key: cord-0806959-n3b597r5
authors: Jafari-Oori, Mehdi; Dehi, Manijeh; Ebadi, Abbas; Moradian, Seyyed Tayeb; jafari, Mojtaba
title: Incidence of cardiac complications following COVID-19 infection: An umbrella meta-analysis study
date: 2022-01-10
journal: Heart Lung
DOI: 10.1016/j.hrtlng.2022.01.001
sha: e7ba835385786c3151de6baf9888bd2874e2b7e9
doc_id: 806959
cord_uid: n3b597r5

BACKGROUND: COVID-19 causes fatal cardiac damages. Despite many overwhelming meta-analysis related to cardiac complications following COVID-19 disease, no umbrella meta-analysis study has been conducted. OBJECTIVES: We aimed to report the summarized pooled incidences of cardiac complications in the overall, critically ill, and deceased patients, compare the cardiac complications between the severe/non-severe or deceased/non-deceased patients, and also compare poor outcomes between patients with/without acute myocardial injury (AMI). METHODS: PubMed, Scopus, web of science, Cochrane, ProQuest, Springer, Sage journals were searched before April 2021. After assessing the quality and duplicate data, data were run by the random/fixed-effect models, I2 heterogeneity index, Egger's test, and sensitivity analysis. RESULTS: After removing duplicate data, in the overall COVID-19 patients, the pooled incidence of AMI, heart failure, arrhythmia, cardiac arrest, and acute coronary syndrome (ACS) were 21%, 14%, 16%, 3.46%, and 1.3% respectively. In the patients with severe disease, the pooled incidence of AMI and shock were 33 and 35%, respectively. Similarly, in the deceased COVID-19 patients, the pooled incidence rate of AMI and arrhythmia were 56% and 47.5%, respectively. The patients with severe disease were at higher risk of AMI (RR = 5.27) and shock (OR = 20.18) compared with the non-severe cases. Incidence of AMI was associated with transfer to the intensive care units (ICU) (RR = 2.92) and mortality (RR=2.57, OR = 8.36), significantly. CONCLUSION: Cardiac complications were found to be increased alarmingly in COVID-19 patients. Baseline and during hospitalization checking with electrocardiography, echocardiography, and measuring of cardiac biomarkers should be applied.

COVID-19 usually presents with symptoms of a respiratory infection, but extra-pulmonary symptoms, including neurological, respiratory, renal, and cardiac complications, are also common 1 . Cardiac diseases are more prevalent in hospitalized patients with COVID-19 2 .

Studies have reported different incidence rates of cardiac injuries from 4.2 to 25 [3] [4] [5] [6] [7] . The incidence is higher in COVID-19 patients with severe disease 7 .

Diverse cardiac complications, including acute myocardial injury (AMI), myocarditis, heart failure (HF), arrhythmia, pericardial effusion, cardiomyopathy, myocardial infarction (MI), cardiogenic shock, and cardiac biomarkers elevation have been addressed in different studies 1, [8] [9] [10] [11] . A wide range of cardiac involvement, from increased blood biomarkers with no clinical symptoms (such as abnormalities on cardiac imaging or asymptomatic cardiac arrhythmias) to severe cardiac injuries, has been reported in COVID-19 patients 12 . In some studies, recovery occurred within a few days [13] [14] [15] , whereas in others, fulminant myocarditis resulted in a prolonged recovery 13, 16 . In patients with severe disease, the newly developed cardiac complications often lead to poor outcomes such as admission to the intensive care unit (ICU) or mortality 17, 18 . Direct myocardial injury from hemodynamic disorders or hypoxemia, inflammatory myocarditis, stress cardiomyopathy, thrombosis due to hypercoagulability, or systemic inflammation, are addressed as a potential mechanism of cardiac injury 19 .

