Impact of Rapid Molecular Diagnostic Method on Antibiotics Exposure Duration in ICU Patients With Postoperative Peritonitis (DIRECTABDO)

Impact of Rapid Molecular Diagnostic Method on Antibiotics Exposure Duration in ICU Patients With Postoperative Peritonitis

Impact on Duration of Antibiotic Therapy of Rapid Molecular Diagnostic Method for Rapid Analysis of Susceptibility of Peritoneal Samples in Postoperative Peritonitis in ICU Patients

Hospitals, and more specifically Intensive Care Units (ICU), face the challenging issue of emergence and rapid spread of multi-drug resistant bacteria (MDR). In some cases, the therapeutic choice is extremely limited. Prevention and adequacy of antibiotic therapy (AB) are the key responses applied toward these threats. A delayed adequate AB is a known factor of poor prognosis. Intra-abdominal infections (IAI) are frequent, polymicrobial and life threatening diseases. Source control and adequate AB are instrumental issues in this setting. Despite technical advances, susceptibility testing of the microorganisms collected from peritoneal samples is not usually available before day two or three after surgery. In this time lapse, empiric AB might be inadequate (not targeting all the pathogens, which leads to a prolonged duration of AB and potential increased morbidity/mortality risk) or too broad (with the two issues of ecology with an increased risk of selection of MDR bacteria and additional costs). In a pilot study evaluating the potential benefit of a direct culture of peritoneal samples from resuscitation patients treated for peritonitis, we observed that conventional treatment made it possible to obtain microbiological results within a median of 3 [extremes 2-7] days whereas a direct microbiological technique by E-test gave results in 1 [1-2] days (p <0.0001). With this technique close to conventional microbiological, a change in antibiotic therapy could have been achieved within an average of 1 [1-2] days versus 4 [1-11] days with conventional management (p = 0, 0006). The development of modern molecular techniques suggests that a large margin of improvement for the rendering and the precision of the results is possible.

