First-Month Adherence to Home Non-Invasive Ventilation in Stable Hypercapnic COPD: Associations with BMI and Bicarbonate Levels Authors: Nezihe Ciftaslan Goksenoglu, Zuhal Karakurt, Sinem Gungor, Ozlem Yazıcıoglu Mocin, Huriye Berk Takır, Emine Aksoy, Birsen Ocaklı, Gokay Gungor, Nalan Adıguzel Published: 2026 ABSTRACT IntroductionLong-term home non-invasive mechanical ventilation (NIV) is widely used in patients with hypercapnic chronic obstructive pulmonary disease (COPD). However, its efficacy is heavily dependent on patient adherence. Body mass index (BMI) can influence tolerance and ventilatory load, but its impact on adherence remains unclear. Additionally, NIV for later use at home is frequently initiated during hospitalization to facilitate early discharge, and its ongoing necessity varies.MethodsThis retrospective cohort study included patients with stable hypercapnic COPD prescribed home NIV between 2017 and 2025 at a tertiary respiratory intensive care unit outpatient clinic. Demographics, arterial blood gases, spirometry, ventilator settings, and first-month NIV usage data were extracted from the patients’ medical records. We used correlation analyses and Poisson regression modeling to evaluate factors associated with daily hours of NIV use.ResultsA total of 413 patients were included. The mean partial pressure of carbon dioxide was 48 ± 7 mmHg. The mean forced expiratory volume in the first second was 28%, indicating severe airflow limitation. Overall adherence was very poor, with approximately 80% of patients using NIV for <4 hours a day, regardless of their BMI category. We found no significant linear correlation between BMI and NIV adherence. A large proportion of the low-adherence group had elevated bicarbonate (HCO3−) levels (≥27 mmol/L), suggesting chronic compensated hypercapnia. Multivariable analysis found BMI to predict slightly higher adherence (risk ratio [RR] 1.01; p = 0.050), whereas higher HCO3− was associated with lower adherence (RR 0.96; p = 0.020).DiscussionBMI was not a major determinant of first-month NIV adherence, whereas elevated HCO3− levels may help identify patients at increased risk of poor early NIV adherence. Introduction Chronic obstructive pulmonary disease (COPD) is a progressive condition characterized by persistent airflow limitation and recurrent exacerbations. It frequently results in chronic hypercapnic respiratory failure.1,2 Long-term home non-invasive ventilation (NIV) is an established therapeutic option for stable hypercapnic COPD. In appropriately selected patients, its potential benefits include improved gas exchange, reduced nocturnal hypoventilation, decreased hospitalization rates, and enhanced survival.3 However, clinical improvements resulting from home NIV use are heavily dependent on adequate adherence. Previous studies indicate that at least 4–5 hours of daily use are required to achieve physiological benefits.4–6 Despite these considerations, adherence in COPD cohorts is frequently suboptimal, and discontinuation or very low nightly use is commonly reported.2,7 Various potential contributors to poor NIV adherence have been suggested, including symptom burden, ventilator settings, mask/interface comfort, patient-perceived benefits, and body mass index (BMI).6,8 In addition, adherencemay also be influenced by patient education, psychological factors, family support, socioeconomic barriers, and structured post-discharge follow-up. However, the relationship between BMI and home NIV adherence in COPD is unclear. Some studies have reported that patients with obesity and COPD, who represent the systemic COPD phenotype, demonstrate better tolerance, more consistent NIV use, and improved survival rates.9 However, others have found no meaningful association between BMI and NIV adherence.5,6 This suggests that BMI alone may not determine ventilator use. In some clinical settings, home NIV may be initiated during hospitalization following an acute hypercapnic exacerbation to facilitate hospital discharge. However, the timing and indications for NIV initiation vary according to individual clinical circumstances.3 Elevated bicarbonate (HCO3−) reflects chronic renal compensation for persistent hypercapnia and may indicate long-standing ventilatory insufficiency. Whether this physiological marker is associated with early adherence to home NIV remains unclear. Therefore, this study aimed to assess the relationship between BMI and first-month NIV adherence in patients with stable hypercapnic COPD and to identify clinical markers associated with