Several systematic reviews and meta-analysis studies related to diverse cardiovascular complications following COVID-19 disease have been published so far [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] . However, as the number of individual review studies increases, clinical practitioners may find it difficult to keep track of information 33 . Additionally, different incidence rates were reported for each cardiac complication in every individual review study. Also, some primary studies included in some reviews were mostly with case report design, which due to their small sample size, the accuracy of their reported effect measure became questionable. In addition, only a few reviews reported cardiac complications based on the disease severity. Additionally, some previous studies have not specified the exact number of cardiac complications and poor outcomes associated with them. Therefore, this umbrella review study attempts to fill the mentioned gaps with three contributions to the existing literature. As a first contribution, we tried to provide a summarized pooled incidence of the cardiac complications based on nonsevere, severe cases, and deceased COVID-19 patients, which can facilitate data understanding for clinicians. Our next contributions are to determine which one of the incidences of cardiac complications is more than others or which one of them has a strong association with poor outcomes, which have remained unclear so far. Therefore, to fill these gaps, the authors aimed to conduct the present study.

This umbrella meta-analysis study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for systematic reviews and meta-analyses 34 and Joanna Briggs Institute (JBI) approach for performing umbrella meta-analysis study 33 . The protocol of this study was registered in PROSPERO with the registration number CRD42021277992.

PubMed, Scopus, web of science, Cochrane Database of Systematic Reviews, ProQuest, Springer, Sage journal, and the internal database such as Magiran, SID, Irandoc, and elmnet were searched for systematic reviews and meta-analysis studies on cardiac complications following COVID-19 infection from December 2019 to April 2021. Google and Google Scholar were used as a searching engines as well. Furthermore, the reference lists of included articles and gray literature were screened to find more relevant studies. Our syntax was formed with Mesh terms of COVID-19, "2019-nCoV Diseases", "SARS-CoV-2 Infections", coronavirus, cardiac, cardiovascular, myocardial, myocarditis, myocardium, heart, hypotension, arrhythmia, hypertension, "blood pressure", "systematic review", meta*, pooled effect, and pooled estimate using operators of "OR", "AND", "NOT", and "*". A sample of search syntax for Scopus was presented in Supplementary Table 1 

We included articles published as a systematic meta-analysis study within peer-reviewed journals by full text in English or Persian languages from December 2019 to April 2021.

Articles must also have a clear literature search strategy, a standard quality assessment tool, and report any cardiac complications after COVID-19 infection using pooled estimated approaches (incidence rate, odds ratio (OR), risk ratio (RR)) in adult patients (≥ 18 years).

The meta-analysis studies with unnecessary data, narrative reviews, or studies in specific groups such as children or neonates/pregnant were excluded.

After developing a data extraction form using the JBI data extraction tool for systematic reviews and research synthesis, 33 primary data were extracted by two authors. The developed data form included items such as the objectives, types of the review articles, year of publication, names of searched databases, the timeframe of literature search, country of studies, sample characteristics, and key findings regarding cardiac complications in COVID-19 patients.

To summarize data, quantitative meta-analysis approaches were used. For the effect sizes (incidence rate or RR/OR) were reported in more than one study, the summarized pooled was calculated using a random/fixed-effects model regarding the level of heterogeneity. The Higgin's I 2 -indices of 0-25%, 26-75%, and 75-100% indicate low, moderate, and high between studies heterogeneity levels, respectively 35 . We used both the fixed-effect model and DerSimonian-Laird random-effects model for homogeneous (I < 50% and P > 0.05) and heterogeneous (I ≥ 50% or P ≤0.05) data, respectively 36 . The DerSimonian-Laird randomeffects model considers both within-study and between-studies variations 37 . According to the number of included studies, we assessed publication bias with Egger's test for AMI incidence. Assessment of publication bias of the rest pooled estimates was not possible because of an insufficient number of studies, since at least 10 studies are required to assess publication bias 38 . A sensitivity analysis with the leave-one-out method was implemented to evaluate the influence of one single study effect on overall pooled estimation. The sub-group analysis was based on the severity of the disease that included the patients with severe disease (ICU admitted, critically ill, or under ventilator patients), the non-severe group (nonadmitted to the ICU, non-critically ill, or non-under ventilator patients), and the deceased patients (deceased patients) accordingly. We also evaluated the primary studies for overlapping or duplicate data, as recommended 39 . The primary studies in the included metaanalysis studies were reviewed for duplicate data, then any existing overlapping data concerning estimated incidence rates was excluded. To run the forest plot, after removing all duplicate data, we first selected a study with larger or more accessible data and then added other non-duplicate study data to get the summarized effect sizes. Stata version 16.0 software (Stata Corp, TX USA) was used for the quantitative analyses. The p-values for statistical significance was set at 0.05.

