Abstract

Introduction. COVID-19, a multisystem disease, has implications for various immunity and infection biomarkers. Yoga (Y), meditation (M), and pranayama (P), and their combinations have shown positive changes on those biomarkers among other than COVID-19 patients and healthy people. So, we aimed to document the evidence of possible implication in a systematic way. Materials and Methods. We screened 84 full texts, published in the last ten years, from three databases, from which only 44 met the eligibility criteria, and then extracted the data related to demographic characteristics, intervention, results, and strengths and limitations in two MS-Excel grids, and then presented them in tables and figures. Furthermore, we carried out meta-analysis including subgroup and sensitivity analysis using a random effects model of 11 RCTs and reported the mean difference, heterogeneity, and value with 95% CI and presented them with forest and funnel plots and the tables. Results. Twenty-five biomarkers of 4023 participants (range, 15–413) from 13 countries, healthy and clinical, from both sexes above 18 years, and from mainly clinical settings, were reported. YMP intervention, in solitary or in different possible combinations with varied durations among clinical and pregnant (range, 960–4800 minutes) and healthy (960–8400 minutes, excluding two studies of 20 minutes only) participants, was reported. It was revealed that 25 biomarkers, nine among the apparently healthy, 14 among the patients, and two among the pregnant, changed favourably (). Furthermore, either in meta- or subgroup-analysis, mean differences of IL-6 (−1.44 pg/ml) (95% CI) (−2.33, −0.55), (, I2 = 82%), Cortisol (−40.75 pg/ml) (95% CI) (−64.13, −17.38), (, I2 = 87%), and TNF-α (−3.40 pg/ml) (95% CI) (−4.83, −1.98), (, I2 = 79%) showed statistically significant changes. Nonetheless, considerable heterogeneity and publication bias were observed among the studies. Conclusion. Although more than two dozens of biomarkers in individual studies showed favourable changes, only IL-6, Cortisol, and TNF-α produced significant combined results, even then with much less certainty. Further meta-analysis of biomarkers of COVID-19 patients is highly recommended. Registration: CRD42021283894.

1. Introduction

Yoga, meditation, and pranayama, or yogic breathing, have been practised worldwide since antiquity. Research, both here and elsewhere, has shown that practises like guided asanas, specific pranayama, and meditation can improve the body’s immune system responses. To develop a holistic framework for individuals’ health, the interrelationships between mind and body have been extensively studied in recent times [15]. With the advent of the pandemic, researchers and clinicians are in search of practises and therapeutics that can reduce the impact of COVID-19 on human health. In this line, recent reviews of complementary and alternative medicine (CAM) and traditional Chinese medicine (TCM) for COVID-19 have provided preliminary evidence of effectiveness [6, 7]. In addition, a recent review also underscored that the effects of natural compounds from Nigella sativa are promising [8]. It has been stated that as a result of COVID-19 disease, leukocytes, neutrophils, cytokine levels [IL-2R, IL-6, IL-8, and IL-10], and infection biomarkers like CRP, PCT, and ferritin are significantly increased [9, 10]. Multiple studies have shown that yoga, meditation, and pranayama interventions are efficacious in regulating an array of biomarkers, including cytokines [4, 5].

Further research in this area is ongoing since the disease is still uncertain in many aspects, especially regarding its long-term complications [10] and the outbreak of possible new mutants due to natural selection [11]. During the cytokine storm that occurs in COVID-19 syndrome, proinflammatory cytokines are released uncontrollably. A positive relationship between the severity of the viral infection and the mortality rate was observed following an uncontrolled and dysregulated secretion of inflammatory and proinflammatory cytokines [12]. Recent evidence also shows that systemic vasculitis and cytokine-mediated coagulation disorders act as the principal actors of multiorgan failure among COVID-19 patients with severe complications [10]. Assessing the effect of different biomarkers in COVID-19 can help in the early diagnosis of the disease, confirming and classifying the disease severity, identifying the high-risk cohort, framing intensive care unit (ICU) admission criteria, rationalising therapy, assessing response to therapies, predicting outcomes, and framing criteria for discharge from the ICU and/or the hospital [9]. So, it is the need of the hour to assess and identify measures that can control the dysregulated secretion of cytokines and other inflammatory mediators with possible implications for COVID-19. Till date, multiple studies have shown that yoga, meditation, and pranayama interventions are efficacious in regulating an array of biomarkers, including cytokines [4,5]. However, these interventions and their outcomes presented over different settings and durations have shown mixed results. Considering the inconclusive evidence present, the heterogeneity in outcomes, and the duration of interventions given in the previous studies, we sought to systematically review the evidence of the effects of yoga, meditation, and pranayama interventions on COVID-19 related inflammatory biomarkers by including only randomized controlled trials (RCTs).

