Non-Invasive Ventilation (NIV) and Homeostatic Model Assessment (HOMA) Index in Stable Chronic Obstructive Pulmonary Disease (COPD) Patients with Chronic Hypercapnic Respiratory Failure: A Pilot Study

Abstract

The effects of Non-invasive Ventilation (NIV) on Insulin Resistance (IR) in stable Chronic Obstructive Pulmonary Disease (COPD) patients have not been fully explored. The aim of this study was to assess the effects of NIV on IR and adiponectin levels during one year application of NIV in stable COPD patients with Chronic Hypercapnic Respiratory Failure.

Twenty-five (25) stable COPD patients with Chronic Hypercapnic Respiratory Failure and with no self-reported comorbidities completed the study. NIV was administered in the spontaneous/timed mode via a full face mask using a bi-level positive airway pressure system. Spirometry, blood pressure, arterial blood gases, dyspnea, daytime sleepiness, serum fasting glucose and insulin levels were assessed. IR was assessed with the calculation of the Homeostatic Model Assessment (HOMA) index. Adiponectin was measured with radioimmunoassay. Study participants were re-evaluated on the first, third, sixth, ninth and twelfth month after the initial evaluation.

There was a significant improvement in FEV₁ values from the first month (34.1 ± 11.6% vs 37 ± 12.3%, p = 0.05). There was a significant decrease in IR by the ninth month of NIV use (3.4 ± 2.3 vs 2.2 ± 1.4, p < 0.0001), while adiponectin levels significantly improved from the first month of NIV use. Stepwise regression analysis revealed that baseline HOMA index was associated with p_aCO₂ (⇓ = 0.07 ± 0.02, p = 0.001), while baseline adiponectin levels were associated with FVC (⇓ = 0.05 ± 0.02, p = 0.035) and the concentration of serum bicarbonate (HCO_3-) (-⇓ = 0.18 ± 0.06, p = 0.002). Insulin sensitivity and glucose metabolism as well as adiponectin levels improved along with the improvements in respiratory failure.

Keywords: :

• adiponectin
• insulin resistance
• glucose metabolism
• systemic inflammation

Introduction

Chronic Obstructive Pulmonary Disease (COPD) is a preventable and treatable disease, with cigarette smoking being the major risk factor (1). It is a growing cause of morbidity and mortality worldwide, and will be the third leading cause of death by the year 2020. The burden of this disease is even greater if we consider the significant impact of COPD on cardiovascular morbidity and mortality. The pathophysiological mechanisms and clinical manifestations of COPD are not restricted only to pulmonary inflammation and airway remodeling (2). On the contrary, in the last decade COPD has been recognized as a systemic disease. Inflammatory mediators in the systemic circulation may contribute to skeletal muscle wasting and cachexia, and may initiate or worsen comorbidities such as ischemic heart disease, heart failure, diabetes mellitus (DM), metabolic syndrome, osteoporosis, anemia and depression (1).

As in many other medical conditions, COPD is associated with low grade systemic inflammation that has been associated with insulin resistance (IR). Various factors can contribute to this finding such as the spill of inflammatory mediators from the pulmonary system into the circulation, hypoxia, effects of obesity and hormonal disturbances (2). Insulin resistance has been implicated in the pathogenesis of the metabolic syndrome (3). Furthermore, there is evidence that insulin resistance predisposes to cardiovascular risk (4), whereas, it is well known that insulin resistance is associated with low serum adiponectin levels (5).

Non-invasive mechanical ventilation (NIV) has been shown to have a positive impact on the metabolic profile of patients with Obstructive Sleep Apnea (OSA) and Obesity Hypoventilation Syndrome (OHS) (6,7). In addition, NIV has been shown to improve blood gases and quality of life in COPD patients with Hypercapnic Respiratory Failure (8,9). Yet, the effects of NIV on insulin resistance of stable COPD patients with Chronic Hypercapnic Respiratory Failure have not been explored. The aim of the present study was to assess the effects of NIV on insulin resistance and adiponectin levels during one year application of NIV in this specific group of patients.