The Assessment of Multiple Systematic Reviews (AMSTAR) tool was used to assess the methodological quality of the included studies 40 . This 11-item tool is scored as 8-11, 4-7, 

Although no ethical approval is required for conducting a review study, our institute required an ethical code for review studies in Covid-19 related studies. Therefore, to conduct this study, the ethical code IR.BMSU.REC.1399.146 was obtained from Baqiyatallah University of Medical Sciences (BMSU).

The literature search provided 810 reviews, 40 of which were eligible for full-text review.

After reviewing the full text, 26 meta-analysis studies were eligible for inclusion [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] .

The study selection process is shown in Fig. 1 . 

In all included meta-analysis articles, overall, 916 observational studies with more than 180,148 COVID-19 patients were analyzed. The sample sizes of the two studies by Pellicori et al. 27 and Momtazmanesh et al. 43 Table 2 . Table 2 . Basic characteristics of included systematic reviews and meta-analysis studies

All included studies had moderate to high methodological quality. Egger's test discovered no publication bias in the reported pooled incidences of AMI (P = 0.260), ( Supplementary Fig.   2 ). The sensitivity analysis showed no significant single study effect for the total summarized pooled incidence of ACS, AMI, cardiac arrest, myocarditis, arrhythmia, shock, elevated CK, CK-MB, and CTnI, and also for total summarized pooled OR/RR. Supplementary Figure 3 illustrates an example of the leave-one-out analysis for incidence of AMI.

A wide range of cardiac complications were identified in COVID-19 patients as follows: arrhythmia 20, 23, 27, 43, 44, 47, [49] [50] [51] HF 20, 27, 30, [43] [44] [45] 51 , pericardial effusion 20, 30, 43 24, 42 were reported in more than one study, we were able to estimate their summarized pooled incidence. In 11 studies, the pooled incidence of AMI was ranged from 15% (95% CI = 11-20%) 29 to 24.4 (95% CI =21.4-27.4) 28 . To calculate the overall AMI incidence in COVID-19 patients, two studies were excluded since their 95% CIs were not available 24, 46 . Also, after removing duplicate data, the summarized pooled incidence of AMI was estimated as 21% (95% CI = 15 -26, I 2 = 80.33). In addition, the summarized pooled incidences of ACS, cardiac arrest, HF, arrhythmia and shock were estimated as 1.3% (95% CI = 0-2, I 2 = 55), 3 .46% (95% CI = 3.1-3.83, I 2 = 35.70), 14% (95% CI = 0-29, I 2 = 86.66), 16% (95% CI = 9-22, I 2 = 93.99), 6.1 (95% CI = -3 -15, I 2 = 96.21), respectively. Further, the summarized pooled incidences of the raised CK 20 43 , CK-MB 20,43 , cardiac Troponin I (CTnI) 20, 29, 30, 43 with their I 2 index are indicated in Table 3 . The forest plots of these estimations by considering the duplicate data are shown in Supplementary Fig. 2 . The incidences of cardiac arrest and the elevated CK were run with a fixed-effect model because their heterogeneity level was lower than 50%, and the others were done with a random effect model. In table 3, we present the summarized incidences with two forms of with and without duplicate data. We observed that after we removed the duplicate data, the summarised pooled effect size (incidence rate, OR, RR) mostly decreased and their 95% CI widened.

Among cardiac complications, the pooled incidences of AMI, shock, and HF were reported in the severe and deceased COVID-19 patients. Among patients with severe disease, the summarized incidences of AMI 29, 52 , and shock 22 patients, so we were unable to estimate their summarized pooled measures ( Table 4 ). The forest plot of each summarized estimate is shown in Supplementary Fig. 6 . Supplementary Fig. 7 . Table 5 . Risk or odds of poor outcomes in COVID-19 patients with AMI/without AMI

This study is the first umbrella review that synthesized the current meta-analysis studies about the incidence of cardiac complications following COVID-19 infection. We estimated the summarized pooled incidence of cardiac complications in the sub-groups of overall, severe, and deceased COVID-19 patients. Our findings demonstrated that the deceased patients or patients with severe COVID-19 disease have a higher incidence of cardiac complications compared to the non-severe groups. Similarly, the literature has shown that a higher incidence of newly developed cardiac complications was significantly associated with poor prognosis and mortality 53, 54 . Although the risk of cardiac complications is high in all patients with COVID-19 such as severe and non-severe ones, however, critically ill patients have a much higher incidence of these complications 12 .

