Ambulatory pulse oximetry monitoring in Japanese COPD outpatients not receiving oxygen therapy

Chronic obstructive pulmonary disease (COPD) is a leading cause of morbidity and disability worldwide, and is predicted to become the third highest cause of death by 2020 [1]. Desaturation profiles during a 6-minute walk test (6MWT) may predict prognosis primarily in patients with severe COPD with a percent predicted forced expiratory volume in one second (FEV₁) of < 50% [2]. Time to desaturation during a 6MWT also predicts desaturation time in 24-hour ambulatory oximetry monitoring primarily in moderately hypoxaemic COPD patients with a resting partial pressure of arterial oxygen (PaO₂) between 60 and 70 mmHg [3]. However, it remains unknown whether desaturation profiles measured by ambulatory oximetry can predict prognosis in COPD patients. Transient desaturations have been observed in patients with moderate to severe chronic pulmonary disorders, even without significant resting hypoxemia [4]. Point measurements of resting oxygen saturation by pulse oximetry (SpO₂) and PaO₂, the conventional parameters used to determine requirements of long-term oxygen therapy, are not sufficient for assessment of desaturation during activities of daily living [5]. Field walking tests such as the 6MWT, which is the standard test used for assessment of functional exercise tolerance, do not always provide a good reflection of variations in oxygen saturation, because most activities of daily living are performed at submaximal levels of effort. 6MWT results did not predict the degree of desaturation during defecation in patients with chronic respiratory failure [6]. Conventional assessment methods are therefore not satisfactory for obtaining a comprehensive understanding of oxygen saturation throughout the day.

Pulse oximetry is a rapid, non-invasive method of monitoring the oxygen saturation of haemoglobin. Recent technological advances have enabled development of low-cost, small, user-friendly, portable pulse oximeters that can record data for over 24 hours. Many studies have evaluated the usefulness of pulse oximetry in COPD patients, but most studies only monitored patients continuously during the night or for a short period during the day. Although several studies monitored patients continuously during the day, most of these included patients with severe COPD on long-term oxygen therapy, or exclusively analysed patients on long-term oxygen therapy.

This study aimed to evaluate the desaturation profiles of Japanese COPD outpatients not receiving supplemental oxygen therapy during activities of daily living at home using continuous 24-hour pulse oximetry monitoring, and analyse the frequency of desaturation and relationships between the desaturation profile and the clinical characteristics and risk of exacerbation. The ability of the desaturation profile to predict exacerbation of COPD was investigated.

A total of 51 Japanese patients were enrolled in the study from February 2011 to September 2011. Thirty-eight patients were ex-smokers, 12 were current smokers, and 1 was a non-smoker who had been chronically exposed to excessive environmental tobacco smoke. We excluded the 12 current smokers from our analyses, because smoking status was not considered at study entry and some current smokers were unexpectedly included. Current smoking may increase SpO₂ values due to a high serum carboxyhaemoglobin (COHb) concentration [15‐17]. The characteristics of the ex-/non-smokers are shown in Table 1.

The daytime and nighttime oximetry monitoring results are shown in Table 2. There were no daytime desaturators and three nighttime desaturators. Although we did not confirm OSA by polysomnography, the results showed possible OSA (3%ODI of more than 15 desaturations/hour) in six patients [9]. Comparison of these 6 patients with the remaining 33 patients found no significant differences in the proportion of SpO₂ values below 90% in the daytime (1.7 ± 2.1 vs 3.2 ± 6.0%, p = 0.28) or nighttime (15.4 ± 9.9 vs 5.9 ± 17.0%, p = 0.08) or in the mean daytime SpO₂ value (94.6 ± 0.8 vs 94.7 ± 1.3%, p = 0.86), but did find a significant difference in the mean nighttime SpO₂ value (92.3 ± 0.9 vs 93.9 ± 1.8%, p = 0.03). In the nine patients with a resting PaO₂ of 60–70 mmHg (resting PaO₂ 66.2 ± 2.8 mmHg), including no daytime and two nighttime desaturators, the proportion of SpO₂ values below 90% was 8.4 ± 9.0% during the day and 20.2 ± 28.5% during the night, and the 3%ODI was more than 15 desaturations/hour in one ex-smoker (15.6 desaturations/hour).

Simple regression analysis showed that higher %FEV₁ (β = -0.12, p = 0.01), higher PaO₂ (β = -0.30, p = 0.002) and lower mMRC dyspnoea grade (β = 2.38, p = 0.002) were significantly associated with a lower proportion of daytime SpO₂ values below 90% (Additional file 1: Data 1). Multivariate analysis identified only PaO₂ as a significant predictor of the proportion of daytime SpO₂ values below 90% (β = -0.21, p = 0.04) (Additional file 1: Data 2).