poor adherence. This study addresses an important gap in the literature regarding early adherence to home NIV in routine clinical practice. Methods This study was approved by our local ethics committee (20.01.2026/2024-122) and was conducted in accordance with the ethical principles of the 1964 Declaration of Helsinki and its later amendments. The informed consent requirement was waived due to the retrospective nature of the study. Patient consent for review of medical records was not required by the University of Health Sciences, Istanbul Sureyyapasa Chest Diseases and Thoracic Surgery Training and Research Hospital. All patient data were anonymized before analysis, and patient confidentiality was strictly maintained throughout the study. Study Design This was a retrospective observational cohort study that aimed to evaluate home NIV adherence in patients with COPD.6 Setting The study was conducted in the respiratory intensive care unit outpatient clinic of a tertiary government teaching hospital. This was an equivalent environment to those of previous NIV monitoring studies of European and international COPD cohorts.9 Patients Eligible patients were those with a confirmed diagnosis of COPD, which was defined as a post-bronchodilator forced expiratory volume in the first second/forced vital capacity (FEV1/FVC) <0.70,1 and chronic hypercapnic respiratory failure. The criteria for a diagnosis of hypercapnia were partial pressure of carbon dioxide (PaCO2) ≥45 mmHg and potential of hydrogen (pH) >7.34. These diagnostic measures are consistent with the guideline thresholds for home NIV.3 The inclusion criteria were prescription of home NIV during hospitalization or outpatient evaluation;3 clinical stability defined as the absence of acute COPD exacerbation, hospitalization, or active respiratory infection requiring treatment at the first-month outpatient follow-up visit;5 available arterial blood gas (ABG), spirometry, and NIV adherence data.6 A total of 830 patients prescribed home NIV between 2017 and 2025 were screened for eligibility. After application of the inclusion and exclusion criteria, 413 patients were included in the final analysis (Figure 1).Figure 1Flow diagram of patient selection and study inclusion. Patients with hypercapnic COPD prescribed home non-invasive ventilation (NIV) between 2017 and 2025 were screened for eligibility. After applying the inclusion and exclusion criteria, 413 patients were included in the final analysis.Flowchart of patient selection for chronic respiratory failure study with inclusion and exclusion criteria.The flowchart outlines the patient selection process for a study on chronic respiratory failure. It begins with 830 patients from an intensive care unit outpatient clinic. Inclusion criteria include age 18 years or older, documented diagnosis of stable hypercapnic COPD and use of home NIV therapy for at least one month. Exclusion criteria include OHS, interstitial lung disease, kyphoscoliosis, tracheostomized patients, active malignancy, unstable cardiac disease, recent acute COPD exacerbation and inadequate data. After applying these criteria, 413 patients are deemed eligible. These are divided into two groups: adherence, with NIV usage of 4 hours or more per day (73 patients) and non-adherence, with less than 4 hours per day (340 patients). The process concludes with a first control after discharge, focusing on demographic and clinical data. Flow diagram of patient selection and study inclusion. Patients with hypercapnic COPD prescribed home non-invasive ventilation (NIV) between 2017 and 2025 were screened for eligibility. After applying the inclusion and exclusion criteria, 413 patients were included in the final analysis. The exclusion criteria were neuromuscular diseases, obesity hypoventilation syndrome, and chest wall disorders.7 Data Collection The data collected were consistent with that collected in previous large observational NIV cohorts.7,9 Demographic information included age, sex, and BMI. Patients were grouped into the standard BMI categories, where a BMI <22 kg/m2 is underweight, a BMI of 22–29 kg/m2 is in the healthy/overweight range, and a BMI ≥30 kg/m2 is obese. The ABG data were those recommended for evaluating chronic hypercapnia and metabolic compensation, ie, pH, PaCO2, partial pressure of oxygen (PaO2), and HCO3−.3 Spirometry data on FEV1 and FVC were gathered, as these are essential markers of COPD severity.1 Arterial blood gas measurements, including bicarbonate (HCO3−) levels, were obtained at the first outpatient follow-up