Disease specific thresholds for determining extubation readiness: The optimal negative inspiratory force for chronic obstructive pulmonary disease patients.

OBJECTIVES: The negative inspiratory force (NIF) has been used to help clinicians predict a patient's likelihood of successful liberation from mechanical ventilation (MV). However, the utility of the traditional threshold of ≤-30 cmH 2O may not be appropriate for patients with chronic obstructive pulmonary disease (COPD). This study aims to define the optimal predictive NIF threshold for COPD patients.
METHODS: A prospective-observational multi-center study was conducted in intensive care units of six academic medical centers. All patients had COPD and were intubated for hypercapnic respiratory failure. The process of weaning from MV was conducted according to the defined hospital protocol. NIF was measured after 120 min of spontaneous breathing trial (SBT). The sensitivity, specificity, positive, and negative predictive value (PPV, NPV), positive and negative likelihood ratios (LR+, LR-) were calculated, and the diagnostic accuracy recorded.
RESULTS: A total of 90 patients with COPD (39 males and 51 females) were included. Of these, 43 patients (47.8%) were successfully extubated whereas 47 patients (52.2%) failed SBT or required re-intubation (P = 0.654). The threshold value of ≤-25 cmH2O offered the optimal performance in COPD patients: area under the receiver operating characteristic (ROC) curves ROC curves 0.836, sensitivity 95.0%, specificity 86.0%, PPV 84.4%, and NPV 95.6%., LR+ 6.79, LR- 0.06, and the diagnostic accuracy 90.7%.
CONCLUSIONS: In mechanically ventilated COPD patients with hypercapnic respiratory failure, the NIF threshold of ≤-25 cmH2O was a moderate-to-good predictor for successful ventilator liberation, and outperforms the traditional threshold of ≤-30 cmH2O. Copyright:

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is a major cause of global morbidity and mortality, affecting 251 million people worldwide.[123] COPD ranked 8th in terms of disability-adjusted life years globally from 1990 to 2015.[34] It is a leading cause of death in the US and is projected to be the third leading cause of death globally by 2020.[56]

COPD is a chronic lower respiratory disease characterized by persistent respiratory symptoms and chronic airflow limitation due to small airways disease and parenchymal destruction.[7] In patients with appropriate symptoms and significant exposures to noxious stimuli, the diagnosis of COPD is made with formal spirometry and is defined by The Global Initiative for Chronic Obstructive Lung Disease (GOLD) schema as the presence of a postbronchodilator forced expiratory volume in one second (FEV1) to forced vital capacity ratio <0.70.[7] Disease severity (GOLD criteria) is graded using FEV1as a percent of predicted.[78]

COPD may be punctuated by periods of acute worsening of respiratory symptoms termed exacerbations, which account for the majority COPD-associated mortality and health-care spending.[5] Up to 80% of COPD exacerbations are caused by respiratory infections, with viral infections generally inducing worse hypoxemia, longer hospital length-of-stay (LOS), and greater deterioration of lung function than bacterial infections.[910111213] Exacerbations may lead to acute respiratory failure requiring noninvasive ventilation (NIV) or, in more severe cases, invasive mechanical ventilation (IMV). Although various prognostication tools have attempted to predict IMV needs,[14] this remains largely a clinical decision. IMV is associated with a myriad of complications for patients, including increased exposure to nosocomial infections,[15] barotrauma and ventilatory-induced lung injury,[1617] hemodynamic compromise,[1819] diaphragmatic and limb muscle weakness,[202122] organ dysfunction,[232425] neurological and psychological sequela,[2627] protracted hospital and intensive care unit (ICU) LOS, and increased mortality.[2829303132] These risks increase with IMV duration. As such, a timely assessment of extubation readiness and ventilatory weaning is of utmost importance.[28] That said, premature weaning and failed extubation are not without their complications, including increased mortality, higher rates of tracheostomy dependence, and increased ICU and hospital LOS, and higher transfer rates to long-term-care facilities.[333435]

Given the importance of accurate assessment of extubation readiness, a multitude of factors have been studied for the prediction of extubation success including clinical gestalt, tidal volume (VT), respiratory rate (RR), minute ventilation, rapid shallow breathing index (RSBI; defined as RR/VTbreaths/min/L), and negative inspiratory force (NIF). NIF, also known as negative inspiratory pressure or maximal inspiratory pressure, is a commonly used clinical parameter to assess extubation readiness by assessing diaphragmatic strength during the inspiratory phase.[36] This metric is simple to obtain at the bedside. An early study indicated that a NIF of ≤−30 cmH2O predicted successful weaning, whereas ≥−20 cmH2O predicted weaning failure.[37] Unfortunately, subsequent studies did not support these thresholds due to high false-positive and false-negative results.[38] As studies across heterogeneous populations have shown, there is not a single threshold that will predict extubation success; however, its utility for homogeneous subpopulations remains unclear and more granular data is needed.[3639] The objective of this study was to determine if in COPD patients on IMV ≥24 h (population), what NIF threshold performs best to identify patients at the highest likelihood of successful extubation (primary outcome).