2. Materials and Methods

We conducted this review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

2.1. Study Selection Criteria

A systematic literature search was carried out after registering the review protocol in the international prospective register of systematic reviews (PROSPERO Regd. No. CRD42021283894). Peer-reviewed RCTs published during the last 10 years (2010–2021) were included. We sought to identify studies that used any version of yoga, meditation, and/or pranayama, intervened as solitary or in combination, among healthy and clinically ill individuals in adults (≥18 years), including pregnant women, to assess the effectiveness of biomarkers related to or influenced by COVID-19. We were also guided by our earlier scoping review on yoga, immunity, and COVID-19 for study selection [13]. The search was conducted mainly in three databases: The Cochrane Library, PubMed, and Google Scholar. Additionally, we also carried out a citation search. Keywords and MeSH terms were informed by previous literature searches. Six sets of search terms were used including terms related to (i) yoga, (ii) meditation, (iii) pranayama, (iv) intervention, (v) biomarker, and (vi) controlled trial (S1_ Search Strategy with Keywords and MeSH).

2.2. Data Extraction and Analysis

Based on our previous scoping review of yoga, immunity, and COVID-19 [13], we developed and then extracted data using two standardized MS-Excel grids. The data extraction form were initially piloted (n = 2 studies) and then refined and finalised by all reviewers. The data were extracted by two reviewers (CA and ShS), and the third reviewer (KS) assessed when there was a discrepancy. Pre-post effectiveness of YMP on biomarkers was found using mean difference, difference in differences of means (∆1-∆2), and effect sizes (Cohen’s d). However, we only reported values along with the direction of effectiveness in interpreting the tests. The basic characteristics and detailed findings of the included studies are given in Tables 1 and 2, respectively.

For data synthesis, we developed an Excel grid for the main outcomes (biomarkers). The evidence was synthesised into a narrative form. A descriptive analysis of the characteristics of study populations was carried out using range values and proportions.

For meta-analysis, we calculated the mean difference when at least three studies consistently reported the primary biomarkers. We produced forest plots, funnel plots, and risk of bias (ROB) assessment charts using the Cochrane risk of bias tool, from RevMan 5.4.1. The effect size of the meta-analysis was estimated through random effects, considering the post-pre mean change in biomarker concentration and standard deviation (SD). Studies that reported standard mean error (SME) were converted to SD, and for those not specifying SME or SD, either of the two was considered. Following the Cochrane guideline, we calculated the pooled SD from the given parameters, including an additional value of the correlation coefficient, assuming 0.8. Regarding three-armed studies, we pooled the mean and SD from the two experimental groups for the outcome measure.

Subgroup analysis was performed post hoc by stratifying the studies based on intervention type such as yoga, meditation, pranayama, or yoga only, or meditation only, and intervention duration in weeks and minutes. For this, we divided IL-6 into six subgroups, cortisol into two, and TNF-a into three.

3. Results

3.1. Search Result

An initial search yielded 174 articles. After removing the duplicates and excluding the articles assessing the title and abstract, 87 articles were retrieved for full text. We removed 40 full-text articles that failed to meet the inclusion criteria, as well as 3 articles that were unavailable in full text. Finally, 44 articles were selected for the review (Figure 1).

