Clinical prediction scores predicting weaning failure from invasive mechanical ventilation: Role and limitations

Abstract

Invasive mechanical ventilation (IMV) has become integral to modern-day critical care. Even though critically ill patients frequently require IMV support, weaning from IMV remains an arduous task, with the reported weaning failure (WF) rates being as high as 50%. Optimizing the timing for weaning may aid in reducing time spent on the ventilator, associated adverse effects, patient discomfort, and medical care costs. Since weaning is a complex process and WF is often multi-factorial, several weaning scores have been developed to predict WF and aid decision-making. These scores are based on the patient's physiological and ventilatory parameters, but each has limitations. This review highlights the current role and limitations of the various clinical prediction scores available to predict WF.

Key Words: Clinical scores; Invasive mechanical ventilation; RSBI; Weaning; Weaning failure

Core Tip: Delay in weaning from invasive mechanical ventilation or weaning failure (WF) may increase patient mortality, morbidity, risk of secondary infections, length of hospital/ICU stay and healthcare costs. Physician’s ability to predict successful weaning has been shown to have low accuracy, with poor positive and negative predictive values. As the pathophysiology of WF is complex and multifactorial, a single parameter may not suffice to predict successful weaning. Hence, several clinical scores, encompassing multiple patients and ventilatory factors have been devised to predict WF. However, none of the current scores is ideal and each have their inherent limitations.

INTRODUCTION

Weaning is an essential component in the care of critically ill patients on invasive mechanical ventilation (IMV). Weaning refers to the process of liberating the patient from IMV and removing the endotracheal tube[1]. Hence, patient cooperation is imperative, especially during the recovery phase from critical illness, to achieve successful weaning.

Delay in weaning hampers patient care and increases the risk of developing complications, including prolonged intensive care unit (ICU) stay, ventilator-associated pneumonia, ICU delirium, diaphragmatic atrophy, venous thromboembolism and eventually the cost of therapy[2-6]. Also, patients needing prolonged ventilation consume more ICU resources, up to nearly 37%[7]. On the other hand, premature extubation may place additional stress on the cardio-respiratory systems, leading to weaning failure (WF). Interestingly, about 40%-50% of the duration of ventilation is spent in the weaning process[8-11]. Further, studies have also shown that mechanical ventilation duration is directly proportional to mortality and shares a linear relationship[11-13].

Assessment for readiness to wean should be done daily in order to shorten the weaning process. This has shown to be an independent predictor for successful weaning and overall survival of the patient[14,15]. However, objective criteria for assessing readiness to wean are not ideal, each having their inherent limitations.

Several factors have been reported to be associated with higher WF rates. As the pathophysiology of WF is complex and multi-factorial, a single parameter may not suffice to predict successful weaning. Physician's ability to predict successful weaning has shown low accuracy, with poor positive (50%) and negative (67%) predictive values[16]. Hence, multiple scores have been devised, incorporating both respiratory and non-respiratory parameters, to accurately identify patients at risk for WF. However, every score has limitations and must be tested in different patient populations and geographical areas and over time to ensure their validity. To date, no single best score has been defined to predict WF. However, knowledge of the various scores to predict successful weaning or failure helps better understand the pathophysiology of the weaning process and what factors play a significant role. Most scores have been developed considering respiratory, ventilatory, and cardiovascular parameters. However, WF could also be secondary to neurological issues, wherein airway reflexes may be hampered, and extubation may be difficult. It is imperative to understand that no ideal weaning score has been developed to date, and different weaning scores may apply to different patient subsets to achieve a successful outcome. In some cases, combining different parameters or scores may be more helpful. This review highlights the current role and limitations of the various clinical prediction scores available to predict WF.

TERMINOLOGY

Readiness to wean

Testing for readiness to wean is to determine whether a patient can be weaned from IMV. Objective clinical criteria are usually used to assess whether a patient can be weaned from the ventilator. In some cases, additional physiological parameters may be considered to decide on the initiation of weaning[1,14,15].

Weaning

Weaning is the process of decreasing ventilator support so that the patient can assume a greater proportion of their own ventilation, either gradually by reducing ventilator support or spontaneous breathing trials (SBTs)[1].