METHODS

A prospective observational study was conducted in the closed mixed medical-surgical ICUs at six academic medical centers in a resource-limited setting from January 1, 2013, to February 28, 2017. The study was approved by the Investigational Review Board at Baqiyatallah University of the Medical Sciences (340/5/5904) and Shahid Beheshti University of the Medical Sciences (SBMU1/REC/1393/89), and these approvals were accepted by each of the enrolling centers: Baqiyatallah Hospital, Shariati hospital, Loghman Hakim Hospital, Besat Hospital, Masih Daneshvari Hospital, and Talaghani Hospital. All study parts were reviewed according to the Strengthening the Reporting of Observational Studies in Epidemiology “STROBE.”[40] Consent was required and covered both study participation and publication of de-identified aggregate findings. Surrogate consent from the patient's legal guardian or designated health proxy was permitted in cases where the subject lacked decision-making capacity. All patients that survived and regained their faculties were informed of the project. All data generated or analyzed during this study are included in this article. De-identified individual subject data may be available from the corresponding author on a reasonable request.

Patients were eligible for study participation if: (1) established COPD diagnosis before hospitalization; (2) admitted with hypercapnic respiratory failure; (3) age ≥50 years; (4) IMV for ≥24 h (any mode except high-frequency percussive ventilation or high-frequency oscillatory ventilation); (5) no sedatives or low dose sedation (1–2 mg midazolam per hour or 25–100 mcg fentanyl per hour);[41] and (6) meeting the American Thoracic Society/European Respiratory Society (ATS/ERS) guidelines to participate in a weaning trial.[42] Pre-defined weaning criteria included: (1) patent upper airway; (2) ability to protect airway (Glasgow Coma Score <10; adequate gag and cough reflexes); (3) ability to clear secretions; (4) secretion burden not requiring suction more frequently than every 2 h; (5) level of support (fraction of inspired oxygen [FiO2] <50%; positive end-expiratory pressure [PEEP] ≤5 cm H2O); and (6) hemodynamic stability not requiring chemical (vasopressors, inotropes) or mechanical (e.g. intra-aortic balloon pump, extracorporeal life support) circulatory support.

Patients were excluded if they did not meet the predefined weaning criteria, self-extubation, history of neuromuscular disease, acute coronary syndrome, overt cardiac failure, end-stage renal disease, hemodynamic instability, or for signs of systemic infection during the weaning process.

All patients were ventilated using Dräger Evita® XL or Evita®4 ventilators (Dräeger Medical, Inc., Lubeck, Germany). Weaning and extubation readiness was determined by the respiratory therapist (RT) using pre-defined criteria and the result of a spontaneous breathing trial (SBT). The SBT consisted of pressure support (PS) ventilation mode with PS ≤12 cmH2O, PEEP ≤8 cmH2O, and FiO2≤0.4, adjusted to achieve an arterial pH 7.35, arterial oxygen saturation (SaO2) ≥92%, RR 10–25 breaths/min, and without arrhythmias. NIF measurements were taken as one measurement following best effort and performed following a hold maneuver on the ventilator.

The decision to return to MV was made by jointly by the RT and the attending physician (both blinded to NIF results) based on signs of poor tolerance: SaO2< 90% with FiO2>0.4; PaCO2>55 mmHg or increased by ≥10 mmHg; arterial pH < 7.33; RR > 35 breaths/min or increased by 50% for ≥5 min; heart rate (HR) >140 beats/min or increased/decreased by >20%; mean arterial pressure >130 mmHg or <70 mmHg; or the presence of agitation, diaphoresis, disorientation, or depressed mental status.[14] The patients who demonstrated one of these signs during the SBT were considered a failed wean and returned to ventilatory support.[16]

Data collection

Patients were selected consecutively among admitted COPD patients meeting inclusion criteria. The data collection tool consisted of a two-part checklist including demographic variables (age, sex, marital status, and smoking) and ventilatory parameters (intubation hours, hospitalization hours, spontaneous RR, spontaneous VT, systolic blood pressure (BP), diastolic BP, HR, static and dynamic compliance).

Statistical analysis

Parametric data were expressed as mean ± standard deviation (SD), while nonparametric data were expressed as number and percentage. Receiver operating characteristic (ROC) curves were used to determine the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), diagnostic accuracy, LR+, and LR − of NIF at select cutoffs for predicting successful ventilator liberation. Statistical significance was set at P < 0.05. Statistical analyses were performed using IBM® Statistical Package for Social Sciences version 22.0 (IBM Corp., Chicago, IL, USA).

