Risk factors and clinical prediction models for prolonged mechanical ventilation after heart valve surgery

The study population consisted of adult patients (age range: 18 years to 75 years) who underwent their initial valve surgery at our hospital between September 2020 and January 2022. Inclusion criteria encompassed patients meeting the specified age range, while patients were excluded if they met any of the following criteria: (1) emergency surgery or prior history of cardiac surgery; (2) presence of severe congenital heart malformations, infective endocarditis, cardiac tumors, or macrovascular disease; (3) coexistence of chronic obstructive pulmonary disease, bronchiectasis, lung infections, or other conditions significantly impacting lung function; (4) presence of severe comorbidities such as liver and kidney failure, autoimmune diseases, malignancies, or infections; (5) significant neurological or psychiatric disorders including delirium, dementia, epilepsy, or a history of schizophrenia; (6) incomplete data. This retrospective study was conducted at a single center and adhered to the principles of the Declaration of Helsinki. The trial received approval from the Ethics Committee of the Second Affiliated Hospital of Nanchang University. The requirement for informed consent was waived due to the observational and retrospective nature of the study, as well as the anonymization of data prior to analysis.

Intraoperative rocuronium bromide 0.4 mg/kg/h was continuously pumped to maintain muscle relaxation. Dexmedetomidine and 1% propofol were employed to maintain sedation. The patient was subjected to controlled mechanical ventilation in the Intermittent Positive Pressure Ventilation mode. The respiratory parameters were set: volume control mode, gas flow rate of 2 L/min, tidal volume of 8 ml/kg, inspiratory/expiratory ratio of 1:2, inspiratory pause time of 20%, and the ventilation frequency per minute was adjusted to maintain the end of carbon dioxide partial pressure not higher than 40mmHg. A fraction of inspired oxygen was set at 80%, and 5 cmH₂O positive end-expiratory pressure(PEEP) was fixed for ventilation. Rocuronium bromide, dexmedetomidine, and 1% propofol were employed to maintain muscular relaxation and sedation. Neuromuscular monitoring was performed by a relaxometry train of four. Sugammadex was administered to reverse the neuromuscular blockage effect of rocuronium. Intermittent sufentanil injections (2 µg/kg) were administered to provide analgesic support, while sevoflurane was intermittently inhaled to ensure the patient remained in general anesthesia.

Patients were administered intravenous heparin sodium at a dose of 3 mg/kg prior to initiation of cardiopulmonary bypass, leading to systemic heparinization. The attainment of an ACT exceeding 480 s served as a pivotal parameter for commencing cardiopulmonary bypass. Employed in this procedure was a JOSPRA apparatus, alongside a membrane oxygenator and single-use cardiopulmonary bypass tubing. Pre-filling was accomplished with Sodium lactate Ringer’s solution, hydroxyethyl starch and sodium chloride injection, and de-whitened suspended red blood cells. Notably, during the final stages of valve surgery and the initial steps of cardiac incision closure, a 1:1 heparin antagonism was executed using protamine sulfate. In instances where the patient’s ACT surpassed their preoperative physiological baseline, an additional dose of 1 mg/kg protamine sulfate was administered. The cessation of cardiopulmonary bypass was executed gradually following the successful resumption of cardiac function. During CPB, mechanical ventilation is stopped. When the heart resumed beating but the CPB had not stopped, the lung resuscitation technique was used first, and three times manual ventilation was used to make the peak airway pressure reach 30 cmH₂O, which was maintained for 10–15 s, and then half of the initial tidal volume was taken for ventilation. The initial ventilation mode was resumed after the CPB stopped.

Subsequent to the conclusion of the surgical procedure, the patient was transitioned to the ICU. The patient was systematically managed in adherence to a standardized protocol. Sustained monitoring encompassed parameters such as non-invasive blood pressure, electrocardiogram, pulse oximetry, central venous pressure, arterial blood pressure, and continuous administration of vasoactive medications to ensure the preservation of stable hemodynamics. In the ICU, postoperative mechanical ventilation was mainly performed in synchronous intermittent command ventilation (SIMV) mode, with an initial setting of tidal volume of 10 ml/kg, oxygen concentration of 60%, PEEP of 5 cmH₂O, inspiratory and expiratory ratios of 1:1.5-2.0, and a frequency of 14 breaths/minute. Specific ventilator parameters were adjusted according to the patient’s condition. The cessation of sedation is contingent upon a safety evaluation conducted by the ICU team, followed by expeditious extubation. The criteria employed for the postoperative extubation of patients who have undergone heart valve surgery encompass the following facets: (1) Attainment of patient consciousness concomitant with the presence of an intact cough reflex; (2) The assurance of hemodynamic stability, with an absence of arrhythmias, discernible chest tightness, dyspnea, or related symptomatic manifestations; (3) Manifestation of spontaneous respiratory activity, maintenance of partial oxygen pressure exceeding 60 mmHg, FiO2 less than 0.5, and a pH level within the range of 7.35 to 7.45; (4) Maintenance of body temperature above 36.5 °C without the manifestation of chills; (5) The pericardial and mediastinal drainage systems exhibit a restricted flow capacity, characterized by a sustained rate not surpassing 100 milliliters per hour for a duration extending beyond two hours; (6) Preservation of normal peripheral circulation and a urine output rate exceeding 1 mL per kilogram of body weight per hour; (7) Attainment of satisfactory muscle strength in the extremities.