Methods

Patients

Twenty-five (25) COPD patients with Chronic Hypercapnic Respiratory Failure who were on stable condition and had stopped smoking for at least two years, were recruited from the outpatient clinic of a tertiary University Hospital. COPD was diagnosed according to the ATS/ERS guidelines (10). Patients’ entry criteria to qualify as suffering from Hypercapnic Respiratory Failure were p_aO₂ < 60 mmHg and a p_aCO₂ > 50 mmHg, while breathing room air. These arterial blood gas values were constant for at least two months for all the patients included in the study, while receiving optimal bronchodilation therapy.

Patients included in the study did not have any self-reported comorbidities (cardiovascular disease, diabetes mellitus, history of cancer, lung disease other than COPD, any other chronic disease). Sixteen (16) patients were receiving antihypertensive drugs. A fasting glucose in the diabetes range was an exclusion criterion. The major problem in the present study was to find patients with a normal fasting glucose. Most of the candidates were excluded for this reason. Probably this is due to the high incidence of the metabolic syndrome in the specific group of patients (COPD with coexisting Chronic Hypercapnic Respiratory Failure).

All patients had undergone a polysomnography (PSG) study during the previous 6 months (Alice 4 Diagnostic Device OBS/G7829, Respironics), and all subjects had apnoea –hypopnoea index (AHI) < 10 episodes/hour without symptoms compatible with OSA. Patients with BMI < 19 and BMI > 35 kg/m² were excluded. No change of medications and no new diagnoses were recorded during the one-year follow-up. No patient was participating in a pulmonary rehabilitation programme before or during the study. The study was approved by the Ethics Committee of the University of Thessaly and all subjects provided informed consent.

Patients were hospitalized for 2–3 days during the initial application of NIV, in order to ensure maximal compliance. NIV was administered in the spontaneous/timed mode via a full face mask using a bi-level positive airway pressure system (VPAP III ST, ResMed, Sydney, Australia). This mode was chosen for the ability to set inspiratory and expiratory times and it has been proven effective in COPD patients (11). Inspiratory and expiratory positive airway pressures (IPAP and EPAP respectively) were adjusted according to the patient's comfort and synchrony with the ventilator and a marked reduction in the use of accessory muscles. For the first 2 hours from the initiation of ventilation the patient was observed by a pulmonologist and a fully trained nurse, in order to assure a decrease in accessory respiratory muscle use and respiratory rate, check the patient's comfort of mask ventilation and possible air leaks of the system.

Supplemental oxygen was added as needed in order to maintain oxygen saturation between 88% and 92%. Arterial blood gases were measured one hour after the initiation of ventilation. A 5% decrease in p_aCO₂ values was considered as adequate ventilation support. Subsequently, the patients and their relatives were trained in-hospital until they were fully confident in using the ventilator. Technically skilled personnel installed the ventilator at the patients’ homes and provided full technical support when required. Only two (2) patients dropped-out of the study (originally, 27 patients began receiving NIV). We believe that the small drop-out rate was due to the close follow-up by the study coordinators. In addition, giving to the patients’ close relatives an active role in the monitoring of their compliance (in each re-evaluation all the patients were interviewed with their spouses or the relatives who lived in the same house) proved to be very effective.

Physiological measurements

Spirometry was performed with a dry spirometer (KoKo Legend, Ferraris, UK), according to the ATS guidelines (12). Arterial blood gases were measured at rest, with the patient in sitting position, while breathing room air. Dyspnea was assessed with the Medical Research Council (MRC) dyspnea scale (13). Subjective daytime sleepiness was evaluated with the Epworth Sleepiness Scale (ESS), a validated eight-item, self-completion questionnaire (14), that has been validated for the Greek language (15).

Fasting morning venous blood samples were collected from each subject, they were immediately centrifuged at 1500 g for 15 min at 4°C and the supernatants (serum) were stored at −80°C for subsequent analysis. Serum fasting glucose was measured with a commercially available analyzer (Olympus 2700 AU, Olympus Life and material Science Europa GmbH, Hamburg, Deutschland). Insulin levels were measured with a commercially available analyzer (Roche E-170 Modular, Indianapolis, USA) and insulin resistance was assessed with the calculation of the Homeostatic Model Assessment (HOMA) index, using the formula: fasting serum insulin (mU/L) × fasting plasma glucose (mg/dL)/405. Adiponectin was measured with radioimmunoassay (LINCO Research, USA) and its sensitivity ranged between 1–200 ng/ml.