Spontaneous awakening trials

Spontaneous awakening trials (SATs) refers to the daily interruption of sedative or narcotic medications to assess for wakefulness and alertness. This is tested objectively by eye-opening to verbal commands, following simple commands, or having a Sedation-Agitation Scale score of 4–7 or a Richmond Agitation and Sedation Scale of –1 to +1.

As per the Pain, Agitation, Delirium, Immobility, and Sleep disruption guidelines, daily sedation interruption is recommended for titrating sedatives in the critical care unit[17].

A study conducted by Kress et al[18] compared daily interruption of sedation to physician-driven titration or discontinuation of sedatives. They concluded that daily interruption of sedation was associated with a shorter duration of mechanical ventilation (> 2 days), a shorter length of ICU stay (> 3.5 days) and overall better ICU outcomes[18].

It has also been found that achieving deeper levels of sedation is associated with delay in extubation, longer time on mechanical ventilation and overall poor ICU outcomes[19,20]. Deep sedation (27.2%, n = 513) is associated with an in-hospital mortality hazard ratio of 1.661 [95% confidence interval (CI): 1.074-2.567; P = 0.022] and a 2-year hazard ratio of 1.866 (95%CI: 1.351-2.576; P < 0.001)[19,20]. These studies highlight the need for sedation protocols to achieve light sedation as it reduces ventilator days and improves mortality.

SAT also forms an integral part of weaning from IMV. The Awakening and Breathing controlled trial[21], also known as the Wake up and Breathe study, compared daily SAT followed by SBT with the usual sedation protocol and SBT alone. The authors observed that patients receiving SAT followed by SBT had more ventilator-free days, shortened hospital and ICU stay length, and improved 1-year survival[21]. Hence, the coordination of SAT followed by SBT plays an integral role in weaning, forming a core feature of the ABCDEF bundle[22].

SBT

SBT is conducted in patients who fulfil the readiness to wean criteria. It can be conducted with the patient on an invasive ventilator on spontaneous breathing modes [continuous positive airway pressure (CPAP)] and minimal ventilator support [positive end-expiratory pressure (commonly known as PEEP) up to 5 cm H₂O and pressure support of < 7 cm H₂O). Another method is disconnecting the patient from the ventilator and using an oxygen device such as a T-piece. SBT usually lasts for 30 min to 120 min. Once the patient passes the SBT, extubation can be planned.

Weaning success

It is defined as the extubation of the patient and the absence of ventilatory support for 48 h post-extubation[1].

WF definition

(1) Failure of SBT; (2) Need for reintubation within 48 h of extubation; or (3) Death within 48 h following extubation[23,24].

SBT failure is defined by objective and subjective criteria, as mentioned in Table 1[23-25]. Various studies have reported different rates of WF. The overall incidence of WF after a single SBT is between 26% and 42%[24,26]. This range is probably because of varied definitions of WF or differences in the population base used in different studies.

Table 1 Criteria for failure of spontaneous breathing trial.

Objective criteria	Subjective criteria
Tachycardia; tachypnoea; hypertension or hypotension; hypoxemia; acidosis; arrhythmias	Agitation or distress; altered or depressed mental status; sweating; increased work of breathing

Extubation

Extubation refers to removing the endotracheal tube and is the final step in liberation from IMV. Extubation can be performed only when a patient has a patent airway with good airway reflexes. Extubation failure is associated with a high mortality rate. Excessive secretions, arterial PaCO₂ > 45 mmHg, duration of IMV > 72 h, upper airway disorders and failed weaning trials are predictors for failed extubation[26-28].

Weaning in progress

Refers to a patient who has been extubated and put on non-invasive ventilator (NIV) support. These patients signify an intermediate category where they do not need NIV but need respiratory support in the form of NIV.

Based on the difficulty and length of the weaning process, intubated patients can be divided into three categories (Table 2).

Table 2 Classification of patients on the basis of weaning process.