visit approximately one month after hospital discharge and home NIV initiation. Arterial blood gas analysis was performed using the institutional blood gas analyzer, and bicarbonate (HCO3−) values were obtained as the analyzer-derived (calculated) values routinely generated from arterial blood gas measurements. We recorded the inspiratory positive airway pressure (IPAP) and expiratory positive airway pressure (EPAP) ventilator settings. High-intensity NIV strategies for hypercapnic COPD typically use IPAP levels of around 22–30 cmH2O.5 We collected NIV adherence data. This was recorded as the average daily hours of use and was obtained from device SD card, which are the gold standard for adherence measurement.6 Days without recorded device use were recorded as 0 hours of NIV use. The device-recorded adherence data were cross-checked with patient-reported daily NIV use during the outpatient visit to ensure consistency. We chose to focus on the adherence data from the first month of use as this has been shown to strongly predict long-term NIV continuation.7 The initial ventilation parameters used were the same for all patients. NIV was initiated with bilevel positive airway pressure (Bi-PAP) in spontaneous/timed (S/T) mode, per the guideline recommendations for hypercapnic COPD.3 IPAP and EPAP were then titrated to a tidal volume of 6 mL/kg of ideal body weight during each patient’s stay in our intensive care unit (ICU) by our ICU ventilators to optimize PaCO2 reduction and patient comfort. This is consistent with high-intensity NIV protocols.5 Before hospital discharge, ventilator pressures were individually adjusted according to serial arterial blood gas measurements and the patient’s clinical status to optimize PaCO2 reduction and patient comfort. The same institutional NIV initiation and titration protocol was applied throughout the study period (2017–2025). Finally, we recorded the interface selection (nasal or oronasal mask). In each patient, selection was based on tolerance, as recommended for optimum adherence.6 Outcomes The primary outcome was first-month NIV adherence, which was defined as the average number of hours of daily usage. We defined the adherence threshold as ≥4 hours a day as clinical improvements and reductions in PaCO2 have consistently been associated with at least 4–5 hours of nightly NIV use.4,6 The secondary outcomes were the associations of adherence with BMI, ABG variables, lung function, and ventilator parameters. These are consistent with the outcomes used in prior adherence prediction models.9 Statistical Analysis SPSS for Windows, v. 22 (IBM Corp., Armonk, NY, USA) was used for statistical analysis. Continuous variables were presented as mean ± standard deviation or median and interquartile range (IQR) based on variable distribution characteristics (Shah – Lung – 2022). Categorical variables were described as frequencies and percentages. Spearman correlation was used to determine the relationships between variables, as NIV adherence data are generally not normally distributed in real-world cohorts.7 We used Poisson regression to model the relationships between continuous explanatory variables and hours of daily NIV use. This was a similar approach to those applied in previous NIV adherence analyses.9 Potential overdispersion was evaluated using the Pearson chi-square/degrees of freedom and deviance/degrees of freedom statistics. In addition, a negative binomial regression model was fitted as a sensitivity analysis to assess the robustness of the findings P-values < 0.05 were considered significant. Results A total of 413 patients with stable hypercapnic COPD were included in the study. The mean age was 65 ± 9, years, and the majority were male (79%). The mean BMI was 27 kg/m2. Overall, 47% of patients had a BMI of 22–29 kg/m2 and 32% classified as obese (BMI ≥ 30 kg/m2) (Table 1). The median daily NIV use during the first month of follow-up was 1 hour a day in our cohort. The median number of cigar packs a year among current smokers was 50. Most patients were ever-smokers (82%). Additional exposures included biomass smoke (13%) and other environmental/occupational factors (67%). Comorbidities were common. The most frequent were diabetes in 18% and anemia (Hb < 10 g/dL) in 22% (Table 1).Table 1Patients Characteristics, Respiratory ParametersAge, mean (SD)65 (10)Male, N (%)324 (79)Body mass index (BMI) BMI, mean (SD)27 (7) BMI [22−29]196 (47) BMI ≥ 30132 (32) BMI < 2285 (21) BMI < 30281 (68)NIV use per/day median (1st-3rd quartile)1 (1–3)Smoking History Pack-years of