Nocturnal Oxygen Desaturation Index is Inversely Correlated with Airflow Limitation in Patients with Chronic Obstructive Pulmonary Disease

Abstract

The concurrent diagnosis of chronic obstructive pulmonary disease (COPD) and sleep apnoea–hypopnoea syndrome (SAHS) (overlap syndrome), can contribute to worsening respiratory symptoms, but whether the severity of COPD is associated with co-morbid SAHS is unknown. We investigated whether the severity of COPD is associated with the complication of SAHS by examination of nocturnal oximetry as an alternative to polysomnography. Patients with COPD concurrently completed nocturnal oximetry, pulmonary function tests, a COPD assessment test, an Epworth sleepiness scale and a hospital anxiety and depression scale to evaluate the severity of COPD and possible concurrent presence of SAHS. We retrospectively analysed the data to assess correlation between the oxygen desaturation index (ODI) and each clinical variables and evaluated the predictors of ODI ≥ 15. This study included 103 patients (91 males, 88%) with a mean age of 72 ± 8 years and body mass index of 22 ± 3 kg/m². ODI was positively correlated with FEV₁, FEV₁/FVC and FEV₁% predicted, which meant that ODI was inversely correlated with airflow limitation. Univariate logistic regression analysis revealed that FEV₁% predicted and FEV₁/FVC were predictors of ODI ≥ 15. ODI is inversely correlated with airflow limitation and milder COPD patients may have co-morbid SAHS.

Keywords:

• airflow limitation
• chronic obstructive pulmonary disease
• oximetry
• oxygen desaturation index
• sleep apnea–hypopnea syndrome

Abbreviations

AHI	=	Apnoea–hypopnoea index
BMI	=	Body mass index
CAT	=	Chronic obstructive pulmonary disease assessment test
COPD	=	Chronic obstructive pulmonary disease
ESS	=	Epworth sleepiness scale
FEV 1	=	Forced expiratory volume in one second
FVC	=	Forced vital capacity
GOLD	=	Global Initiative for Chronic Obstructive Lung Disease
HADS	=	Hospital anxiety and depression scale
ODI	=	Oxygen desaturation index
PSG	=	Polysomnography
REM	=	Rapid eye movement
SAHS	=	Sleep apnoea–hypopnoea syndrome

Introduction

Chronic obstructive pulmonary disease (COPD) is a common condition characterized by persistent airflow limitation that is usually progressive and associated with systemic co-morbid conditions (1, 2). Sleep apnoea–hypopnoea syndrome (SAHS) is a common disorder, although often unrecognized, and is associated with considerable functional impairment (3). The concurrent diagnosis of COPD and SAHS was first described as the “overlap syndrome” in 1985 (4). These patients have particularly severe disease, with a lower partial pressure of oxygen in arterial blood (PaO₂), higher partial pressure of carbon dioxide in arterial blood (PaCO₂) and higher pulmonary artery pressures than patients with SAHS or COPD alone (5–7). Patients with overlap syndrome also have increased mortality compared with patients with SAHS alone (8, 9), and continuous positive airway pressure treatment improves survival in them (10, 11).

Polysomnography (PSG) is currently the gold standard test in the diagnosis of SAHS. However, PSG is costly and time-consuming, and therefore difficult to perform for the majority of patients suspected of having SAHS. Thus, nocturnal oximetry has been proposed as a simpler alternative to PSG in the diagnosis of SAHS because it is readily available, relatively inexpensive, and easily administered. The oxygen desaturation index (ODI) obtained by nocturnal oximetry correlates well with the apnoea–hypopnoea index (AHI) and performs well in the diagnosis of SAHS (12–15); hence, nocturnal oximetry might meet the demand for screening of SAHS in the community, and is a possible diagnostic tool.

We hypothesized that patients with severe COPD suffer from more severe SAHS. Therefore, we conducted this retrospective study to investigate whether the severity of COPD is associated with the complication of SAHS by examination of nocturnal oximetry as an alternative to PSG.

Methods

Subjects

We conducted this retrospective study in a 400-bed hospital that has a respiratory referral centre. The subjects had been referred to or were being treated at our hospital between July 2014 and April 2015. We performed pulmonary function tests and nocturnal oximetry to evaluate the severity of COPD and to screen for the co-morbidity SAHS for stable COPD outpatients. The COPD assessment test (CAT), Epworth sleepiness scale (ESS) and hospital anxiety and depression scale (HADS) were completed concurrently. We included patients with a diagnosis of COPD supported by a history of past or current smoking and obstructive lung disease with a forced expiratory volume in one second (FEV₁)/forced vital capacity (FVC) < 70%. However, patients who suffered from acute illness, dementia and poor performance status were excluded because it was difficult for them to perform the tests. Patients who received domiciliary oxygen therapy were also excluded.

