This was a prospective, observational, single-center study conducted over a 1-year period in a university hospital intensive care unit (ICU). The protocol was approved by our institutional ethics committee and written informed consent was obtained from patients or their next of kin. The study included adult patients recruited from those under MV for more than 48 h who underwent a planned SBT. Patients with a tracheostomy, inadequate transthoracic echocardiographic window, atrial fibrillation, or do not reintubate orders were excluded.

Demographic data, causes of MV and illness severity (SAPS III) were recorded at ICU admission. Daily fluid balance was recorded from admission until the discharge from ICU. Weight was recorded using beds with built-in patient weighing scales (TotalCare, Hill-Rom, Batesville, Ind) at ICU admission, and subsequently, at 8 am every day during ICU stay. Patients were considered ready to undergo a SBT with a T-tube piece when they fulfilled all the classical criteria.12 At the start of SBT, the Vt and RR were measured in triplicate with a portable spirometer and the RR/Vt ratio was calculated.13 SBT was performed with the patient in a semirecumbent position, breathing through a T-tube with oxygen at a concentration prescribed by the attending physician. Patients were reconnected to MV if any of the standard criteria were met.12 SBT lasted 30–120 min, depending on patient tolerance or physician’s decision. Heart rate, invasive arterial blood pressure, and arterial oxygen saturation were continuously recorded during SBT. Clinical and echocardiography data were obtained immediately before and at the end of the SBT trial in all patients.

Weaning failure was defined as the need to reconnection to MV during SBT, or if the patient was extubated but needed mechanical ventilation (invasive or non-prophylactic non-invasive ventilation) or died within the following 48 h.5 Patients were reassessed daily and underwent a new SBT when weaning criteria were subsequently met until final extubation or death.

All echocardiography were digitally stored, and measurements were performed off-line by a single operator (LZ, with competence in advanced critical care echocardiography) blinded to the SBT results. Echocardiographic variables were measured before and at the end of SBT using a CX50 ultrasound device (Philips Healthcare, DA Best, the Netherlands). Systolic function was assessed by the monoplane Simpson method in the apical four-chamber view. Left ventricular diastolic function was assessed by Pulsed-wave Doppler measuring early (E) and late (A) diastolic wave velocities at the mitral valve, and tissue Doppler measuring early (e′) diastolic wave velocity at the external mitral annulus. All measures were averaged over a minimum of three cardiac cycles. Isolated diastolic dysfunction was defined using the 2016 European Society of Cardiology guidelines (LVEF ≥ 50% with plasma NT-proBNP > 1250 ng/L and E/e′ ratio ≥ 13 or e′ < 9 cm/s).14

Due to the observational nature of the study, a formal calculation of the study sample size was not applicable.

Categorical data are presented as number (%), while continuous variables are presented as mean (95% confidence interval) or median [interquartile range] if non-normally distributed. To compare univariate data, we employed the Student’s t, ANOVA, Mann–Whitney U, or Kruskal–Wallis tests according to data distribution. Categorical data were compared using Fisher’s exact test. Comparisons between pre-SBT and end-SBT data were evaluated using a paired Student’s t-test or a Wilcoxon signed-rank sum test, based on data distribution. Statistical significance was set at p < 0.05, with all tests being two-sided. To evaluate the effect of fluid balance and diastolic dysfunction in weaning failure, Cox regression analysis was conducted with adjustments for potential confounding variables. A first model was constructed, including factors related to weaning failure in prior studies8,9 such as diastolic dysfunction, pneumonia as the reason for intubation, the RR/Vt ratio, age, and SAPS III. Furthermore, a second model was created by adding the cumulative fluid balance from ICU admission to SBT day.

A total of 497 patients were admitted to our medical-surgical ICU during the study period (Fig. 1), of which 410 (82%) were mechanically ventilated. Among these, 347 (85%) remained under invasive MV for more than 48 h, and 221 patients underwent a planned SBT. Finally, as presented in Table 1, we obtained complete data from 89 patients (58 men, mean age 62 years (59−65), SAPS III 64 (61−68)). Weaning failed in 33 (37%) patients and was successful in 56 (63%). All baseline characteristics were similar between groups. Patients in the failure group needed more days of MV (9 [7–21] vs. 7 [4–10], p = 0.001) and ICU (12 [9–23] vs. 10 [7–16], p = 0.048) than the success group, although there were no differences between groups in days of MV before the first SBT.

Figure 1.

Study flow chart.

The flow chart shows the exclusion criteria. Weaning failure was defined as the need to reconnection to mechanical ventilation during spontaneous breathing trial (SBT), or if the patient was extubated but needed mechanical ventilation (invasive or non-invasive) or died within the following 48 h.

(0,24MB).