smoking, median (IQR)50 (35–70) Ever-smokers, n (%)338 (82) Biomass, N (%)55 (13) Other Exposure, N (%)274 (67)Diabetes, N (%)73 (18)Anemia (Hb <10 g/dL), n (%)87 (22)HCT (%), mean (SD)40 (6)Arterial Blood Gases on first month pH, mean (SD)7.41 (0.04) PaCO2 mmHg, mean (SD)48 (7) PaO2 mmHg, (1st-3rd quartile)65 (55–75) HCO3 mmol/L, mean (SD)30 (4) FiO2 (%), mean (SD)23 (5) Oxygen saturation (%), mean (SD)91 (7) PaO2/FiO2 (1st–3rd quartile)295 (243–343)Pulmonary Function Tests, median (1st-3rd quartile) FEV1 mL720 (530–960) FEV1%28 (21–38) FVC mL1370 (1010–1760) FVC%41 (33–52) FEV1/FVC57 (48–64)IPAP, (1st, 3 rd quartile)25 (22–28)EPAP, mean (SD)5 (1)Abbreviations: NIV, noninvasive ventilation, SD, standard deviation, BMI, body mass index, Hb, hemoglobin, HCT, Hematocrit, FEV1, Forced expiratory volume in 1 second, FVC, Forced vital capacity, IPAP, Inspiratory Positive Airway Pressure, EPAP, Expiratory Positive Airway Pressure. Table 1Patients Characteristics, Respiratory ParametersAge, mean (SD)65 (10)Male, N (%)324 (79)Body mass index (BMI) BMI, mean (SD)27 (7) BMI [22−29]196 (47) BMI ≥ 30132 (32) BMI < 2285 (21) BMI < 30281 (68)NIV use per/day median (1st-3rd quartile)1 (1–3)Smoking History Pack-years of smoking, median (IQR)50 (35–70) Ever-smokers, n (%)338 (82) Biomass, N (%)55 (13) Other Exposure, N (%)274 (67)Diabetes, N (%)73 (18)Anemia (Hb <10 g/dL), n (%)87 (22)HCT (%), mean (SD)40 (6)Arterial Blood Gases on first month pH, mean (SD)7.41 (0.04) PaCO2 mmHg, mean (SD)48 (7) PaO2 mmHg, (1st-3rd quartile)65 (55–75) HCO3 mmol/L, mean (SD)30 (4) FiO2 (%), mean (SD)23 (5) Oxygen saturation (%), mean (SD)91 (7) PaO2/FiO2 (1st–3rd quartile)295 (243–343)Pulmonary Function Tests, median (1st-3rd quartile) FEV1 mL720 (530–960) FEV1%28 (21–38) FVC mL1370 (1010–1760) FVC%41 (33–52) FEV1/FVC57 (48–64)IPAP, (1st, 3 rd quartile)25 (22–28)EPAP, mean (SD)5 (1)Abbreviations: NIV, noninvasive ventilation, SD, standard deviation, BMI, body mass index, Hb, hemoglobin, HCT, Hematocrit, FEV1, Forced expiratory volume in 1 second, FVC, Forced vital capacity, IPAP, Inspiratory Positive Airway Pressure, EPAP, Expiratory Positive Airway Pressure. Patients Characteristics, Respiratory Parameters BMI, mean (SD) BMI, mean (SD) BMI [22−29] BMI [22−29] BMI ≥ 30 BMI ≥ 30 BMI < 22 BMI < 22 BMI < 30 BMI < 30 Pack-years of smoking, median (IQR) Pack-years of smoking, median (IQR) Ever-smokers, n (%) Ever-smokers, n (%) Biomass, N (%) Biomass, N (%) Other Exposure, N (%) Other Exposure, N (%) pH, mean (SD) pH, mean (SD) PaCO2 mmHg, mean (SD) PaCO2 mmHg, mean (SD) PaO2 mmHg, (1st-3rd quartile) PaO2 mmHg, (1st-3rd quartile) HCO3 mmol/L, mean (SD) HCO3 mmol/L, mean (SD) FiO2 (%), mean (SD) FiO2 (%), mean (SD) Oxygen saturation (%), mean (SD) Oxygen saturation (%), mean (SD) PaO2/FiO2 (1st–3rd quartile) PaO2/FiO2 (1st–3rd quartile) FEV1 mL FEV1 mL FEV1% FEV1% FVC mL FVC mL FVC% FVC% FEV1/FVC FEV1/FVC Abbreviations: NIV, noninvasive ventilation, SD, standard deviation, BMI, body mass index, Hb, hemoglobin, HCT, Hematocrit, FEV1, Forced expiratory volume in 1 second, FVC, Forced vital capacity, IPAP, Inspiratory Positive Airway Pressure, EPAP, Expiratory Positive Airway Pressure. The results of our correlation analysis of the factors associated with daily hours of NIV use are presented in Table 2. We found no significant linear association between NIV use and BMI (r = 0.04, p = 0.43). Similarly, we found no significant correlations between pH, PaCO2, PaO2/fraction of inspired oxygen (FiO2), FEV1, FVC, IPAP, or EPAP and adherence. The only significant correlation identified was between NIV use and HCO3− levels, with higher HCO3− negatively associated with daily NIV use (p = 0.010).Table 2Correlation of NIV Use Hours per Day of Body Mass Index, Respiratory Variations and PressuresVariableRhop valueBMI kg/m20.040.43Ph−0.080.11PaCO2−0.090.08HCO3−0.140.01PaO2/FiO20.060.19FEV1%0.050.36FVC%0.060.22IPAP−0.070.14EPAP−0.030.52 Table 2Correlation of NIV Use Hours per Day of Body Mass Index, Respiratory Variations and PressuresVariableRhop valueBMI kg/m20.040.43Ph−0.080.11PaCO2−0.090.08HCO3−0.140.01PaO2/FiO20.060.19FEV1%0.050.36FVC%0.060.22IPAP−0.070.14EPAP−0.030.52 Correlation of NIV Use Hours per Day of Body Mass Index, Respiratory Variations and Pressures Higher HCO3− levels were associated with fewer hours of daily NIV use. The daily NIV use of participants grouped by BMI is shown in Figure 2. Median NIV adherence was similarly low across all three BMI weight groups. The overall distribution of adherence hours did not differ significantly