Study protocol

Study participants were re-evaluated in the NIV outpatient clinic on the first, third, sixth, ninth and twelfth month after the initial evaluation. If the patient had an exacerbation at any of those time-points, measurements were obtained one month after the event. Overall, six (6) patients had a single exacerbation during the one-year follow-up. In every appointment physical examination was performed and arterial blood gases, blood pressure, spirometry, MRC dyspnea scale and ESS were assessed.

Additionally, the settings and the hours of ventilator use (as obtained from the machines’ time counters) were evaluated, in order to determine proper compliance. During the follow-up period, patients underwent adjustments in NIV masks or ventilator settings as needed, in order to maintain patient-ventilator synchrony and optimize gas exchange and functional status. NIV compliance was confirmed by checking the records in the SD cards of the ventilators. In order to be considered compliant with treatment, each patient had to use the ventilator properly (air leaks less than 10lt/min) for at least 6 hours daily. Nine (9) patients were receiving supplemental oxygen (from 1 L/min to 3 L/min).

Finally, fasting morning venous blood samples were collected from each subject at every appointment. Physical activity was monitored during the study. That is to say, in each re-evaluation the patients were thoroughly interviewed, regarding all aspects of their daily activities. No major alterations in daily activities were reported. No patient was participating in a pulmonary rehabilitation programme before or during the study.

Statistical analysis

Continuous variables are presented with mean ± standard deviation (SD). Normality of distribution of the variables was assessed using the Shapiro-Wilk test. Differences in changes of study variables during the follow up period were evaluated using repeated measurements analysis of variance (ANOVA). Bonferroni correction was used in order to control for type I error in multiple comparisons. HOMA index was log-transformed for the analysis due to its skewed distribution. To define factors associated with changes in Insulin Resistance and adiponectin levels, stepwise mixed linear regression analysis was conducted. P_aO₂, p_aCO₂, pH, HCO_3-, FEV₁, FVC, BMI and age were the independent variables. p values < 0.05 were considered statistically significant. Analyses were performed using SPSS 17 (SPSS, Chicago, IL, USA) and STATA 8.0.

Results

Twenty-five (25) COPD patients with Hypercapnic Respiratory Failure, aged 67.6 (± 6.8 yr), participated in the study. They exhibited an apnoea –hypopnoea index (AHI) of 5.1 units (± 3.9), whereas the mean pack-years value was 50.6 (± 19.0).

Physiological outcomes

There was a constant improvement in spirometric parameters throughout the study period. Specifically there was a significant improvement in FEV₁ values from the first month (34.1 ± 11.6% vs 37 ± 12.3%, p = 0.05) with a further improvement from the third to sixth month (36.1 ± 13.3% vs 38.6 ± 13.1%, p = 0.001). Concerning FVC there was a sustained improvement after the sixth month of NIV use (48.8 ± 12% vs 54.4 ± 11.7%, p = 0.010) which was maintained thereafter. Dyspnea improved from the first month (MRC from 3.7 ± 1 to 3.3 ± 0.9, p = 0.009) and thereafter, and so was ESS (7.5 ± 4 vs 4.9 ± 1.7, p < 0.0001) (Table 1).

Table 1.   Participants’ dyspnea index, daytime sleepiness and pulmonary volumes during follow-up period

	FEV1 (%)	FVC (%)	FEV1/FVC (%)	MRC	ESS
	Mean value ± SD	Mean value ± SD	Mean value ± SD	Mean value ± SD	Mean value ± SD
Before intervention	34.3 ± 11.6	48.8 ± 12	54 ± 9.6	3.7 ± 1	7.5 ± 4
1 month	37 ± 12.3	51.9 ± 11.4	54.3 ± 10.9	3.3 ± 0.9	4.9 ± 1.7
3 months	36.1 ± 13.3	52.2 ± 10.9	53.1 ± 11.6	2.9 ± 1	3.6 ± 1.3
6 months	38.6 ± 13.1	54.4 ± 11.7	53.8 ± 10.9	2.8 ± 0.9	3.6 ± 1.8
9 months	39.4 ± 15.2	55.5 ± 14	54.7 ± 12.6	2.8 ± 0.9	3.2 ± 1.3
12 months	39.8 ± 16.5	55.9 ± 15.4	55 ± 12.3	2.6 ± 1	3.1 ± 1.1
P** before vs 1st month	0.050	0.140	0.792	0.009	<0.001
P** before vs 3rd month	0.179	0.064	0.624	<0.001	<0.001
P** before vs 6th month	0.002	0.010	0.939	<0.001	<0.001
P** before vs 9th month	0.008	0.014	0.709	<0.001	<0.001
P** before vs 12th month	0.012	0.010	0.615	<0.001	<0.001
P** 1st month vs 3rd month	0.437	0.788	0.208	0.004	0.001
P** 3rd month vs 6th month	0.001	0.016	0.324	0.538	0.903
P** 6th month vs 9th month	0.408	0.426	0.426	0.491	0.119
P** 9th month vs 12th month	0.639	0.643	0.733	0.356	0.503
P^‡	0.014	0.014	0.608	<0.001	<0.001