Data collection

All clinical data were collected from medical records. Exacerbation of COPD was defined as “chest problems requiring treatment with antibiotics and/or oral corticosteroids” based on a previous report (16). We counted exacerbations of COPD within a year of the examinations, including nocturnal oximetry, pulmonary function tests and questionnaires. We used a de-identified database for the analysis in order to account for confidentiality; therefore, this study did not pose any risk to patients. The Institutional Review Board of Shinko hospital approved this study (approval number: 1431).

Questionnaires

We evaluated CAT, ESS and HADS concurrently with nocturnal oximetry and pulmonary function tests. CAT is a questionnaire for people with COPD that is designed to measure the impact of COPD on a person's life (17). The CAT scale ranges from 0 to 40, with higher scores indicating worse symptoms and control. ESS is an instrument used to measure average daytime sleepiness (18). A score of 11 or more reflects greater than normal daytime sleepiness. HADS is a screening tool for assessing the severity of symptoms of depression and anxiety (19). The HADS consists of two subscales assessing depression and anxiety. A score of 0–7 denotes the absence of depression/anxiety, 8–10 indicates possible depression/anxiety, and 11 or higher indicates probable depression/anxiety. If it was difficult for patients, especially for the elderly, to answer the questions, trained staff interviewed the patients and completed the questionnaires.

Pulmonary function tests

Pulmonary function tests were conducted by trained operators, in accordance with American Thoracic Society/European Respiratory Society guideline (20), using a CHESTAC-8800 system (Chest, Tokyo, Japan).

Nocturnal oximetry

Patients underwent home nocturnal oximetry using a pulse oximeter (PULSOX-Me300, Konica Minolta Sensing Inc., Osaka, Japan). A technician visually checked the oximetry recordings and obvious artefacts were deleted, followed by analysis of the recordings using a computer program (DS-Me, Konica Minolta Sensing Inc., Osaka, Japan). Desaturation events are defined as at least a 4% decrease in the percutaneous oxygen saturation (SpO₂) from the mean level lasting for 4–120 seconds. The oximetry tracings show a sawtooth pattern and the dips of ≥ 4% are detected as desaturation events. The total number of desaturations was divided by the total time sleeping in hours as reported by the patient, and calculated as the median ODI. The cut-off point of ODI was ≥ 15. The total time of SpO₂ < 90% was divided by the hours of sleep to give the percentage of time with SpO₂ < 90% (%T90).

Statistics

Continuous variables were expressed as numbers (%), means (standard deviation [SD]) or medians (interquartile range [IQR]). Correlations between ODI and other clinical parameters were assessed using the Spearman correlation coefficient. The Kruskal–­Wallis test was used to compare the value of ODI among the four stages of the GOLD (Global Initiative for Obstructive Lung Disease) classification. Univariate logistic regression analysis was used to identify significant variables predicting ODI≥15. A P value < 0.05 was deemed statistically significant. Statistical analysis was performed using JMP 9 software (SAS Institute Inc., Cary, NC, USA).

Results

Patients

This study included 103 patients with a mean age of 72 ± 8 years and 91 males (88%). Table 1 shows baseline characteristics of the patients who participated in this study. The BMI was 22 ± 3 kg/m². All patients received some inhaled medication; long-acting methacholine antagonist was the most frequent choice. Table 2 shows medical conditions associated with COPD. Twenty-nine patients (28%) suffered from exacerbations of COPD within a year. The median (IQR) value of CAT, ESS, HADS anxiety and HADS depression were 13 (7–18), 5 (3–8), 3 (1–7) and 6 (2–9), respectively. Pulmonary function tests showed FEV₁/FVC of 51±12% and FEV₁% predicted of 61±21%.

Table 1.  Baseline characteristics of subjects

	All patients, n = 103
Age, mean (SD), years	72 (8)
Male, n (%)	91 (88)
BMI, mean (SD), kg/m^2	22 (3)
Underlying co-morbidity
Hypertension	40 (39)
Diabetic mellitus	15 (15)
Heart failure	16 (16)
Coronary artery diseases	12 (12)
Cerebral infarction	9 (9)
Smoking history, median (IQR), pack-years	53 (41–76)
Treatment
LAMA, n (%)	84 (82)
LABA, n (%)	32 (31)
ICS/LABA, n (%)	41 (40)
Oral corticosteroids, n (%)	6 (6)
Theophyline, n (%)	7 (7)

BMI = Body mass index; LAMA = Long-acting methacholine antagonist; LABA = Long-acting beta agonist; ICS = Inhaled corticosteroid.