between BMI groups (p = 0.56). Although obese patients (BMI ≥ 30 kg/m2) displayed slightly higher median and upper-range NIV use than the other two groups, the difference did not reach statistical significance. Within all three BMI groups, there was a high proportion of patients with very low NIV adherence, clustering around 1 hour a day. There was also a wide range of outliers in each group. This pattern indicates that BMI was not a major determinant of first-month NIV usage in this real-world cohort, with consistent suboptimal adherence across all BMI groups.Figure 2Daily NIV use hours across BMI categories (BMI <22 kg/m2, 22–29 kg/m2, and ≥30 kg/m2). Median NIV usage did not significantly differ among BMI groups (p = 0.56). Box plots demonstrate median values, interquartile ranges, and extreme outliers. Circles indicate outliers, and stars indicate extreme outliers.A box plot showing daily NIV use across BMI categories. Daily NIV use hours across BMI categories (BMI <22 kg/m2, 22–29 kg/m2, and ≥30 kg/m2). Median NIV usage did not significantly differ among BMI groups (p = 0.56). Box plots demonstrate median values, interquartile ranges, and extreme outliers. Circles indicate outliers, and stars indicate extreme outliers. To identify independent predictors of daily NIV use, we performed multivariable Poisson regression analysis. The results are shown in Table 3. Each 1 kg/m2 increase in BMI was associated with a 1% increase in average daily NIV use (risk ratio [RR] 1.01, 95% CI: 1.00–1.02).Table 3Association Between NIV Use Hours and Associated Factors Was Assed Using Poisson Regression ModelParameterRisk ratio95% Wald Confidence Interval for Risk Ratiop valueLower LimitUpper LimitIntercept4.651.8811.520.010Body mass index kg/m21.011.001.020.050HB0.990.961.020.29CO2 mm/Hg0.980.961.010.73HCO3 mmol/L0.960.950.990.020FEV10.970.950.990.010FVC1.021.011.030.010IPAP1.000.991.010.34EPAP1.030.941.130.57 Table 3Association Between NIV Use Hours and Associated Factors Was Assed Using Poisson Regression ModelParameterRisk ratio95% Wald Confidence Interval for Risk Ratiop valueLower LimitUpper LimitIntercept4.651.8811.520.010Body mass index kg/m21.011.001.020.050HB0.990.961.020.29CO2 mm/Hg0.980.961.010.73HCO3 mmol/L0.960.950.990.020FEV10.970.950.990.010FVC1.021.011.030.010IPAP1.000.991.010.34EPAP1.030.941.130.57 Association Between NIV Use Hours and Associated Factors Was Assed Using Poisson Regression Model Although modest in magnitude, higher BMI was independently associated with greater daily NIV use after adjustment for the other variables in the model. We found a significant negative association between serum HCO3− levels and NIV usage (RR 0.96, 95% CI: 0.95–1.00, p = 0.020). Thus, patients with higher HCO3−, which reflects more pronounced metabolic compensation for chronic hypercapnia, used NIV for fewer hours per day. Analysis of patients’ lung function parameters showed opposite trends. While higher FEV1 predicted less NIV use (RR 0.97, p = 0.010), higher FVC predicted higher NIV use (RR 1.02, p = 0.010). We found no independent associations between ventilator pressures (IPAP and EPAP) and NIV adherence (p > 0.05). Overall, the regression model indicates that BMI and FVC contribute positively to early NIV adherence, whereas higher HCO3− levels and lower FEV1 are associated with diminished adherence. Discussion This study revealed markedly low early adherence to home NIV in patients with stable hypercapnic COPD, with almost 80% of patients using the device for fewer than 4 hours a day during the first month. This poor adherence pattern was consistent across all BMI categories, indicating that weight status alone does not explain early NIV use. Patients with elevated HCO3− levels, reflecting chronic compensated hypercapnia, were predominantly among the low-adherence group. In an adjusted regression analysis, we found a higher BMI to be modestly associated with greater NIV adherence. However, reduced adherence demonstrated much stronger relationships with higher HCO3− and lower FEV1. These findings suggest that physiological compensation and clinical status may play more important roles in early NIV use than BMI-defined phenotypes. The overall demographic and clinical profile of our cohort is broadly consistent with previously reported real-world COPD populations treated with home NIV. The trend toward older age and a predominance of men reflects typical epidemiologic patterns of COPD of sufficient severity to require ventilatory support.2 Our cohort’s mean BMI of 27 kg/m2 falls within the overweight-to-obese