Simple weaning	Initiation of weaning to successful extubation in the first attempt without encountering any difficulty
Difficult weaning	Failed initial weaning and requirement of up to 3 SBTs or as long as 7 days from the first SBT to achieve weaning success
Prolonged weaning	Failed at least 3 weaning attempts or require 7 days or more of weaning after the first SBT

SBT: Spontaneous breathing trial.

The overall prognosis of patients in the simple weaning group is good, with an ICU mortality of about 5%. However, patients in the difficult-to-wean or prolonged weaning group have a high mortality rate of 25%[24,26].

CAUSES OF WF

All patients who fail the weaning trial should be evaluated for potential reversible causes of WF. The causes of WF are broadly divided as follows: Respiratory factors[29-32]; cardiac factors[33-35]; neuromuscular factors[36-42]; critical illness neuromyopathies[43-45]; and psychological factors, and metabolic, nutrition and endocrine disorders[41,46,47] (Table 3).

Table 3 Causes of weaning failure.

Respiratory factors[29-32]	Increased airway resistance - bronchospasm, excessive secretions, ET block, kinks, blood in ET Decrease compliance - pulmonary edema, fibrosis, atelectasis, acute respiratory distress syndrome (ARDS) Ventilator induced lung injury - ventilator associated pneumonia, pneumothorax, hemothorax Increased work of breathing - dynamic hyperinflation, patient - ventilator dyssynchrony/asynchrony
Cardiac factors[33-35]	Pre-existing cardiovascular disorders – ischemic heart disease, valvular heart disease, pericardial diseases; stress cardiomyopathy; unresolved primary systemic disease
Neuromuscular factors[36-42]	Primary neuromuscular disorders
	Myasthenia gravis
	Guillain-Barre syndrome
	Myopathies
	Peripheral neuropathies
	Secondary causes
	Ongoing sedatives
	On neuromuscular paralyzing agents
	Critical illness neuromyopathies
	Ventilator induced diaphragmatic dysfunction
	Long term corticosteroid use
Neuropsychological[43-45]	ICU delirium; anxiety
Metabolic & endocrine[41]	Dyselectrolytemias; hypokalaemia, hypomagnesemia, hypophosphatemia; hypo/hypernatremias; hypo/hyperglycaemia; hypothyroidism; hypoadrenalism; long-term corticosteroid use
Nutrition[46,47]	Pre-existing malnutrition; underweight; Inadequate calorie intake; refeeding syndrome

ARDS: Acute respiratory distress syndrome; ET: Endotracheal tube; ICU: Intensive care unit.

READINESS TO WEAN

Since prolonged mechanical ventilation can lead to significant morbidity and mortality, every patient on IMV should be assessed for weaning daily[14,15]. The first step is to assess readiness to wean, followed by the actual weaning process, which includes an SBT to determine the likelihood of successful extubation. Various studies have identified multiple criteria which can be used to assess readiness to wean. These criteria are broadly classified into clinical and objective criteria. The most important and foremost criterion for weaning is that the primary disease, which leads to respiratory compromise and the need for IMV, is improving. Weaning may fail until there is no resolution of the acute phase of the disease. The other criteria used to assess for weaning are mentioned in Table 4. These criteria should be considered a relaxed recommendation rather than deliberations for weaning.

Table 4 Criteria to assess for readiness to wean.

Clinical criteria	Objective criteria
Primary disease for which patient needed IMV is improving; good cough reflex; absence of secretions	Respiratory; SaO2 > 90% with FiO2 < 0.4 (Pf > 150), PEEP < 8 cm H2O; RR < 35/min; MIP < 20-25 cm H2O; Vt > 5 mL/kg on SBT; VC > 10 mL/kg; No acidosis. Cardiovascular; HR < 140/min; SBP 90-160 mmHg with minimal to no vasopressors; neurological; conscious or adequate mental status with good airway reflexes

FiO₂: Fraction of inspired oxygen; IMV: Invasive mechanical ventilation; SBT: Spontaneous breathing trial; HR: Heart rate; MIP: Maximum inspiratory pressure; PEEP: Positive end-expiratory pressure; RR: Respiratory rate; SaO₂: Oxygen saturation of arterial blood; SBP: Systolic blood pressure; Vt: Tidal volume; VC: Vital capacity.

