Both obesity and obstructive sleep apnea (OSA) are recognised independent risk factors for the development of atrial fibrillation (AF)1, and a large proportion of people living with OSA have comborbid AF2. In the recent years, the effectiveness of Continuous Positive Airway Pressure (CPAP) in reducing AF recurrence has been shown3,4. The mechanism by which CPAP is effective into reducing that burden is however not well understood. One of the underlying pathological changes that occur in OSA patients is altered atrial electrical conduction, predisposing them to cardiac re-entry circuits and thus to the development of AF5-9. After conducting a broader literature search to put our research into context, I aim to specifically explore the literature to date looking at the effect of CPAP in reversing those atrial electrical changes that occur in patients with OSA, in the context of reducing the risk of developing AF. A systematic literature search across four databases yielded five papers answering our specific research question10-14, which we then qualitatively analysed and appraised. While the literature to date seems to show that CPAP is effective in reversing atrial electrical conduction predisposing OSA patients to the development of AF, the research is scarce and the methods and results are to date very heterogeneous and specific to certain populations, limiting the generalisability and applicability of results. This prompts the need for further research in the area and provides a rationale for the parallel research project conducted by the rest of my MD research group.
Section IIB – Context, Rationale, Study questions
Atrial fibrillation (AF) is the commonest sustained human arrhythmia15. It is associated with substantial morbidity, mortality and health economic costs16. Obesity and obstructive sleep apnea (OSA) are recognised independent epidemiological risk factors for the development of AF1. This is of considerable importance from a public health perspective with an estimated 66% of Australian adults classified as overweight or obese17. OSA is defined as partial or complete obstruction of airflow during sleep, resulting in repetitive episodes of apneas or hypopneas, and a subsequent decrease in oxygen saturation. There is a direct relationship between OSA and Body Mass Index (BMI), and it is estimated that 40% of obese people live with comorbid OSA18. Furthermore, the large scale Sleep Heart Health Study demonstrated that, compared to subjects without sleep disordered breathing, patients with sleep-disordered breathing were four times more likely to develop AF2. Importantly, patients with AF and underlying OSA have worse symptoms and outcomes than AF patients without OSA19. With the increasing trends in obesity prevalence, the burden of AF is expected to grow in parallel20. This prompts the need for further research into treatment options and efficacy, in an attempt to reduce the anticipated AF-related burden of disease. Specifically, the recognition that OSA represents an independent risk factor for the development of AF, merits investigating whether appropriate treatment of OSA could reduce the burden of AF. The most common treatment for moderate to severe OSA is Continuous Positive Airway Pressure (CPAP)21, and thus in the recent years, research looking at the potential for CPAP to reduce AF burden has increased. Specifically, two large scale meta-analyses have found that CPAP is effective in reducing the risk of AF recurrence in an OSA population3,4. However, the mechanism by which it does so is not well understood, and several potential mechanisms of how OSA predisposes to the development of AF have been put forward. Our research group aims to specifically investigate whether CPAP is effective in reversing the atrial electrical changes that occur in patients with OSA and predispose them to AF, using ECG markers. My part in this project is to first conduct an extensive literature review to put our research question into context and gain an understanding of the proposed mechanistic links between OSA and AF, and then to conduct a systematic search across four databases to unveil existing studies that also aim to answer our specific research question: Is CPAP effective in reversing pathological atrial electrical changes that predispose OSA patients to the development of AF?
415
Section IIC: Review of Existing Knowledge
I first synthesise findings from the background literature search in this introduction, and report findings of the systematic search in the results section, as it is the main focus of my project. I focus on two particular aspects for my background search, namely, the proposed mechanisms by which OSA may lead to the development of AF, and the pathological atrial electrical changes observed in OSA patients.