Our findings showed that in the overall COVID-19 patients, the incidence rate of cardiomyopathy, shock, AMI, HF, CA, myocarditis, MI, and pericardial effusion ranged from the highest (33.33 %) to the lowest (2.62), respectively. The cardiac injury was marked with the raised cardiac blood markers of NT-pro BNP, CK-MB, cTnI, and CK, and manifested mainly with chest pain/tightness and palpitation, respectively. Different studies revealed that increased cardiac blood markers are indicators of cardiac damage during COVID-19 infection 55, 56 . Evaluation of dynamic changes in the cardiac biomarkers shows that these indicators of heart damage increase abnormally from the middle of hospitalization and peak immediately before the patient's death 57 . Cardiac biomarkers can be systematically measured over hospitalization to assist in the early detection of cardiac complications, which in turn can reduce mortality with preliminary intervention.

Another finding of this study was the incidence of other organ injuries such as liver, kidney, or lung in COVID-19 patients who suffered from cardiac complications. Due to the localization of angiotensin converting enzyme II (ACE2) protein in the human organs such as lung tissue, liver, kidney, and digestive system, COVID-19 simultaneously damages all organs by direct binding to ACE2 or by triggering a cytokine storm which results in cytokine release syndrome (CRS) 58 . Cytokine storm is characterized by the secretion of proinflammatory mediators, such as IL-6, and an invasion of monocytes and macrophages that can cause multi-organ damage 59 . CRS response has been also reported in patients infected with SARS-CoV and MERS-CoV 16 60 .

Our summarized meta-analysis demonstrated that COVID-19 patients with cardiac complications were several times more likely to develop severe disease, or need to be admitted to the intensive care unit, or die than those without it. In accordance with our findings, a study showed that COVID-19 patients with cardiac damage are more likely to develop a severe form of disease or to be transferred to the intensive care unit, or to die 52 . In addition, the researchers showed that there was a significant increase in TnI in patients with severe disease compared to non-severe individuals 11, 61 . Further, one study showed that from the tenth day of hospitalization, a rapid increase in troponin levels was reported in the deceased patients, which was not the same as in the survivors 5 accelerates the conversion of angiotensin II to angiotensin 1-7, which has a protective effect on the cardiac system. Subsequently, by binding to ACE2, viruses block the release of angiotensin 1-7, in this way, they cause myocardial damage. Viruses also damage the cardiac system directly by binding to ACE2 and initiating the inflammatory response process 68, 69 .

Among our reported effect sizes, mostly overlapping data were found. After removing duplicate data, most effect sizes were reduced slightly and their 95% confidence intervals increased. In an umbrella meta-analysis study, the use of initial study data repeatedly reduces the 95% confidence interval and increases the effect size 39 , which also was reflected in our study.

Our study has limitations that must be disclosed to readers. Since, compared to other countries, most of the initial studies were mainly from China, therefore, the findings of the present study should be interpreted with caution. In addition, our study did not examine the prevalence of cardiac diseases and was limited to their incidence in COVID-19 patients.

While the prevalence of cardiac disease in COVID-19 patients can be higher than its incidence. Prevalence can also show the importance of the issue more seriously. We were unable to estimate a summarized pooled incidence for some cardiac complications, such as pericardial effusion, cardiac insufficiency, cardiomyopathy, and MI, therefore, additional umbrella meta-analysis studies are required with new upcoming review reports on these cardiac complications.

Based on the mentioned limitations, we suggest some recommendations for future research.