3.2. Description of Included Studies

The study consisted of randomized control trials ranging from less than one week (including two studies which had an intervention of 20 minutes) to 28 weeks (including one study which had 12 months of intervention). The majority of the interventions ranged from 5 to 16 weeks (n = 35) [1448], either solo (Y or M or P) or in combinations (YM, YP, MP, and YMP). There were studies providing yoga (n = 2) (Y) [29, 44], pranayama (n = 2) (P) [3, 28], meditation (M) (n = 11) [1420, 4952], meditation and pranayama (n = 2) (MP) [21, 53], yoga and meditation (n = 4) (YM) [32, 33, 35, 36], yoga and pranayama (n = 8) (YP) [2224, 27, 30, 31, 54, 55], and yoga, meditation, and pranayama (n = 15) (YMP) [25, 26, 34, 3743,4548, 56] as interventions. Most of the studies had two groups (intervention and control), while 6 studies were reported to have 3-experimental arms. The studies were carried out in 13 countries, with the maximum number of studies coming from the USA (n = 19, 43%), followed by India (n = 13, 30%), Iran (n = 2, 5%), and one each in China, Thailand, Spain, Netherlands, Sweden, Republic of Korea, Brazil, Australia, Taiwan, and Portugal. The interventions given included Iyengar, Patanjali Raj, Hatha yoga, and other forms of yoga, mindfulness-based stress reduction (MBSR), and pranayama (Tables 1 and 2).

3.3. Study Quality

Of the 44, twenty-three RCTs were open-label, followed by 19 single-blinded and 2 double-blinded trials. Among the 37 full RCTs, one was a crossover design. The comparator arms included 23 controls or wait-list or attention or observational controls, 11 usual or standard care, and 10 were given placebo or active control. Placebo interventions included health education (HE) with or without supportive therapy, counselling, exercise, rehabilitation, relaxation, healthy living workshop (HLW), dietary intervention, physical exercise, sleep hygiene education (SHE), and stress reduction. Risk of bias (ROB) of 11 studies included in meta-analyses, three were found to be low, one medium, and seven with high risk (Table 3).

3.4. Participants

A total of 4023 participants were included in 44 trials, ranging from 15 to 413 participants. Most participants were patients, including those either from in-patient or out-patient clinics or hospital settings. Only a few studies (n = 4) were carried out in community and educational institutions. Almost all studies included both men and women aged 18 years or above, whereas some trials were exclusively conducted only on females (n = 13) [14, 20, 26, 27, 2931, 46, 48, 51, 52, 54, 56] (Tables 1 and 2).

3.5. Intervention

The trials encompass various forms of asanas, deep relaxation, and other techniques of yoga, meditation, and pranayama. The participants (intervention group only) learned the poses and techniques in the presence of an instructor, and then they were asked to practise the learning at home or in their free time. In only one study, participants learned through an online platform. Except for two studies [16, 19], the daily duration of intervention was less than an hour (n = 10) [3, 18, 2123, 25, 37, 41, 53], one to two hours (n = 25) [14, 20, 24, 2632, 35, 3840, 4249, 52, 55] and more than two hours. The intervention duration of the trial by Banasik and colleagues [31] was taken from the authors’ previous study [57] (n = 7) [1517, 33, 34, 36, 50].

3.6. Outcomes
3.6.1. Yoga (Y)

Two studies that included only yoga as an intervention reported soluble tumour necrosis factor receptor II (sTNF RII) (a cell surface receptor for the proinflammatory cytokine), C-reactive protein (CRP), interleukin 1 receptor antagonist (IL-1RA) [29], interleukin 2 (IL-2) [44], cortisol [29], and Interleukin 6 (IL-6) [29, 44]. The duration of intervention ranged between 1440 and 2160 minutes.

3.6.2. Pranayama (P)

Two studies [3, 28] used pranayama as an intervention with a duration ranging from 20 to 720 minutes. The following biomarkers were studied: IL-6, CRP, 6 minute walk distance (6MWD), diffusing capacity of the lungs for carbon monoxide (DLCO), forced expiratory volume in 1 sec (FEV1), inspiratory capacity (IC), residual volume to total lung capacity ratio (RV/TLC), alveolar volume to total lung capacity ratio (VA/TLC), and inspiratory time to total breathing cycle time ratio (Ti/Ttot) [28]. A second study reported monocyte chemoattractant protein 1 (MCP-1), interleukin 8 (IL-8), interleukin 1 beta (IL-1β), IL-1RA, IL-6, interleukin 10 (IL-10), interleukin 17 (IL-17), interferonγ induced protein-10 (IP-10), macrophage inflammatory protein 1-beta (MIP-1b), and tumor necrosis factor (TNF-α) [3].

