Hindawi Canadian Respiratory Journal Volume 2019, Article ID 3478968, 10 pages https://doi.org/10.1155/2019/3478968 Research Article Noninvasive Ventilation Weaning in Acute Hypercapnic Respiratory Failure due to COPD Exacerbation: A Real-Life Observational Study Paola Faverio ,1 Anna Stainer,1 Federica De Giacomi,1 Grazia Messinesi,1 Valentina Paolini,1 Anna Monzani,1 Paolo Sioli,1 Irdi Memaj,1 Oriol Sibila,2 Paolo Mazzola,3 and Alberto Pesci1 1 School of Medicine and Surgery, University of Milan Bicocca, Respiratory Unit, San Gerardo Hospital, ASST Monza, Monza, Italy 2 Respiratory Department, Hospital de La Santa Creu I Sant Pau, Autonomous University of Barcelona and Biomedical Research Institute Sant Pau (IIB Sant Pau), Barcelona, Spain 3 School of Medicine and Surgery, University of Milan Bicocca, Acute Geriatrics Unit, San Gerardo Hospital, ASST Monza, Monza, Italy Correspondence should be addressed to Paola Faverio; paola.faverio@gmail.com Received 27 September 2018; Revised 1 February 2019; Accepted 26 February 2019; Published 25 March 2019 Academic Editor: Pierachille Santus Copyright © 2019 Paola Faverio et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The most recent British Thoracic Society/Intensive Care Society (BTS/ICS) guidelines on the use of noninvasive ventilation (NIV) in acute hypercapnic respiratory failure (AHRF) suggest to maximize NIV use in the first 24 hours and to perform a slow tapering. However, a limited number of studies evaluated the phase of NIV weaning. The aim of this study is to describe the NIV weaning protocol used in AHRF due to acute exacerbation of chronic obstructive pulmonary disease (AE-COPD), patients’ characteristics, clinical course, and outcomes in a real-life intermediate respiratory care unit (IRCU) setting. We performed a retrospective study on adult patients hospitalized at the IRCU of San Gerardo Hospital, Monza, Italy, from January 2015 to April 2017 with a diagnosis of AHRF due to COPD exacerbation. The NIV weaning protocol used in our institution consists of the interruption of one of the three daily NIV sessions at the time, starting from the morning session and finishing with the night session. The 51 patients who started weaning were divided into three groups: 20 (39%) patients (median age 80 yrs, 65% males) who completed the protocol and were discharged home without NIV (Completed Group), 20 (39%) did not complete it because they were adapted to domiciliary ventilation (Chronic NIV Group), and 11 (22%) interrupted weaning ex abrupto mainly due to NIV intolerance (Failed Group). Completed Group patients were older, had a higher burden of comorbidities, but a lower severity of COPD compared to Chronic NIV Group. Failed Group patients experienced higher frequency of delirium after NIV discontinuation. None of the patients who completed weaning had AHRF relapse during hospitalization. While other NIV weaning methods have been previously described, our study is the first to describe a protocol that implies the interruption of a ventilation session at the time. The application of a weaning protocol may prevent AHRF relapse in the early stages of NIV interruption and in elderly frail patients. 1. Introduction once acidosis and respiratory distress are solved, proposing different weaning methods but not clearly establishing the Bilevel noninvasive ventilation (NIV) is a milestone in the superiority of one over the others [3–6]. A recently pub- treatment of acute hypercapnic respiratory failure (AHRF) lished randomized controlled trial (RCT) questioned the in patients with acute exacerbation of chronic obstructive utility of the weaning phase [5]. In this RCT, Sellares et al. pulmonary disease (AE-COPD) [1, 2]. Unfortunately, only a compared the prolongation of NIV for three nights after limited number of studies evaluated the weaning from NIV recovery from an AHRF episode to direct NIV 2 Canadian Respiratory Journal discontinuation in COPD patients without previous do- corticosteroids, and increased bronchodilator therapy [13]. miciliary ventilation [5]. No differences were observed in In our study population, AE-COPD was the primary ad- regard to prevention of AHRF relapse, hospital readmission, mission diagnosis according to the physician in charge. or mortality [5]. Furthermore, the application of a weaning Patients who also had pneumonia or congestive heart failure phase resulted in a prolonged intermediate respiratory care as secondary diagnoses were not excluded from the study, unit (IRCU) stay. These results seem to suggest that NIV can and these variables were reported in the Results section. be directly discontinued once the acute episode is solved. However, as in all RCTs, very elderly patients, not fully cooperative and with several comorbidities, have rarely been 2.3. NIV Initiation Criteria and Indication