range. Again, this is common in hypercapnic COPD cohorts, and it is thought that obesity may increase ventilatory load and partially overlap with sleep-disordered breathing phenotypes.9 The distribution of patients between the BMI categories in our study, particularly the substantial proportion of obese patients (32%), aligns with contemporary data showing rising obesity rates among patients with advanced COPD.7 The median NIV usage of only 1 hour a day during the first month is notably lower than the adherence levels reported in many European and North American cohorts. These studies typically report adherence rates (with adherence defined as ≥4 hours a day of nightly NIV use) of 40–60%.6 This striking discrepancy emphasizes the real-world environment in which our data were gathered and suggests that early post-discharge adherence issues may be more pronounced in settings where NIV initiation aims to facilitate earlier discharge rather than occurring within a structured outpatient treatment plan. The high burden of comorbidities observed in our cohort, particularly diabetes and anemia, reflects the clinical fragility of this patient population, among whom factors such as fatigue, mask intolerance, or limited perceived benefits may contribute to poor early adherence. Overall, the baseline characteristics of our cohort are consistent with the expected profile of patients receiving home NIV, yet the exceptionally low first-month adherence rate distinguishes our findings and underscores the need for closer early follow-up and optimization of patient support strategies. In our study, the majority of the respiratory and ventilatory parameters, including pH, PaCO2, PaO2/FiO2, IPAP, EPAP, FEV1, and FVC, showed no significant linear correlations with daily NIV usage. This is concordant with previous findings from real-world NIV cohorts, which have similarly found that objective gas exchange variables and ventilator pressures do not predict adherence behavior.6,7 Previous studies have demonstrated that physiological severity does not reliably translate into greater NIV use, as measurable respiratory abnormalities do not necessarily increase perceived symptomatic improvements. The only significant correlation observed in our cohort was a negative association between HCO3− levels and NIV adherence. Elevated HCO3− levels represent chronic renal compensation for longstanding hypercapnia. They have been proposed as an indicator of chronic ventilatory insufficiency; however, their relationship with NIV adherence has received little research attention. In our study, higher HCO3− levels were independently associated with lower NIV adherence. Because arterial blood gas measurements and device-recorded adherence data were obtained during the same follow-up visit, this finding should be interpreted as an association rather than evidence of a causal relationship. This finding suggests that patients with compensated hypercapnia may represent a subgroup at increased risk of early NIV non-adherence; however, prospective studies are needed to clarify the temporal and causal nature of this association. This notion has previously been implied, but to our knowledge, the present study is among the first to quantitatively evaluate this association.2,5 The lack of correlation between ventilator pressures (IPAP/EPAP) and NIV adherence aligns with existing evidence that factors such as mask comfort, patient education levels, and behavioral factors affect adherence more than pressure levels. Similarly, the lack of significant associations between adherence and either FEV1 or FVC is consistent with research showing that spirometry-based measures of severity do not reliably predict NIV adherence.9 Collectively, the results of our correlation analysis demonstrate that adherence is driven less by objective physiological markers and more by the perceived necessity, clinical stability, and compensatory metabolic state of the patient. The distribution of NIV use across BMI groups (Figure 1) showed uniformly low adherence in all groups, with no statistically significant difference between underweight, normal/overweight, and obese patients. This pattern contrasts with that found in several previous studies, which have suggested that obese patients with COPD may demonstrate better tolerance and more consistent use of nocturnal NIV.7,9 Obesity has been described as the systemic COPD phenotype. Earlier studies have suggested that the increased ventilatory load, comorbid sleep-disordered breathing, and higher inspiratory effort often