**Difference between measures.

^‡Repeated measures ANOVA. Total time effect.

Boldfaced values in the table represent statistically significant changes.

A significant reduction in p_aCO₂ was observed by the first month (54.0 ± 6.2 vs 46.8 ± 6.5 mmHg, p < 0.0001), Moreover, there was a significant increase in p_aO₂ from the first month (56.9 ± 3.3 vs 62.2 ± 9.4 mmHg, p = 0.013). Both improvements in arterial blood gases were maintained throughout the study period. Overall there was not any significant change in BMI during the one year of the trial (apart from a significant increase seen by the sixth month (25.6 ± 5.8 vs 26.4 ± 4.7 kg/m², p = 0.030) which was not maintained thereafter (Table 2).

Table 2.   BMI and acid-base balance during follow-up period

	paO2 (mmHg)	paCO2 (mmHg)	pH	HCO3- (mmol/l)	BMI (kg/m^2)
	Mean value ± SD	Mean value ± SD	Mean value ± SD	Mean value ± SD	Mean value ± SD
Before intervention	56.9 ± 3.3	54.0 ± 6.2	7.40 ± 0.03	32.4 ± 4.3	25.6 ± 5.8
1 month	62.2 ± 9.4	46.8 ± 6.5	7.41 ± 0.02	29.9 ± 3.8	26 ± 5.2
3 months	62.4 ± 7.7	46 ± 5.9	7.42 ± 0.02	29.4 ± 3.7	26.1 ± 5
6 months	65 ± 7.7	43.4 ± 5.2	7.41 ± 0.03	28 ± 2.2	26.4 ± 4.7
9 months	65.6 ± 6.4	43 ± 5.4	7.41 ± 0.02	28.7 ± 2.6	26.3 ± 4.6
12 months	64.7 ± 6.4	43.8 ± 4	7.41 ± 0.03	27.8 ± 3.1	26.2 ± 4.2
P** before vs 1st month	0.013	<0.001	0.357	0.002	0.058
P** before vs 3rd month	0.004	<0.001	0.027	0.003	0.081
P** before vs 6th month	<0.001	<0.001	0.076	<0.001	0.030
P** before vs 9th month	<0.001	<0.001	0.107	<0.001	0.062
P** before vs 12th month	<0.001	<0.001	0.079	<0.001	0.255
P** 1st month vs 3rd month	0.918	0.224	0.066	0.372	0.582
P** 3rd month vs 6th month	0.081	0.006	0.714	0.082	0.263
P** 6th month vs 9th month	0.626	0.544	0.850	0.104	0.536
P** 9th month vs 12th month	0.248	0.243	0.994	0.148	0.734
P^‡	<0.001	<0.001	0.111	<0.001	0.019

**Difference between measures.

^‡Repeated measures ANOVA. Total time effect.

Boldfaced values in the table represent statistically significant changes.

Concerning HOMA index there was a significant reduction in insulin resistance by the ninth month of NIV use (3.4 ± 2.3 vs 2.2 ± 1.4, p < 0.0001) and the difference remained significant until the twelfth month. As for adiponectin levels there was a significant improvement from the first month of NIV use (11.4 ± 4.9 vs 13.3 ± 5.5 ng/ml, p = 0.003) (Table 3).