to Chronic Do- included, although they often represent a consistent part of miciliary NIV. In accordance with the protocol adopted by the COPD population with AHRF admitted to IRCUs [5]. In our institution, bilevel NIV was initiated in patients with such a frail group, many factors may play a role in AHRF AHRF (arterial carbon dioxide partial pressure (PaCO2) relapse after NIV discontinuation, including other re- > 45 mmHg) due to AE-COPD in case of respiratory acidosis spiratory comorbidities, delirium, and sarcopenia [7]. (pH ≤ 7.35) and/or respiratory distress. Respiratory distress Therefore, the actual need of weaning from NIV in a real-life was defined as the respiratory rate (RR) > 25 associated with cohort of patients with AHRF and AE-COPD and the best use of respiratory accessory muscles. In case of pH < 7.25, way to perform it still need to be determined. decision regarding NIV initiation was shared with the in- Our aim is to present and test the feasibility of the NIV tensive care unit (ICU) team. weaning strategy used in patients with AHRF due to AE- All patients were ventilated with pressure support COPD admitted in our IRCU. Furthermore, we describe the ventilation modality. The ventilators available in our IRCU characteristics, clinical course, and outcomes of the cohort of were Respironics Esprit and Resmed Elisée 150. Full face and patients who underwent the NIV weaning protocol. oronasal interfaces were used. In our institution, indications for domiciliary NIV in COPD patients were persistent daytime hypercapnia after at 2. Materials and Methods least two episodes of AHRF requiring NIV in the prior year; 2.1. Study Design. This was a retrospective cohort study on and comorbidities such as obstructive sleep apnea or adult patients hospitalized in the IRCU of San Gerardo obesity-hypoventilation syndrome, in accordance with Hospital, Monza, Italy, from January 2015 to April 2017 with guidelines for each disease [1, 14, 15]. a diagnosis of AE-COPD. The study was approved by the local Institutional Review Board (#684-May 2017) and did 2.4. NIV Disconnection and NIV Weaning Initiation Criteria. not require patients’ informed consent due to the retro- NIV disconnection was attempted when all the following spective design. were concurrently met: The following data were retrospectively collected from patients’ charts: demographics, comorbidities and Charlson (a) Adequate oxygenation defined as arterial oxygen comorbidity index [8], COPD characteristics (including partial pressure to fraction of inspired oxygen ratio pulmonary function tests performed at the previous avail- (PaO2/FiO2) > 200 mmHg during NIV with fraction able outpatient visit during clinical stability, number of of inspired oxygen (FiO2) < 0.5 exacerbations and hospitalizations during the prior year, and (b) pH > 7.35, RR < 25 without use of respiratory ac- COPD phenotype—emphysema, chronic bronchitis, cessory muscles asthma-COPD overlap, and bronchiectasis-COPD overlap (c) Hemodynamic stability (assessed through heart rate [9, 10]), severity scores at admission (acute physiology and and blood pressure) chronic health disease classification system II (APACHE II) [11] and sequential organ failure assessment (SOFA) [12]), (d) Kelly score [16] ≤ 2 radiological imaging during hospitalization (either chest The weaning phase began when all the following criteria X-ray, computerized tomography scan, or both), arterial were reached after at least 1 hour of disconnection from NIV blood gases (ABG), days of NIV application and hospital and by administering oxygen through a venturi mask with a length of stay, presence of delirium during hospitalization, FiO2 titrated in order to maintain peripheral oxygen satu- and NIV parameters. Follow-up data, including three- ration (SpO2) 88–92%: month mortality and hospitalization rate, were also retro- spectively collected from outpatient clinic charts. All pa- (a) pH > 7.35, RR < 25 without use of respiratory ac- tients discharged from our IRCU with a diagnosis of AE- cessory muscles COPD were followed-up after 3 months from discharge at (b) Hemodynamic stability our COPD outpatient clinic; those who did not present at the (c) Kelly score [16] ≤ 2 visit received a phone call to reprogram it. 2.5. NIV Weaning Protocol. The weaning protocol used in 2.2. AE-COPD Definition. AE-COPD was defined as acute our institution is summarized in Figure 1. The term “NIV worsening of respiratory symptoms that resulted in addi- session” was used to indicate the continuous application of tional therapy, including antibiotics, systemic NIV for at least 3 hours to the maximum duration tolerated Canadian Respiratory Journal 3 Acute phase: respiratory distress and/or AHRF (b) Hemodynamic instability (c) Kelly score [16] > 2 Maximize NIV use with 3 sessions/day (morning, afternoon, and