Glycated Hemoglobin (HbA1c) as a Predictor of Outcomes during Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Abstract

Systemic inflammation may be the common denominator between COPD and type 2 diabetes and may explain the correlation in both diseases’ development and progress. The aim of this prospective observational study is to examine the prognostic value of glycated hemoglobin levels (HbA1c) and HbA1c-adjusted glycemic variables (glycemic gap, stress hyperglycemia ratio και modified stress hyperglycemia ratio) in an acute exacerbation of COPD (AECOPD) as well as in COPD disease's morbidity and mortality during the following year. We evaluated patients hospitalized only for COPD exacerbations. Levels of HbA1c and HbA1c-adjusted glycemic variables were recorded upon admission. The study outcomes included duration of hospital stay, need for mechanical ventilation and exacerbation outcome. All subjects were followed up for one year. A total of 156 patients were included in the study (74.4% men, age [mean ± SD] 72 ± 7 years). Patients (21.8%) had type 2 diabetes and 67.9% of patients were receiving ICS treatment. The median value of HbA1c was 5.9 (IQR: 5.4, 6.5). Necessity for mechanical ventilation was significantly higher for patients with lower values of HbA1c [median: 5.3 (IQR 5.02, 6.3) vs. 5.9 (IQR 5.5, 6.5), p = .038]. However, duration of hospitalization, death during hospitalization as well as the number of new exacerbation events, time to next exacerbation and mortality during the following year did not differ significantly. Moreover, none of the HbA1c-adjusted glycemic variables examined, demonstrated any statistical significance. In conclusion neither the preceding nor the present glycemic state exhibit a predictive value regarding short- or long-term outcomes of an AECOPD.

Keywords:

• COPD
• type 2 diabetes
• HbA1c
• Hba1c adjusted variables

Introduction

Chronic obstructive pulmonary disease (COPD) often coexists with other diseases (co morbidities) that may have a significant impact on disease course [1]. The implicating pathogenetic mechanisms behind this relation are not yet fully understood but it is likely to be multifactorial including tissue hypoxia, oxidative stress, physical inactivity and systemic inflammation [2]. Increasing evidence supports the ‘spill over’ theory of bronchial and alveolar inflammation into the systemic circulation which contributes to the development of comorbidities [3,4].

Patients with COPD are known to be at increased risk of developing diabetes compared to the general population [5,6] and the risk seems to increase with disease severity [7]. Low-grade chronic inflammation is part of the insulin resistance syndrome and is associated with the development of type 2 diabetes [8,9]. Thus, it is possible that chronic inflammation in combination with hypoxia which also seems to have a negative effect on glucose metabolism and insulin resistance [10], could be common denominators for both COPD and diabetes [11].

Glycated hemoglobin (HbA1c) is a well-validated measure of glycemic status over the previous 8–12 weeks [12]. HbA1c levels are characterized by lower biological variability, and they are relatively unaffected by acute stress [13]. Recent studies have shown that several hemoglobin A1c (HbA1c)-adjusted glycemic variables, including an elevated glycemic gap, the stress hyperglycemia ratio and the modified stress hyperglycemia ratio are better biomarkers for identifying patients at a higher risk of critical illnesses [14–17] including those experiencing a COPD exacerbation [13].

We hypothesized that in COPD patients experiencing an AECOPD, it is possible that HbA1c may reflect the existing inflammatory status prior to the onset of the exacerbation and that it might be a better indicator of the existing inflammatory milieu which influences the severity of an exacerbation. The objective of this prospective observational study was to evaluate the impact of HbA1c and secondarily of blood glucose and HbA1c – adjusted glycemic variables in COPD morbidity and mortality during both an acute exacerbation and the following year.

Methods

Patients

This study enrolled consecutive patients admitted for AECOPD in four Respiratory Medicine departments of tertiary referral hospitals in Greece. Patients included were over 40-year old and current or ex-smokers (≥20 pack-years). All were previously diagnosed with COPD according to a respiratory physician and all had available spirometric data prior to the exacerbation. A history of type 2 diabetes (T2D) was recorded upon admission. Severity of COPD disease or whether T2D regimens included or not insulin therapy did not affect patient inclusion criteria. They were both admitted and discharged to the hospital according to GOLD criteria [1] as well as the respiratory physician’s judgment. During hospitalization all patients received standard care for the acute exacerbation [1] and for blood glucose control [18]. None of the HbA1c or the HbA1c – adjusted glycemic variables values affected patients’ treatment options or decision for discharge. Moreover, none of patients’ regimens for COPD or T2D affected inpatient and outpatient treatment options. Exclusion criteria included type 1 diabetes (T1D), hospitalization due to other reasons, hospitalization due to an AECOPD or outpatient therapy with oral corticosteroids (OCS) during the past 3 months. All patients were invited to participate willingly and were assured that their care would not be affected by their decision to participate in the study. The study protocol was approved by the local research ethics committee (501/17-07-2018) and all participants provided written informed consent.