Mechanistic links between OSA and AF: Firstly, patients with OSA have recurrent episodes of nocturnal hypoxia during their sleep, which is in itself an independent predictor for the development of AF22. The severity of nocturnal hypoxemia further predicts increased inter-ventricular septum thickness, eventually leading to impaired filling of the left ventricle23. This in turn increases the pressure in the left atrium, over time leading to an increase in left atrial size and a substrate for the development of AF24. Secondly, the increased effort of breathing against an obstructed airway leads to an abrupt increase in intra-thoracic negative pressure in OSA patients. Studies have demonstrated that an increase in negative intra-thoracic pressure further results in an increase in left atrial volume25. Furthermore patients with OSA have higher left atrial volume indices than BMI- and blood pressure-matched controls without OSA, suggesting that this effect may be independent of obesity and hypertension, and more specific to acute changes that occur during nocturnal apnoeic episodes26,27. An enlarged atrium is furthermore a risk factor for the development of AF28,29, providing a mechanistic link between increased intra-thoracic negative pressure and the development of AF (Appendix – Figure 1). Lastly, altered autonomic activity has been demonstrated to occur in OSA patients, and be related to severity of disease30. This increase in sympathetic drive is mediated by both the increase in intra-thoracic pressure31 and episodes of hypoxia32, and eventually leads to an increase in blood pressure33, thereafter resulting in elevated left ventricular pressures and diastolic dysfunction, further exacerbating left atrial enlargement34,35. Put together, it is thus thought that transient acute changes that occur in episodes of sleep apnea, such as increase episodes of hypoxia and hypercapnia, as well as changes in intra-thoracic pressure and autonomic regulation, can in the long term lead to cardiac tissue ischemia and fibrosis, remodelling of the cardiac wall structure and enlargement of the atrium36 (Appendix – Figure 1).
Pathological atrial electrical changes in OSA patients: Cardiac structural and functional changes that occur in obstructive sleep apnea patients can in turn result in disturbed electrical conduction in the atrium, which provides a substrate that favours electrical re-entry mechanism (when the signal does not complete its usual conduction circuit and instead loops back on itself), and thus the development of AF. In particular, two electrical changes are important in the predisposition to AF37: First, an increase in total atrial conduction time (TACT)38, which is ideally measured by Signal-Averaged P-Wave Duration39,40, or alternatively shown by increased P-Wave Duration (PWD)41 on ECG traces. Second, a heterogenous conduction between cardiac chambers also predicts the development of AF, which is represented by an increase in inter and intra-atrial conduction time, as measured by either Tissue Doppler Imaging (TDI), as well as P-wave dispersion (Pdisp) on ECG traces8,9,37. Observational studies comparing OSA patients with AF to non-OSA patients with AF have found that OSA patients have prolonged TACT5. Both P-wave duration, as well as P-wave dispersion are higher in patients with OSA compared to normal controls, and are also correlated with both the duration of decrease in nocturnal oxygen saturation7, as well as with disease severity6,42.
As our research group aims to investigate the effect of CPAP treatment on electrocardiographic (ECG) markers of altered atrial conduction in an OSA population, I then focus my systematic search to unveil papers answering this specific research question. I then aim to appraise those papers, paying particular attention study methods, limitations and potential confounding variable, as well as the implication for our broader research project.
647
Methods
With the recognition that ideally, a proper systematic review requires at least two independent assessors, as per the Cochrane Handbook for Systematic Reviews43, which wasn’t feasible here, I nonetheless use a methodical systematic search to capture the entirety of the exiting literature relating to our specific research question: What is the effect of CPAP on atrial electrical conduction changes that predispose OSA patient to the development of AF?
Four databases were searched for relevant papers: MEDLINE, EMBASE, Cochrane Library, and CINAHL. Key search words included mapped subject headings for CPAP, AF and OSA, as well as additional phrases such as “Atrial Fibrillation”, “AF”, “OSA”, “obstructive sleep apnea, “CPAP”, “Continuous Positive Airway Pressure” (Appendix – Table 1). I opted not to focus our key words for the outcome to “atrial electrical remodelling”, as the different wording combinations for this phrase is wide, and studies may have been missed in the search.
My initial search yielded 316 publications. 98 papers were repeats across database, and, after a first screen by title and abstracts, 76 papers were deemed relevant to our broad topic of interest (any publication that examined the link between OSA and AF, and discussed CPAP therapy as the main treatment for OSA) and were initially included. 142 papers were excluded as per the following criteria: any paper that wasn’t a review paper, journal article, or research study (i.e.: case reports, conference abstracts, editorials, comments, letters), publications in another language, or any paper that was deemed irrelevant as per our PICO formulation (i.e.: papers not looking at OSA population, CPAP as main treatment, or AF as an outcome). There were no restrictions on year of publication or other population factors.