The exact mechanism of death in COVID-19 patients with cardiac complications must be precisely determined by answering the question of whether the death of COVID-19 patients is solely due to cardiac complications or whether they die from other causes such as multiple organ failure?. Additionally, in addition to measuring the incidence, measuring the prevalence of cardiac disease in COVID-19 patients can provide more important information about the number of COVID-19 patients who have previously had a cardiac disease and the number of patients who have recently developed cardiac complications. Further, the incidence rates of some cardiac complications, such as cardiomyopathy and cardiac insufficiency, have been less studied, and by more assessing them, a wider range of data regarding cardiac complications will be available. Our study also has an important clinical recommendation, which is that regular cardiac examinations and check-ups should be carried out daily on COVID-19 patients. Because as our findings showed, a high incidence of diverse cardiac complications have prevailed among COVID-19 patients and both the incidence rate and the poor outcomes increase with the progression of the disease.

Our study demonstrated widespread newly developed cardiac complications following COVID-19 infection. Cardiac complications were associated with increased disease severity and mortality. The main signs and symptoms of cardiac injury in COVID-19 patients were chest pain, palpitation, and raised cardiac biomarkers, respectively. Although a wide incidence of cardiac complications has been reported in COVID-19 patients, however, the incidence rates of HF, shock, AMI, arrhythmia, cardiomyopathy, and cardiac insufficiency were more than others. Also, the incidence rates of cardiac complications, such as AMI, HF, and arrhythmia in severe and deceased cases were several times of non-severe non-deceased cases. Furthermore, the severity of disease, ICU admission, mortality, and other organ injuries in the COVID-19 patients with cardiac injuries were several-fold higher than the 

This work was supported by Baqiyatallah University of Medical Sciences (BUMS).

Reference Table 3 . The pooled incidence of cardiac complication in total, non-severe, severe, and deceased COVID-19 patients 

Organ-specific manifestations of COVID-19 infection

Cardiac manifestations in COVID-19 patients-A systematic review

Clinical Characteristics of 138 Hospitalized Patients with 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China

Potential Effects of Coronaviruses on the Cardiovascular System: A Review

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Prevalence and impact of cardiovascular metabolic diseases on COVID-19 in China

Clinical characteristics of 113 deceased patients with coronavirus disease 2019: Retrospective study

Association of Cardiac Injury with Mortality in Hospitalized Patients with COVID-19 in Wuhan, China

Vander Heide RS. Cardiac Endotheliitis and Multisystem Inflammatory Syndrome After COVID-19

Myocardial injury and COVID-19: Possible mechanisms

COVID-19 and cardiac injury: clinical manifestations, biomarkers, mechanisms, diagnosis, treatment, and follow up

Acute Cardiac Injury in Coronavirus Disease 2019 and Other Viral Infections-A Systematic Review and Meta-Analysis

The Variety of Cardiovascular Presentations of COVID-19

Myocardial localization of coronavirus in COVID-19 cardiogenic shock

Successfully Treated With Temporary Mechanical Circulatory Support

First case of COVID-19 complicated with fulminant myocarditis: a case report and insights

Clinical characteristics of hospitalized patients with SARS-CoV-2 infection: A single arm meta-analysis

Early detection of elevated cardiac biomarkers to optimise risk stratification in patients with COVID-19

Inflammation, Immunity, and Infection in Atherothrombosis: JACC Review Topic of the Week

Incidence of Myocardial Injury in COVID-19-Infected Patients: A Systematic Review and Meta-Analysis

Clinical characteristics of 2,459 severe or critically ill COVID-19 patients: A meta-analysis

Clinical characteristics of COVID-19 with cardiac injury: A systematic review and meta-analysis

Clinical features and outcomes of adults with coronavirus disease 2019: A systematic review and pooled analysis of the literature

Clinical features of severe patients infected with 2019 novel coronavirus: a systematic review and meta-analysis

Meta-analysis investigating the relationship between clinical features, outcomes, and severity of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia

COVID-19 and its cardiovascular effects: a systematic review of prevalence studies

Cardiac Injury and COVID-19: A Systematic Review and Metaanalysis

Acute cardiac injury in COVID-19: A systematic review and meta-analysis

Cardiac complications and pertaining mortality rate in COVID-19 patients; a systematic review and metaanalysis