BACKGROUND Antiinfective agents are aimed at eradicating microorganisms, or at least significantly reducing the size of inoculum to facilitate the natural activity of host defence mechanisms. One of the most challenging issues faced by prescribers is the emergence of multidrug resistant bacteria (MDR) and fungi, especially in the ICU setting. In some instances, the therapeutic options are extremely limited or even totally inadequate. The responses to this threat are improvements in prevention of nosocomial infection, antiinfective stewardship policy, and a better use of available resources including optimization of antibiotic therapy from the empiric phase of treatment. The optimization of antibiotic and antifungal therapy is a key issue. In a recent paper analysing the antibiotic prescription in a tertiary care hospital over a 9 to 10 weeks period, the authors reported an inappropriateness in 37% of the cases involving indication of therapy (17.5%), choice of the spectrum (7.6%) and application of the treatment (dosage, timing, duration…)(9.3%). Delayed adequacy for antibiotic therapy is a known factor of prolonged duration of anti-infective therapy and consequently of selection pressure. Adequacy is defined as the microbiologically effective antimicrobial treatment against the causative pathogens. A decreased duration of antibiotic therapy is an easy to implement strategy to decrease the emergence of resistance both at individual and community levels. This strategy of a reduced duration of therapy has additional potential advantages such decreased incidence of adverse effects and reduced costs. The selection of an empiric antibiotic (AB) therapy is often a difficult challenge. Despite advances in conventional microbiologic techniques, identification and susceptibility testing of the microorganisms collected from a suspected infectious site are not usually available before day two or three after sampling. In this time lapse, empiric AB therapy might not target all the pathogens, which leads to a prolonged duration of AB and potential increased morbidity/mortality rates. To overcome this uncertainty, physicians frequently prescribe very broad-spectrum empiric regimens to target a maximum number of potential pathogens. This is specifically the case in ICUs where life-threatening infections justify this policy. However, as mentioned above, the consequences in terms of selection pressure, cost and adverse effects are of major importance. To overcome the disadvantage of delayed microbiologic results, some authors in ICU medicine tried to shorten the delay of results of susceptibility testing in using direct E-test. The E-test antimicrobial susceptibility procedure is a quantitative method for antimicrobial susceptibility testing that consists of a reagent strip with a predefined gradient of AB. The stable gradient provides an inoculum tolerant system that allows its application directly to clinical specimens. This method has been shown to be reliable for predicting the susceptibility of microorganisms directly from positive blood cultures. It has also been applied directly to sputum samples from patients with cystic fibrosis. However, the physicians in charge of the patient still have to wait for at least 24 hours before having the first results of bacterial growth and susceptibility. However, in case of slow-growth this delay could be enlarged. In addition, this technique is inaccurate for antifungal therapy. Major advances in molecular biology over the last decade led to the development of rapid diagnosis and identification techniques. Techniques of genomic analyses bring important information on pathogen identification and on presence of certain resistance genes in hours rather than days and allow an early adaptation of antibiotic treatment,allowing the physician to choose the most accurate therapeutic strategies. Promising results have been obtained in the field of pneumonia. On the opposite, few data are available in intra-abdominal infections. The Unyvero (Curetis) platform is a turnkey rapid diagnostic system that can be used to diagnose severe infections in 4 to 5 hours. The Unyvero IAI panel (Intra-Abdominal Infection) covers the bacteria involved in IAI, such as Gram negative bacilli (Enterobacteriaceae, Pseudomonas aeruginosa, Acinetobacter baumannii), Gram positive bacteria (Streptococcus spp., Staphylococcus spp., Enterococcus spp.), anaerobic bacteria (Bacteroides spp., B fragilis, Prevotella spp.) and fungi (Candida albicans, C glabrata, C tropicalis, Candida spp.), as well as some antibiotic resistance genes such as certain ESBL (CTX-M) and carbapenemases. These infections are among the most frequent complications in surgical patients and are marked by a severe prognosis. They are characterized by high proportions of MDR isolates, reaching 85% of all the cultured organisms. In addition, the characteristics of bowel flora with mixed Gram positive, Gram negative and anaerobes bacteria lead to the constant threat of inadequate empiric antibiotic therapy. These IAI need an early efficient antibiotic therapy combined to source control, especially in the most severe cases admitted in ICU. Postoperative IAI are the cases where the needs for an adequate antibiotic treatment from the empiric phase are the most significant. In community-acquired infections, guidelines have been published that give satisfactory results in a vast majority of the cases. On the opposite in postoperative cases, many unsolved questions remain depending of the case mix and local epidemiology, leading to high proportion of inadequate empiric antibiotic therapy, and consequently increased duration of treatment, delayed adequacy and increased mortality and morbidity rates. Fungal infections, especially candidas are another issue in these cases of IAI. These microorganisms are reported in up to 30% of the operative samples in postoperative intraabdominal infections. However, due to the lack of relevant clinical criteria and the slow growth of these microorganisms, more than 70% of all antifungal empiric therapies given for these patients are inappropriate, and not necessary in most of them. In line with the French recommendations, in patients at risk of fungal infections an empiric antifungal therapy will be started. Every approach aimed at improving the empiric antiinfective therapy in postoperative IAI is welcomed. To our knowledge, few data are available addressing this issue. Techniques derived from conventional cultures are of interest but the delay to reach bacterial growth remains high and obviously >24 hours. Due to the capacities of the rapid molecular diagnostic test, therapeutic decisions (initiation of empirical therapy, interruption of unnecessary empirical therapy or adjunction of specific additional regimens such as antifungal agents) could be performed in a few hours. The potential benefit of these techniques and gain of time could be illustrated in an experimental pilot study performed in our unit analysing peritoneal samples of ICU patients with secondary peritonitis when these new molecular techniques were not yet available. The conventional approach brought results in a median delay of 3 [extremes 2-7] days while the use of direct antibiotic-enriched agar plates were available in 1 [1-2] days (p<0.0001). A change in AB regimens would have been made in a median delay of 4 [1-11] days versus 1 [1-2] days if direct cultures were applied, respectively (p=0.0006). Recently, an external evaluation of a multiplex PRC rapid test for detection of intra-abdominal infections was assessed in 4 laboratories of microbiology. Three hundred clinical samples were evaluated including 107 peritoneal fluids, 47 ascites, 36 peritoneal cavity drain fluids, 32 swabs (eswab fluid as well as charcoal swabs), 27 pus, 29 bile, 7 tissue samples, 7 aspirates, 5 pancreatic fluids, and 3 samples were from blood culture bottles inoculated with peritoneal fluid. Microbiology results were negative for 86 samples. Of these, 62 samples (72%) were also negative with the Unyvero IAI test while pathogens were detected in the remaining 24. In 16/24 samples, the pathogens detected by Unyvero IAI were confirmed by sequencing. Overall Sensitivity was 91.2% and Specificity 99.5%. Most pathogens detected by the Unyvero IAI test achieved a sensitivity of at least 80.0% while 12/27 panel targets achieved a sensitivity of 100%. Specificity ranges from 97.3% to 100% over all panel targets.The Unyvero IAI cartridge was able to detect additional microorganisms in particular anaerobes with most detections confirmed by sequencing. Time to identification was reduced by an average of about 17h (39:06 ± 16:09 h for microbiology vs 22:02 ±4:12 h Unyvero IAI (ID +resistance markers)) and time to full microbiological results was reduced by an average of about 41h (64.19 ± 12.10 h for microbiology vs. 23.44 ± 3. 58 for Unyvero IAI). The most frequent resistance markers detected were mecA/mecC (n=25), aacA4 (n=20), blaCTX-M (n=17) and nine instances of carbapenemase were identified. PPV range from 62.5 to 93.8% while NPV range from 70.2 to 100%. HYPOTHESIS The use of a rapid molecular diagnostic test by multiplex PCR on the peritoneal samples from ICU patients admitted for an intra-abdominal infection is feasible and makes it possible to obtain the identification of the germs present in the clinical sample and detect the presence of certain resistance genes in the first 24 hours after surgery, making it possible to reduce exposure to antibiotics and antifungals without increasing the frequency of clinical or microbiological failures