night) for at least 3 hours per session 2.6. Delirium Definition and Criteria for Identification. Resolution of respiratory distress and/or AHRF The chart-based method (CBM) was used to retrospectively identify delirium on patients’ charts [17, 18], searching for Start NIV weaning with interruption of morning session key terms for identification of delirium (inattention, dis- orientation, hallucinations, agitation, and inappropriate Recurrence of respiratory distress and/or AHRF? behavior) and for any evidence of acute mental status change. No Yes We temporally divided the presence of delirium into two subgroups: delirium during NIV and delirium after NIV Continue weaning with Consider to reincrease NIV discontinuation, in order to evaluate the possible role of this interruption of afternoon session sessions and, for last, night session complication in association with NIV weaning and clinical outcomes. Figure 1: NIV weaning protocol. AHRF, acute hypercapnic re- spiratory failure; NIV, noninvasive ventilation. 2.7. Outcomes. The following outcomes were considered for all patients who initiated the weaning phase: by the patient, with interruption only for oral therapy ad- ministration and/or feeding. During the acute initial phase, (i) Primary outcomes: new worsening of gas exchange the treatment is maximized with 3 NIV sessions per day (pH ≤ 7.35, RR > 25, and/or respiratory distress) (morning, afternoon, and night) (Figure 1). In our NIV after completion of weaning protocol and in-hospital weaning protocol, the patient starts NIV weaning with in- mortality terruption of the morning session. If there is no recurrence (ii) Secondary outcomes: three-month mortality and of respiratory distress or AHRF, the patient continues three-month hospitalization rate weaning with interruption of the afternoon session and, for last, of the nocturnal session. The NIV sessions were 2.8. Statistical Analysis. Data were analyzed using SPSS 21.0 interrupted one per day (e.g., day 1: interruption of the for MAC OS (SPSS Inc., Chicago, IL, USA). Baseline morning session; day 2: interruption of the morning and characteristics of the study population, ABGs, ventilation afternoon session; day 3: interruption of all ses- parameters, and outcomes, were considered for statistical sions—morning, afternoon, and night). The number of days analysis. We assessed normality distribution for continuous at each stage of weaning (NIV during afternoon and night or variables with the Kolmogorov–Smirnov test. Continuous nocturnal NIV only) was at physician discretion. During variables were expressed as median (interquartile range weaning, NIV pressures, both inspiratory positive airway (IQR): 25th–75th percentile) and compared using the pressure (IPAP) and expiratory positive airway pressure Kruskal–Wallis nonparametric ANOVA test. Categorical (EPAP), were maintained on the same personalized settings data were expressed as frequencies and percentages and used to reverse AHRF. compared using the chi-square test. All tests were 2-tailed, During NIV weaning, every patient underwent pul- and a P value < 0.05 was considered statistically significant. monary rehabilitation between NIV sessions with positive expiratory pressure (PEP) and incentive spirometry (IS) devices. In patients with residual mucus encumbrance, we 3. Results used PEP devices (i.e., PEP bottle or TheraPEP, Smiths 3.1. Characteristics of Study Population. Among the 94 pa- Medical) set at low pressure (8 cm H2O) asking the patient to tients hospitalized in the IRCU with a diagnosis of COPD blow as long as possible repeatedly without achieving fa- exacerbation during the study period, we focused on 51 tigue, in order to improve airway clearance. In patients with patients (median age 79 yrs, 53% males) who were started on poor thoracoabdominal movements, we used IS devices in NIV weaning after an episode of AHRF (Figure 2). association with stretching exercises aimed at preventing Of these 51 patients, 20 (median age 80 yrs, 65% males) atelectasis or local hypoventilation. The patient was invited completed the protocol (Completed Group), 20 (median age to repeatedly inspire slowly through the device (Coach II, 65 yrs, 35% males) did not complete the weaning because Smiths Medical) lifting a piston up to the limit set at 70–80% they were adapted to chronic domiciliary ventilation of inspiratory capacity and then hold breath for five seconds. (Chronic NIV Group), and 11 (median age 82 yrs, 65% Pulmonary rehabilitation was set and carried out by a re- males) did not complete the protocol for other clinical spiratory therapist up to two times per day; additional conditions (Failed Group) (Figure 2). sessions were kept by nurses. Clinical conditions for which Failed Group did not NIV weaning failure occurred if any of the following complete weaning included NIV