3.6.3. Meditation (M)

A total of 11 studies [1420, 4952] had only meditation as the intervention. The duration of the intervention ranged from 20 minutes to 12 months. The following biomarkers were reported in the trial: CRP [14, 15, 49], IL-1 beta, TNF-α [49], interferon gamma (IFN- γ) [16, 49], IL-8 [17], sTNFR II [14], cortisol [1820], IL-6 [14, 15, 18, 4952], immunoglobulin A (IgA) [16], IL-10 [16, 5052], high-sensitive C-reactive protein (hsCRP) [5052], and IP-10 [15].

3.6.4. Meditation and Pranayama (MP)

Two studies had a combination of meditation and pranayama [21, 53], ranging from a duration of 336 to 1200 minutes. IL-6, TNF-α [53], and hsCRP [21] were studied in these trials.

3.6.5. Yoga and Meditation (YM)

With four studies having both yoga and meditation [32, 33, 35, 36] intervention, ranging over a duration of 720 to 1740 minutes, including both learning and practising. Biomarkers included in the study were hsCRP [32, 33], IL-6 [32, 33, 35], extracellular superoxide dismutase (EC-SOD/SOD) [32], CD4+ [33, 36], and TNF-α [35].

3.6.6. Yoga and Pranayama (YP)

Eight studies were spanned from 1 hour to 28 weeks of yoga and pranayama intervention [2224, 27, 30,31, 54, 55]. The following markers were studied: IL-6 [23, 24, 27], hsCRP [2224], cortisol [22, 30, 31, 54], TNF-α [23, 24, 27, 55], TBARS and EC-SOD/SOD [23], IL-1 beta [27], IL-8, IL-10, IL-2, IL-12, IFN-γ, IL-4, tidal volume (TV), forced vital capacity (FVC), and FEV1 and peak expiratory force (PEF) [55].

3.6.7. Yoga, Meditation, and Pranayama (YMP)

A total of fifteen studies had a combination of three forms of intervention [25, 26, 34, 3743, 4548, 56]. The following biomarkers were found studied in the trial: IL-1β and IL-10 [37], IL-1β [26], CRP [3841], IL-17A [40, 42], IL-6 levels [25, 26, 3941, 43, 45, 46], TNF-α [25, 26, 39, 41, 42, 45, 47], cortisol [25, 47, 48, 56], TBARS, SOD [45, 47], IFN-γ [47], hsCRP [34, 43, 46], IgA [56], IL-1α [25], IL-8, and MCP-1 [26] (Tables 1 and 2).

3.6.8. Meta-Analysis

For meta-analysis, eleven studies [14, 18, 19, 23, 25, 29, 39, 40, 42, 44, 52] only focused on three biomarkers among the patients met the inclusion criteria, so we calculated the mean differences. The mean differences of IL-6, TNF-α, and cortisol were −0.58 pg/mL (95% CI (−1.37, 0.17)) [14, 23, 25, 29, 39, 40, 42, 44, 52], −2.62 pg/mL (95% CI (−4.29, −0.96)) [23, 25, 39, 40, 42], and −26.71 ng/ml (95% CI (−59.41, 5.99)) [18, 19, 25], respectively. Considerable heterogeneity was observed between the studies (IL-6 (I2 = 95%, ), TNF-α (I2 = 93%, ), and cortisol (I2 = 98%, ) (Figure 2).

3.6.9. Subgroup Analysis

In subgroup analysis (Table 4), yoga-pranayama-meditation (), 6–12-week (1000–2000 min) intervention (), and 8–12-week (3000–4800 min) were found with an overall effect significant in IL-6, cortisol, and TNF-α, respectively. In the meditation subgroup of IL-6, a marginal overall effect () with no heterogeneity (I2 = 0%) was observed.

3.6.10. Sensitivity Analysis

In a sensitivity analysis, we found a significant overall effect () by removing three studies of high risk of bias (C1.2) in IL-6. Similarly, in TNF- α, by removing one moderate risk study (B1.3) () and two high risk studies (C1.2) (p<.00001), overall effects remained significant (Table 5).

3.6.11. Publication Bias

Publication bias was assessed using funnel plots. The majority of the studies were found outside of the 95% CI as visualised, signifying a high publication bias [14, 23, 25, 29, 39, 40, 42, 44] (Figure 2).