Study design

Patient demographics, medical history (including co-morbidities such as hypertension, coronary artery disease, congestive heart failure, arrhythmias or stroke) and medication use were recorded. Clinical parameters recorded on admission included vital signs and arterial blood gases. Standard laboratory measurements such as complete blood count, C-reactive protein (CRP) as well as blood glucose values were recorded on admission and prior to the initiation of any treatment and levels of HbA1c were obtained within 24 h from admission. HbA1c-adjusted glycemic variables (glycemic gap, stress hyperglycemia ratio and modified stress hyperglycemia ratio) were calculated to evaluate stress induced hyperglycemia (SIH). After discharge, all patients were followed up for one year via monthly telephone interviews. Information concerning the patients’ vital status and the number of new AECOPD (defined as need for antibiotics and/or oral corticosteroids, visits to the emergency department and/or hospitalizations) were recorded.

Calculation of HbA1c-adjusted glycemic variables

HbA1c-adjusted glycemic variables were calculated as follows:

1. glycemic gap: [glucose - ADAG (28,7× HbA1c - 46,7)]

2. stress hyperglycemia ratio: [glucose/ADAG (28,7× HbA1c - 46,7)]

3. modified stress hyperglycemia ratio: [glucose/HbA1c]

Study outcomes

The study was designed to investigate the relationship mainly between HbA1c and secondarily between HbA1c – adjusted glycemic variables and COPD morbidity and mortality during an acute exacerbation and in the following year. The primary end-points were the severity of the acute exacerbation defined as length of hospital stay, necessity for mechanic ventilation (invasive or noninvasive) severity of hypoxemia [PaO₂/inspiratory oxygen fraction (FiO₂) ratio], and all-cause mortality at 30 days. Furthermore, COPD morbidity defined as time to the first AECOPD and hospitalization for AECOPD, frequency of those events in 1 year and mortality during the following year were also examined. Comparison between patients depending on the presence of type 2 diabetes, active smoking or the previous use of ICS was also made for all the above parameters.

Statistical analysis

Categorical variables are presented as n (%), whereas numerical variables are presented as mean ± standard deviation (SD) or median (interquartile ranges) for normally distributed and skewed data, respectively. Normality of distributions was checked with Kolmogorov–Smirnov test. Comparisons between groups were performed using chi-square tests for categorical data, as well as unpaired t-tests or Mann–Whitney U-tests for normally distributed or skewed numerical data. Correlations were assessed using Spearman’s and Pearson’s correlation coefficients for skewed and normally distributed variables, respectively. Cox regression analysis was used to evaluate the predicting value of HbA1c and all adjusted glycemic variables for the time of next COPD exacerbation, severe COPD exacerbation and death. Statistical significance was established at p ≤ .05. Data were analyzed using SPSS 17.0 for Windows (SPSS Inc., Chicago, IL, USA).

Results

One hundred fifty-Six patients (74.4% males) with a median age of 73 years, were included in the analysis. Demographic characteristics of the study participants presented on Table 1

Table 1. Demographic, functional and laboratory characteristics of the study participants on admission for COPD exacerbation.

Variable	All n = 156
Age (years)	73.0 (67.0, 78.0)
Sex (Male), N(%)	116 (74.4)
BMI (kg/m^2)	27.05 (23.9, 31.6)
Smoking (current), N(%)	62 (39.7)
Smoking (pack-years)	70.0 (49.0, 90.0)
Glucose at admission (mg/dL)	116.5 (99.0, 141.75)
PO2/FiO2	247.6 (187.5, 281.0)
PCO2 (mmHg)	45.0 (38.5, 58.0)
pH	7.40 (7.36, 7.43)
CRP mg/L	24.3 (5.8, 104.0)
Blood Eosinophils (N)	53.5 (20.0, 167.5)
Blood Neutrophils (N)	7434 (5335, 11330)
HbA1c	5.9 (5.4, 6.5)
Glycemic gap	–6.69 (–24.29, 23.23)
Stress hyperglycemia ratio	0.93 (0.81, 1.19)
Modified stress hyperglycemia ratio	19.38 (16.99, 24.87)
Type 2 diabetes, N(%)	34 (21.8)
Antidiabetic pills, N(%)	30 (19.2)
Insulin therapy, N(%)	4 (2.5)
CAD, N(%)	30 (19.2)
Arrythmia, N(%)	41 (26.3)
Cerebrovascular disease, N(%)	3 (1.9)
Heart failure, N(%)	22 (14.1)
AH, N(%)	66 (42.3)
ICS prior to admission, N(%)	106 (67.9)
LABA/ICS, N(%)	22 (14.1)
LABA/LAMA, N(%)	12 (7.7)
LABA/LAMA/ICS, N(%)	38 (24.4)
Length of stay (days)	10.0 (7.0, 14.8)

