Journal of Diabetes & Clinical Practice
Open Access

COVID-19; Hyperthyroidism; Hypothyroidism; Outcome; Thyroid disease

Introduction

The growing number of cases and the rapidly evolving SARSCoV- 2 has caused coronavirus disease 2019 (COVID-19) to become a global burden that affected people from different groups of age. As the spectrum of signs and symptoms among COVID-19-infected individuals is broad–ranging from mild to severe illness requiring intensive care, identifying risk factors that may predict the patients’ outcome is imperative for a better resource allocation in mitigating this current COVID-19 outbreak [1].

The interaction between thyroid hormone and immune regulation has been explored extensively in both physiological and pathological settings. As the regulation of the immune system plays an integral part in determining COVID-19 patients’ disease progressivity, evidence is required to accentuate the correlation between thyroid hormone and COVID-19 outcomes. In light of this, several studies have been carried out to emphasize the relationship between thyroid status or thyroid abnormalities and the outcome of COVID-19. However, the correlation between the two remains elusive as indicated by the variability of reported study results. Thus, in this present systematic review and meta-analysis [2], we determined to summarize recent findings on the relationship between thyroid disease and COVID-19 outcome.

Material and Methods

This systematic review was presented consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.

Eligibility Criteria

Inclusion and exclusion criteria were defined according to the PECO structure. The population of this study was COVID-19 patients confirmed with polymerase chain reaction (PCR), serum antibody or antigen testing, or highly suspected cases based on clinical presentation.

Exposure and comparator were defined as patients with known thyroid disorders, hypothyroidism defined as measured thyroid levels below normal range during admission, and hyperthyroidism defined as measured thyroid levels above normal reference during admission [3].

The inclusion criteria were research articles and letters that report COVID-19 patients with information on thyroid disorders that were presented as categorical data along with measurable composite poor outcomes–mortality/severity/critically ill/intensive care unit (ICU) admission/patients required hospitalization.

The following types of articles were excluded: case report, nonresearch letter, editorial, invited commentary, review, and abstractonly article [4]. Also, studies that reported only continuous variables of thyroid hormone level were excluded. There was no language restriction applied in this study.

The primary exposure was thyroid abnormalities preexisting thyroid disease from baseline characteristics or a group of patients with lower- or higher-than-normal measured thyroid hormone levels during admission. The type of thyroid measurement during admission to categorize the patient's type of thyroid level abnormality was thyroid-stimulating hormone (TSH), free triiodothyronine (fT3), or free thyroxine (fT4). The patient was considered hypothyroidism if the level of thyroid hormone was lower than normal reference according to the specified cut-off defined by each study. Conversely, the patient was considered hyperthyroidism if the level of measured thyroid hormone was higher than the normal range [5]. Furthermore, we also performed subgroup analyses on each type of thyroid abnormalities–undefined, hypothyroidism, and hyperthyroidism.

We used Newcastle-Ottawa Scale (NOS) to assess the quality of included observational cohort studies. Assessment of the included studies was done independently by four authors. The following aspects were taken into consideration in the assessment: cohort selection, the comparability of cohort-based design or analysis, the way exposure is determined, and the way of outcomes of interest is evaluated. Discrepancies of perception in determining study quality were resolved by discussion [6].

Statistical Analysis

All statistical analysis was performed using R. We used the Mantel- Haenszel method formula to generate pooled effect estimates in a form of risk ratio (RR) along with their 95% confidence interval. The inconsistency index and subgroup analysis using the Chi-square test was used to investigate potential sources of heterogeneity. To consider inter study variability, a random-effects model was used regardless of study heterogeneity.

Subgroup analyses were performed only when at least two studies on each at least two subsets were available. Subgroup analysis was carried out on each component of composite poor outcome and the type of thyroid abnormalities specified in the studies [7]. Further, we also conducted a restricted-maximum likelihood random-effects metaregression analysis to study the influence of the following covariates– gender, age, hypertension, coronary artery disease/cardiovascular disease (CAD/CVD), diabetes mellitus (DM), obesity, and chronic obstructive pulmonary disease (COPD).

Analyses of publication bias were presented qualitatively using a contour-enhanced funnel plot and quantitatively assessed using Egger's linear regression test to indicate small-study effects. Sensitivity analysis was performed under the leave-one-out method to single out the cause of study heterogeneity and statistical significance. When publication bias was detected, trim-and-fill analysis was carried out to evaluate the potential sources of publication bias [8].