4. Discussion

The review included 44 studies from 13 countries (including the USA, followed by India, which conducted the maximum RCTs), comprising 4023 people. The studies were conducted in different populations (healthy, diseased, and pregnant) ranging from 20 to 4800 minutes. Our main findings demonstrate that yoga, meditation, and pranayama, either alone or in combination, are effective in improving immunity in healthy and clinical populations (including pregnant women) by regulating anti- and pro-inflammatory biomarkers. Key findings (which were statistically significant) are (i) among clinical participants, there was a decrease in IL-6 [18, 23, 32, 40, 42, 53], IL-1β [31], IL 17/17A [40, 42], IL- 1α [25], CRP [39, 41], hsCRP [32], TNF/TNF- α [23, 25, 35, 40, 42, 46, 53, 55], sTNFR II [29], cortisol [18, 25, 31, 53, 56], TBARS [23]; and an increase in CD4+ T cells (nonsignificant) [33, 36], IL-2 [44], FEV1, PEF, VT, and FVC [55]; (ii) among healthy participants, an increase in IL-10 [37], IFN-γ, cortisol [47], and IgA [51]; and a decrease in IL-6 [24, 45], CRP [15], TNF-α [24, 26], hsCRP [24], EC-SOD/SOD [32], TBARS [45, 47], MCP-1 [49], and IL-1β [26, 37] was observed (iii) among the pregnant women, the intervention was effective in increasing IgA and decreasing cortisol levels [56] (Figure 3).

Many studies have reported that fluctuations in biomarkers among COVID-19 patients play vital roles in immunity [5867] (Table 6).

In a case report, a 55-year-old COVID-19 positive with comorbidities had his COVID-19 symptoms alleviated as well as his blood sugar level improved after treatment with ayurvedic medicine, yoga, dietary suggestions, lifestyle modifications, and allopathic medications [68]. Another ongoing study can strengthen the importance of yoga intervention in improving the health of COVID-19 patients [69].

Many studies have explored various immunity biomarkers in COVID-19 patients. T cells, an important component of lymphocytes responsible for robust immunity, are reduced significantly during severe infections, including COVID-19 [58, 59]. Cytokines (IL-6, IL-10, IL-8, IL-2R, and so on) consist of various biomarkers produced by both innate and adaptive immune cells [70], which play a vital role in inflammation [71]. A fatal cytokine storm is usually observed in COVID-19 patients who are critically ill [72]. Cytokine storms lead to damaged tissues, resulting in thrombotic tendency and multiple-organ failure [73], suggesting cytokine control necessary. Both individual studies and systematic reviews indicate the importance of IL-6, reporting it to be an independent predictor of disease severity and survival in COVID-19 patients [7477]. CRP, a prominent marker of systemic inflammation, was elevated in the majority of COVID-19 patients with severe illness compared with mild or nonsevere patients [7880]. Cortisol regulation acts as an adaptive immunity, but extreme cortisol responses (relative adrenal insufficiency) are associated with a higher mortality rate [8183]. Cortisol levels were high in COVID-19 positive patients compared to negative [84] and nonfatal [85] COVID-19 patients.

Although meta-analyses showed an overall decrease in stress biomarkers (IL-6, TNF-α, and cortisol), only TNF-α was significant. However, subgroup analysis showed TNF-α (8–12 wk, 3000–4800 minutes), IL-6 (yoga-meditation-pranayama), and cortisol (6–12 wk, 1000–2000 minutes) statistically significant reductions favouring intervention. Our findings are in line with earlier conducted meta-analyses in which both IL-6 and TNF-α were reduced in the intervention group following the yoga and meditation [86, 87]. The change was significant in one study reporting IL-6 [86], whereas there was an insignificant decrease in IL-6 and TNF-α [87], which could have been due to smaller number of studies (n = 2) included in the analysis. Recent reviews on the effects of phytosterols and the effect of an herb, Nigella sativa, have also been shown to reduce the cytokine storm; however, these are beyond the scope of this review and so, further suggested [8, 88].