intolerance, high risk of appeared during the execution of the NIV weaning protocol: pneumothorax due to severe bullous emphysema, and NIV (a) pH ≤ 7.35 and/or respiratory distress weaning failure, as defined in Section 2, which required an 4 Canadian Respiratory Journal 94 patients admitted to the IRCU with a diagnosis of AE-COPD 29 excluded due to absence of respiratory distress and/or AHRF: (i) 13 not treated with NIV (ii) 12 continued usual domiciliary NIV (iii) 2 patients with tracheostomy in usual domiciliary invasive mechanical ventilation (iv) 2 iatrogenic cause of AHRF 65 patients with respiratory distress and/or AHRF treated with NIV 14 cases of NIV failure: (i) 11 patients with worsening of underlying disease (ii) 3 patients due to NIV intolerance 51 patients with resolution of respiratory distress and/or AHRF who started NIV weaning 20 patients completed NIV 20 patients in domiciliary NIV: 11 patients did not complete NIV weaning weaning as per protocol (i) 13 patients readapted to protocol: usual domiciliary NIV (i) 9 early discontinuation due to NIV «Completed Group» (ii) 7 patients adapted de novo to intollerance domiciliary NIV (ii) 1 early discontinuation due to high risk of pneumothorax «Chronic NIV Group» (iii) 1 due to worsening of respiratory failure which required increase in NIV sessions «Failed Group» Figure 2: Flow chart of study population. IRCU, intermediate respiratory care unit; COPD, chronic obstructive pulmonary disease; AHRF, acute hypercapnic respiratory failure; NIV, noninvasive ventilation. increase in NIV sessions (Figure 2). None of the 51 patients 3.2. Severity on Admission and Arterial Blood Gases developed facial skin lesion; however, three patients de- Evaluation. APACHE II score was greater in the Failed veloped gastric distension associated with nausea in one case. Group compared to the other two study groups. In contrast, Demographics, comorbidities, and COPD characteristics there was no between-group difference in the SOFA score of the three study groups are summarized in Table 1. (Table 2). Also, pH values of the ABG performed at ad- Completed Group patients were older and had a higher mission were lower in Completed Group and Failed Group burden of comorbidities, evaluated through the Charlson in comparison with Chronic NIV Group (7.30, 7.29, and index. In regard to specific comorbidities, Completed Group 7.35, respectively; P value <0.01) (Figure 3). There were no presented the highest frequency of cardiovascular diseases, between-group differences in concomitant causes of AHRF while Chronic NIV Group showed more frequently obesity- (Table 2). hypoventilation syndrome and mood disorders. Patients in Furthermore, we analyzed the ABGs performed at dif- Completed Group presented more frequently a chronic ferent stages both during and after NIV discontinuation bronchitis phenotype and had a lower number of COPD (Figure 3). No difference was observed between groups with exacerbations and hospitalizations in the prior year. How- the exception of lower PaCO2 values in Completed Group ever, this finding is related to the definition itself of patients and Failed Group compared to Chronic NIV Group at ABGs in “Chronic NIV Group” that had to present at least two performed at 24 and 48 h after NIV discontinuation, but not episodes of AHRF, requiring NIV in the prior year to receive at the ABG performed at discharge. chronic nocturnal domiciliary NIV. Only two patients in Completed Group received a do-not-resuscitate/do-not- intubate (DNR/DNI) order, which was significantly less 3.3. NIV Duration and Parameters. Compared to the frequent than in Failed Group. Completed Group, the duration of NIV was shorter in the There was no between-group difference in the prevalence Failed Group and longer in the Chronic NIV Group (for the of delirium during NIV and during hospitalization as a latter group, days on nocturnal domiciliary NIV were in- whole (Table 1). The prevalence of delirium after NIV cluded in the computation of total NIV duration, because it discontinuation, however, was greater in the Failed Group was artificial to exactly locate the moment when weaning than in the Completed Group. was finished) (Table 3). However, no differences were found Canadian Respiratory Journal 5 Table 1: Demographics, comorbidities, COPD characteristics, and presence of delirium in study population. Completed Group Chronic NIV Group Failed Group P (N � 20) (N � 20) (N � 11) value Demographics, n (%) Age (yrs), median (IQR) 79 (74–85) 65 (55–75) 82 (77–84) <0.01 Males 13 (65) 7 (35) 7 (64) 0.65 BMI (kg/m2), median (IQR) 24.9 (21.1–27.1) 26.2 (24.6–29.1) 23.9 (21–23.4) 0.12 Current or prior smokers 20 (100) 20 (100) 11 (100) — Do-not-resuscitate order 2 (10) 1 (5) 5 (45) 0.02 Comorbidities, n (%) Charlson comorbidity index, median (IQR) 6.75 (6-7) 4 (3–5) 7 (6–8) <0.01 Diabetes mellitus 4 (20) 2 (10) 2 (18) 0.09 Chronic heart diseases 12 (60) 4 (20) 4 (36) 0.09 Prior myocardial infarction 9 (45) 0 4 (36) 0.26 Cerebrovascular diseases 2 (10) 3 (15) 2 (18) 0.51 Chronic liver disease 2 (10) 1 (5) 0 0.25 Solid tumor 2 (10) 1 (5) 2 (18) 0.59 Hematologic malignancy 1 (5) 0 0 0.28 Obstructive sleep apnea syndrome 1 (5) 4 (20) 1 (9) 0.055 Kyphoscoliosis 0 0 2 (18) 0.03 Obesity-hypoventilation syndrome 0 4 (20) 0 0.63 Dementia 0 0 0 — Psychiatric disorders 2 (10) 9 (45) 3 (27) 0.16 Chronic renal failure 2 (10) 0 2 (18) 0.63 COPD characteristics, n (%) FEV1 (L), median (IQR) 0.79 (0.62–0.99) 0.49 (0.36–0.82) 0.89 (0.69–1) 0.03 FEV1 (%), median (IQR) 42 (35–50) 33 (17–45) 40 (32–49) 0.21 GOLD 4∗ 16 (80) 15 (75) 7 (64) 0.60 Emphysema phenotype 3 (15) 0 5 (45) 0.87 Chronic bronchitis phenotype 15 (75) 9 (45) 4 (36) 0.03 Asthma-COPD overlap phenotype 2 (10) 7 (35) 2 (18) 0.39 Bronchiectasis-COPD overlap phenotype 0 3 (15) 0 0.68 No. of exacerbations in the previous year, 1 (0-1) 3 (2–5) 1 (0–2) 0.07 median (IQR) No. of hospitalizations in the previous year, 1 (0-1) 2 (0–5) 1 (0–2) 0.01 median (IQR) Long-term oxygen therapy 10 (50) 17 (85) 5 (45) 0.81 Prior domiciliary NIV 0 17 (85) 0 0.27 Delirium, n (%) Delirium according to CBM 4 (20) 4 (20) 4 (36) 0.36 Delirium during NIV 4 (20) 4 (20) 3 (27) 0.68 Delirium after NIV discontinuation 0 — 3 (27) — BMI, body mass index; CBM, chart-based method; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in the 1st second; GOLD, global initiative for chronic obstructive lung disease; IQR, interquartile range 25th–75th percentile; NIV, noninvasive ventilation. ∗ Classification of airflow limitation severity based on postbronchodilator FEV1 [13]. Table 2: Severity scores on admission and concomitant causes of AHRF. Completed Group Chronic NIV Group Failed Group P value (N � 20) (N � 20) (N � 11) Severity scores, median (IQR) Apache II 19 (17–21) 16 (15–19) 20 (14–21) 0.02 SOFA score 3 (2–4) 3 (2-3) 3 (2-3) 0.80 Concomitant causes of AHRF, n (%) Pneumonia 5 (25) 5 (25) 4 (36) 0.55 Congestive heart failure 14 (70) 8 (40) 7 (64) 0.48 Vital parameters on admission, median (IQR) Respiratory rate 23 (20–30) 22 (20–34) 26 (24–32) 0.46 Heart rate 99 (90–117) 97 (81–109) 95 (87–114) 0.78 Mean arterial pressure 95 (86–123) 97 (91–106) 101 (90–130) 0.55 AHRF, acute hypercapnic respiratory failure; APACHE II, acute physiology and chronic health evaluation II; IQR, interquartile range 25th–75th percentile; NIV, noninvasive ventilation; SOFA, sequential organ failure assessment. 6 Canadian Respiratory Journal 100 90 80 70 60 50 40 Admission 1h NIV 3 h NIV 24 h after NIV 48h after NIV Discharge interruption/1st interruption/2nd day domiciliary day domiciliary NIV NIV Median (IQR) 52 (43–66) 66 (62–72) 65 (57–72) 64 (56–69) 61 (55–66) 63 (60–65) 51 (46–60) 67 (61–70) 59 (52–73) 61 (56–63) 63 (55–68) 62 (59–67) 53 (41–61) 65 (61–69) 63 (61–78) 62 (55–66) 67 (60–75) 60 (58–61) Completed Group (N = 20) Chronic NIV Group (N = 20) Failed Group (N = 11) (a) 70 65 60 55 50 45 P value = 0.01 40 P value = 0.04 35 30 Admission 1h NIV 3 h NIV 24h after NIV 48h after NIV Discharge interruption/1st interruption/2nd day domiciliary day domiciliary NIV NIV Median (IQR) 65 (58–73) 60 (51–69) 56 (47–66) 50 (48–53) 51 (47–54) 50 (45–54) 63 (57–73) 62 (51–70) 63 (56–68) 62 (49–66) 57 (51–63) 53 (49–61) 65 (54–78) 56 (46–68) 59 (43–65) 47 (43–51) 45 (44–56) 52 (48–55) Completed Group (N = 20) Chronic NIV Group (N = 20) Failed Group (N = 11) (b) Figure 3: Continued. Canadian Respiratory Journal 7 7.45 7.4 P value = 0.02 7.35 P value < 0.01 7.3 7.25 P value < 0.01 7.2 Admission 1h NIV 3h NIV 24 h after NIV 48 h after NIV Discharge interruption/1st interruption/2nd day domiciliary day domiciliary NIV NIV Median (IQR) 7.30 (7.26–7.34) 7.32 (7.29–7.36) 7.34 (7.32–7.38) 7.42 (7.39–7.46) 7.42 (7.40–7.44) 7.41 (7.39–7.43) 7.35 (7.30–7.38) 7.35 (7.31–7.37) 7.34 (7.32–7.37) 7.38 (7.36–7.41) 7.39 (7.37–7.41) 7.41 (7.38–7.42) 7.29 (7.27–7.35) 7.34 (7.31–7.38) 7.36 (7.31–7.42) 7.40 (7.36–7.42) 7.38 (7.33–7.40) 7.40 (7.36–7.44) Completed Group (N = 20) Chronic NIV Group (N = 20) Failed Group (N = 11) (c) 300 250 200 150 100 50 0 Admission 1h NIV 3h NIV 24h after NIV 48h after NIV Discharge interruption/1st interruption/2nd day domiciliary day domiciliary NIV NIV Median (IQR) 225 (185–287) 244 (214–269) 240 (208–282) 250 (227–292) 236 (230–270) 273 (228–296) 217 (184–247) 255 (214–278) 233 (216–271) 250 (238–266) 252 (231–266) 263 (243–293) 189 (132–223) 248 (224–290) 241 (225–339) 263 (214–280) 285 (200–226) 271 (218–321) Completed Group (N = 20) Chronic NIV Group (N = 20) Failed Group (N = 11) (d) Figure 3: Arterial blood gases at different stages both during NIV and after NIV discontinuation. ABG, arterial blood gas; IQR, interquartile range 25th–75th