An anti-infective treatment is adequate if the anti-infective agents administered correspond to the simplest effective anti-infective treatment as determined by conventional microbiology

De-escalation is defined as a switch to a narrower spectrum anti-infective agent or a switch from bi to monotherapy. A de-escalation of β-lactams is considered to be a prescription change from one molecule of one rank to that of a lower rank

The time in hours between randomization and adequate anti-infective treatment (antibiotic and antifungal)

The proportion of patients with resumption or prolongation of the antibiotic until discharge from hospital or until D28.

Resumption is defined as a resumption of antibiotic therapy beyond the 48th hour after discontinuation of treatment; extension is defined as antibiotic treatment for more than 8 days

Organ failure is defined by a SOFA score ≥ 3 points for each organ (for example 3 points for respiratory failure). Organ failure may improve and then reappear during the observation period.

The proportion of patients with emergence of at least one multidrug-resistant bacteria on clinical or hygiene samples (rectal and oropharyngeal swabs).

Emergence is defined as the isolation of a BMR on weekly hygiene samples OR clinical isolates not detected at the inclusion. Bacteria with resistance to at least one agent in three or more antimicrobial categories are considered BMR.

Onset of clinical failure until D28, defined as one of the following criteria regardless of the cause: need for additional antibiotic therapy, additional surgical surgery or drainage to cure the infection, or death due to infection beyond the first 48 hours of antibiotic therapy

Occurrence of microbiological failure until D28, defined as persistence of infection or superinfection (i.e. the emergence of a new documented germ at the site of infection with signs and symptoms of worsening of infection). Microbiological responses, at both patient and germ levels, will require follow-up samples obtained by culture

Evaluate the diagnostic performance of the Unyvero IAI test in the intervention group compared to the conventional technique

Diagnostic performance of the Unyvero IAI test as compared to the conventional technique (gold standard) : Se;Sp;NPV;PPV

Inclusion Criteria: Patients will be eligible providing that they meet all the following criteria: having a surgery for a suspicion of post-operative peritonitis; free, informed and written consent of the patient, or his relatives or the trusted person previously designated or failing that, inclusion in an emergency situation Exclusion Criteria: Patients will not be eligible if they meet at least one of the following criteria age under 18 years; current pregnancy; poor chance of survival (SAPS II > 65 points within 12 hours before inclusion); patient for whom a limitation of care has been decided patient included in another therapeutic trial involving antibiotics agents patient who has already participated in this study patients unable to give consent (under guardianship or curatorship) patient not affiliated to social security Randomization criteria: surgical samples collected during reoperation for microbiological purpose; the two surgical samples can be processed by the microbiology laboratory within a maximum period of 8 hours between the start of the two techniques