Data are presented as median (IQR) unless otherwise indicated.

Abbreviations: BMI: Body Mass Index; PO₂/FiO₂: partial pressure of arterial oxygen/fractional inspired oxygen; PCO₂: partial pressure of carbon dioxide; CRP: C-Reactive Protein; HbA1c: glycated hemoglobin; CAD: coronary artery disease; AH: arterial hypertension; ICS: inhaled corticosteroids; LABA: long-acting β₂-agonists; LAMA: long-acting muscarinic antagonists.

HbA1c levels, exacerbation severity and outcomes.

Major correlations related to exacerbation severity are presented on Table 2. Patients who required invasive mechanical ventilation (IMV) had lower HbA1c levels on admission compared to those who did not require IMV median (IQR), [5.3 (5.02, 6.35) vs. 5.9 (5.5, 6.5)], p = .038 for patients requiring and not requiring IMV respectively, while no significant difference was observed between patients who required or not, noninvasive mechanical ventilation (NIMV), [6.1 (5.6, 7.1) vs. 5.9 (5.4, 6.5)], p = .21 respectively. 12 (7.7%) patients died during hospitalization. No significant difference was observed between HbA1c levels on admission in patients who died during hospitalization and those who were discharged from the hospital [5.6 (5.1, 6.4) vs. 5.9 (5.5, 6.5), p = .21] respectively.

Table 2. Correlation of HbA1c levels with exacerbation characteristics.

Variable	p	r
PO2/FiO2 ratio	.78	–0.03
PCO2	.07	0.19
pH	.05	–0.20
FEV1	.18	0.11
CRP	.41	–0.08
Blood neutrophil levels	.64	–0.04
Blood eosinophils	.01	0.19
Length of hospital stay	.49	0.05

Abbreviations: PO₂/FiO₂: Partial pressure of arterial oxygen/Fractional inspired oxygen; PCO₂: Partial pressure of carbon dioxide; FEV₁: Forced Expiratory Volume in one second; CRP: C-Reactive Protein.

HbA1c adjusted glycemic variables exacerbation severity and outcomes

Correlations related to HbA1c adjusted glycemic variables and exacerbation severity are shown on Table 3. No significant differences were found on the levels of either of the HbA1c adjusted glycemic variables between patients requiring or not invasive mechanical ventilation, NIMV or those deceased during hospital stay compared to those who survived, p > .05 for all comparisons (Table 4).

Table 3. Correlations of HbA1c adjusted glycemic variables and exacerbation characteristics.

Variable	Glycemic gap		Stress hyperglycemia ratio		Modified stress hyperglycemia ratio
	p	r	p	r	p	r
PO2/FiO2 ratio	.70	–0.04	.76	–0.03	.66	–0.04
PCO2	.91	0.01	.99	0.0	.66	0.04
pH	.78	0.03	.91	0.01	.71	–0.04
FEV1	.45	–0.08	.53	–0.06	.64	–0.05
CRP	.46	0.08	.49	0.07	.48	0.07
Blood neutrophil levels	.005	0.287	.009	0.271	.007	0.279
Blood eosinophils	<.001	–0.358	.001	–0.374	.001	–0.336
Length of hospital stay	.31	0.10	.26	0.11	.1	0.16

Abbreviations: PO₂/FiO₂: Partial pressure of arterial oxygen/fractional inspired oxygen; PCO₂: Partial pressure of carbon dioxide; FEV₁: Forced Expiratory Volume in one second; CRP: C-Reactive Protein.

Bold indicate significant correlations.

Table 4. Differences between levels of HbA1c adjusted glycemic variables.