Table 2.  Medical conditions associated with COPD

	All patients, n = 103
Resting SpO2, mean (SD), %	96 (2)
Exacerbation of COPD within a year, n (%)	29 (28)
CAT, median (IQR)	13 (7–18)
ESS, median (IQR)	5 (3–8)
HADS
Anxiety	3 (1–7)
Depression	6 (2–9)
Pulmonary function test
FEV1, mean (SD), L	1.59 (0.63)
FEV1/FVC, % (SD)	51 (12)
FEV1, mean (SD), % predicted	61 (21)
FVC, mean (SD), L	3.1 (0.9)
FVC, mean (SD), % predicted	95 (23)

SpO₂ = Percutaneous oxygen saturation; COPD = Chronic obstructive pulmonary disease; CAT = COPD assessment test; ESS = Epworth sleepiness scale; HADS = Hospital anxiety and depression scale; FEV₁ = Forced vital capacity in one second; FVC = Forced vital capacity.

Nocturnal oximetry

Table 3 shows results of nocturnal oximetry. The mean (SD) value of nocturnal SpO₂ was 93 (4)%. The median (IQR) value of ODI was 4 (2–7). Thirty-one patients (30%) showed ODI ≥ 5, 15 patients (15%) showed ODI ≥ 10 and 9 patients (9%) showed ODI ≥ 15.

Table 3.  Nocturnal oximetry

	All patients, n = 103
Mean SpO2, mean (SD), %	93 (4)
%T90, median (IQR), %	3.3 (0.4–10.4)
ODI, median (IQR)	4 (2–7)
ODI≥5, n (%)	31 (30)
ODII≥10, n (%)	15 (15)
ODII≥15, n (%)	9 (9)

SpO₂ = Percutaneous oxygen saturation; %T90 = Percentage of time of SpO₂ < 90%; ODI = Oxygen desaturation index.

Correlations between results of ODI and clinical variables

Table 4 shows the correlation between ODI and clinical variables. ODI was positively correlated with FEV₁, FEV₁/FVC and FEV₁% predicted, which meant that ODI was inversely correlated with airflow limitation. The mean (SD) values of ODI of GOLD 1, GOLD 2, GOLD 3 and GOLD 4 were 9.9 (10.7), 4.4 (4.3), 4.8 (6.2) and 2.8 (3.6), respectively, and ODI was significantly different among GOLD classifications (Figure 1). Univariate logistic regression analysis showed that ODI≥15 was predicted by FEV₁% predicted and FEV₁/FVC. It was revealed age was the positive predictor of ODI≥15 and CAT was the negative predictor of ODI ≥ 15 (Table 5).

Table 4.  Correlations between ODI and clinical variables

	ODI
	Coefficient	P value
Age	0.11	0.3
BMI	0.05	0.6
Pack-years of smoking	−0.08	0.4
Resting SpO2	0.15	0.1
CAT	−0.15	0.1
ESS	0.05	0.6
HADS
Anxiety	0.05	0.6
Depression	−0.11	0.3
FEV1	0.38	<0.001
FEV1/FVC	0.36	<0.001
FEV1% predicted	0.32	0.001
FVC	0.18	0.07
FVC% predicted	0.15	0.1

Data are Spearman correlation coefficients. ODI = Oxygen desaturation index; BMI = Body mass index; SpO₂ = Percutaneous oxygen saturation; CAT = Chronic obstructive pulmonary disease assessment test; ESS = Epworth sleepiness scale; HADS = Hospital anxiety and depression scale; FEV₁ = Forced vital capacity in one second; FVC = Forced vital capacity.

Figure 1.  Oxygen desaturation index (ODI) by the GOLD (Global Initiative for Chronic Obstructive Lung Disease) classification. ODI was significantly different among GOLD classifications and tended to be higher in the milder GOLD classification groups.

