This case–control study was performed from January 2020 until December 2020 in Khorshid Hospital, affiliated with Isfahan University of Medical Sciences, Isfahan, Iran. A power calculation of 0.8 with a significance threshold of 5% indicated that it would be suitable to use 30 patients in each group (opium-dependent patients with OHS vs. non-opium users with OHS):

$$n=\left(\frac\right)\left(\frac-\frac+Z_-\beta )^}}\right)+\left(\frac_^-\frac}\right)$$

The ∆ represents the statistical power of 0.8 and φ is the ratio of case to control, which is 1. The remaining variables were extracted from a previous study on a similar topic. Specifically, the apnea/hypopnea index (AHI) and the hypopnea index output were extracted and used in this power calculation [2].

Patients were included if they had PaCo2 > 45 whilst awake in an upright position, BMI ≥30 Kg/m2, and were not suspected of having other causes of hypoventilation such as hypothyroidism and neuromuscular disease. These patients had already been diagnosed with OHS and, thus, pulmonary function testing was not needed to confirm this. Patients with congestive heart failure, severe chronic obstructive pulmonary disease, interstitial lung disease, and incomplete documents were excluded from this study. Therefore, both groups contained patients with OHS and the patients in the case group had an opium dependence with a mean consumption of 0.8 g per day. A cohort study of 50,000 Iranians from 2012 indicated that the range of opium consumption was 0.2–1–2 g per day, with a median of 0.6 g per day [19]. This means our current sample is within range of the standard opium-dependent population in Iran. The Research Committee of Isfahan University of Medical Sciences approved the study protocol, and the Ethics Committee confirmed it.

Twenty healthy people without any history of hypoventilation, pulmonary disease, and opium dependence performed breathing exercises whilst awake during the day. This was to estimate the rib cage and abdominal circumference coefficient factors.

Two devices were used to measure breathing. First, two transducer coils (Pneumobelts) from the polysomnography device (Löwenstein Medical Technology GmbH, Hamburg, Germany) were wrapped around the rib cage and abdomen. The respiratory volume could be determined through the expansion and contraction of these areas during breathing. Specifically, one transducer coil is placed around the rib cage, just under the armpits, and the other around the abdomen, with its upper edge below the lowest rib. Secondly, each person was instructed to breathe into the Spirobag in order to measure VT. A Spirobag is a bag that holds 800 cc when fully expanded and it was calibrated so that a 1.0-volt signal equals a 1.0-litre volume. Therefore, the VT was 800 cc due to the Spirobag being 800 cc. This volume remained fixed throughout usage [20]. See Fig. 1 below of a healthy subject using a Pneumobelt and the Spirobag during calibration.

A healthy subject using a Pneumobelt and the Spirobag during calibration

The following breathing maneuvers were performed: five deep breaths in the supine position followed by a three-minute recovery time, then five deep breaths in the standing position. After this, the chest and abdomen frequency changes were evaluated. The relationship between rib cage, abdominal signals, and VT can be modelled as follows:

In the supine position, breathing occurs more through abdominal expansion whilst in the standing position, breathing occurs through chest expansion. Therefore, we determined the changes in both positions. For estimating the x and y calibration factor, the following equations were used:

$$\mathrm=\frac\right)\left(vol\right)-\left(ab\right)\left(VOL\right)}$$

$$Y=\frac\right)\left(vol\right)-\left(rc\right)\left(VOL\right)}$$

Uppercase letters represent volumes obtained from a standing patient and lowercase letters represent those from a patient in a supine position [15]. After estimating the x and y calibration factor for each of the 20 healthy participants, the mean range of x and y was calculated.

After the coefficients were calculated, both groups of OHS patients (opium-dependent and non-opium users) were admitted to the sleep lab for one night of polysomnography. Here, they received the Pneumobelt without the Spirobag. Demographic data, neck circumference, STOP-Bang (snoring history, tired during the day, observed stop breathing while sleeping, high blood pressure, BMI more than 35 kg/m2, age more than 50 years, neck circumference more than 40 cm, and sex), Epworth Sleepiness Scale (ESS), and the Bilevel positive airway pressure or continuous positive airway pressure treatment data were collected from the patients. Polysomnography was evaluated in two ways: first, the dimensional changes of the rib cage and abdomen were determined according to our previously mentioned equation. This presented us with an estimated VT. Secondly, other polysomnographic data such as average and minimum oxygen saturation and apnea–hypopnea index (AHI) were evaluated in both groups.

Continuous and categorical data were reported as mean ± standard deviation (SD) and frequency (percentage). The normality of continuous variables was checked using the Kolmogorov–Smirnov test and the Q‑Q plot. Categorical data were compared between groups using the chi-squared test and independent-samples t-test for normally distributed continuous data. We compared polysomnographic variables with independent-samples t-test as well as analysis of covariance (ANCOVA) for adjusting potential confounding variables. P ≤ 0.05 was considered statistically significant. All statistical analyses were performed using SPSS statistics for Windows v.16 (SPSS Inc., Chicago, IL, USA).