Abstract
Purpose The rapid spread of the COVID-19 pandemic has prompted researchers from all over the world to share their experience. The results were numerous reports with variable quality. The latter has provided an impetus to examine all published meta-analyses and systematic reviews on COVID-19 to date to examine available evidence. Methods: Using predefined selection criteria, a literature search identified 43 eligible meta-analyses and/or systematic reviews. Results: Most (N=17) studies addressed clinical manifestations and associated comorbidity, 6 studies addressed clinical manifestations in pregnant women and younger individuals, 8 studies addressed diagnostic data, 9 studies addressed various interventions, and 9 studies addressed prevention and control. The number of studies included in the various systemic reviews and meta-analyses ranged from 2 to 89. While there were some similarities and consistency for some findings, e.g. the relation between comorbidities and disease severity, we also noted occasionally conflicting data. Conclusion: As more data are collected from patients infected with COVID-19 all over the world, more studies will undoubtedly be published and attention to scientific accuracy in the performance of trials must be exercised to inform clinical decision-making and treatment guidelines.
Introduction
In December 2019, an outbreak of a new infectious disease in Wuhan in the Hubei Province of China was announced (1). Little did we know; this disease would change the world. The disease was caused by a beta-coronavirus called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and the disease was recently named coronavirus disease 2019 (COVID-19). On March 11, 2020, when there were around 120,000 confirmed COVID-19 cases in over 10 countries, the World Health Organization declared COVID-19 a pandemic. As of 30 April 2020, 3,250,267 cases have been reported across 185 countries, resulting in 234,701 deaths, and 1,052,550 people have recovered (2).
The rapid spread of the COVID-19 pandemic, its alarming severity and mortality, and the resultant detrimental impact on world economy and social and financial needs of millions of people across the globe, have prompted researchers and clinicians from all over the world to share their experience. For better or for worse, a mass of scientific research has been published, as COVID-19 has created the need for fast dissemination of information about the pandemic. Published and preprint manuscripts are shared widely via news and social media outlets, allowing the experiences to be shared across the globe. Nevertheless, there has been recent concern about the quality of COVID-19 research, and how rigorous the peer review processes were in this era of rapid dissemination (3, 4). While the experiences should be shared rapidly and broadly since this is a novel disease process, care must be taken to distinguish between anecdotal experiences and data from randomized controlled trials- the standard clinical trials based on which treatment guidelines are usually created.
In this review, we examined all published meta-analyses and systematic reviews about COVID-19 to date. While we did not intent to evaluate the quality of published studies, we planned to thoroughly examine the data provided and we attempted to compare studies that addressed similar research questions to establish consistency or discrepancy between reports.
Methods
Search strategy
Between January 1st, 2020 and April 30th, 2019, we identified eligible studies using electronic literature search of the following databases: MEDLINE, EMBASE, and the Cochrane Library. We used Medical Subject Heading terms or keywords using a combination of the following search terms: “2019 novel coronavirus,” OR “COVID-19,” OR “Coronavirus,” OR “SARS-CoV-2,” OR “Chinese Coronavirus”. The search terms were combined with the publication types “systematic review,” OR “meta-analysis”. Following screening of retrieved records for the relevant titles, the relevant abstracts were reviewed, subsequently, the full text articles were obtained to determine appropriateness for final inclusion.
Selection criteria
We included all studies that met the following criteria: (1) published in English language between January 1st, 2020 and April 30th, 2020; (2) only published as systematic review, or meta-analysis, or both; (3) examining any of the following: clinical features, diagnosis, associated comorbidity, treatment and intervention, or prevention and control of COVID-19. Studies that addressed any combination of those features were also considered eligible; and (4) including relevant data for any age, gender, race, or a specific risk group. We also intended to include duplicate articles if they provided additional relevant data.