Adjusted glycemic variable	Patients not required invasive MV	Patients who required invasive MV	p-value	Patients not required NIMV	Patients who required NIMV	p-value	Patients discharged from the hospital	Patients deceased during exacerbation	p-value
Glycemic gap	–7.76 (–24.9, 18.8)	20.33 (–20.3, 57.9)	.13	–7.15 (–22.4, 18.8)	–2.76 (–32.4, 29.4)	.79	–6.52 (–23, 25)	–8.08 (–35, 5.4)	.53
Stress hyperglycemia ratio	0.93 (0.8, 1.1)	1.20 (0.9, 1.5)	.14	0.93 (0.8, 1.1)	0.97 (0.8, 1.2)	.75	0.93 (0.8, 1.2)	0.92 (0.7, 1)	.43
Modified stress hyperglycemia ratio	19.23 (17, 24.3)	23.52 (18.4, 30.7)	.14	19.24 (17.2, 23.7)	19.80 (16.5, 25.9)	.57	19.38 (17, 25)	18.47 (15.6, 21)	.36

Abbreviations: MV: mechanical ventilation; NIMV: noninvasive mechanical ventilation.

HbA1c levels and one-year follow up outcomes

During the following year, the mean number of new AECOPD events in total (moderate and severe) was 1.15 (SD ± 1.15) and the median time for the first AECOPD event was 4 months (IQR) (2, 7). In Cox regression analysis, the levels of HbA1c on admission was not a predictor neither of the time to next moderate or severe COPD exacerbation HR: 1.00, 95%CI: 0.80–1.25, p = .98 nor of the time to next severe COPD exacerbation HR:0.90, 95%CI: 0.74–1.33, p = .96. Finally, HbA1c on admission was not a predictor of mortality HR: 0.97, 95%CI: 0.59–1.58, p = .9.

Regarding the adjusted glycemic variables, none of them was found to be a predictor to either moderate or severe COPD exacerbation in the following year, the time to the next severe exacerbation or death. All aforementioned results are presented on Table 5.

Table 5. Cox regression analysis for the prediction of the next COPD exacerbation, severe COPD exacerbation and death during the one year of follow up.

	HR	95% CI	p-value
COPD exacerbation (moderate or severe)
HbA1c	1.00	0.80–1.25	.98
Glycemic gap	2.00	0.99–1.01	.80
Stress hyperglycemia ratio	0.90	0.44–1.86	.77
Modified stress hyperglycemia ratio	0.99	0.96–1.03	.73
Severe COPD exacerbation
HbA1c	0.90	0.74–1.33	.97
Glycemic gap	1.00	0.99–1.01	.76
Stress hyperglycemia ratio	1.00	0.96–1.06	.87
Modified stress hyperglycemia ratio	1.15	0.45–2.96	.77
Death
HbA1c	0.97	0.59–1.58	.90
Glycemic gap	1.00	1.00–1.02	.25
Stress hyperglycemia ratio	1.84	0.57–5.93	.30
Modified stress hyperglycemia ratio	1.04	0.97–1.10	.26

Abbreviations: HbA1c: Glycated hemoglobin; HR: Hazard Ratio; CI: Confidence Interval; COPD: Chronic Obstructive Pulmonary Disease.

Finally, 20 patients (12.8%) died during the next year of follow up. The levels of HbA1c or adjusted glycemic variables did not differ in patients who died compared to those who survived during the next year as presented on Table 6.

Table 6. Differences on HbA1c levels and adjusted glycemic variables between patients who died and those who survived during the year of follow up.

Variable	Patients who survived	Patients deceased	p-value
HbA1c	5.90 (5.52, 6.50)	5.75 (5.50, 6.72)	.74
Glycemic gap	–7.52 (–25.9, 21.31)	13.26 (–13.11, 44.49)	.06
Stress hyperglycemia ratio	19.23 (16.67, 24.62)	22.77 (19.20, 28.26)	.14
Modified stress hyperglycemia ratio	0.93 (0.80, 1.15)	1.09 (0.90, 1.43)	.10

Data are presented as median (IQR).

Abbreviations: HbA1c: glycated hemoglobin.

