rate of surgical and ICU-patients, some of the invasive procedures that the patients underwent and in the number of complications (acute respiratory distress syndrome and liver failure) (Table 2). Susceptibility data of the isolates are shown in Table 3.
A. baumannii: susceptibility patterns (N=77).
|Antibiotics||SN (%)||IN (%)||RN (%)||No informationN (%)|
|Amikacin||44 (57)||8 (10)||10 (13)||15 (19)|
|Tobramycin||44 (57)||6 (8)||24 (31)||3 (4)|
|Gentamicin||23 (30)||2 (3)||44 (57)||8 (10)|
|Sulbactam||23 (30)||29 (38)||16 (21)||9 (12)|
|Tigecycline||54 (70)||9 (12)||1 (1)||13 (17)|
Mortality according to antibiotic treatment can be seen in Table 1. As for the prognosis of the patients with infections due to MDR-C AB, in the bivariant analysis a statistically significant relation was found between mortality and presence of underlying conditions, McCabe 1, heart disease, Winston≤2, as well as the presence of bacteraemia, anaemia (hemoglobin<10mg/dl) and the development of complications such as kidney failure, acute respiratory distress syndrome and shock (Table 4). Inadequate antibiotic treatment, both empirical and definite ones, and the use of monotherapy versus combined definitive treatment, were significantly associated with a higher mortality (Table 4). Seven patients did not receive antibiotic treatment due to early death before microbiological data were available or because the physician in charge did not consider the possibility of an infection.
Prognostic factors of patients with A. baumannii infections. Bivariant analysis.
|Characteristics||Recoveryn=39 (%)||Exitusn=38 (%)||p|
|Age years (range)||50 (18–83)||60 (21–88)||ns|
|Men||30 (77)||26 (68)||ns|
|Women||9 (23)||12 (32)|
|Yes||23 (59)||36 (95)||<0.05|
|Diabetes mellitus||12 (31)||12 (32)||ns|
|Arterial hypertension||15 (39)||22 (58)||ns|
|Alcoholism||3 (8)||7 (18)||ns|
|Smoking||6 (15)||12 (32)||ns|
|Dyslipidemia||4 (10)||6 (16)||ns|
|Lung disease||8 (20)||8 (21)||ns|
|Kidney failure||6 (15)||5 (13)||ns|
|Liver disease||3 (8)||6 (16)||ns|
|Heart disease||9 (23)||21 (55)||<0.05|
|Vasculopathy||4 (10)||5 (13)||ns|
|Immunosuppression||3 (8)||11 (29)||ns|
|Trasplantation||3 (8)||8 (21)||ns|
|Systemic disease||1 (3)||3 (8)||ns|
|Cancer||3 (8)||7 (18)||ns|
|2||13 (33)||11 (29)|
|3||24 (61)||1 (3)|
|Medical||6 (15)||6 (16)||ns|
|Surgical||33 (85)||32 (84)|
|ICU admission||30 (77)||34 (90)|
|Medical ICU||19 (49)||21 (55)||ns|
|Surgical ICU||11 (28)||13 (34)|
|Invasive procedures||34 (87)||35 (92)||ns|
|Mechanical ventilation||26 (67)||33 (87)||ns|
|Tracheotomy||22 (56)||23 (60)||ns|
|Central venous catheter||30 (77)||34 (90)||ns|
|Nasogastric tube||24 (61)||28 (74)||ns|
|Enteral nutrition||19 (49)||18 (47)||ns|
|Parenteral nutrition||19 (49)||25 (66)||ns|
|Red blood cells trasfusions||30 (77)||34 (90)||ns|
|Urinary catheter||31 (79)||35 (92)||ns|
|Surgery||30 (77)||35 (92)||ns|
|Previous antibiotics||29 (74)||32 (84)||ns|
|≤2||17 (44)||35 (92)||<0.05|
|>2||22 (56)||3 (8)|
|<12,000cells/mm3||17 (44)||16 (42)||ns|
|>12,000cells/mm3||22 (56)||22 (58)|
|<10mg/dl||28 (72)||35 (92)||<0.05|
|>10mg/dl||11 (28)||3 (8)|
|Respiratory||21 (54)||26 (68)||ns|
|Blood culture||9 (23)||18 (47)||<0.05|
|Skin and soft tissue (surgical wounds)||18 (46)||16 (42)||ns|
|Intra-abdominal fluids||3 (8)||2 (5)||ns|
|Urine||5 (13)||3 (8)||ns|
|Central venous catheter tip||10 (26)||12 (32)||ns|
|Central nervous system catheter tip||1 (3)||0 (0)||ns|
|Yes||8 (20)||28 (74)||<0.05|
|Kidney failure||5 (13)||16 (42)||<0.05|
|Acute respiratory distress syndrome||2 (5)||21 (55)||<0.05|
|Hepatic failure||2 (5)||7 (18)||ns|
|Shock||3 (8)||22 (58)||<0.05|
|- Adequate||22 (56)||13 (34)||<0.05|
|- Inadequate||17 (44)||25 (66)|
|- Adequate||37 (95)||20 (53)||<0.05|
|* Monotherapy||10 (26)||13 (34)|
|* Combined treatment||27 (69)||7 (18)|
|- Inadequate||2 (5)||18 (47)|
In the multivariant analysis including the 77 cases of infection, the prognosis factors associated with mortality were age (OR 1.09; 95% CI 1.02–1.2), McCabe 1 (OR 33.98; 95% CI 4.33–266.85), bacteraemia (OR 9.89; 95% CI 1.13–86.13), inadequate empiric treatment (OR 16.7; 95% CI 2.15–129.79), and inadequate definitive treatment (OR 26.29; 95% CI 1.45–478.19).