Data extraction
All authors reviewed the full text of potential articles and discussed the data intended for extraction and decisions were documented. Extracted data included the following fields: first author last name, the purpose of the study, the type of the study being a systematic review; a meta-analysis; or both, number of included articles, number of included patients, pertinent cross references, and the relevant findings.
Results
We identified 7,330 potentially relevant articles. After exclusion of duplicate references, non-relevant literature, and those that did not satisfy the inclusion criteria, 43 systematic reviews/meta-analyses were included (5-47).
Table 1 depicts the summary of the 43 included systematic reviews/meta-analyses. Most (N=17) studies addressed clinical manifestations and associated comorbidity, 6 studies addressed clinical manifestations in pregnant women and younger individuals, 8 studies addressed diagnostic data, 9 studies addressed various interventions, and 9 studies addressed prevention and control. Several studies were designed to report on more than one aspect. The number of studies included in the various systemic reviews and meta-analyses ranged from 2 to 89 and some systemic reviews and meta-analyses included no COVID trials. In Table 2 highlights the most significant findings derived from each systematic review/meta-analysis classified according to the research objective(s).
Table 2 highlights the most significant findings derived from each systematic review/meta-analysis classified according to the research objective(s). The systematic reviews/meta-analyses we reviewed included studies with occasionally disparate outcomes. The systematic review by Vardavas et al (6) included studies that demonstrated no relationship between COVID-19 infection severity and smoking and others that did. Three systematic reviews/meta-analyses reported patients were predominantly middle-aged men (10) (19) (40), and 2 systematic reviews/meta-analyses reported the prevalence of hypertension and diabetes to be around 16% and 8%, respectively (17) (20). Fever (>80%) and cough (~60%) were the most common symptoms in 2 systemic reviews and meta-analyses (10) (19) (40). Hypertension and diabetes were reported to be associated with more severe disease in several systematic reviews/meta-analyses and ARDS was reported as complication in 9.4% to ~30% of COVID-19 infections. COVID-19 mortality varied anywhere from 3.2 to 14% (10) (20) (39). In pregnant women with COVID-19 infection, the caesarean section rate was >80% and 2 systematic reviews/meta-analyses showed no vertical transmission. Children with COVID-19 presented with mild to moderate symptoms and 2 systematic reviews/meta-analyses reported no deaths in children aged 0 to 9 years (Table 2).
The most frequent chest CT findings were GGO and bilateral infiltrates in several systematic reviews/meta-analyses and 1 meta-analysis showed increased procalcitonin concentrations were associated with higher risk for severe COVID-19 infections (23). With regards to medication interventions, 3 systematic reviews/meta-analyses reported no benefit for using antiviral drugs (15) (29) (35). One systematic review concluded neither benefit or harm with use on non-steroidal anti-inflammatory drugs and that interleukin-1 and −6 have no modulatory effects on the course of COVID-19 infections (12). Systematic reviews/meta-analyses on the effects of chloroquine and hydroxychloroquine showed varying results. One systematic review noted superior symptom control and improvement in radiographic imaging in one study and superior virological clearance compared to control in another study (13). Another systematic review/meta-analysis also showed reduction in radiologic progression of disease, but no significant benefit regarding virological cure or mortality reduction (24) (Table 2).
Systematic reviews/meta-analyses on prevention and control of COVID-19 provided occasionally conflicting results as well. For example, one review concluded that school closures did not contribute to control of SARS and modeling for COVID-19 predicted school closures would only prevent 2-4% disease mortality (45); while another review that combined COVID-19 and non-COVID-19 studies reported that quarantine combined with school closures and other measures was effective (5). One systematic review/meta-analysis examining medical masks vs N95 respirators for preventing COVID-19 in healthcare workers did not include any COVID-19 studies (26).