Discussion

To our knowledge, this is the first study to examine directly the predictive value of HbA1c and the second one examining the predictive value of glycated-hemoglobin adjusted glycemic variables during an AECOPD and during the following year. This study is a prospective multicenter observational study and we showed that neither HbA1c nor HbA1c-adjusted glycemic variables were able to predict the severity of an AECOPD or short-term mortality. Only patients with low levels of HbA1c exhibited a poor inpatient course in terms of mechanical ventilation necessity during the acute exacerbation event. Also, all three HbA1c adjusted glycemic variables studied were weakly but significantly associated with blood neutrophil and blood eosinophil levels. Moreover, HbA1c levels as well as values of HbA1c adjusted glycemic variables were not significantly correlated with the risk of exacerbations, time to first exacerbation event or mortality during the following year. Our data suggest that neither HbA1c nor HbA1c adjusted glycemic variables are predictors of exacerbation outcomes in COPD patients admitted to the hospital due to an AECOPD.

In our study, there was no association between HbA1c levels or HbA1c adjusted glycemic variables and the length of hospital stay. Our results are in contrast with a previous study in which it has been shown that increased levels of HbA1c was an independent predictor of prolonged hospital stay and that COPD patients hospitalized for AECOPD were 1.6 times more likely to have a prolonged hospital stay for every 1 standard deviation above the mean HbA1c level [19]. However, the above study was a retrospective one and is lacking information related to several important variables like spirometry, arterial blood gas analysis, details on home oxygen therapy and mortality outcomes.

Interestingly, in our study, patients who required intubation and IMV had lower levels of of HbA1c compared to patients who did not require IMV respectively. A possible explanation derives from the fact that low levels of glucose as well can have a negative impact on COPD disease [20]. Indeed hypoglycemia has been reported to reduce VEGF-A production [15]. Low levels of VEGF have been associated with chronic airway obstruction by enabling the apoptosis of lung alveolar epithelial cells [20–22]. Another possible mechanism is that hypoglycemia via increasing platelet aggregation and fibrinogen formation enhances vascular changes of the lung [20]. Presumably, patients with low HbA1c levels experience hypoglycemia through long periods of time and due to their impaired respiratory capacity, they require more support during the exacerbation event.

In our study, neither the levels of HbA1c nor the levels of HbA1c adjusted glycemic variables correlated significantly with PaO₂/FiO₂ ratio or PCO₂ levels. Likewise Yang et al. did not document a significant difference in HbA1c levels and the presence of acute respiratory failure in patients admitted due to AECOPD [13]. Similarly, a previous retrospective study on diabetic patients admitted due to community-acquired pneumonia did not exhibit a significant correlation between levels of HbA1c and acute respiratory failure, development of ARDS or in-hospital mortality (41).

To date there is a lack of evidence regarding the possible prognostic value of HbA1c levels in short-term mortality following an AECOPD as well as long-term morbidity and mortality in COPD patients. In critically ill patients, it was found that HbA1c levels per se failed to predict in hospital mortality[23]. However, a previous study has shown that co morbid diabetes during hospitalization for an AECOPD increases the risk of death (OR 1.93) [24]. Furthermore, regarding the long term prognosis of COPD patients the coexistence of diabetes either preexisting or incident has a negative impact in terms of necessity for hospitalization and death over long periods of follow up [5,7,25]. Indeed according to the Emerging Risk Factors Collaboration, the HR for COPD-related death was 1.27 compared to subjects without diabetes [26]. Nevertheless, in our study HbA1c levels failed to exhibit any prognostic value for survival or future exacerbation risk. The whole issue remains controversial since many co-existing parameters may affect the predictive role of HbA1c in terms of short- and long-term mortality.

Our study is the first to examine in such extent the impact of HbA1c levels on short-term and long-term morbidity and mortality of COPD patients. Our negative results can be partly attributed to the complexity of which HbA1c levels, current glycemic state and inflammatory pathways interact during an AECOPD. Numerous studies have examined the impact of hyperglycemia on the outcomes of an AECOPD concluding that increasing blood glucose concentrations are associated with longer length of hospital stay and mortality [27], failure of noninvasive ventilation and infectious pulmonary complications [28]. Blood glucose is increased in acute illness due to a combination of metabolic effects such as release of catecholamines, cortisol, glucagon and increased peripheral insulin resistance [29]. Moreover the standard use of corticosteroids either inhaled or systemic during an AECOPD affects even more the present glycemic state. It is probable that preceding glycemic state reflected by HbA1c levels little contributes to the outcome of the current event since many immediate changes take place.

In this study, none of the HbA1c-adjusted glycemic variables significantly correlated with the severity of the initial acute exacerbation event and mortality within 30 days from admission. Recent data imply but not fully establish a prognostic role of these variables in critically ill patients. It has been demonstrated that in diabetic patients higher glycemic gap levels are associated with an increased risk of multiorgan dysfunction syndrome (MODS) and development of acute respiratory failure (ARF) [30]. Moreover, in diabetic patients admitted due to an AECOPD specifically a higher glycemic gap and modified stress hyperglycemia ratio were associated with the development of acute respiratory failure (ARF). Despite the fact being superior than acute hyperglycemia and HbA1c those glycemic variables were not able to predict ARF independently [13].