RESULTS

Ninety patients were included in the final analysis, of which 51 (56.7%) were women. The mean (±SD) age of the patients was 69.81 (±10.66) years. Patient demographics are depicted in Table 1. Forty-three (47.8%) patients were extubated successfully (P = 0.654).

Table 1

Patient demographics and characteristics

Variable	Value (n=90)
Age*	61.40±12.83
Duration of intubation (h)*	108.50±24.37
Duration of hospitalization (day)*	15.07±16.41
Spontaneous respiratory rate (bpm)*	29.54±3.56
Spontaneous tidal volume (ml)*	294.18±34.91
Systolic blood pressure (mmHg)*	126.00±13.84
Diastolic blood pressure (mmHg)*	73.72±10.17
Heart rate (bpm)*	89.91±11.36
Temperature (°C)	36.98±0.32
Static compliance (ml/cmH2O)	75.50±3.94
Dynamic compliance (ml/cmH2O)	59.94±5.04
Gender, n (%)
Male	39 (43.3)
Female	51 (56.7)
Marital status, n (%)
Single	8 (20.00)
Married	72 (80.00)
Smoking, n (%)
>1 pack/month	61 (67.8)
≤1 pack/month	29 (32.2)

*Mean±SD. SD: Standard deviation

The area under the ROC curves (AUC) of 0.836 for the NIF threshold of ≤−25 cmH2O offered the optimal performance in COPD patients : sensitivity 95.0%, specificity 86.0%, PPV 84.4%, NPV 95.6%., LR+ 6.79, LR− 0.06, and diagnostic accuracy 91.7% [Table 2]. The influence of intubation duration and hospital LOS on NIF index performance at the ≤−25 cm H2O threshold is shown in Table 3.

Table 2

Performance of negative inspiratory force index at cutoff ≤−25 cmH2O in chronic obstructive pulmonary disease patients

Parameter	NIF
Sensitivity (%)	95.00
Specificity (%)	86.00
PPV (%)	84.44
NPV (%)	95.56
+LR	6.79
−LR	0.06
DA (%)	90.67
Probability of the weaning success if test is negative (P [W+/T−]) (%)	4.44
Probability of the weaning success if test is positive (P [W+/T+]) (%)	84.44
AUC±SE (%)	0.836±0.06

NIF: Negative inspiratory force, PPV: Positive predictive value, NPV: Negative predictive value, +LR: Likelihood ratio of positive test, −LR: Likelihood ratio of negative test, DA: Diagnostic accuracy, AUC: Area under the curve, SE: Standard error

Table 3

Effect of intubation duration and length-of-hospitalization on the accuracy of negative inspiratory force index in predicting weaning outcome

Variables	Intubation duration (h)		ICU length-of-stay (days)
	≤105	>105	≤9	>9
Sensitivity (%)	66.67	35.48	45.00	65.22
Specificity (%)	64.71	92.31	70.37	60.00
PPV (%)	75.42	86.51	85.17	92.80
NPV (%)	93.60	76.14	87.24	94.34
+LR	1.89	4.61	1.52	1.63
−LR	0.52	0.70	0.78	0.58
DA (%)	65.37	78.19	65.73	63.98
Probability of the weaning success if test is negative (P [W+/T−]) (%)	6.4	23.86	12.76	5.66
Probability of the weaning success if test is positive (P [W+/T+]) (%)	75.42	86.51	85.17	92.80
Prevalence (%)	44.4	44.4	44.4	44.4
AUC±SE (%)	0.689±0.080	0.654±0.093	0.565±0.084	0.630±0.086

PPV: Positive predictive value, NPV: Negative predictive value, +LR: Likelihood ratio of positive test, −LR: Likelihood ratio of negative test, DA: Diagnostic accuracy, AUC: Area under the curve, SE: Standard error, ICU: Intensive care units

A comparison was made between patients with a duration of intubation (DoI) of ≤105 versus >105 h. The sensitivity was lower both in patients with DoI ≤105 h (by ≤28.3 h) and DOI >105 h (by ≤59.52 h) when compared to the uncategorized NIF. The specificity was lower in patients with DoI ≤105 h (by ≤ 21.29), but higher in patients requiring >105 h of intubation (by ≥6.31) when compared to the uncategorized NIF. The LR + was lower both in patients with DoI ≤105 h (by ≤4.9) and DOI >105 h (by ≤2.18) when compared to the uncategorized NIF. The LR − was higher both in patients with DoI ≤105 h (by ≥0.46) and DOI >105 h (by ≥0.64) when compared to the uncategorized NIF. The AUC was lower both in patients with DoI ≤105 h (0.689) and DOI >105 h (0.654) when compared to the uncategorized NIF (0.836).