Discussion
Of the 7850 potentially relevant articles, 43 systematic reviews/meta-analyses met our inclusion/exclusion criteria were included (Figure 1). The various studies included addressed clinical manifestations and associated comorbidity, clinical manifestations in pregnant women and younger individuals, diagnostic data, various interventions, and prevention and control of COVID-19. As noted, we did not intent to evaluate the quality of published studies but noted occasionally conflicting data, large ranges for COVID-19 associated acute respiratory distress syndrome (ARDS) and mortality estimates, systematic reviews/meta-analyses that included non-COVID-19 trials, similarities for comorbidities increasing disease severity such as diabetes and hypertension, and predominance of middle-aged male patients with COVID-19 infections.
Following the outbreak of COVID-19 in Wuhan in the Hubei Province of China in December 2019 (1) and the subsequent declaration by the WHO of a pandemic that has caused alarming mortality and disruption of the world economy affecting several millions of people worldwide, there was need for rapid dissemination of information regarding this novel infection as more clinicians and scientists dealt with the reality of COVID-19. Technology and social media played a large part in the sharing of clinical information and experiences with disease management-Chinese clinicians communicated their experiences with European and American clinicians via virtual conferences, for example, because the virus had affected their community first. The information disseminated this way was based on clinical experiences and the patients seen in their hospitals who may differ in comorbidities from patients seen in Europe and North America, however, the dissemination of information was critical to guide clinicians.
We also witnessed a flood of published manuscripts in several journals, including highly impactful ones. Some published studies included small numbers of patients and it is likely the studies were published to share as much information as possible given the devastating effects of COVID-19 on millions globally. As more patients became ill with COVID-19, as hospitals became overwhelmed with sick patients with various ranges of disease severity and ran out of beds and ventilators, as personal protective equipment was challenging to secure in some countries, as various medications were attempted for prophylactic use and treatment, as financial institutions suffered major loses, as weaknesses in healthcare systems were exposed, and most important, as millions of lives were lost, the internet served as an integral space to share both accurate and inaccurate information; including published and retracted studies.
There were common clinical manifestations and comorbidities in the systematic reviews/meta-analyses we included. Notably, patients were predominantly middle-aged men and comorbidities such a diabetes, hypertension, cardiovascular disease, and smoking were prevalent and associated with more severe disease. The most common clinical manifestations were fever, cough, and dyspnea. Rates of ARDS and mortality varied between studies. In pregnant women with COVID-19 infection, the caesarean section rate was >80% and children with COVID-19 presented with mild to moderate symptoms and no deaths were reported in children aged 0 to 9 years (Table 2). The most frequent chest CT findings were GGO and bilateral infiltrates. With regards to medication interventions, no benefit for using antiviral drugs was found, no benefit or harm with use of non-steroidal anti-inflammatory drugs was found, and no benefit with interleukin-1 and −6 was found. Radiographic improvement and better symptom control were reported with chloroquine and hydroxychloroquine. Finally, quarantining, social distancing, and school closures may contribute to reduction of disease spread and may be more effective if used in combination (Table 2).
The ability to rapidly disseminate information via the rapid review process some journals implemented is a double-edged sword. It remains of utmost important to maintain the integrity of published science while keeping in mind the devastation this COVID-19 has caused and the desperate need for a cure, medications that will lessen disease severity, and/or a vaccine. Additionally, care must be taken to distinguish between anecdotal experiences and data from randomized controlled trials- the standard clinical trials based on which treatment guidelines are usually created.
As more data is collected from patients infected with COVID-19 all over the world, more studies undoubtedly will be published. As clinicians, scientists, and members of society, we will look forward to the sharing of information. Scientists, reviewers, editors, and journals alike must work to maintain the integrity of the science shared. The internet, social media, the ability to hold virtual conferences across the globe, and the yearning to find a cure will promote dissemination of clinical experiences and science. This pandemic will not only change the way we live, but also encourage the use of non-conventional methods to share science.
Data Availability
Not applicable
Footnotes
Emails Nasrien E. Ibrahim: nibrahim{at}mgh.harvard.edu
There is no conflict of interest to disclose