found in obese patients might translate into greater perceived benefit from NIV. However, our real-world data did not reflect this pattern. Although obese patients showed slightly higher upper-range usage, overall adherence remained low and did not differ significantly from that in the other BMI groups. This lack of divergence across BMI strata suggests that BMI is not a reliable determinant of early NIV adherence. Instead, the low usage across all BMI groups may reflect systemic factors such as a low education level in our cohort, early post-discharge instability, poor interface comfort, or insufficient symptom severity to motivate nightly NIV use. This highlights the importance of in Another important consideration is patient selection for long-term home NIV. The HOT-HMV trial demonstrated that patients with persistent hypercapnia reassessed 2–4 weeks after recovery from an acute hypercapnic exacerbation derived significant benefit from home NIV in addition to home oxygen therapy, with a reduced risk of hospital readmission or death. In contrast, our study reflects routine clinical practice during the study period, in which patients were discharged with home NIV according to clinical judgment and were not systematically reassessed for persistent hypercapnia before treatment initiation. Consequently, our cohort may represent a more heterogeneous real-world population than that included in the HOT-HMV trial. This difference in patient selection should be considered when interpreting our findings and may partly explain the variability in early NIV adherence observed in our study. Dividualized adherence support strategies, irrespective of the patient’s BMI category.10 Our multivariable Poisson regression model provided additional insights into the determinants of early NIV non-adherence. One of the most noteworthy findings was the independent but modest positive association between BMI and daily NIV use. This aligns with the hypothesis that greater ventilatory load in obese patients may translate to greater perceived symptomatic benefits from nocturnal NIV, potentially supporting treatment adherence.7,9 However, the effect size we observed was small, and BMI did not meaningfully differentiate between adherence categories. Therefore, while BMI may exert some influence, it is insufficient as a stand-alone predictor of NIV use. In contrast, the independent association between higher HCO3− levels and reduced adherence was strong. This finding suggests that patients with more pronounced metabolic compensation for chronic hypercapnia may be less symptomatically motivated to use their NIV device. This result constitutes a novel contribution to the literature, as previous NIV trials and registries have rarely examined HCO3− as an adherence predictor.2,5 Our results imply that the compensated hypercapnia reflected by elevated HCO3−can be used to identify a subgroup at risk of early non-adherence. The likely explanation for this is that these patients experience less severe symptoms and may have been prescribed NIV during hospitalization to facilitate early discharge, rather than to address a persistent symptomatic burden. The opposite effects of FVC and FEV1 on adherence further highlight the complexity of NIV adherence behavior. Greater FVC was associated with increased adherence, possibly reflecting preserved lung volumes and better tolerance of ventilatory support. Conversely, higher FEV1 predicted lower NIV use. This is a counterintuitive but previously noted pattern in real-world cohorts. It appears to reflect a reduced perceived need for nightly ventilation among patients with milder airflow obstruction.6 Importantly, ventilator pressures (IPAP and EPAP) were not independent predictors. This was also consistent with prior data showing that pressure settings are far less influential than patient comfort, education levels, and perceived benefit. Taken together, the findings of our Poisson model demonstrate that early NIV adherence is shaped less by disease severity metrics and ventilatory settings and more by physiological compensation and individual patient characteristics. These findings reinforce the critical need for structured early follow-up. Even with adjustments for multiple clinical variables, the overall adherence level observed in this study remained extremely low. This poor first-month usage was not simply attributable to demographic or physiological factors, but reflects the broader real-world challenge of ensuring early NIV implementation. This study had several limitations that should be