The prolonged β-lactam Antibiotics intravenous infusion strategy has emerged as the standard treatment for sepsis despite its unknown efficacy. The investigators will conduct a prospective, multi-center, cluster randomized controlled clinical trial. The investigators aimed to compare the clinical efficacy and prognosis of prolonged β-lactam antibiotics intravenous infusion versus short-term intravenous infusion in ICU patients with early sepsis. The investigators expect to recruit 40 branch centers and enroll at least 2600 patients with sepsis.

Sepsis and septic shock can lead to high morbidity and mortality. The mortality of sepsis is related to inappropriate antibiotic treatment strategy. Due to the pathophysiological characteristics of patients with sepsis, the pharmacokinetics of antibiotics have changed, and antibiotic treatment strategies applied to general mild and severe infections may not be suitable for those patients. The relevant international guidelines recommend that antibiotic treatment for patients with sepsis should base on pharmacokinetic/pharmacodynamic principles, but this recommendation is based on low-level clinical evidence.

β-lactam antibiotics, including penicillins, cephalosporins, carbapenems and so on, are the most widely used antibacterial drugs in clinical practice. The best predictive parameter of those antibiotic bactericidal activity is the time during which the free drug concentration exceeds the target microbial MIC value (fT >MIC). According to Monte Carlo approach and clinical studies, as a PK/PD target value, an effective bactericidal effect can be achieved if fT>MIC reaches more than 40%; fT>MIC reaches more than 60%-70% can achieve the maximum bactericidal effect, which can be used in the treatment of severe cases. For severe infection and prevention of bacterial resistance, fT>MIC needs to reach 90%-100%.

The results of clinical studies have proved that improper or inadequate initial empiric antibiotic treatment is an independent risk factor that affects the therapeutic effectiveness and prognosis of severe infections. Important reasons for the lower-than-expected antibiotic treatment effect in severely ill patients include at least the following two factors: (1) Changes in the pathophysiological state of severely ill patients on drug metabolism, such as capillary leakage leading to increased drug distribution volume. As well as the high excretion and low obstruction hemodynamic characteristics of sepsis, increased renal blood flow leads to high excretion of water-soluble drugs, which often reduces the effective plasma concentration of the drug; (2) ICU infection of pathogenic bacteria has increased drug resistance and increased MIC. The above factors lead to the decrease of drug fT>MIC, which affects the efficacy of antibiotics. In recent years, the optimization of PK/PD-guided time-dependent antimicrobial treatment programs have confirmed that the administration method of prolonging the infusion time or continuous infusion can maintain a good steady-state blood drug concentration, prolong fT> MIC, and improve clinical curative effect, and can reduce the amount of antibiotics.

The main problems with PK/PD-guided antimicrobial therapy are: (1) Lack of convincing large sample clinical research results; (2) It has not been confirmed which subgroup (such as the sepsis severity, drug-resistant patterns of pathogens, immunocompetence) can be benefited from this strategy; (3) It is not clear whether the method of prolonged infusion can be applied in all kinds of β-lactam antibiotics .

This study adopts multi-center, openness, cluster randomization method to group, and eliminates the bias caused by factors such as the treatment environment in a single ward through the multi-center study; through Uniform training realizes the standardization of drug delivery methods to eliminate researchers' human bias in treatment operations and observations. At the same time, the regional randomization method sets the drug delivery method for a certain research center in a certain research phase to be determined, which eliminates the operational error and observation bias when the researcher needs to face multiple drug delivery programs at the same time. It can greatly reduce research costs and human bias, and more reliably obtain the impact of PK/PD-guided antibiotic treatment on the prognosis of patients and the impact of bacterial resistance in the entire ward.

Inclusion Criteria:

1. Patients in ICU who need to be treated with β-lactam antibiotics for clinical diagnosis of infection;
2. Meet the diagnostic criteria of sepsis 3.0 in the previous 24h;
3. At assessment of eligibility, treating doctor expects patient to need treatment in ICU beyond the next calendar day.

Exclusion Criteria:

1. The infection is diagnosed clinically, but the acquired pathogens are not sensitive to the study drug;
2. Has a history of allergies to study drugs;
3. Those who have a serious condition and the expected survival time is less than 72 hours.
4. Receipt of potential study medication for > 24 hours before randomization.
5. Pregnancy
6. Death is deemed imminent and inevitable.
7. Receiving palliative or supportive treatment only at the time of assessment for eligibility.
8. Treating doctor not committed to provision of advanced life-support, including any of mechanical ventilation, dialysis and vasopressor administration for at least the next 48 hours.
9. Consent not gained for study participation and entry under a waiver-of-consent not approved by the jurisdictional human research ethics committee.

Study protocol and ICF will be shared with other researchers when start this trial.

Study protocol and ICF will be shared with other researchers when start this trial for five years.

Ever researchers can access our study protocol and ICF from the web of clinical trials.gov.