Table 3.   Hormonal and HOMA changes during follow-up period

	Insulin (μU/ml)	Glucose (mg/dl)	Adiponectin (ng/ml)	HOMA
	Mean value ± SD	Mean value ± SD	Mean value ± SD	Mean value ± SD
Before intervention	14.5 ± 10.5	93.9 ± 11.6	11.4 ± 4.9	3.4 ± 2.3
1 month	12.5 ± 5.7	91.2 ± 14.9	13.3 ± 5.5	2.9 ± 1.5
3 months	11.8 ± 8.7	91.7 ± 13.5	13.4 ± 6.2	2.9 ± 2.3
6 months	10.5 ± 7.2	91.8 ± 13.4	14 ± 6.3	2.5 ± 1.9
9 months	9.4 ± 5.7	90.5 ± 13	14.2 ± 6.4	2.2 ± 1.4
12 months	8.8 ± 5.1	89.5 ± 12.6	13.4 ± 6.3	2 ± 1.3
P** before vs 1st month	0.225	0.192	0.003	0.959
P** before vs 3rd month	0.205	0.280	0.005	0.680
P** before vs 6th month	0.063	0.265	<0.001	0.083
P** before vs 9th month	0.010	0.087	<0.001	0.007
P** before vs 12th month	0.011	0.023	0.012	0.006
P** 1st month vs 3rd month	0.596	0.725	0.886	0.901
P** 3rd month vs 6th month	0.175	0.919	0.136	0.997
P** 6th month vs 9th month	0.050	0.121	0.468	0.706
P** 9th month vs 12th month	0.504	0.397	0.021	0.986
P^‡	0.003	0.043	0.007	0.001

**Difference between measures.

^‡Repeated measures ANOVA. Total time effect.

Boldfaced values in the table represent statistically significant changes.

Results from univariate regression analysis for HOMA index and adiponectin are shown in Table 4. Baseline p_aCO₂ was associated with HOMA levels, while baseline pH, HCO₃ and FVC were associated with adiponectin levels. Stepwise multiple regression analysis revealed that baseline p_aCO₂ was the only factor associated with HOMA index, while adiponectin levels were associated with baseline FVC (β ± SE = 0.05 ± 0.02, p = 0.035) and baseline HCO₃₋(β ± SE = −0.18 ± 0.06, p = 0.002).

Table 4.   Results from univariate regression analysis for HOMA index and adiponectin

	HOMA		Adiponectin
	β ± SE*	p	β ± SE*	p
PaO2	−0.004 ± 0.006	0.542	0.04 ± 0.03	0.136
paCO2	0.03 ± 0.008	<0.001	−0.05 ± 0.05	0.195
pH	−1.39 ± 1.96	0.476	−18.03 ± 8.71	0.038
HCO3	0.0004 ± 0.01	0.978	−0.20 ± 0.06	<0.001
FEV1	−0.005 ± 0.006	0.436	0.002 ± 0.03	0.939
FVC	−0.004 ± 0.005	0.526	0.06 ± 0.02	0.009
BMI	0.03 ± 0.02	0.229	0.18 ± 0.13	0.156
Age (years)	−0.01 ± 0.02	0.478	0.19 ± 0.17	0.272

*Regression coefficient ± standard error.

Discussion

In this pilot study, we have evaluated the effects of long-term application of Non-invasive mechanical ­Ventilation (NIV) in stable COPD patients with Chronic Hypercapnic Respiratory Failure on insulin resistance and adiponectin levels. Specifically we examined the effects of NIV on glucose metabolism in stable hypercapnic COPD patients and we found that insulin resistance improved during the follow-up period and the same was observed for adiponectin levels. Improvements in glucose metabolism were accompanied by improvements in spirometry and arterial blood gases, whereas there was not any change on BMI.

It is well known that COPD patients have an increased risk of developing metabolic syndrome which is associated with an increase in the levels of systemic inflammatory markers and physical inactivity in these patients irrespective of lung function impairment (16). The metabolic syndrome represents a cluster of risk factors (abdominal obesity, atherogenic dyslipidaimia, hypertension and insulin resistance) that predispose affected patients to systemic inflammation and cardiovascular risk (17). Low grade systemic inflammation in COPD has been associated with insulin resistance (2). Yet, to our knowledge there is not any study addressing the issue of the change in systemic inflammation and insulin resistance after the initiation of a therapeutic regime.