WEANING SCORES AND INDICES

A summary of various scores employed for predicting WF is given in Table 5.

Table 5 Summary of various scores for predicting weaning failure in patients receiving invasive mechanical ventilation.

Score	Components	Value	Advantages	Disadvantages
A: RSBI	Respiratory rate; Vt; RSBI = RR/Vt	< 105 – weaning success; > 105 – weaning failure	Simple; easy to calculate	Affected by multiple factors – fever, infection, anxiety, restrictive disorders, weaning technique and suctioning; not the best predictor for COPD/neuro-medical and neurosurgical pt.
B: CROP index	Compliance (thoracic); respiratory rate; oxygenation (Arterial); pressure (P1 Max)	> 13 mL/breath/min – successful weaning	Incorporates respiratory and ventilatory parameters	Complex calculations; not tested in neuro-medical/neurosurgical pt.
C: ExPres score	RSBI; dynamic lung compliance; days of IMV; GCS; muscle strength; haematocrit; creatinine; neurological comorbidity	Score; > 59 – extubate pt.; 45-58 with no risk factors – extubate pt.; 45-58 with risk factors – extubate to NIV; < 44 – extubation failure	More robust score; simple tool; shown to reduce extubation failure rates	Needs large scale studies for validation
D: HACOR	Heart rate; acidosis; consciousness; oxygenation; respiratory rate	Score > 5 – weaning failure	Simple bedside tool	Limited studies; large scale trials required to assess for validity
E: WEANS NOW	Weaning parameters; endotracheal tube; ABG; nutrition; secretions; neuromuscular blocking agents; obstructive airway disease; wakefulness	Score of 1 or more – weaning failure	Incorporates multiple parameters	Very complex score; widespread application may be difficult
F: BWAP	Pulmonary; physiological; psychological	Score > 50 – weaning success	Comprehensive weaning checklist	Limited literature; modified-BWAP may be more practical
G: Morganroth scale	Adverse factor score; ventilator score	Score < 55 – successful weaning; range – 27 variables max score 75	Can be applied to pt. requiring short-term and long-term IMV	Small study; multiple variables considered
H: Persian weaning tool	Respiratory; cardiovascular; general status of pt.	26 parameters; range 26-75; score > 50 –readiness to wean	Comparable to BWAP score	Limited literature regarding its utility
I: Gluck & Corgian score	RSBI; ratio of dead space/Vt; static lung compliance; airway resistance; CO2 pressure	Score < 3 – weaning success; score of 3 – not equivocal; > 3 failure to wean	Simple tool; easy bedside measurement	Small study; large trials warranted

BWAP: Burns wean assessment program; CO₂: Carbon dioxide; COPD: Chronic obstructive pulmonary disease; GCS: Glasgow coma scale; IMV: Invasive mechanical ventilation; NIV: Non-invasive ventilation; pt.: Patient; RR: Respiratory rate; RSBI: Rapid shallow breathing index; Vt: Tidal volume; ABG: Arterial blood gas.

Rapid shallow breathing index

Rapid shallow breathing index (RSBI) is a simple tool to predict successful weaning in patients on IMV. It is one of the most commonly used indices for predicting liberation from IMV. RSBI is calculated by dividing the respiratory rate by the patient's tidal volume (Vt). In 1991, Yang and Tobin were the first to describe RSBI in a prospective study of 100 patients[48]. In their study, they reached a cut of 105 for RSBI. They concluded that an RSBI of > 105 breaths/L/min was associated with WF and a value of < 105 breaths/L/min was associated with successful weaning and extubation with a sensitivity of 97% and a specificity of 64%[48]. The calculated positive and negative predictive values were 78% and 95%, respectively. Studies conducted by Epstein et al[49] and Jacob et al[50] have also corroborated similar findings on RSBI. A study by Frutos-Vivar et al[51] also identified RSBI as one of the best predictors of extubation failure.

However, there is much scepticism about RSBI and its widespread use. Though a simple tool, RSBI values can be affected by secondary factors such as fever, infection, patient position, anxiety disorders, restrictive lung diseases, any other pre-existing disorders, small size endotracheal tube, female sex, weaning technique and suctioning[52-54].

