Identifying patterns of reported findings on long-term cardiac … – BMC Medicine

This systematic review and meta-analysis provides a comprehensive and in-depth examination of findings from studies on long-term cardiac complications of COVID-19. Up to July 2023, at least 150 studies examined 49 different long-term cardiac complications of COVID-19. Chest pain and arrhythmia were the two most widely reported complications. These studies varied substantially in different aspects of study design, and only a quarter of them were of high quality based on our quality assessment. Meta-analysis identified high heterogeneity across studies for almost all cardiac complications, and subgroup analyses showed systematic differences in reported prevalence by the quality and characteristics of included studies. Most strikingly, we observed that studies of high quality reported much lower prevalence of different cardiac complications compared to studies of medium and low quality. To our knowledge, this is the first meta-analysis that quantitatively examined how reported findings of studies on long-term cardiac complications of COVID-19 differ by study quality and characteristics.

It is evident that many COVID-19 survivors have experienced lasting cardiac complications, even those who did not have previous cardiovascular diseases or comorbidities, and who had a low risk of cardiovascular diseases before the pandemic. To date, multiple reviews have examined the long-term cardiac complications of COVID-19 (Additional file 1: Table S6) [4, 5, 9,10,11,12,13,14,15, 28,29,30,31,32]. For example, several previous systematic reviews and meta-analyses also quantified chest pain and arrhythmia as two of the most common long-term cardiac complications, with prevalence estimates for chest pain ranging from 5 to 16% and prevalence estimates for arrhythmia ranging from 10 to 11% [9, 29, 30, 32]. Our prevalence estimates of these two complications broadly agree with these previous findings. A probable source of minor differences in prevalence estimates comparing our study to previous studies is that different systematic reviews and meta-analyses used different inclusion criteria to select studies.

Besides being more updated and comprehensive in terms of article search, our study examined whether there are systematic differences in reported findings by the quality and characteristics of included studies. Meta-analysis stratified by these factors shows that studies of low quality, small sample size, unsystematic sampling method, and cross-sectional design are more likely to report higher prevalence estimates of long-term cardiac complications. For example, the prevalence of chest pain among studies of low quality (22.17%) was five times higher than that among studies of low quality (3.89%). Such patterns were also observed for the prevalence of arrhythmia (studies of low quality vs. high: 24.09% vs. 2.68%) and other less examined long-term cardiac complications. This observation shows how sensitive reported findings can be depending on the quality and characteristics of studies on cardiac complications of COVID-19. Therefore, it is important to take these factors into account and better interpret the findings from these studies.

In our quality assessment, we determined that around 25% of 150 included studies were of high quality and the remaining 75% were of medium or low quality. The small number of high-quality studies demonstrates the urgent need to improve the quality of studies investigating the long-term cardiac complications of COVID-19. Due to the large number of studies included in this systematic review and meta-analysis, we did not enumerate their references. A list of these studies ordered by total quality assessment score can be accessed in Additional file 1: Table S2. A key feature among included studies of medium or low quality was that they were predominantly based on clinical or hospital samples. These studies usually had small sample sizes and had no or only one point of follow-up. While relatively small clinical or hospital-based studies, which are often easier to conduct in shorter periods, can be useful to establish preliminary evidence of an association, especially in an emergent situation such as a pandemic, they often lack population representativeness and statistical power to make broader conclusions about the hypothesized relationships. Furthermore, cross-sectional studies cannot establish the temporality required to infer any causal relationship and prohibit examinations of changes in complications over time.

Interestingly, we observed that studies published in later years (2021 to 2023) had a higher quality assessment score than those earliest studies published in 2020. This shows a general trend of improvement in study quality over time. However, similar quality scores were observed for studies published in 2022 and 2023 (average quality score 9.9 vs. 9.8), which may indicate a recent stagnation in the improvement of study quality on this topic. It is interesting to examine the trend of study quality as more related studies are coming out.

Based on the above findings, we formulated some recommendations for the design and analysis of future studies on long-term cardiac complications of COVID-19 (Table 2). Many studies adopted convenience sampling schemes, which hinders the interpretability and generalizability of their findings. Therefore, it is important to conduct systematic sampling, which can facilitate a continuing and meaningful exploration of the data collected and underpin clinical research. The majority of included studies only assessed long-term complications at a single time point. It is therefore challenging to examine how the long-term complications may change over time. Many studies did not distinguish between long-term complications following COVID and pre-COVID complications at baseline level. Most COVID-19 studies on other types of long-term complications have used similar data sources and analytical methods and will have similar methodological problems as we discussed above. Our study for the first time quantified how reported findings can differ by study quality and selected characteristics. This demonstrates the importance of addressing these methodological problems for COVID-19 studies reporting on long-term cardiac complications and other complications as well.