Adiponectin, which is an anti-inflammatory adipokine, regulates insulin sensitivity and decreases in obesity induced insulin resistance (18). It has been shown that in patients with COPD, plasma adiponectin levels are higher than in controls and negatively correlate with BMI (19). Moreover, in COPD exacerbations adiponectin levels have been found increased (20,21). However, COPD patients have been shown to have impaired glucose metabolism as expressed by increased HOMA-IR (22). Insulin resistance on the other hand is associated with low adiponectin levels (5). Plasma adiponectin levels are lower in obesity and OHS (23,24).

Various disease states associated with lower serum adiponectin concentrations (such as asthma, systemic inflammation, and diabetes mellitus) (25–28) are associated with reduced lung function (29,30). It is therefore possible that lower serum concentrations of adiponectin may be associated with decreased lung function in humans. This hypothesis is supported by a recent study in normal-weight mice with genetic deficiency of ­systemic adiponectin (31). These mice demonstrated local (lung) adiponectin deficiency, increased systemic and local inflammation, and “alveolar simplification and/or enlargement due to abnormal post-natal alveolar development.”

Thyagarajan et al. have shown that independent of obesity, lower serum adiponectin concentrations are associated with lower lung function (32). These associations were no longer significant after adjustment for insulin resistance and C-reactive protein. The attenuation of this association after adjustment for insulin resistance and systemic inflammation suggests that these covariates may be associated in a pathway linking adiponectin and lung function.

Our initial hypothesis was possibly oriented towards the opposite direction, considering that the improvement of tissue oxygenation and lung function through the use of NIV in stable hypercapnic COPD patients will have an impact on serum adiponectin levels which in turn will reduce systemic inflammation and improve insulin resistance. Adiponectin increases the sensitivity to insulin through several mechanisms. It also exerts direct effects on vascular endothelium, diminishing the inflammatory response to mechanical injury and enhancing endothelium protection (33).

A plausible explanation for our findings may be that the mechanical injury, caused by NIV, on the alveolar-capillary membrane has a causal association with the increase of adiponectin levels and in that case adiponectin may play a protective role. On the other hand, the increase of serum adiponectin levels in our study, may not be directly related to NIV. In that case, higher adiponectin levels may be associated with improved pulmonary function through their anti-inflammatory effects. This might be an additional explanation of the more delayed improvement in spirometric parameters.

Insulin resistance and adiponectin levels of the patients in our study, improved after the application of NIV and the improvements were maintained throughout the trial. Specifically, adiponectin levels seem to be more sensitive as they improved from the first month of NIV use, whereas for insulin resistance it took a longer period for any improvement to be seen. Yet the improvements were maintained thereafter.

One possible explanation of this finding is that adiponectin is not purely related to adipose tissue metabolism and previous studies have shown that adiponectin levels are associated with systemic inflammation in patients with COPD. For example, Krommidas et al. (21) have found increased levels of adiponectin at the onset of an acute exacerbation of COPD (AECOPD) that was reduced after the successful management of the exacerbation. These authors have also provided evidence for an association of adiponectin levels with biomarkers of systemic inflammation (including CRP, interleukin-6 and tumor necrosis factor-α). Improvements in systemic inflammation due to the overall improvement of our patients may account at least in part for the early changes in adiponectin. Further studies are needed to support this hypothesis.

In addition, systemic and airway adiponectin concentrations are higher in COPD patients than controls, as demonstrated by case-control studies in humans. Systemic adiponectin is also positively associated with lung function in healthy adults but inversely associated with lung function in subjects with COPD (34). It is therefore possible that pro-inflammatory effects of adiponectin dominate under certain physiologic conditions and anti-inflammatory effects under others. The literature on the associations between adipokines and lung disease has critical gaps that include a lack of adequately powered longitudinal studies. It is uncertain whether adipokine changes precede pulmonary disease or are a consequence of it (34). Future research will determine whether modulation of adipokines, independent of BMI, may allow novel ways to prevent or treat inflammatory pulmonary conditions and our study offers some new data towards this direction.