Results

Study Selection

The initial search pooled 459 articles from PubMed and Cochran Collaboration Central Register of Controlled Clinical Trials. Among 459 studies, two articles were removed automatically due to duplication, yielding 457 articles that were eligible to be screened and sought for retrieval. From all 457 screened articles, we assessed 30 studies for each individual's eligibility.

Discussion

The outcomes of COVID-19 patients worldwide are diverse and can be caused by any comorbidities, one of which is thyroid abnormalities. This systematic review presented a summary of hospital-based data, describing the comparison of primary outcomes, including mortality rate, ICU admission, severity, and hospitalization among patients with euthyroid and thyroid abnormalities. In this present study, the presence of thyroid abnomalities– hypothyroidism, hyperthyroidism, or unspecified thyroid disorder– increased the risk of COVID-19 composite poor outcomes that consist of disease severity, ICU admission, mortality, and hospitalization. Contrary to the cumulative result, Zhang et al. and Lui et al found that the presence of thyroid disease was not associated with more severe COVID-19. In addition, Lui et al. also showed a similar result of the incidence of thyroid abnormalities between mild, moderate, and severe COVID-19 patients. The authors suggested that the result may be affected by the level of SARS-CoV-2 viral load that ought to be confirmed through qRT-PCR. Muller et al. and Gong et al. in their study did not demonstrate a positive and significant association between abnormal thyroid and ICU admission and mortality [9].

Furthermore, meta-regression analyses showed that the correlation between thyroid abnormalities and poor outcomes was affected by the increasing age, not due to other comorbidities. This association because the levels of TSH and thyroxine will decrease gradually and deteriorated physiological response that occurs with increasing age. In addition, the iodine nutritional status of the populations may result in different thyroid function patterns during aging.

Although the interplay between thyroid hormones and immune response has been largely investigated, the exact mechanism involved in the setting of COVID-19 infection remains elusive. Like other viral infections, the immune response has an important role in the outcome of SARS-CoV-2 infection. Previous hypotheses have postulated that thyroid hormones play a role in modulating the cellular level of both innate and adaptive immune responses. The thyroid hormone demonstrates its effects on the immune system at both nuclear levels by activating transcription factors that are responsible for intracellular signaling and at cellular levels by modulating cytokine release on multiple cells when the innate immune response is taking place. Several studies have shown that in the acute phase of critically ill patients, the concentration of T3 was markedly decreased. This abnormal release of thyroid hormone affects the robustness of the immune response that is integral in maintaining COVID-19 outcomes. Moreover, during the process of cytokine storm, the body overproduces cytokines and chemokines triggered by SARS-CoV-2 infection. If this period is prolonged, suppression of the thyroid axis can occur in critically ill patients. It can lead to a worsening of outcomes in patients with thyroid abnormalities [10]. In addition, thyroid hormones also influence other metabolic diseases such as hypertension, diabetes, and cardiovascular disease–those comorbidities are strongly related to an increased risk of poor COVID-19 composite poor outcomes.

Our results showed abnormal thyroid and hypothyroidism were significantly associated with a higher risk of composite COVID-19 poor outcomes. In contrast, hyperthyroidism patients did not pose different risks in COVID-19 poor outcomes. This discrepancy between these findings can be potentially affected by several factors. It has been emphasized that thyroid hormone affects the renin-angiotensinogen system (RAS). Thus, patients with thyroid abnormalities might have dysregulated RAS which contributes to increased angiotensinconverting enzyme 2 expression. ACE2 plays an essential role in SARSCoV- 2 infection as the key receptor of the virus to enter the host. The relationship between thyroid hormones and immune response may be bidirectional. However, cautious considerations should be considered when interpreting the relationship between thyroid hormones and COVID-19 outcomes. First, the effect of excessive thyroid hormone in hyperthyroidism patients may not significantly alter the physiologic regulation of RAS. Second, the dysregulated RAS caused by hyperthyroidism may also not significantly influence the expression of ACE2–yielding to a similar amount of viral entry and disease outcome. Third, the association between hyperthyroidism and poor COVID-19 outcome may be affected by multiple factors. This was demonstrated in the study by Brix et al. that shows a significantly higher risk of poor COVID-19 outcome in hyperthyroidism patients with crude analysis but posed a similar risk ratio of poor outcomes in propensity score weighted analysis. Lastly, a relatively smaller sample and studies involving hyperthyroidism patients may also be one of the possible causes of these findings.