In the scarcity of RCTs reporting the effectiveness and efficacy of yoga, meditation, and pranayama in COVID-19 patients, our findings further add to the yoga-based interventions and their effects on the inflammatory biomarkers. Through our findings, it may be possible to recommend treatment strategies to promote the health of both mild to moderate and severe cases, including symptomatic and asymptomatic COVID-19 patients.

The Indian Public Health Standards (IPHS), in its revised set of standards released in 2012, has integrated yoga as one of the components to be prescribed for AYUSH facilities [89]. Our findings strengthen the usefulness of yoga and its fraternity, which is in tandem with the recommendation of IPHS to implement yoga in primary and secondary health facilities. A minimum of an hour of intervention following a higher duration of yoga followed by pranayama and meditation in healthy and asymptomatic COVID-19 patients can have a positive impact on their health. As COVID-19 patients become more severe, pranayama should be practised for longer periods, followed by meditation and yoga, as various asanas in yoga might be complicated to perform in such patients. Our findings suggest that the intervention can have a long-term positive effect even if practised at home, but beyond 8 weeks, participants might feel unmotivated to continue. Therefore, post-8-week refresher training might be granted as food for thought.

Despite this, the review has certain limitations. Because the studies and participants were from various settings and countries, significant heterogeneity was observed in the meta-analyses, which may cast doubt on the summary effects to some extent and should be approached with caution. Moreover, due to the low number of studies included in meta-analyses, it was unwise and so restricted to carry out quantitative analysis such as Egger’s or Begg’s tests for publication bias. In the systematic review, we included seven pilot and feasibility RCTs, which may have compromised the internal validity of the review. Furthermore, only a single study was respiratory-related (acute respiratory distress syndrome, ARDS) as a disease outcome; it should be extrapolated in COVID-19 patients with solicitude.

5. Conclusion

The systematic review of 44 studies revealed 25 immunity biomarkers, among healthy, clinical, and pregnant women, with possible implications for yoga-related interventions in COVID-19. Total doses of yoga, meditation, or pranayama ranged between 20 and 8400 minutes. However, a smaller dose of YMP was found to make significant changes among pregnant or clinical patients in comparison to healthier ones. Clearly, the combined action of YMP may be effective in lowering IL-6, cortisol, and TNF-α levels in patients, specifically with 2625, 3900, and 1500-minute interventions, respectively, spanned over 8–12 weeks. Also, the YMP package is warranted for pilot RCT among COVID-19 mild to moderate patients with standard care before reaching a firm conclusion. Furthermore, meta-analyses of biomarkers and phytosterols of COVID-19 patients are highly recommended.

Data Availability

Data extracted from the studies, of systematic reviews and meta-analysis, are provided by the corresponding author upon request.

Additional Points

Highlights of the review. What we knew previously: Some practices of complementary and alternative medicine (CAM), traditional Chinese medicine (TCM), and herbal medicine possibly have positive effects on COVID-19 patients. Biomarkers (around 25) are found to be favourably changed by yoga, meditation, and pranayama (breathing exercises) (YMP) among different clinical (other than COVID-19) and healthy participants. Different biomarkers, leading to cytokine storm, are affected in COVID-19. What this review added on: YMP interventions can change the levels of biomarkers favouring health status of healthy, clinical, and the pregnant women with nine, fourteen, and two biomarkers, respectively, altogether 25, which may have possible implications in COVID-19. Clearly, combined action of YMP with specific durations might be effective in reducing IL-6, cortisol, and TNF-α among the COVID-19 patients.

Disclosure

The funder had no role in designing the study, collecting, analysing, interpreting the data, preparing the manuscript, and deciding to publish.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

Komal Shah, Chiranjivi Adhikari, and Shubham Sharma contributed equally. All the authors reviewed and approved the final manuscript for submission.

Acknowledgments

The authors duly acknowledge the Govt. of India, Department of Science and Technology (GOI, DST) for financial support. The authors are also thankful to the authors and coauthors of the included studies. The authors also acknowledge Dr. Varna VP, MPH student of IIPHG 2020–2022 batch, GJ, India, for her contribution. This work was supported by the Government of India, Department of Science and Technology (grant numbers DST/SATYAM/COVID-19/2020/72).

Supplementary Materials

S1_Search Strategy with Keywords and MeSH. (Supplementary Materials)