Additionally, RSBI may not be the best predictor for weaning in specific patient populations such as chronic obstructive pulmonary disease (COPD) and neuro-medical/ neurosurgical patients[55-57]. In COPD patients’ inspiratory efforts may be ineffective in triggering the ventilator, which could lead to falsely low RSBI values. Hence, RSBI may not predict a successful outcome. This was confirmed in a study by Purro et al[55], where they found that about 56% of patients with COPD with an RSBI of < 80 failed weaning trials. Similarly, neurological patients are usually intubated for airway protection and not for a primary lung disease. Hence, RSBI, which incorporates respiratory parameters, may not be an accurate index for predicting weaning success. Hence, studies have shown that RSBI is not a good predictor of weaning outcomes in neurosurgical and traumatic brain injury patients[56,57].

An important technical consideration in calculating RSBI is the ventilator setting used to calculate it. Medical centres worldwide use diverse settings to calculate RSBI, so variations in values may be seen. Studies have shown that RSBI calculated in CPAP or pressure support ventilation (referred to as PSV) mode is lower than the value calculated using SBT with a T-piece[58].

The threshold value of 105 breaths/L/min for RSBI has also been questioned. A randomized controlled trial on 208 patients reported that the cut-offs for RSBI should differ for patients weaned on PSV and T-piece[59]. They concluded that the threshold values of RSBI should be 75 breaths/L/min for PSV and 100 breaths/L/min for T-piece with a diagnostic accuracy of 87% and 82%, respectively, suggesting more accurate values as compared to the traditional cut-off of 105 breaths/L/min[59]. A similar study was done by Danaga et al[60] to evaluate the diagnostic accuracy of RSBI in predicting extubation failure and to assess the traditional cut-off value of 105 breaths/L/min. They conducted a prospective trial on 73 patients and found that the traditional cut-off offered a sensitivity of only 20% with a specificity of 95%. As per the receiver operating characteristic (commonly known as ROC) curve analysis, a cut-off of 76.5 showed a higher sensitivity of 66% with a specificity of 74%. Hence, they concluded that the classical value was inadequate to predict extubation failure[60].

A recent systematic review and meta-analysis assessing the predictive value of RSBI included 79 studies involving 13170 patients with IMV[61]. They reported a pooled sensitivity of 0.6 (95%CI: 0.59-0.61), specificity of 0.68 (95%CI: 0.66-0.70) and the area under the receiver operating characteristic (commonly known as AUROC) curves area of 0.8144. They concluded that RSBI was moderately accurate, with poor pooled sensitivity and specificity in predicting successful extubation[61].

In conclusion, RSBI is a time-tested tool for predicting weaning outcomes. However, the interpretation of RSBI should also consider other factors, such as patient comorbidities, the reason for IMV, and ventilator settings. RSBI, though widely used, cannot be universally applied to all patient populations and may serve as a supplement to the clinical decision-making process.

Serial RSBI

Some authors evaluated the utility of serial RSBI, stemming from the observation that RSBI measured at the beginning of SBT may be normal but may worsen subsequently. This finding could be attributed to fatigue or poor lung mechanics, which may not be evident at the beginning of the trial. Several studies have shown that RSBI measured at 30 min of SBT, the end of SBT, or serial measurements of RSBI were a better predictor of weaning[54,62,63]. Contradictory results were shown by Shah et al[64], who found that serial RSBI measurements were not significantly different and held no statistical significance. Also, serial measurements could not detect extubation failure; hence, its use is questionable[65].

RSBI rate

It has been hypothesized that, as respiratory failure is a dynamic phenomenon, change in RSBI expressed as a rate would be more predictive of weaning success or failure. This concept of RSBI rate, defined as the change in RSBI over time, was tested in a prospective observational study by Segal et al[66] who measured RSBI every 3 min during the SBT in 72 patients on IMV. They identified a threshold value of 20%. They found that RSBI rate < 20% had a sensitivity and specificity of 90.4% and 100%, respectively, with a positive predictive value of 100% and a negative predictive value of 81%. They concluded that the percent change of RSBI during SBT was a better predictor of successful extubation than a single value of RSBI[66].

