Predictors of nocturnal oxygen desaturation in chronic obstructive pulmonary disease in a South Indian population.VD Thomas, S Vinod Kumar, B Gitanjali
Department of Tuberculosis and Respiratory Diseases and Sleep Disorders Laboratory, Jawaharlal Institute of Postgraduate Medical Education and Research, Pondicherry - 605 006, India., India
Correspondence Address: Source of Support: None, Conflict of Interest: None PMID: 12215689
Source of Support: None, Conflict of Interest: None
CONTEXT: Nocturnal Oxygen Desaturation (NOD) in patients with chronic obstructive pulmonary disease (COPD) is associated with a poor prognosis. AIMS: To document the occurrence of NOD in COPD and to identify factors which might predict NOD in COPD patients with daytime arterial oxygen tension (PaO2) > or =60 mm Hg. SETTINGS AND DESIGN: This prospective study was conducted in patients attending the special respiratory clinics or admitted in JIPMER, Pondicherry. METHODS AND MATERIAL: Thirty consecutive patients with COPD were divided into two groups, desaturators and non-desaturators on the basis of nocturnal oxygen saturation. Statistical analysis used: Unpaired 't' test, Karl Pearson's correlation equation and stepwise multiple regression analysis was done. RESULTS: Fourteen patients (46.6%) experienced NOD. Desaturators had lower awake oxygen saturation, PaO2, forced vital capacity, forced expiratory volume in one second and peak expiratory flow rate. There was positive correlation between nocturnal O2 saturation and PaO2 (r=0.638 and P value =0.014) and awake SPO2 (r=0.901 and P value <0.001). CONCLUSIONS: The rate of occurrence of NOD in COPD was 46.6%. Awake SPO2 is the only single predictor of nocturnal oxygen desaturation in these patients.
Keywords: Adult, Aged, Anoxia, diagnosis,epidemiology,Circadian Rhythm, Comparative Study, Female, Human, India, Male, Middle Age, Oxygen, blood,Polysomnography, Predictive Value of Tests, Probability, Prognosis, Prospective Studies, Pulmonary Disease, Chronic Obstructive, blood,physiopathology,Risk Assessment, Sensitivity and Specificity, Severity of Illness Index, Sleep Apnea Syndromes, blood,epidemiology,
The existence of lowered oxygen saturation, abnormal breathing pattern and cardiac arrhythmia has been well documented in chronic obstructive pulmonary disease (COPD)., During normal sleep in healthy individuals, alveolar ventilation transiently decreases which is reflected by higher alveolar and arterial PCO2. Maximum increases in pulmonary arterial pressures occur during rapid eye movement sleep. The quality of sleep in patients with COPD is rather poor and has been attributed to nocturnal cough, hypoxaemia and hypercapnia. The physiological changes of the sleep state superimposed on a pathological condition such as COPD worsens the alveolar hypoxia during sleep in these patients, with transient decreases in oxygen saturation causing concomitant increases in pulmonary arterial pressure. The recurrent alveolar hypoxia secondary to hypoventilation is the primary cause of pulmonary hypertension which leads to cor-pulmonale. Nocturnal oxygen supplementation in order to prevent Nocturnal Oxygen Desaturation (NOD) showed a beneficial effect by reducing pulmonary arterial pressure. Hence, the detection of NOD and its correction by oxygen supplementation is critical in the treatment of COPD.
Since the detection of oxygen desaturation during sleep in non-apnoeic COPD patients requires an all night sleep polysomnography which is expensive, time- consuming as well as labour- intensive, several parameters have been studied in order to find daytime predictors of NOD.,,, It was reported that patients who had the lowest oxygen saturation (SpO2) when awake were those whose saturations became lowest during sleep. Others observed that NOD could not be predicted from respiratory function tests or clinical history and that awake PCO2 was an important determinant. Daytime SpO2 was used to predict nocturnal desaturation in stable COPD patients. Pulmonologists from northern India observed that oxygen saturation was low during sleep in COPD patients when compared to normal controls but no attempt was made to identify predictors. Patients from India are known to differ from their Western counterparts mainly by having a lesser body mass index and a variable smoking history. As obesity hypoventilation is known to occur, Western data may not necessarily reflect patterns of NOD in Indian subjects. Since there was no consensus regarding the predictors of NOD in the above studies and there were no reports from India on this aspect, this study was done to document the occurrence of NOD in COPD and to list its daytime predictors.
Patients between the ages of 30-65 years of both sexes with a clinical diagnosis of COPD, diagnosed by symptoms of chronic cough, exertional dyspnoea and/ or wheezing, spirometry consistent with irreversible or ?15 percent reversible expiratory airflow obstruction and FEV1 <70% predicted were included. Inclusion criteria also specified that they have an average daytime resting PaO2 ? 60 mm Hg. Those COPD patients diagnosed with obstructive or central sleep apnoea (after all night polysomnography), or who were in exacerbation / with history of exacerbation within four weeks or who had severe co-existing illnesses like tuberculosis, diabetes mellitus, and corpulmonale were excluded. Thirty consecutive patients were planned to be enrolled for the study. All patients signed a written informed consent form agreeing to participate in this study.