COVID-19 Related Cardiovascular Comorbidities and Complications in Critically Ill Patients: A Systematic Review and Meta-analysis

A meta-analysis on the role of pre-existing chronic disease in the cardiac complications of SARS-CoV-2 infection. iScience

Summarizing systematic reviews: Methodological development, conduct and reporting of an umbrella review approach

The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration

Assessing heterogeneity in metaanalysis: Q statistic or I 2 Index? Psychol Methods

Quantifying heterogeneity in a meta-analysis

Meta-analysis in clinical trials

How many studies do you need? A primer on statistical power for meta-analysis

Managing overlap of primary study results across systematic reviews: practical considerations for authors of overviews of reviews

Development of AMSTAR: A measurement tool to assess the methodological quality of systematic reviews

Clinical characteristics, laboratory outcome characteristics, comorbidities, and complications of related COVID-19 deceased: a systematic review and metaanalysis

Clinical features, comorbidities, complications and treatment options in severe and non-severe COVID-19 patients: A systemic review and meta-analysis

Malakan Rad E. Cardiovascular disease in COVID-19: a systematic review and meta-analysis of 10,898 patients and proposal of a triage risk stratification tool

Cardiovascular complications of SARS-CoV-2 infection (COVID-19): a systematic review and meta-analysis

Coronavirus disease 2019 (COVID-19) and cardiovascular risk: A meta-analysis

Incidence and mortality risk in coronavirus disease 2019 patients complicated by acute cardiac injury: systematic review and metaanalysis

Arrhythmia in patients with severe coronavirus disease (COVID-19): A meta-analysis

Arrhythmia risk profile and ventricular repolarization indices in COVID-19 patients: A systematic review and meta-analysis

Incidence and impact of cardiac arrhythmias in coronavirus disease 2019 (COVID-19): A systematic review and meta-analysis

Incidence rate and clinical impacts of arrhythmia following COVID-19: a systematic review and meta-analysis of 17,435 patients

Cardiovascular complications in COVID-19: A systematic review and meta-analysis

Prevalence and clinical outcomes of cardiac injury in patients with COVID-19: A systematic review and meta-analysis

Clinical Characteristics of 138 Hospitalized Patients with 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Cardiac troponin I in patients with coronavirus disease 2019 (COVID-19): Evidence from a meta-analysis

Life-threatening cardiac tamponade complicating myopericarditis in COVID-19

The impact of 2019 novel coronavirus on heart injury: A Systematic review and Meta-analysis

Protein-Driven Mechanism of Multiorgan Damage in COVID-19. Med Drug Discov

Cytokine release syndrome in severe COVID-19. Science (80-)

Extrapulmonary manifestations of COVID-19

Analysis of heart injury laboratory parameters in 273 COVID-19 patients in one hospital in Wuhan

Cardiovascular Implications of Fatal Outcomes of Patients with Coronavirus Disease 2019 (COVID-19)

The cardiovascular manifestations of influenza: A systematic review

Cardiovascular complications of severe acute respiratory syndrome

An experimental model for dilated cardiomyopathy after rabbit coronavirus infection

Clinical aspects and outcomes of 70 patients with Middle East respiratory syndrome coronavirus infection: A single-center experience in Saudi Arabia

Role of the ACE2/angiotensin 1-7 axis of the reninangiotensin system in heart failure

Angiotensin-converting enzyme 2 and angiotensin 1-7: Novel therapeutic targets

Conflict of interest: The authors have no conflicts of interest to disclose

Abbreviations: ACS: acute coronary syndrome; AMI: acute myocardial injury.*troponin I above the 99th percentile upper reference limit(>28 pg/ml), or new abnormalities in electrocardiography and echocardiography; **Elevated CK (upper limit of normal of 170,200,185,310, U/L); ***Elevated Creatine kinase-MB(CK-MB) ((more than 5,18,24 ng/ml); ****Elevated cardiac Troponin I (CTnI) (more than 28,40,15.6 ng/L), *defined as a pneumonia condition with a respiratory rate more or equal than 30 times/min or oxygen saturation at resting state less or equal than 93% or partial pressure of arterial oxygen to fraction of inspired oxygen ratio less than 300 18