I examined these 76 initial papers, and after extracting relevant background information to put our research question in context, I further excluded all the ones that didn’t look at electrical markers of atrial conduction as one of the main study outcomes. Reference lists of relevant review papers were also scrutinised for possible publications missed in the search, with no further additions. This helped to provide some assurance that my search strategy was appropriate to capture the entirety of the evidence to date, and that no relevant studies were missed in our analysis, as well as ensure that selection of the relevant papers was not subject to bias.
This left me with a total of five studies looking at the effect of CPAP on atrial electrical remodelling in an OSA population. Those papers were qualitatively analysed and appraised, paying particular attention to the study methods, inclusion and exclusion criteria, results, potential confounding variables, and proposed mechanism for observed effects. I discuss the heterogeneity in methods that could explain the differences in results between studies, as well as attempt to pool the results together to reach a conclusion as to what the evidence shows to date.
Word count: 458
Results
A rigorous search strategy has revealed that to our knowledge, only five studies have sought to investigate the effect of CPAP in reversing the abnormal atrial electrical changes that occur in patients with OSA and predispose them to AF10-14. The methods and results of these studies are summarised in Table 2 (Appendix). We will first describe those studies individually and then appraise them as a whole.
Three studies have looked at the short term effect of CPAP, over the course of 4-6 weeks. Two of those found that CPAP usage was associated with a significant decrease in SAPWD10,12 in patients with moderate-severe OSA, with an increase effect size associated with increase CPAP usage compliance12. Interestingly, those changes in SAPWD were not associated with a decrease in BMI or blood pressure after intervention10,12, suggesting that the effect of CPAP in reversing atrial electrical changes is independent of changes in weight and blood pressure, and more a result of actual alterations of the heart’s electrical conduction. Another study used a different measure of total atrial conduction, PA-Tissue Doppler Imaging (PA-TDI), to measure the effect of CPAP on electrical conduction in OSA patients13. Again, CPAP was found to be effective in reducing total atrial-conduction time after 30 days of treatment, and this was accompanied by a decrease in serum BNP levels, which is a measure of atrial and ventricular stretch, suggesting a potential mediator for this effect13.
Only two studies have looked at the long term effect of CPAP, over a 6 months period. One found that CPAP was effective in reducing total atrial conduction time, as well as reversing markers of inter- and intra-atrial heterogenous conduction in patients with moderate-severe OSA. This was further accompanied by a reduction in left atrial size, again suggesting a mediating effect11. Finally, one study focused on the effect of long-term CPAP therapy compared to standard care in minimally-symptomatic OSA patient only14. This was the only randomised, multi-centre trial, and it was found that CPAP was not effective in reducing either Pmax or Pdisp in minimally symptomatic patients, implying that CPAP has limited value in prophylactically preventing the electrical changes that predispose patients to AF, when they are in an early stage of disease. It was however unclear if these patients have higher Pmax or Pdisp compared to a control non-OSA population, as there was no such control group14.
In conclusion, while the literature looking at the of the effect of CPAP in reversing the atrial electrical changes that predispose OSA patients to developing AF is scarce, results to date seem to show that CPAP is effective in doing so to some extent, both in the short-term and the long-term. Particularly, CPAP has been shown to have an effect on both total atrial conduction time as well inter and intra-atrial heterogenous conduction. This effect is associated with some potential mediating structural changes, such as a reduction in atrial volume as measured by Left Atrial (LA) size on echocardiography11, and atrial stretch as measured by BNP13, and a positive correlation with CPAP compliance was also described12, suggesting a dose-responsive effect. The effect on LA size is however not consistent from study to study, with one study describing no effect on atrial size13, a discrepancy that has interestingly, previously been outlined in studies looking at the effect of CPAP on cardiac structure44-47. Surprisingly, those changes were not associated with a change in blood pressure or BMI10,12, challenging this potential mechanistic link, and implying that CPAP has additional beneficial effects on the heart’s electrical conduction on top of potential weight loss and reduction in blood pressure.