All patients underwent a thorough physical examination and basic investigations which included a haematological profile, blood biochemistry, electrocardiogram (ECG), chest radiography, and ultrasound examination of the heart. Pulmonary function testing was done using MedGraphics PF/DX Pulmonary Function System (Minnesota-55127; USA). The following parameters were measured: Forced vital capacity (FVC); Forced expiratory volume in one second (FEV1); Ratio of FVC to FEV1; Peak expiratory flow rate (PEFR) and Slow vital capacity (SVC). An arterial blood sample was drawn from the radial artery and analysed immediately using an Acid Base and Co-Oximeter Evaluating System (CIBA Corning Diagnostics (H.K.) Ltd., Hongkong) .
Patients underwent all night polysomnography using a computerized Alice-3 Polysomnographic System (Healthdyne Technologies, Marietta, USA). The montage used included electroencephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), ECG, snoring sounds, chest movements, nasal and oral airflow. Nocturnal oximetry was recorded and the data stored in magneto-optical disks. Baseline awake SpO2 was monitored for all patients, which was defined as the mean oxygen saturation during the first 15 min of the recording with the patient awake and supine, just prior to the lights out. Desaturation had been defined as >4% fall, from the baseline SpO2 for ? 5 min.8
As an essential preliminary step, NOD and Nocturnal Non-Desaturators (NND) were systematically compared for any significant differences in distribution of variables studied using a two sample ‘t’ test. Simple linear regression (Karl Pearson’s correlation equation) was used to find out the possible correlation between mean nocturnal saturation and single daytime parameters, which showed a significant difference between the two groups. Only those factors that showed a positive correlation with mean nocturnal oxygen saturation was taken up for stepwise multiple regression analysis, which was applied to evaluate the contribution of awake SpO2 and PaO2 in the prediction of NOD. Fisher’s exact test was used to test for significance after preparing contingency tables.
Thirty patients (28 males and 2 females) were recruited, and subdivided into those who had NOD and those who did not on the basis of nocturnal oximetry. Fourteen were desaturators giving the rate of occurrence as 46.6%. No significant differences were found in the demographic characteristics between NOD and NND [Table - 1]. All arterial blood gas values except pH and four of the five pulmonary function test variables were found to be significantly different in the two groups [Table - 2]. Patients with predicted FEV1 percentage ? 41 and FVC ? 59.5%, had a 3 and 4 fold greater risk of developing nocturnal desaturation respectively.
When correlation analysis was done to study the possible association between the mean nocturnal oxygen saturation and variables showing statistical significance [Table - 3], it was revealed that mean nocturnal O2 saturation had a significant positive correlation with PaO2 (r = 0.638 and P value = 0.014) and awake SPO2 (r=0.901 and P value = <0.001) Further analysis of the desaturators compared with nondesaturators revealed that no patient with an awake SPO2 >95% experienced nocturnal desaturation (a negative predictive value of 100% with a sensitivity of 100% and a specificity of 81.2%).
The occurrence of desaturators in this sample of COPD patients is slightly lower than the reported figures. Western data have patients who have a larger body mass index which is usually associated with a decrease in ventilation resulting in a higher percentage of non-apnoeic desaturators. In some studies, all-night polysomnography was not done and patients who have sleep apnoea may have got included. The number of cigarettes smoked may be independently related to the incidence of nocturnal desaturation though it is not evident from our data. Sixty four percent of desaturators had a clinical picture of chronic bronchitis which is in accordance with others’ findings. Data regarding arterial blood gas values from our study are in accordance with others.,, The occurrence of sustained nocturnal hypoxaemia was most closely related to presence of daytime hypercapnoea and Fletcher et al observed that desaturators had significantly higher PaCO2 and lower daytime PaO2 than non-desaturators. Though this is in agreement with our study others have shown that none of the daytime blood gas values discriminated desaturators from non-desaturators. In our study, awake baseline SPO2 was found to positively correlate with mean nocturnal saturation (r=0.901, P value =<0.001). This is the only parameter, where interstudy comparisons are more robust with considerable agreements across various studies.,
Awake SPO2 is the only independent predictor of nocturnal desaturation. No patient with awake SPO2 >95% had nocturnal desaturation (negative predictive value of 100%, sensitivity 100% specificity 81.2%). These are in accordance with another study which came to similar conclusions. Though desaturators tend to have lower PaO2, higher PaCO2, and more chronic bronchitis than non-desaturators, these parameters are not highly predictive of nocturnal desatura-tion. It should also be noted that we may be underestimating nocturnal oxygen desaturation due to factors involved in monitoring patients in the sleep laboratory but the extent of desaturation would remain the same.
In conclusion this study shows that 46% of patients with COPD will have NOD and that awake baseline SPO2 is the only variable that positively correlates with nocturnal desaturation. This being an inexpensive test, the expense of a nocturnal polysomnography test can be avoided in those patients who have awake baseline SPO2 values above 95%, and hence the therapeutic decision to start on nocturnal oxygen therapy can be deferred.
Ms. Sunitha T.T. for technical assistance during nocturnal polysomn-ography.
[Table - 1], [Table - 2], [Table - 3]