percentile; PaO2, arterial oxygen partial pressure (mmHg); PaCO2, arterial carbon dioxide partial pressure (mmHg); P/ F � PaO2/FiO2, arterial oxygen partial pressure to fraction of inspired oxygen ratio; NIV, noninvasive ventilation. (a) ABG: pO2, (b) ABG: pH, (c) ABG: pCO2, and (d) ABG: P/F. between groups in regard to number of days with maximized For patients in Completed Group and Failed Group, NIV before weaning (3 sessions per day) with a median of lower NIV pressures (both IPAP and EPAP) were used in one day for each group and length of hospital stay. comparison with Chronic NIV Group. In this latter group, 8 Canadian Respiratory Journal Table 3: Hospital length of stay and NIV duration, NIV parameters, and presence of delirium. Completed Group Chronic NIV Group Failed Group P value (N � 20) (N � 20) (N � 11) Hospital length of stay and NIV duration, median (IQR) Hospital length of stay, days 10 (8–14) 11 (10–14) 11 (7–15) 0.76 NIV duration, days 4 (3–6) 11 (10–14) 2 (1–3) <0.01 Number of days with 3 NIV sessions 1 (1-2) 1 (0–3) 1 (0–3) 0.96 Number of days with 2 NIV sessions 1 (1-2) 2 (1–9) 0 (0-1) <0.01 Number of days with 1 NIV session 1 (1-2) 4 (0–5) 0 (0-1) < 0.01 NIV parameters, median (IQR) Maximum IPAP (cm H2O) 16 (14–19) 21 (18–23) 17 (15–19) <0.01 Maximum EPAP (cm H2O) 6 (5–7) 8 (7-8) 6 (5–7) <0.01 IPAP (cm H2O) with domiciliary NIV — 19 (17–23) — — EPAP (cm H2O) with domiciliary NIV — 8 (5–8) — — EPAP, expiratory positive airway pressure; IQR, interquartile range 25th–75th percentile; IPAP, inspiratory positive airway pressure; NIV, noninvasive ventilation. NIV pressures used during the acute phase were similar to failed the first weaning attempt because of worsening re- those titrated for the chronic domiciliary NIV use, as shown spiratory failure. In the studies by Sellares et al. and Lun in Table 3. et al., 19% and 44% of patients, respectively, in the direct discontinuation groups failed the first withdrawal attempt [5, 6]. These percentages were higher than in the stepwise 3.4. Outcomes. Three patients, all in Failed Group, experi- withdrawal (5% and 26%, respectively) [5, 6]. In spite of this enced relapse of AHRF during the first 48 hours from NIV variable not being included among study outcomes, it is discontinuation (P < 0.01), and two of them died during the important to consider that the application of a weaning hospitalization (both of them had received a DNR/DNI protocol may prevent respiratory function deterioration order due to the severity of comorbidities) (Table 4). No even in the early stages of NIV interruption. difference was observed between groups in regard to sec- Furthermore, both the aforementioned studies recruited ondary outcomes. only fully cooperative patients with a limited burden of comorbidities. In a real-life setting, it is often necessary to 4. Discussion apply NIV to elderly patients with a high burden of comorbidities (median Charlson comorbidity index in our While other NIV weaning methods have been previously population ranged between 4 and 7), not fully cooperative described [3–6], our study is the first to describe a protocol (as demonstrated by the prevalence of delirium in our co- that implies the interruption of a ventilation session at the hort) and who received a DNR/DNI order (16% of cases in time. None of the patients who completed weaning expe- our cohort). These frail patients often excluded from clinical rienced AHRF relapse or died in the course of hospitali- trials could be the ones most benefiting from weaning zation. A considerable proportion of patients, particularly protocols. those younger, with a lower burden of comorbidities, but a Only a minority of patients in our study completed the higher severity of COPD, was initiated on chronic domi- weaning protocol, while a considerable proportion (39%) ciliary NIV. Delirium should be taken into account among was initiated on chronic domiciliary NIV. These patients the factors that can complicate the clinical course of ven- were younger, with a lower burden of comorbidities but a tilated patients with AHRF due to AE-COPD even in the higher severity of COPD compared to patients who com- weaning phase. pleted weaning. These data suggest that patients who are Despite the wide range of studies available on weaning most likely to receive chronic domiciliary NIV are those with from invasive mechanical ventilation [19], few studies have a longer life expectancy and more severe COPD. investigated the strategies to wean patients from NIV [3–6]. According to the CBM, 23% of patients in our cohort had Similarly to Damas et al., who described a weaning protocol a diagnosis of delirium. A meta-analysis published in 2012 that maintains the same number of sessions per day but reported a prevalence of delirium of 37% in patients on NIV, provides a progressive reduction