Interestingly in this study an inverse relationship, weak but statistically significant, between the blood eosinophil count and values of HbA1c-adjusted glycemic variables was observed. Despite the fact that elevated values of glycemic variables were not associated with the exacerbation’s severity, the negative correlation with the blood eosinophil count may be an indirect sign. It is known that patients with higher eosinophil counts at the onset of an AECOPD present with lower disease severity in terms of length of hospital stay [31,32] and need for noninvasive ventilation [33]. In addition COPD exacerbations with acute respiratory failure requiring ICU admission appear to have a better outcome when the peripheral eosinophil level is >2% [34]. Conversely, eosinopenia is a well-recognized feature of higher hospital mortality [33,35,36]. On the other hand, diabetic patients with elevated values of HbA1c glycemic variables and in particular values of the glycemic gap, as aforementioned are at an increased risk of developing acute respiratory failure during an AECOPD [13], developing septic shock, acute respiratory failure and be admitted to the ICU during hospitalization due to community acquired pneumonia [15]. Nevertheless, more data are needed to establish this correlation and examine the underlying mechanism.

The strengths of this study include a well-characterized COPD patient population from four Respiratory Medicine Departments and its prospective design. A possible explanation for our negative results is that according to guidelines all patients received 40 mg of methylprednisolone for 5 days once admitted and this may have suppressed the inflammatory process. As a result, the impact of stress induced hyperglycemia (SIH) on AECOPD outcomes may have been impeded. Indeed in a previous study, [13] it has been reported that glycemic gaps and stress hyperglycemia ratios were not significantly different among the corticosteroid users and possibly indicated diminished physiological stress because of AECOPD. In contrast patients who had not received corticosteroids and developed acute respiratory failure had significantly higher glycemic gaps and modified stress hyperglycemia ratios, higher risks of pulmonary infection and longer hospital stays compared to patients receiving corticosteroids [13].

An important limitation of this study is that patients were not studied separately whether type 2 diabetes was present or not, whether diabetes was controlled or not and whether they received antidiabetic treatment (oral and/or insulin) or not. Interpretation of results is complex when the history of diabetes is taken into account. It is known that during an AECOPD a history of diabetes is strongly associated with increased long-term mortality (adjusted HR 3.03). Interestingly in this study screening diabetes, defined as HbA1c ≥ 6.5%, in patients without a known history was not associated with late mortality. Perhaps ‘non-diabetic mechanisms’ exist that affect levels of HbA1c but are not sufficient to induce diabetes adverse effects on patients [37]. It is also known that the use of oral antidiabetic treatment and especially metformin in patients with COPD is associated with fewer COPD related emergency department visits and hospitalizations during the following 2 years possibly due to its anti-inflammatory effect [38]. It is likely that our diverse study population and the small size of each population did not allow any significant correlations to reveal.

However, it should be mentioned that direct comparison with findings of other studies is very difficult since many differences exist mainly as the study population and its characteristics is concerned. For example, most studies included only diabetic patients and in general, the median HbA1c levels as well as values of glycemic variables appeared to be much higher in other studies compared to ours [mean HbA1c: 7.9 ± 2.1 [13], 9.09 ± 2.14 [30] vs. 6.12 ± 0.98 and mean glycemic gap: 45.9 ± 123.3 [13] vs. 2.21 ± 41.44]. Another limitation is the small final number of participants. In order to include patients hospitalized exclusively for AECOPD who had not received oral corticosteroids during the past 3 months or prior to the emergency department arrival it was necessary to exclude a great number of patients. Finally, the one year of follow up period may be short in order to evaluate mortality rates.

Conclusion

Chronic obstructive pulmonary disease and type 2 diabetes mellitus appear to have much in common in terms of disease development and a bidirectional impact on each other’s progress. Nevertheless, in our study neither the glycemic state (HbA1c) nor glycemic variability (HbA1c-glycemic variables) appear to be directly representative of the exacerbation’s evolution. In addition, these variables do not have any predictive value over the following year. Future studies are needed in order to examine the short- term and long-term prognostic value of glycemic variables if any.

Declaration of interest

The authors declare that they have no conflict of interest related to the present study.