Word count: 596
Discussion, limitations and conclusion
In conclusion, our background literature search revealed that OSA patients have an increase in pathological atrial electrical structure compared to the standard population, which favours re-entry mechanism leading to the development of AF. Particular measures of these atrial electrical disturbances are P-Wave Duration (PMax), Signal-averaged P Wave Duration (SAPWD), two measures of total atrial conduction time; as well as P-wave dispersion (Pdisp), and doppler imaging (PA-TDI) measures of inter- and intra-atrial conduction heterogeneity. These changes have all been linked with OSA and are to some extent correlated to disease severity.
One thing to note in appraising those papers is the heterogeneity in both patient population and methods used. Baseline OSA severity, exclusion criteria, patient characteristics (age, sex, weight…) were widely dissimilar between those 5 studies. Moreover, the methods used to measure atrial electrical conduction as well as the main outcome variables were all different in those studies, except for two studies looking at SAPWD10,12. While it is validating that the same effects were found using different measures of atrial electrical conduction, the lack of standardisation of methods make it hard to reach a general conclusion, or pool results together to get an effect size.
Only one of those studies was a double-blinded randomized control trial14, and the rest of the studies were cohort studies with serious methodological flaws. Firstly, the control and patient population were often not matched, with the patient population having higher BMI, HT and atrial size at baseline10,13. The exploration of potential confounding variables was also poorly assessed in those studies, with only one study looking at potential reduction in BMI and blood pressure12. Moreover, while the usage of CPAP was monitored, the correlation between compliance to treatment and effect size was only described in two studies12,14, and no correlation with baseline OSA severity was described either. Further methodological limitations of these studies include high intra- and inter-observer variability when measuring ECG intervals14, patient population mostly consisting in male subjects, and small sample size in most studies.
Nonetheless, while the methods used in these papers were heterogeneous, the results are in fact consistent from study to study, as well as in line with background proposed mechanisms and observed atrial changes in OSA patients, which provide some validity and reliability in the results. It is however interesting to note that the only RCT in these papers did not find a significant effect size, which questions the amplitude of the effect when stricter methodological designs are used. This difference could nonetheless be due to the different population studied, as patients in this particular study only had minimally-symptomatic OSA14. Excluding this study however, CPAP does show promising effects in reversing atrial electrical changes that could predispose moderate-severe OSA patients to developing AF. It is important to take into consideration that the in 4 out of these 5 studies, any patient with previous cardiovascular disease, AF event, or abnormal baseline ECG changes were excluded from these studies. This means that these results are only applicable to a population that has otherwise little cardiovascular health burden. It also means that the effect seen is only relevant to the risk of developing first onset AF, and not recurrent AF.
The analysis of these five existing papers has multiple implications for the conduction of our proposed study. Firstly, the need for more research on this topic area is made evident, both by the paucity of studies to date as well as the heterogeneity in methods and results. Importantly, our study would be one of the first to look at patients with possible other cardiovascular complications, and potential past AF events, as they do not form part of our exclusion criteria. Ideally, a randomised controlled trial comparing CPAP therapy to no-CPAP would be conducted, but for ethical reasons, it is unfeasible to do that in a population suffering from moderate-severe OSA. A thorough methodical approach will however be necessary, with particular attention to baseline covariates and potential confounders, as those are often overlooked in the previous papers. Another member of my MD group is specifically and exclusively looking at the effect of CPAP on potential covariates, which would be a valuable addition to this field of research. Finally, as our proposed method of looking at atrial electrical conduction in our selected population consists of manual measurement of ECG variables, it is important to acknowledge the high intra- and inter-observer differences described in previous studies using this method52 as a limitation of our proposed project.
My particular project also carries some limitations, most important of which is the lack of a second independent assessor, which is an essential requirement for a well-conducted systematic review43. As a result, both extraction and reporting of the data are subject to biases, and analysis of the results may be obscured by personal gaps in knowledge in this complex field. Secondly, our analysis does not involve any statistical calculation of effect size, due to small sample size and wide heterogeneity in outcome variables, limiting our ability to accurately estimate the extent of the effectiveness of CPAP in reversing atrial electrical changes. Lastly, due to the limited number of papers yielded in our search, we were not able to tighten our exclusion criteria to select papers that fit most with our actual study population or methods, limiting the generalisability and applicability of findings to our overarching project.