This present study supports the previous meta-analyses exploring the relationship between thyroid abnormalities with COVID-19 poor outcomes. Hariyanto et al. found in their meta-analysis that the presence of thyroid abnormalities was associated with severe COVID-19 outcomes. However, the study only included eight studies with unspecified thyroid abnormalities. On the one hand, this current study included more studies yielding a relatively higher sample, subgroup analyses based on disease outcomes and thyroid abnormalities, and further meta-regression which was not previously conducted. Ultimately, the evidence of the association between thyroid abnormalities and COVID-19 poor outcomes is strengthened by our present results.

Although this present analysis addresses the association between thyroid hormones and COVID-19 outcomes in a larger population, this study has several limitations. Studies that were included in our analyses were largely conducted in a retrospective fashion that may impose a relatively greater potential risk of bias. The association between COVID 19 severity and the degree of thyroid dysfunction could not be assessed due to the paucity of reported studies on each subset. In addition, one cross-sectional study with a relatively larger sample was included that potentially affects cumulative effect estimates. It is important to note that iodine intake may vary across the country. Similarly, ethnicity may also take place to the varied iodine intake levels, especially in a large country. However, we did not perform subgroup analysis under the sociodemographic subset due to study paucity and underreported characteristics on each country or ethnic group of each study. Nevertheless, our study demonstrated a moderate-to-high quality of evidence illustrating those thyroid abnormalities and hypothyroidisms were associated with COVID-19 poor outcomes. Moreover, the generalizability of our findings can be ascertained as the populations included in the analyses were diverse.

Conclusion

Thyroid abnormalities are associated with a higher risk of disease severity, mortality, ICU admission, and hospitalization among COVID-19 patients. This association was significantly influenced by an increase in age. Also, the presence of thyroid disorder and hypothyroidism, but not hyperthyroidism, is associated with a higher risk of poor outcomes.

1. AwanMJ, Rahim MMS, Salim N, Mohammed MA, Garcia Zapirain B, et al.(2021)Efficient Detection of Knee AnteriorCruciate Ligament from Magnetic Resonance Imaging Using Deep LearningApproach.Diagnostics 11: 105.

Indexed at, Crossref, Google Scholar

2. Acharya UR, Oh SL, Hagiwara Y, Tan JH, Adam M, et al.(2017)A deep convolutional neural network modelto classify heartbeats.Comput Biol Med 89:389-396.

Indexed at, Crossref, Google Scholar

3. Dantas G, Siciliano B, França BB, da Silva CM, ArbillaG, et al. (2020)The impact of COVID-19 partial lockdownon the air quality of the city of Rio de Janeiro, Brazil.Sci Total Environ 729.

Indexed at, Crossref, Google Scholar

4. Lanciotti RS, Kerst AJ, Nasci RS, Godsey MS, MitchellCJ, et al (2000)Rapid detection of West Nile virus fromhuman clinical specimens, field-collected mosquitoes and avian samples bya TaqMan reverse transcriptase-PCR assay.J Clin Microbiol 38:4066-4071.

Indexed at, Crossref, Google Scholar

5. Pérez Ruiz M, Pedrosa Corral I, Sanbonmatsu Gámez S,Navarro Marí M (2012)Laboratory detection of respiratoryviruses by automated techniques.Open Virol J 6:151-159.

Indexed at, Crossref, Google Scholar

6. Cadogan A, McNair P, Laslett M, Hing W(2013)Shoulder pain in primary care: diagnosticaccuracy of clinical examination tests for non-traumatic acromioclavicularjoint pain.BMC Musculoskelet Disord 14: 156-166.

Indexed at, Crossref, Google Scholar

7. Li L (2020)Artificial Intelligence DistinguishesCOVID-19 from Community Acquired Pneumonia on Chest CT.Radiology.

Indexed at, Crossref, Google Scholar

8. Rahimzadeh M, Attar A (2020)A modified deep convolutional neuralnetwork for detecting COVID‐19 and pneumonia from chest X‐ray images based on the concatenation ofXception and ResNet50V2.IMU 19: 100360.

Indexed at, Crossref, Google Scholar

9. Bai HX, Hsieh B, Xiong Z, Halsey K, Choi JW, et al.(2020)Performance of radiologists indifferentiating COVID-19 from viral pneumonia on chest CT.Radiology.

Indexed at, Crossref, Google Scholar

10. Caruso D, Zerunian M, Polici M (2020)Chest CT features of COVID-19 in Rome,Italy.Radiology 2020: 201237.

Indexed at, Crossref, Google Scholar