Table 5.  Logistic regression analysis for predictors of ODI ≥ 15

		ODII ≥15
	OR	95% CI	P value
Age (per year)	1.22	1.08–1.42	< 0.001
BMI (per kg/m^2)	1.09	0.87–1.38	0.5
Pack-years of smoking (per pack-years)	0.99	0.96–1.01	0.3
Resting SpO2 (per percent)	1.39	0.87–2.49	0.2
CAT (per score)	0.88	0.77–0.99	0.02
ESS (per score)	1.02	0.86–1.18	0.8
Exacerbation of COPD within a year	0.81	0.10–3.16	0.7
HADS
Anxiety (per score)	0.94	0.73–1.18	0.6
Depression (per score)	0.97	0.79–1.16	0.7
FEV1 (per liter)	2.59	0.91–7.83	0.08
FEV1/FVC (per percent)	1.07	1.00–1.16	0.04
FEV1% predicted (per percent)	1.03	1.00–1.07	0.03
FVC (per liter)	1.42	0.66–3.15	0.4
FVC% predicted (per percent)	1.01	0.98–1.04	0.4

OR = Odds ratio; CI = Confidence interval; ODI = Oxygen desaturation index; BMI = Body mass index; SpO₂ = Percutaneous oxygen saturation; CAT = Chronic obstructive pulmonary disease assessment test; ESS = Epworth Sleepiness Scale; HADS = Hospital Anxiety and Depression Scale; FEV₁ = Forced vital capacity in one second; FVC = Forced vital capacity.

Discussion

The principal results of this study are (1) the ODI is inversely correlated with airflow limitation, and ODI is higher in the milder GOLD classification groups; (2) FEV₁% predicted and FEV₁/FVC are significant predictors of ODI ≥ 15. We hypothesized that severe COPD patients would tend to have the complication of SAHS, but precisely the opposite result was obtained: milder COPD patients tended to have a higher ODI.

This paradoxical correlation between airflow limitation and ODI is a surprising observation. In a large epidemiological study conducted in the United States with 5,954 participants, the relationship between AHI and FEV₁/FVC was evaluated by analysis of variance after adjusting for age, race, sex, BMI and smoking status (21). The multivariate models revealed no significant interaction between BMI and FEV₁/FVC in their relationship with AHI, but there was a significant positive relationship between AHI and FEV₁/FVC in that study. It is unknown whether airflow limitation itself is associated with AHI, or why severe COPD patients show lower AHI. In a similar way, it is challenging to explain why patients with severe airflow limitation tend to show a lower ODI, but several possible causes may contribute to the finding.

The first assumption is that diminished fat and soft tissue around the neck after losing weight during COPD progression may result in upper airway patency and reduce the possibility of nocturnal desaturation. It has been reported neck circumference is a positive independent risk factor for high AHI regardless of BMI (22). The second assumption is that diminished inspiratory force resulting from muscle weakness and a flattened diaphragm during COPD progression may have an effect on upper airway patency and may lead to a lower ODI, because negative intrathoracic pressure mainly derived from contraction of the diaphragm during inspiration may collapse the upper airway. Finally, diminished rapid eye movement (REM) sleep, which is observed with progression of COPD, could reduce the risk of SAHS (23), and therefore severe COPD patients may also have a lower ODI. These assumptions are largely speculative and there is also a need to determine whether age and CAT score predict the ODI levels or not.

PSG is the gold standard test in the diagnosis of SAHS. However, PSG is costly and time consuming, so nocturnal oximetry can be used as an alternative or the screening test (12). ODI obtained by nocturnal oximetry correlates well with AHI (13, 14). Nakano et al. reported a diagnostic sensitivity and specificity for ODI calculated at a 4% threshold for SAHS defined as AHI ≥ 15 (15). The sensitivity/specificity of ODI ≥ 15 for patients with a BMI < 25 was 54%/100%. Moreover, Chiner et al. reported a sensitivity/specificity for ODI calculated at a 4% threshold for SAHS defined as AHI ≥ 15. The sensitivity/specificity of ODI ≥ 15 were 62%/93% (12). These studies suggested that nocturnal oximetry performed well in screening for SAHS. However, a final diagnosis of SAHS by oximetry alone is not an established method, therefore we suggest more extensive diagnostic tools such as polygraphy or polysomnography for quantification of severity before deciding on therapy.

The major limitations of the study were its retrospective design and small size. The differential diagnosis of obstructive SAHS or central SAHS could provide important information, but was not possible by nocturnal oximetry. In addition, we could not evaluate important factors such as neck circumference and cephalometric abnormalities, which are associated with the pathophysiology of SAHS (22, 24). Because the study was retrospective, we were unable to evaluate PaO₂ and PaCO₂.

In conclusion, ODI is inversely correlated with airflow limitation, and FEV₁% predicted and FEV₁/FVC are significant predictors of ODI ≥ 15. Milder COPD patients may be likely to have the co-morbidity of SAHS.

Declaration of Interest Statement

The authors declare that they have no conflict of interest. The authors alone are responsible for the content and writing of the paper.