acknowledged. First, patients were drawn from a single tertiary respiratory center. Nevertheless, this center is among the largest outpatient clinics for chronic respiratory failure and home NIV follow-up in the region, and its medical personnel have substantial experience and expertise in this field. Moreover, the real-world data obtained strengthens the relevance of our findings. Second, the available adherence data were limited to that obtained at the first-month follow-up visit. As a result, we were unable to assess whether NIV usage improves with time, repeated education on its use and benefits, or adjustment of ventilator settings during longer-term follow-up. Future studies evaluating adherence trajectories over several months would clarify whether early non-adherence persists over time or responds to structured interventions. Third, we did not investigate the underlying reasons for poor adherence. We did not assess factors such as low patient education levels, socioeconomic limitations, mask intolerance, psychological barriers, and a lack of family support. Yet, all of these could potentially contribute to early disengagement from home NIV. Systematic evaluation of these behavioral and contextual determinants would provide valuable insights that might be utilized in the development of targeted strategies for adherence improvement. Fourth, although elevated bicarbonate levels were independently associated with reduced NIV adherence, information regarding potential confounding factors that may influence bicarbonate levels, such as loop diuretic use or other causes of metabolic alkalosis, was not systematically available in our retrospective database. Therefore, adjustment for these potential confounders was not possible, and the observed association should be interpreted with appropriate caution. Fifth, our study reflects routine clinical practice during the study period, and patients were not systematically reassessed for persistent hypercapnia 2–4 weeks after an acute exacerbation before initiation of long-term home NIV, as recommended by the HOT-HMV trial. Consequently, some degree of heterogeneity in patient selection cannot be excluded and should be considered when interpreting our findings. Another limitation of this study is the relatively short follow-up period. As the primary objective was to evaluate first-month home NIV adherence, longer-term clinical outcomes, including changes in PaCO2, hospital readmissions, quality of life, and mortality, could not be assessed. Therefore, our findings should be interpreted as reflecting early NIV adherence rather than the long-term effectiveness of home NIV. Finally, we were unable to evaluate the clinical consequences of adherence patterns, such as changes in hypercapnia, symptoms, or quality of life, and hospital readmission rates, due to a lack of longitudinal outcome data. Future prospective studies that incorporate patient-centered outcomes and structured follow-up protocols are needed to better understand the clinical significance of early NIV adherence in stable hypercapnic patients with COPD with and without obesity. Conclusions In this real-world cohort of patients with stable hypercapnic COPD, first-month adherence to home non-invasive ventilation (NIV) was strikingly poor, with nearly 80% of patients using the device for less than the recommended 4 hours per day. The consistently low adherence observed across all BMI categories suggests that BMI is not a major determinant of early NIV use. In contrast, higher HCO3− levels were independently associated with lower NIV adherence, indicating that physiological markers of chronic compensated hypercapnia may help identify patients at increased risk of poor early adherence. Because arterial blood gas measurements and device-recorded adherence data were obtained during the same follow-up visit, these findings should be interpreted as associations rather than evidence of causality. The first month after hospital discharge appears to be a critical period for identifying patients who may require additional support to achieve adequate NIV use. Structured early follow-up, targeted patient education, and regular reassessment of the ongoing need for long-term home NIV may help optimize adherence and avoid unnecessary continuation of therapy. Future prospective studies with standardized patient selection and longer follow-up are warranted to clarify the temporal relationship between HCO3− levels and NIV adherence and to identify effective strategies for improving long-term engagement with home NIV.