Five ex-smokers were lost to follow-up. Two ex-smokers died of acute heart failure and unknown cause, respectively. Thirty-two ex-/non-smokers were alive and still followed up at the time of data cut-off. One third of the patients experienced exacerbations during the follow-up period (Table 3). The median time from oximetry monitoring to the first exacerbation was not reached (Figure 1). Univariate (Table 4) and multivariate (Table 5) analyses did not detect any factors that were significantly associated with exacerbation.

This study of Japanese COPD outpatients not receiving supplemental oxygen therapy included no daytime desaturators and few nighttime desaturators. The proportions of daytime and nighttime SpO₂ values below 90% did not predict exacerbation.

To the best of our knowledge, only four previous studies have reported the results of continuous 24-hour oxygen saturation monitoring during activities of daily living using a portable pulse oximeter (Table 6) [3, 5, 11, 18]. The population of the present study differed from the populations of these previous studies in terms of race, disease severity (%FEV₁ and resting PaO₂), physique (BMI), and SpO₂ values. Patients with more severe COPD generally tend to be more underweight [19‐21]. However, the patients in this study had milder COPD but were more underweight than patients in previous non-Asian studies. Additionally, our oximetry results were much better than in previous studies. Patients in this study had less frequent daytime and nighttime desaturations than those in the previous studies. Although even the nine patients with a resting PaO₂ of 60–70 mmHg had better oxygen saturation profiles than patients in the previous non-Asian studies, their desaturation profiles and pulmonary function were comparable to those of the 55 non-desaturators in the study by Casanova et al., which also found that non-desaturators were less hypercapnic and less hypoxemic during the daytime than desaturators in spite of having similar pulmonary function and Saint George’s Respiratory Questionnaire scores [11]. Our nine patients with a resting PaO₂ of 60–70 mmHg had lower resting PaCO₂ values than patients in the previous studies. These comparisons suggest that desaturators are likely to have hypercapnic respiratory failure.

In this study, oximetry results did not predict exacerbation. This is consistent with the findings of Trauer et al., who reported that ambulatory oximetry results in patients with a resting PaO₂ of 56–70 mmHg did not predict exacerbation or survival [12]. They proposed four reasons for the lack of association between oximetry results and survival: 1) oximetry results are not fundamentally important for predicting prognosis, 2) progression of COPD is unpredictably heterogeneous, 3) the sample size was not large enough to detect differences, and 4) low ambulatory oxygen saturation reflected greater levels of physical activity. In the present study, the lack of association between oximetry results and the risk of exacerbation may be related to the small sample size, as the patients had mild COPD and therefore had lower rates of exacerbation and death than in the study by Trauer et al.[12]. However, the CAT score may be an independent predictor of exacerbation, as also recently reported in other studies [22, 23].

This study has some limitations. First, despite using an activity log, it was not possible to accurately determine the patients’ physical activities at specific times. It is difficult for patients to accurately record activities by the minute, or accurately recall activities over longer periods of time [24]. Most patients in this study had inaccurate recordings, and patients frequently forgot to record any activities at all. The resulting activity logs were not useful for comparison with pulse oximetry results. It has been reported that patients with COPD are markedly inactive in daily life [25], and that the reduction in physical activity starts in the early stages of the disease, even before COPD is diagnosed [26]. It is therefore possible that the minimal desaturations recorded in this study reflected reduced activity levels. Use of motion sensors such as accelerometers seems to be a promising alternative to patient-reported activity logs. Only one previous study reported concomitant measurement of SpO₂ and physical activity by oximeter, accelerometry and actigraphy, respectively [27]. However, the analysis in this study divided activities into only four categories: walking, slow-intermittent-walking, active-not-walking, and resting. These broad classifications are not sufficient for linking desaturation profiles with activities of daily living. Second, the influence of COHb on SpO₂ values in the 12 current smokers was not assessed, because the COHb concentrations were not measured. Therefore, we excluded the current smokers from our analyses, and it remains unknown whether ambulatory oximetry monitoring is reliable in smokers. Finally, oximetry monitoring was performed only once, as in the previous studies. Only the study by Casanova et al. reported repeat 24-hour oximetry monitoring, at 1–3 weeks after the initial monitoring in 11 patients [11]. They did not find significant differences between the results of the two monitoring periods. Although it is necessary to investigate whether a single 24-hour monitoring period is sufficient for evaluation of the desaturation profile, it is difficult to perform monitoring over multiple periods because of patient discomfort and disturbance of activities of daily living while the monitor is attached.

Our 24-hour ambulatory oximetry monitoring provided precise data regarding the desaturation profiles of Japanese outpatients with COPD. Both daytime and nighttime desaturations were infrequent. The proportion of SpO₂ values below 90% did not predict exacerbation of COPD.

The data sets supporting the results of this article are included within the article and its additional files.