A recent meta-analysis showed that constant positive airway pressure (CPAP) significantly improved insulin resistance in nondiabetic patients with moderate to severe OSA, while no significant change in body mass index was detected. Compared with fasting blood glucose at baseline, there was no change in glycemic control after 6 months of CPAP use (6). In another study concerning patients with OSA, there was not any significant change on insulin resistance after 3 months of treatment with CPAP (35). However, a four month treatment with CPAP of 10 patients with diabetes and OSA resulted in a reduction of fasting insulin despite maintenance of BMI (36). Probably, the modification of insulin sensitivity is time dependent and the length of effective treatment has to be determined in the future.

The improvements in the aforementioned parameters were seen although there was no significant change in the BMI of our patients. As a consequence there have to be other parameters that influence the changes in these variables. Indeed in the regression analysis we found that HOMA index presents a positive association with p_aCO₂, whereas adiponectin levels are associated with FVC and HCO_3-. It has been proposed that hypoxia can mediate its detrimental effects on glucose metabolism and insulin resistance at least through the effects on insulin sensitivity at the level of adipose tissue. A major reduction of oxygen supply may lead to adipose tissue inflammation and be potentially aggravated via a misbalance between adipocyte size and neovascularisation that leads to local hypoxic areas or intradipocytic hypoxia (37).

Oltmanns et al. found that hypoxia caused glucose intolerance, whereas hypoxia mediated increase in HIF-1a can induce adipose tissue fibrosis and resistance to insulin (38,39). Yet, in COPD patients, mild-to-moderate disease does not per se enhance adipose tissue inflammation (22). However, recent data support that obese patients, despite a lower adipose tissue blood flow, exhibit adipose tissue (AT) hyperoxia, which seems to be explained by lower AT oxygen consumption. This is accompanied by insulin resistance, impaired AT capillarization, and higher AT gene expression of inflammatory cell markers. It is proposed that the metabolic signatures of human AT do not support the notion of a hypoxic state in obesity. This needs to be studied in non-obese patients with COPD in order to extract safe conclusions (40,41).

In our population there was a significant improvement in p_aO₂ from the first month of NIV's application, but regression analysis revealed that the most significant determinant of adiponectin in our patients was HCO₃₋. Improvements in bicarbonate levels depict and result from the improvement of respiratory acidosis, and therefore it can be hypothesized that the determinant of the change in adiponectin levels is the change in bicarbonate which is related to the improvements of p_aCO₂. Hypercapnia has not been studied until now as a potent determinant of metabolic syndrome. A previous study in neonatal piglets failed to show any significant effect of exposure to intermittent hypercapnic hypoxia on glucose metabolism and insulin sensitivity (42). Further studies are needed to speculate the effects of hypercapnia on adipose tissue hormones.

FVC was the second major predictor of the improvements seen in adiponectin levels. The improvements in lung function seen due to the application of NIV resulted in the regulation of the metabolic profile. Many studies have addressed this issue in the past. Low vital capacity or restrictive pattern is associated with insulin resistance (43,44), and FEV₁ and FVC values are inversely correlated with Insulin Resistance (45,46). In COPD the results are more conflicting. It seems that COPD patients have an increased risk of type 2 DM (47,48). Concerning adiponectin, two studies failed to show any significant correlation between spirometric lung function and serum adiponectin correlations, yet in one study there was a positive association between adiponectin concentrations and residual volume (19,20).

Limitations

Our findings are subject to some limitations. The study did not include a control group and the follow-up period was restricted to one year, whereas the sample size was rather small. Consequently, generalization of findings regarding NIV superiority in terms of outcome should be done cautiously. However, we believe that the strict patient inclusion criteria and the meticulous assessment of adipokines and insulin resistance in our population contribute to further clarification of the impact of NIV on many aspects of COPD. Another possible limitation of the present study is the fact that we did not measure circulating triacylglycerol (TAG), HDL/LDL- cholesterol or pro-inflammatory markers, which could provide a better insight of the metabolic profile of our patients. Further studies are needed in that direction.

Conclusions

In conclusion, in this study we have examined the effects of non-invasive ventilation on insulin resistance and serum adiponectine levels of stable COPD patients with Chronic Hypercapnic Respiratory Failure and we found that insulin sensitivity and glucose metabolism as well as adiponectin levels improve along with the improvements in respiratory failure.

Declaration of Interest Statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.