Although pathophysiological mechanisms underlying COVID-19 cardiac complications remain unclear, studies suggest that the chronic inflammatory response may be hyperactivated by persistent viral reservoirs in the initial acute phase, which may lead to post-acute COVID-19 cardiovascular sequelae [4, 5, 13]. Studies have shown that over 20% of patients with acute COVID-19 had evidence of cardiac injury, even if they did not have underlying cardiovascular diseases or pre-existing comorbidities [33,34,35,36]. It is hypothesized that viral invasion through binding angiotensin-converting enzyme-2 (ACE-2) causes a cytokine storm and triggers systemic hyper-inflammation, which can affect multiple organ systems and induce cardiac injury as one of the severe complications [4, 15]. Persistent chest pain and arrhythmia may be indicative of underlying cardiac abnormalities and damage resulting from systematic hyper-inflammation and/or viral myocarditis affecting the cardiac conduction system. It is critical for clinicians to thoroughly examine patients with long-term cardiac complications of COVID-19, especially for survivors with pre-existing cardiac conditions and other high-risk comorbidities.

Our systematic review and meta-analysis have multiple strengths. First, to our knowledge, this is the most comprehensive systematic review focusing on long-term cardiac complications of COVID-19. It included preprints and articles published in different languages and the global network. Second, in an effort to ensure our results were up to date, we regularly updated our search to capture articles published from the early phases of the pandemic to the most recently published studies. Our review will serve as an invaluable resource for updating researchers and clinicians on key discoveries around long-term cardiac complications of COVID-19. Third, we assessed the quality of included articles from the perspective of study design and epidemiologic principles and provided detailed recommendations on future long-COVID epidemiologic research. The NOS tool assessed the quality of each included study and potential risk of bias, and the GRADE approach determined the level of evidence. Fourth, we performed meta-analysis and subgroup analyses to examine patterns of reported findings, and we observed systematic patterns of reported findings of existing studies.

Our study also has several limitations. First, studies included in our systematic review and meta-analysis are highly heterogeneous. We, therefore, performed subgroup analyses by multiple characteristics, and we believe that existing heterogeneity across studies makes it difficult to generalize our results to the general population. Second, we were unable to stratify our meta-analysis by the length of follow-up because of widely varying follow-up times and different index dates of follow-up across studies. We intended to report the meta-analysis results by hospitalization status; however, most studies have a mixed cohort of inpatients and outpatients, and some studies did not report this information. Such variations in design and lack of detailed data made the stratified results hard to interpret. Finally, we could not stratify our analyses based on prior comorbidities, history of cardiovascular diseases, treatment or medication use for cardiac complications, or COVID-19 vaccination status due to limited reporting of such information, particularly in the studies published during the initial stages of the pandemic. This is because much of the related information was not clearly given in most existing studies. We plan to conduct these analyses once more data on these factors becomes available.

As the pandemic comes to an end worldwide, we may live together with COVID-19 in the coming years, and the epidemiology of long-term cardiac manifestations of COVID-19 might change over time. We think that multiple factors may strongly influence the prevalence or rate of long-term cardiac complications of COVID-19, including a shift in the demographic affected from primarily older people with comorbidities at the beginning to the general population, the availability of vaccination, treatment, and in-home testing, and the emergence of new COVID-19 variants [5, 37]. In future studies, how these factors may influence long-term cardiac complications of COVID-19 should be carefully examined.

In conclusion, we found there were diverse manifestations of cardiac complications, and many can last for months and even years. There is substantial heterogeneity in terms of study design and systematic differences in the reported prevalence of complications by study quality and characteristics. Specifically, we found that studies with low-quality, small sample size, unsystematic sampling method, or cross-sectional design were most likely to report a higher prevalence of complications among individuals who survived COVID-19. We believe that a deeper understanding of long COVID is currently prevented by the limitations of the published literature. Our study underscores the need to conduct high-quality studies on long COVID and the importance of long-term cardiac surveillance of COVID-19 survivors.

Read more:

Identifying patterns of reported findings on long-term cardiac ... - BMC Medicine