Similarly, a comparison was made between patients with ICU LOS ≤9 days and >9 days. The sensitivity was lower both in patients with ICU LOS ≤9 days (by ≤ 50.0) and ICU LOS >9 days (by ≤29.8) when compared to the uncategorized NIF. The specificity was lower both in patients with ICU LOS ≤9 days (by ≤15.63) and ICU LOS > 9 days (by ≤26.0) when compared to the uncategorized NIF. The LR+ was lower both in patients with ICU LOS ≤9 days (by ≤5.27) and ICU LOS >9 days (by ≤5.16) when compared to the uncategorized NIF. The LR − was lower both in patients with ICU LOS ≤9 days (by ≥0.72) and ICU LOS >9 days (by ≥0.52) when compared to the uncategorized NIF. The AUC was lower both in patients with ICU LOS ≤9 days (0.565) and ICU LOS > 9 days (0.639) when compared to the uncategorized NIF (0.836).

DISCUSSION

COPD is a complex disease with high socio-economic impact, morbidity, and mortality.[123456] Increased severity, as measured by GOLD class, is associated with increased complications related to ventilator associated pneumonias,[43] IMV duration[44] and overall ICU LOS.[45] In-hospital mortality rates for ventilated COPD patients have been reported as high as 24%, with 1-year mortality rates up to 59%.[324647] For those on IMV, ventilator liberation and recovery rates range 60%–90%.[3246] Early and aggressive ventilator weaning is paramount for limiting complications.

Various weaning parameters have been proposed to predict extubation success in ICU populations. The RSBI is among the most commonly used indices in non-homogeneous populations for predicting successful ventilator liberation.[38484950] The accuracy of this measurement, however, has been questioned in the setting of sepsis, fever, supine position, restrictive physiology, and patient anxiety.[51] One study reported that including the RSBI into a weaning protocol after a decision was made to attempt an SBT prolonged weaning and IMV for an additional day without reducing the number of failed extubations or tracheostomy placements.[52] Unfortunately, RSBI underperforms as a predictor of extubation readiness in patients with COPD, with several studies citing low sensitivity and specificity for detecting extubation readiness in COPD patients.[535455] Additionally, RSBI may overestimate the ability to liberate COPD patients from IMV using the traditional threshold of <105 breaths/min/L.[5556] Moreover, in the COPD population, an RSBI threshold of ≤85 breaths/min/L outperformed the widely used threshold <105 breaths/min/L, yielding a 95.5% probability of extubation success.[50]

NIF is a relatively simple and noninvasive test to assess for extubation readiness. Patients are asked to perform a forceful inspiration after expiring to a residual volume level against an occluded mouthpiece.[3657] NIF hypothetically reflects diaphragmatic strength and hence global inspiratory muscle strength with resultant alveolar ventilation.[3236] In a COPD patient with exacerbation, this may then represent an early surrogate of respiratory failure and recovery.

Early studies that established NIF normative values had little standardization of equipment and patient population, making it difficult to generalize values to specific subgroups such as COPD patients.[5859] In an attempt to standardize respiratory muscle testing, the ATS/ERS released a joint statement to this effect in 2002.[60] A subsequent meta-analysis reported that many of the studies that have been used to derive NIF datasets have suffered from similar methodologic limitations,[36] and to date, a validated NIF threshold for COPD patients has not been reported.

NIF is determined by the force-length of the diaphragmatic muscles, and as such, has a degree of variability at different residual volumes. Inspiratory efforts are maximized at low lung volumes, whereas expiratory efforts are maximized at higher lung volumes.[60] As COPD-patients have inherently higher functional residual capacities (FRC), a low NIF may more reflect a shortened inspiratory fiber length than a reduction in diaphragmatic strength.[60] This may explain in part why the optimal NIF threshold in this COPD population varied from prior reports in heterogeneous populations that may be functioning at a lower FRC. This study demonstrated that the NIF threshold of ≤−25 cmH2O performed optimally in COPD patients to predict successful weaning and ventilator liberation. More negative thresholds may not be as discriminatory in COPD patients.

CONCLUSIONS

In mechanically ventilated COPD patients with hypercapnic respiratory failure, the NIF threshold of ≤−25 cmH2O performed optimally as a moderate-to-good predictor for successful ventilator weaning and liberation. More negative thresholds may not be as discriminatory in COPD patients.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Research quality and ethics statement

The authors of this manuscript declare that this scientific work complies with reporting quality, formatting and reproducibility guidelines set forth by the EQUATOR Network (http://www.equator-network.org/). The study was approved by the Investigational Review Board at Baqiyatallah University of the Medical Sciences (340/5/5904) and Shahid Beheshti University of the Medical Sciences (SBMU1/REC/1393/89), and these approvals were accepted by each of the enrolling centers: Baqiyatallah Hospital, Shariati hospital, Loghman Hakim Hospital, Besat Hospital, Masih Daneshvari Hospital, and Talaghani Hospital. Trial registration was not indicated for this study type.