CROP index

The CROP index uses four parameters, namely thoracic compliance, respiratory rate, arterial oxygenation, and P1max to predict WF[48]. It was developed by Yang and Tobin as another weaning tool to be compared to RSBI to assess weaning outcomes[48]. In a Chinese study by Li et al[67] CROP index was studied and evaluated to predict the success of weaning in patients with acute exacerbation of COPD on IMV. They included 215 patients and achieved weaning success in 182 patients. A CROP value of > 13.521 mL/breath/min had a sensitivity of 87.9% and specificity of 91.9% in predicting successful weaning and extubation. The positive and negative predictive values were 0.97 and 0.58, respectively, with an odds ratio < 1, making it a strong predictor for successful weaning[67].

Generally, a CROP index above 13 has been shown to be associated with successful weaning[63]. However, CROP is a less studied and evaluated index for weaning success in view of its complexity of calculation, and RSBI, a simpler alternative, is widely used. Also, CROP only considers respiratory parameters and non-respiratory parameters are not considered. The validity of CROP in weaning patients suffering from neurological and neurosurgical problems has yet to be studied, warranting further large-scale trials for the same.

Extubation predictive score

The Extubation Predictive Score (ExPreS) score, was developed by Baptistella et al[68] to predict the likelihood of weaning success. The authors wanted to combine the respiratory and non-respiratory parameters since WF is multi-factorial and not confined to respiratory parameters. They studied 110 patients who underwent extubation with success in 101 patients (91.8%). Their primary outcome was successful extubation at 48 h. ROC were analysed for the parameters which showed statistically significant association. The AUROC values were used to identify the parameters for inclusion in the ExPreS score. The following parameters, namely RSBI during SBT, dynamic lung compliance, duration of IMV, muscle strength, Glasgow coma score, haematocrit, and serum creatinine, were significantly associated with extubation outcome. With these parameters along with neurological comorbidity, they formulated the score with the total values indicating low (57.1%), intermediate (88.3%), and high (98.7%) probability of extubation success based on sensitivity and specificity. Patients with a score > 59 were extubated, and those with a score of < 44 were continued on IMV. Those with scores between 45–58 and no risk factors for extubation were extubated, while those with risk factors (COPD, obesity, cardiomyopathy) were extubated to NIV. ExPreS was a more robust score, simple and easily applicable to most of the population. ExPreS has been shown to decrease the extubation failure rate from 8.2% to 2.4%. However, more studies are required to assess its validity[68].

HACOR score

The HACOR (Heart rate, Acidosis, Consciousness, Oxygenation and Respiratory rate) score was initially proposed and developed by Duan et al[69] to formulate a simple bedside tool to predict NIV failure in acute hypoxemic respiratory failure patients. The authors concluded that a HACOR score of > 5 predicted a high risk of NIV failure and the need for intubation and IMV[69]. Later, they studied the validity of the HACOR score to predict NIV failure in COPD patients and reported it to have good predictive power for determining NIV failure in this patient subset, too[70]. Subsequent studies also assessed its utility in predicting NIV failure in different patient subgroups, and HACOR < 6 was also shown to predict high-flow nasal oxygenation success[71-73]. Since the original HACOR score did not take baseline variables such as the presence of pneumonia, cardiogenic pulmonary oedema, acute respiratory distress syndrome, immunosuppression, septic shock and the Sequential Organ Failure Assessment (commonly referred to as SOFA) score, an updated HACOR score was proposed by Duan et al[74] to study the validity and predictive power for NIV failure. It was found to have a higher predictive power than the original score.

The role of HACOR in predicting WF was initially studied by Chaudhuri et al[75] in 120 patients. One-time HACOR score was calculated at 30 min of a 120-min SBT trial to predict WF, and they found that all the variables of HACOR were statistically and significantly different between successful and failed weaning groups. The HACOR score had a sensitivity of 83.8% and a specificity of 96.4% with an AUC of 0.95%. They concluded that a HACOR score > 5 predicted WF and may be used to assess weaning patterns[75]. The studies of HACOR scores in predicting WF are limited and more clinical trials may be required to assess their validity. Studies may also be needed to assess the validity of the updated HACOR score as they consider baseline parameters that are missing from the original HACOR.