Data regarding the duration of postoperative mechanical ventilation were collected for the study participants. Patients with less than 24 h of postoperative ventilation were classified into the non-postoperative mechanical ventilation (non-PMV) group, while those with more than 24 h of mechanical ventilation were classified into the postoperative mechanical ventilation (PMV) group. Baseline characteristics and clinical data, including demographic information, comorbidities, preoperative variables, and surgical variables, were collected for both groups. If a clinical factor is missing in 20% of patients, this factor will be excluded from the analysis. The missing values of the data will be processed using multiple imputations.

Postoperative mechanical ventilation exceeding 24 h was classified as PMV. Hypertension was determined by systolic blood pressure ≥ 140 mm Hg, diastolic blood pressure ≥ 90 mm Hg, or antihypertensive medication. Diabetes was defined as fasting blood glucose ≥ 7.0 mmol/L, random blood glucose ≥ 11.1 mmol/L, or the use of diabetic medication. Left atrial diameter index (LADi), left ventricular end-diastolic diameter index (LVDdi), and left ventricular end-systolic diameter index were calculated by dividing left atrial diameter, left ventricular end-diastolic diameter, and left ventricular end-systolic diameter by body surface area (BSA), respectively. Current smokers were defined as patients who had smoked at least 100 cigarettes in their lifetime and were currently smoking. Current alcohol drinkers consumed alcohol more than once per week during the previous six months.

This single-center retrospective study screened 849 patients undergoing valve surgery between September 2020 and January 2022; 299 patients were excluded, and 550 were included in the final analysis (Fig. 1). The median duration of mechanical ventilation after surgery was 19.5 h. Among them, 62 patients (11.27%) developed postoperative PMV, with a median mechanical ventilation time of 26.75 h in the PMV group. The remaining 488 patients did not experience PMV postoperatively, with a median mechanical ventilation time of 19 h in the non-PMV group. The median age of patients in the PMV group was 61.5 years, compared to 54 years in the non-PMV group. In the PMV group, 61.3% of patients were male, compared to 50.4% in the non-PMV group. Hypertension was present in 25.8% of the PMV group, coronary atherosclerosis in 9.7%, diabetes mellitus in 8.1%, and a history of cerebral infarction in 14.5%. Conversely, in the non-PMV group, 20.5% had hypertension, 7% had coronary atherosclerosis, 5.1% had diabetes mellitus, and 7.8% had a history of cerebral infarction. The two groups’ baseline characteristics, preoperative data, and procedure-related data are presented in Tables 1 and 2, and 3, respectively.

Univariate analyses were performed on 550 patients to identify potential risk factors for PMV using baseline characteristics (Table 1), preoperative data (Table 2), and procedure-related data (Table 3). The results indicated that the following 12 factors might be associated with the risk of PMV in patients undergoing valve surgery: age, body mass index (BMI), heart rate (HR), current smoking, left atrial diameter index, NT-proBNP, preoperative white blood cell count (WBC), red blood cell count (RBC), platelet count, intraoperative blood transfusion, cardiopulmonary bypass time, and aortic clamping time. Subsequently, multivariate logistic regression analysis was performed on the aforementioned potential risk factors with a significance level of P < 0.05. The analysis revealed that age (OR = 1.082, 95% CI = 1.042–1.125; P = 0.000), current smoking (OR = 1.953, 95% CI = 1.007–3.387; P = 0.047), left atrial internal diameter index (OR = 1.04, 95% CI = 1.002–1.081; P = 0.041), preoperative red blood cell count (OR = 0.049, 95% CI = 0.275–0.876; P = 0.016), and aortic clamping time (OR = 0.017, 95% CI = 1.005–1.057; P < 0.017) were identified as independent risk factors for the development of PMV. Other characteristics did not demonstrate significant associations (Table 4).

BMI, body mass index; HR: Heart Rate; LADi, Left atrial diameter index; NT-proBNP, N-terminal pro-brain natriuretic peptide; WBC, white blood cell; RBC, Red blood cell; CPB, Cardiopulmonary bypass.

A nomogram was constructed using a multivariate logistic regression model that screened five independent risk factors for PMV. Each variable was assigned a score based on its regression coefficient, as illustrated in Fig. 2. By scoring the variables above and aggregating the scores of all factors in patients, we could forecast the probability of PMV occurrence. Internal validation was conducted through 1000 bootstrap resamples to assess the prediction model’s performance. The model exhibited a C-index of 0.782, as determined by the bootstrap analysis. The X-axis of the calibration curve represents the predicted probability of the model, and its Y-axis represents the actual probability. The diagonal dashed line represents a perfect prediction of the ideal model. The solid line presents the performance of the nomogram, and the closer the two lines fit to the diagonal dashed line, the better the predictive consistency of the nomogram. The calibration curve (Fig. 3A) notably demonstrated good concordance between the predicted and observed outcomes. The ROC curve (Fig. 3B) also displayed an AUC of 0.782 and an accuracy of 0.884. Furthermore, DCA (Fig. 4A) provided insight into the potential clinical benefits of the prediction model, while the clinical impact curve (Fig. 4B) underscored its robust predictive capabilities and clinical utility.