In the multivariate analysis including the 57 cases of infection with adequate definitive treatment, the prognosis factors associated with mortality were McCabe 1 (OR 24.08; 95% CI 3.67–157.96) and monotherapy versus combined treatment (OR 7.11; 95% CI 1.63–30.99).Discussion
In our cohort of patients with infection due to MDR-C AB, the mortality rate was 49%. When comparing our series of patients with bacteraemia due to MDR-C AB with a historical cohort of patients10 with bacteraemia due to A. baumannii carbapenem-susceptible (ABCS), the former presented a higher mortality (66.7%) than the latter (25.6%). In previous studies5,11–14 global mortality ranged from 17% to 62%, depending on differences in the type of patients, the prognosis of the underlying diseases or the initial clinical severity within different cohorts. These mortality rates are similar to ours (49.3%), though early mortality (10.39%) in our series was lower than that noticed by other authors3 (19–35%). This difference could be related to the fact that in our sample both patients with and without bacteraemia were included, while in other studies only cases of bacteraemia were selected (and bacteraemia is in our cohort a factor associated with mortality). The question “is mortality related to MDR-C AB or due to underlying conditions?” remains difficult to answer. In our series, McCabe 1 was a prognostic factor of mortality; however, the fact that the bacteraemia, inadequate empiric or definite treatment are associated with mortality in the multivariate analysis leads us to think that the presence of infection with MDR-C AB also plays a role in the clinical evolution of the patient. These results are similar to those pointed out by other authors. Thus, Cisneros and Rodríguez-Baño15 also found disseminated intravascular coagulation and inadequate antibiotic treatment as factors associated with death. However, other authors have not found an association between inadequate antibiotic treatment and mortality.5,14,15 The fact that in some studies inadequate treatment is a factor associated with mortality, while in others only the variables related to severity at onset and those of the underlying conditions are associated with death, prevents from universally accepting that infection with MDR-C AB is per se a mortality factor. Therefore, severity of all underlying conditions would itself determine the critical situation of the patients, in whom infection with MDR-C AB would not be more than an epiphenomenon in a very complex clinical condition.
Unfortunately, we could not carry out in our study, a multivariate study, comparing the historical cohort10 (ABCS) with the group of patients with infections due to MDR-C AB. This represents a restriction which has not allowed us to assess the influence of resistance to carbapenems in the clinical evolution of the patients. If we look through the literature, there are few papers that analyse this aspect, but the results are equally contradictory, since in some cohorts4 the influence of carbapenem resistance in mortality is obvious while in others,13 this variable does not affect patients’ clinical evolution.
There are still not enough data in the literature about which is the most adequate antibiotic treatment in patients with infections due to MDR-C AB. In our series, combined treatment was associated with a lower mortality than monotherapy, both in the bivariant and the multivariant analyses. Research on in vitro synergy,16–18 some experimental studies19–21 and some small clinical cohorts,22–24 also seem to support the idea of prescription of combined treatment against monotherapy, just as probably common sense or knowledge about other multiresistant serious infections recommend. Our clinical experience and this paper, in spite of its limitations, make us think, together with other authors25 that monotherapy is not the best treatment option; in fact, resistance and therefore therapeutic failure have been described during tigecycline treatment.26–28 We have also appreciated this phenomenon: in two patients with bacteraemia secondary to lung infection and treated with tigecycline, initial MIC was 1mcg/ml and after three days of treatment (and no clinical response) blood cultures remained positive and MIC had gone up to 8mcg/ml; both patients died. Likewise, the appearance of subpopulations of A. baumannii developing resistance during colistin treatment has been described29; the effectiveness of this antibiotic in monotherapy in ventilator-associated pneumonia and severe infections due to MDR-C AB30,31 ranges between 52% and 61%, being side effects a main disadvantage (mainly nephrotoxicity).32 Sulbactam use, another recommended option in MDR-C AB infections, is limited by the increasing number of resistant strains (more than 50% of the isolates31; 68% in our cohort). Nevertheless, experts recommend the use of sulbactam high doses to overcome this limitation (dose-related side effects are uncommon).
As for combined treatment, in vitro studies and research on experimental20–24 models have found that the addition of rifampicine to another active antibiotic, such as tigecycline, colistin or sulbactam, has a synergistic effect and limits the development of resistances during treatment; however, studies are controversial and in another experimental work colistin activity was lessened by rifampicine.29 In 26 cases of MDR-C AB severe infections Motaouakkil et al.22 observed an effectiveness of 100% with the combination of colistin and rifampicine and in a cohort of 29 patients treated with high doses of rifampicine (10mg/kg/day) and colistin, Bassetti et al.23 achieved an effectiveness of 76% and only 10% developed nephrotoxicity, despite the fact that patients suffered from previous kidney failure. With the same combination, Petrosillo et al.24 obtained 50% positive results, although they used lower rifampicine doses. We used a standard dose of 600mg/day.
Nowadays, according to some experts, a triple combination of colistin, tigecycline and rifampicine could probably be the choice, though there is not enough clinical experience.
Some of the limitations of our study are (1) small sample size; (2) identification of A. baumannii did not include molecular techniques and (3) we used a Winston score to assess severity, as this is a subjective measure of gravity.
To sum up, the role of MDR-C AB as an infectious agent in the prognosis of severely compromised patients is not clarified and more studies looking deeply into the true clinical impact of resistance are needed. However, meanwhile, better epidemiological knowledge and the establishment and adequate fulfilment of nosocomial infection control programs and antibiotics’ prescription policies (mainly carbapenems) might contribute to reduce the incidence of MDR-C AB infections, as well as achieve a more adequate clinical and therapeutic approach.33Conflicts of interest
The authors have no relevant financial interest in this article or financial conflicts of interest. None to declare.