of hours for each NIV which was higher than the one generally presented in session, we did not observe recurrence of AHRF during hospitalized nonventilated patients (10–31%) [20]. Fur- hospitalization after weaning [4]. Two recently published thermore, the risk of NIV failure in patients with delirium RCTs by Sellares et al. and Lun et al. comparing immediate was 2.12 times higher than in those without delirium. Our vs. stepwise NIV withdrawal did not show any benefit from study groups showed similar prevalence of delirium during the use of a weaning strategy in preventing relapse of AHRF NIV, whereas, after NIV discontinuation, patients who after NIV discontinuation, suggesting that NIV can be di- discontinued NIV ex abrupto showed a higher prevalence rectly discontinued in COPD patients once the episode of compared to those who completed the weaning protocol. We AHRF is solved [5, 6]. However, when looking at the postulate that delirium, on a case-by-case basis, might be weaning phase itself, in our population, only one patient both a cause and a consequence of NIV intolerance and Canadian Respiratory Journal 9 Table 4: Study outcomes. Completed Group Outcome, n (%) Chronic NIV Group (N � 20) Failed Group (N � 11) P value (N � 20) Primary AHRF relapse after NIV discontinuation 0 0 3 (27) 0.01 In-hospital mortality 0 0 2 (18) 0.03 Secondary Readmission rate at 3 months 6 (30) 11 (55) 2 (22) 0.91 Mortality at 3 months 4 (20) 4 (20) 4 (44) 0.23 IQR, interquartile range 25th–75th percentile; AHRF, acute hypercapnic respiratory failure; NIV, noninvasive ventilation. discontinuation. It is important to emphasize that patients Abbreviations who had a sudden NIV interruption were also the oldest and had the highest burden of comorbidities and the highest ABG: Arterial blood gases percentage of DNR/DNI orders. All these factors, together AE- Acute exacerbation of chronic obstructive with the higher incidence of delirium after NIV discon- COPD: pulmonary disease tinuation, may have played a role in the worse clinical AHRF: Acute hypercapnic respiratory failure outcomes (AHRF relapse after NIV discontinuation and in- APACHE Acute physiology and chronic health disease hospital mortality) observed in this group of patients. II: classification system II However, as a potentially preventable and treatable condi- CBM: Chart-based method tion, an early diagnosis is fundamental in frail patients, such COPD: Chronic obstructive pulmonary disease as those ventilated for AHRF, not only in the acute phase but DNR/ Do-not-resuscitate/do-not-intubate also in the weaning process. DNI: Furthermore, although patients in the “Failed group” EPAP: Expiratory positive airway pressure interrupted NIV earlier compared to other groups, no FEV1: Forced expiratory volume in the 1st second differences were observed in regard to the length of hospital FiO2: Fraction of inspired oxygen stay, once again indicating that these patients had a com- ICU: Intensive care unit plicated clinical condition that required a prolongation of IPAP: Inspiratory positive airway pressure the hospitalization regardless of NIV use. IRCU: Intermediate respiratory care unit Limitations of the present study to be acknowledged IS: Incentive spirometry mainly address the small sample size and the monocentric NIV: Noninvasive ventilation retrospective design, as well as the absence of a control group PaCO2: Arterial carbon dioxide partial pressure to compare different weaning methods. Delirium was PaO2/ Arterial oxygen partial pressure to fraction of assessed through a retrospective, although validated [18], FiO2: inspired oxygen ratio method that may have under- or overestimated the actual PEP: Positive expiratory pressure incidence of this complication. Since our purpose was to RCT: Randomized controlled trial picture the use of a NIV weaning method in a real-life RR: Respiratory rate scenario, we included in the study patients with AE-COPD SpO2: Peripheral oxygen saturation as primary diagnosis and secondary diagnosis of pneumonia SOFA: Sequential organ failure assessment. and congestive heart failure, as declared in Section 2. However, the study population was too small to allow an- Data Availability alyses and speculations on the different application and utility of a NIV weaning protocol according to the con- The data used to support the findings of this study are comitant causes of AHRF. available from the corresponding author upon request. Future studies should address the question of the utility of NIV weaning and the best way to apply it. Assuming that Conflicts of Interest the paradigm “one size fits all” is incorrect, it would be of paramount importance to determine which population The authors declare that there are no conflicts of interest might benefit the most from the application of NIV weaning regarding the publication