WEANS NOW score

The WEANS NOW score was developed by Lin et al[76] in 2020. They considered the following variables to develop the score: Weaning parameters; endotracheal tube; arterial blood gas analysis; nutrition; secretions; neuromuscular-affecting agents; obstructive airway problems; and wakefulness. Lin et al[76] conducted a retrospective study on 205 patients with acute respiratory failure and concluded that a WEANS NOW score of 1 or more and the need for prolonged IMV of > 21 days was associated with failure of extubation. To calculate weaning parameters, maximum inspiratory pressure, maximum expiratory pressure, spontaneous Vt, minute ventilation, and RSBI are required. As a result, it is a very complex score, and its widespread application across multiple ICUs may be complex.

Nayak et al[77] compared ExPreS, HACOR, and WEANS NOW scores to evaluate the best score for weaning prediction. They concluded that HACOR (score > 5) was comparable to ExPreS (score > 69) in predicting weaning with 70% accuracy. However, calculating the WEANS NOW score took a lot of work and effort to apply.

Burns wean assessment program

Burns wean assessment program (BWAP) is a scoring instrument designed and developed to reduce the variability in the clinical management of mechanically ventilated patients[78]. It consists of 26 factors recorded within 24 h of the weaning trial. A score > 50 was associated with successful weaning (P = 0.001)[79]. Overall, it was a simple score that could be measured quickly and included various parameters related to the patient's pulmonary, physiological and psychological conditions. Jiang et al[80] developed a modified version of the BWAP score (mBWAP) and used it to predict weaning outcomes in patients requiring long-term mechanical ventilation of > 21 days[80]. Successful extubation was achieved in 78.5% of the 527 patients. The mBWAP scores were higher in the successfully extubated patients with a sensitivity and specificity of 81.4% and 82.1%, respectively, when a cut-off value of > 60 was used.

Similarly, Jeong and co-workers studied the clinical application of the score in 103 patients on IMV[81]. They concluded that higher mBWAP scores were associated with successful weaning after the first SBT, with a reported sensitivity of 76% and specificity of 64%. Another study by Abdelaleem et al[82] reported that weaning success was associated with higher m-BWAP scores (mBWAP > 55) with a sensitivity and specificity of 73.77% and 84.85%, respectively. Overall, the mBWAP score has shown to be a better predictor of weaning success than other traditionally used scores.

Morganroth scale

In 1984, Morganroth et al[83] developed a scale to assess for weaning from IMV. This scale consists of the “Adverse factor score” and “Ventilator score”. Both the scores together consisted of 27 variables and the total score was the sum of the individual variables. The maximum possible score is 75, with a score < 55 suggestive of successful weaning with a sensitivity and specificity of 93% and 86%, respectively. However, the study included only 11 patients on prolonged mechanical ventilation[83]. Other studies have also reported similar results using the Morganroth scale[84].

Persian weaning tool

Persian weaning tool (referred to as PWT) is a nationalized protocolized weaning tool used in Iran to assess readiness to wean in patients on IMV[85]. It evaluates three major areas, namely respiratory, cardiovascular and general status of the patients. A total of 26 parameters are assessed, with the lowest and highest scores being 26 and 75, respectively. Each parameter is scored from 1 to 3 depending on the status of the patient, i.e. score 1 (critical condition needing intervention), score 2 (condition necessitating care but no major intervention) and score 3 (patient appropriate condition for the parameter). A 'not applicable' option is also available for the parameter if a clear, definitive answer is unavailable. A score of > 50 indicates readiness to wean[85].

Bazrafshan et al[86] tested the validity of PWT and compared it to BWAP and Morganroth's criteria for weaning. They reached a cut-off point 57 with a sensitivity and specificity of 0.679 and 1, respectively. With this cut-off point, they showed that there was a statistically significant correlation between PWT and BWAP (P < 0.05) with no difference in identifying patients for readiness to wean (P = 0.453)[86].