Table 2 shows values observed before and at the end of the SBT in each weaning group. Before the weaning process, there were no differences between groups in any of the respiratory, hemodynamic or echocardiographic variables analysed.

The partial pressure of oxygen in arterial blood decreased significantly in the failure group and was lower than in the success group at the end of the SBT (75 mmHg (67−84) vs. 87 mmHg (78−95), p = 0.025).

Both groups showed significant increases in heart rate and the pressure-rate product during the SBT. However, at the end of the SBT, systolic blood pressure (134 mmHg (126−143) vs. 125 mmHg (121−130), p = 0.032) and heart rate (96 beats/min (90−102) vs. 83 beats/min (79−87), p = 0.033) were higher in the failure group than in the success.

The prevalence of isolated diastolic dysfunction was similar between the groups before the SBT (30% in the failure group and 25% in the success group, p = 0.586). However, at the end of the SBT, there was a significant difference in the prevalence of diastolic dysfunction between the weaning failure and success group (39.3% vs. 17.8%, p = 0.025).

Transmitral E wave maximal velocity significantly increased during the SBT in the failure group (85 cm/s (76−95) vs. 93 cm/s (80−107), p = 0.042), while there was no change in the success group (85 cm/s (77−92) vs. 87 cm/s (80−94), p = 0.806). There were no significant differences between groups in the maximal velocity of the lateral e′ wave and E/e′ ratio, neither at the start nor at the end of the SBT.

Differences in fluid balances and weight between patients who experienced weaning failure and those who succeed are showed in Table 3 and Fig. 2. Patients who failed the weaning had a significantly greater cumulative positive fluid balance than successful weaning patients in the first 48 h of ICU admission (216 mL [−1151 to 2357] vs. −1495 mL [−3506 to 1346], p = 0.017). Average daily fluid balance from ICU admission until SBT was less negative in patients who failed the weaning than in those who succeeded (−648 mL [−884 to −138] vs. −893 mL [−1284 to −501], p = 0.007). However average daily fluid balance from the first SBT until the ICU discharge was more negative in the weaning failure than in the success group (−973 mL [−1493 to −201] vs. −425 mL [−1065 to 12], p = 0.034).

Figure 2.

Fluid balances according to weaning outcome.

Box-whisker plots showing fluid balances depending on the weaning outcome: success (grey bars) or failure (white bars). p Values in each group are shown (Mann–Whitney U test). #< p 0.05 compared to success group. SBT: spontaneous breathing trial. ICU-SBT day: fluid balance from ICU admission to spontaneous breathing trial day. SBT-ICU dis: fluid balance from spontaneous breathing trial day to intensive care unit discharge. ICU stay: cumulative fluid balance of entire ICU stay.

(0,2MB).

Patients who failed the weaning had no weight differences from ICU admission to SBT, while those who succeeded had a loss of body weight relative to admission in ICU (0 kg [−2.6 to 2.2] vs.−1.25 kg [−4.3 to 0.5], p = 0.015).

Cox regression analyses (Table 4) showed that RR/Vt at the beginning of the SBT was associated with weaning failure in both models, regardless of whether the fluid balance was included (hazard ratio (HR) 5.015 (95% CI: 1.434−18.53); p = 0.012) or not (HR 4.702 (95% CI: 1.344−16.453); p = 0.015). Otherwise, in the model 2, with fluid balance included, age (HR 1.03 (95% CI: 1.005–1.05); p = 0.019) and diastolic dysfunction (HR 2.55 (95% CI: 1.123−5.812); p = 0.012) were associated with weaning failure together with fluid balance from ICU admission to SBT (HR 1.02 for each 250 mL of positive fluid balance (95% CI: 1.009−1.032); p < 0.001).

This study has several limitations. The first limitation is related to the observational nature of the study, it is important to note that our results only indicate an association rather than causality. Second, the lack of a gold standard definition of diastolic dysfunction complicated the analysis of our data. Third, no echocardiographic data were recorded in the consecutive weaning trials in patients failing the first SBT. Fourth, all the patients underwent the SBT through a T-tube with oxygen, no other modalities of SBT were studied. Finally, this study was conducted in a single center and our findings might not apply to other ICUs with different fluid or weaning protocols.

Nonetheless, the main strengths of our study include its prospective design, the meticulous compilation of daily fluid balances from ICU admission to ICU discharge, and the detailed echocardiographic assessment at the beginning and the end of the SBT.

Weaning failure due to diastolic dysfunction is highly related to fluid balance, and the deleterious effect of fluid balance on diastolic function is associated with age. Therefore, future investigations of diastolic dysfunction in the context of weaning failure should consider the role of fluid balance. Continuous and comprehensive monitoring of fluid status is essential in ICU patients to initiate de-resuscitation or active fluid removal as soon as possible.