of this paper. protocols, including frail elderly patients. References 5. Conclusions [1] A. C. Davidson, S. Banham, M. Elliott et al., “BTS/ICS guideline for the ventilatory management of acute hyper- In our cohort, the completion of a NIV weaning protocol capnic respiratory failure in adults,” Thorax, vol. 71, no. 2, based upon the interruption of one of the three daily NIV pp. ii1–ii35, 2016. sessions at the time is associated with the absence of negative [2] B. Rochwerg, L. Brochard, M. W. Elliott et al., “Official ERS/ outcomes, including AHRF relapse and in-hospital mortality. ATS clinical practice guidelines: noninvasive ventilation for Patients with AHRF should be carefully evaluated for pres- acute respiratory failure,” European Respiratory Journal, ence of delirium both during and after NIV discontinuation. vol. 50, no. 2, article 1602426, 2017. 10 Canadian Respiratory Journal [3] J. Duan, X. Tang, S. Huang, J. Jia, and S. Guo, “Protocol- interviewer ratings using the confusion assessment method,” directed versus physician-directed weaning from noninvasive Journal of the American Geriatrics Society, vol. 53, no. 2, ventilation,” Journal of Trauma and Acute Care Surgery, pp. 312–318, 2005. vol. 72, no. 5, pp. 1271–1275, 2012. [19] J.-M. Boles, J. Bion, A. Connors et al., “Weaning from me- [4] C. Damas, C. Andrade, J. P. Araújo, J. Almeida, and chanical ventilation,” European Respiratory Journal, vol. 29, P. Bettencourt, “Desmame de ventilação não invasiva: no. 5, pp. 1033–1056, 2007. experiência com perı́odos de descontinuação,” Revista Por- [20] M. Charlesworth, M. W. Elliott, and J. D. Holmes, “Non- tuguesa de Pneumologia, vol. 14, no. 1, pp. 49–53, 2008. invasive positive pressure ventilation for acute respiratory [5] J. Sellares, M. Ferrer, A. Anton et al., “Discontinuing non- failure in delirious patients: understudied, underreported, or invasive ventilation in severe chronic obstructive pulmonary underappreciated? A systematic review and meta-analysis,” disease exacerbations: a randomised controlled trial,” Euro- Lung, vol. 190, no. 6, pp. 597–603, 2012. pean Respiratory Journal, vol. 50, no. 1, article 1601448, 2017. [6] C.-T. Lun, V. L. Chan, W.-S. Leung et al., “A pilot randomized study comparing two methods of non-invasive ventilation withdrawal after acute respiratory failure in chronic ob- structive pulmonary disease,” Respirology, vol. 18, no. 5, pp. 814–819, 2013. [7] J. James, “Comprehensive geriatric assessment during emergency admission,” Nursing Older People, vol. 28, no. 2, pp. 16–22, 2016. [8] M. E. Charlson, P. Pompei, K. L. Ales, and C. R. MacKenzie, “A new method of classifying prognostic comorbidity in longitudinal studies: development and validation,” Journal of Chronic Diseases, vol. 40, no. 5, pp. 373–383, 1987. [9] J. R. Hurst, J. S. Elborn, and A. D. Soyza, “COPD- bronchiectasis overlap syndrome,” European Respiratory Journal, vol. 45, no. 2, pp. 310–313, 2015. [10] A. Segreti, E. Stirpe, P. Rogliani, and M. Cazzola, “Defining phenotypes in COPD: an aid to personalized healthcare,” Molecular Diagnosis & Therapy, vol. 18, no. 4, pp. 381–388, 2014. [11] W. A. Knaus, E. A. Draper, D. P. Wagner, and J. E. Zimmerman, “Apache II: a severity of disease classifi- cation system,” Critical Care Medicine, vol. 13, no. 10, pp. 818–829, 1985. [12] J.-L. Vincent, R. Moreno, J. Takala et al., “The SOFA (Sepsis- related Organ Failure Assessment) score to describe organ dysfunction/failure,” Intensive Care Medicine, vol. 22, no. 7, pp. 707–710, 1996. [13] Global Initiative for Chronic Obstructive Lung Disease, GOLD 2017 Global Strategy for the Diagnosis, Management and Prevention of COPD, Global Initiative for Chronic ObstructiveLung Disease-OLD, 2018, https://goldcopd.org/ gold-2017-global-strategy-diagnosis-management-prevention- copd/. [14] L. J. Epstein, D. Kristo, P. J. Strollo et al., “Clinical guideline for the evaluation, management and long-term care of ob- structive sleep apnea in adults,” Journal of Clinical Sleep Medicine, vol. 5, no. 3, pp. 263–276, 2009. [15] A. Chanda, J. S. Kwon, A. J. Wolff, and C. A. Manthous, “Positive pressure for obesity hypoventilation syndrome,” Pulmonary Medicine, vol. 2012, Article ID 568690, 9 pages, 2012. [16] B. J. Kelly and M. A. Matthay, “Prevalence and severity of neurologic dysfunction in critically III patients,” Chest, vol. 104, no. 6, pp. 1818–1824, 1993. [17] A. Morandi, L. M. Solberg, R. Habermann et al., “Docu- mentation and management of words associated with de- lirium among elderly patients in postacute care: a pilot investigation,” Journal of the American Medical Directors Association, vol. 10, no. 5, pp. 330–334, 2009. [18] S. K. Inouye, L. Leo-Summers, Y. Zhang, S. T. Bogardus, D. L. Leslie, and J. V. Agostini, “A chart-based method for identification of delirium: validation compared with