Gluck and Corgian scoring system

In 1996, Gluck and Corgian developed a scoring system to determine if a patient's ability to be weaned can be predicted at the time of admission[87]. They conducted the study on adult patients with long-term IMV of more than 3 weeks with no active sepsis or neurological issues. Out of the multiple parameters they evaluated, 5 were significant: RSBI; the ratio of dead space to Vt; static lung compliance; airway resistance; and CO₂ pressure. A score > 3 was associated with failure to wean, a score of 3 was not predictive and a score < 3 was associated with weaning success. The score demonstrated sensitivity, specificity, and positive and negative predictive values of 1.0, 0.91, 0.83, and 1.0, respectively[87]. The parameters evaluated were simple and easily measurable at the bedside. However, the study was conducted on only a small sample size of 20 patients. Bruton et al[88] evaluated this score and concluded that this scoring system did not show any false negative results but had a few false positive results.

NEWER MODALITIES TO PREDICT WEANING FROM IMV

Ultrasound-based indices

Bedside ultrasonography has become integral to modern ICUs. It is increasingly employed for diagnosis, therapeutic interventions and even prognostication in critically ill patients. As the diaphragm is the major respiratory muscle responsible for 60%-80% of the respiratory workload, assessing its function to predict weaning success seems prudent. The most commonly used diaphragmatic parameters, assessed by bedside ultrasonography, are the diaphragmatic excursion (DE, measuring the distance it moves during the respiratory cycle) and the diaphragm thickening fraction (DTF, variation in its thickness during a respiratory effort). Other diaphragmatic parameters evaluated include diaphragm thickness (depicting atrophy) and the thickening fraction (representing diaphragm inspiratory effort)[89]. Evaluation of other respiratory muscles using ultrasonography has also been done to evaluate readiness to wean. Ultrasonic measurement of parasternal intercoastal muscle thickness is an effective marker for assessing extubation failure[90].

Ultrasound may also help assess the fluid status by assessing the inferior vena cava diameter, pleural/pericardial effusion, and extravascular lung water. Lung ultrasound score has also been shown to accurately predict successful weaning[91]. Further, integrating different ultrasound variables (cardiac, lung and diaphragmatic) may provide a more comprehensive approach to predicting WF[92,93]. However, the high cost and availability of the equipment, the need for operator training, and logistical constraints in conducting ultrasonography in critically ill patients may restrict its widespread utility.

Artificial intelligence

Like with many other healthcare fields, the role of artificial intelligence (AI) is currently being evaluated in the weaning of mechanically ventilated patients. Early reports have suggested that AI-based algorithms may correctly identify the right time to wean patients off IMV[94]. Recent studies have also shown that machine learning algorithms can accurately predict chances of successful weaning even before intubation[95]. This may help rationalize intubation and aid in patient triaging. However, these AI-based tools need to be tested in prospective large-scale trials and compared with other established tools before they are incorporated into our routine clinical practice.

FUTURE PERSPECTIVES

The presently available clinical scores may not be ideal, but they serve as a valuable guide to determine readiness to wean and predict chances of WF. However, large-scale multi-centre randomized control trials are required to evaluate their utility in different ICU populations and assess their efficacy over time. Further, comparative studies are required to compare the accuracy of different scores encompassing different patient and ventilatory factors. Using bedside ultrasonography and AI-based algorithms may enable a more patient-centric approach to wean from IMV and help improve clinical outcomes. Their integration with already available clinical scores may open up newer research and clinical care avenues.

CONCLUSION

To conclude, the intensivist must start preparing for weaning as early as possible while trying to identify and mitigate the risk factors associated with WF. As multiple demographic, physiological and clinical parameters may be responsible, the use of clinical scoring systems encompassing multiple factors may be beneficial in identifying patients at higher risk for WF. RSBI still retains its utility due to ease of application and repeatability. HACOR score, although not originally developed to predict WF, is a promising tool for further research in patients on IMV. However, identifying a single bedside clinical score, which is easy to compute and can be applied to different patient populations, requires large-scale comparative studies. Bedside ultrasonography may provide a more clinical approach to identifying the risk of WF in a particular patient. In the future, using AI-based protocols may provide a more customized approach per the patient's requirements.