Effect of Tibetan herbal formulas on symptom duration among ambulatory patients with native SARS-CoV-2 infection: A retrospective cohort study ☆

;


A B S T R A C T
Background: Despite abundant data regarding factors that influence COVID-19 symptom severity and need for hospitalization, few studies examine time to resolution of symptoms and potential complementary and Abbreviations: SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; VOCs, variants of concern; COVID-19, coronavirus disease of 2019; CAM, complementary and alternative medicine; PCR, polymerase chain reaction; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology; SAS, Statistical Analysis System; SD, standard deviation; COPD, chronic obstructive pulmonary disease; ILD, interstitial lung disease; HIV, human immunodeficiency virus infection; RA, rheumatoid arthritis; Hep B, hepatitis B; Hep C, hepatitis C; y, years; Q1, first quartile; Q3, third quartile; IQR, interquartile range; CI, confidence interval; WHO, World Health Organization; BMI;, body mass index; NF-κB, nuclear factor kappa B; MAPK3, mitogen-activated protein kinase 3; CCL2, chemokine (C-C motif) ligand 2; PTGS2, prostaglandin-endoperoxide synthase 2; IL-1β, interleukin-1β; IL-6, interleukin-6; CDC, Centers for Disease Control and Prevention; NCIRD, National Center for Immunization and Respiratory Diseases.☆ Special Issue on Global Traditional and Indigenous Medical Systems.Guest editors, Drs.Gang Chen and Sarah Spencer

Introduction
In native infections with the initial Wuhan-Hu-1 strain of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) through Delta variants of concern (VOCs), data pooled from international sources suggest that 85% of unvaccinated people have mild illness and 14% develop severe disease requiring hospitalization (Cheng et al., 2021;Wu and McGoogan, 2020;Liguoro et al., 2020).These earliest COVID-19 infections demonstrated greater symptom severity than later VOCs, including higher rates of hospitalization, intensive care admissions, and death, especially among unvaccinated individuals (Lewnard et al., 2022).For severe and high risk cases, antivirals and steroids have been administered to facilitate rehabilitation and increase survival rates (Sanders et al., 2020).Yet for most mild-and medium-severity cases, at the time of our study, an effective treatment did not exist.Though current vaccination rates are high and antivirals are now widely administered for mild-and medium-severity ambulatory cases, only remdesivir has exhibited some evidence for reducing time to recovery, yet still with inconsistent results (Vegivinti et al., 2022).Likewise, symptom rebound has been reported for oral antivirals, particularly among the vaccinated (Ranard et al., 2023).Thus, the role of traditional therapeutics in reducing time to recovery for an emerging viral infection is timely and pertinent and may be instructive for future novel infectious agents.
Despite a paucity of biomedical treatments early in the COVID-19 pandemic, traditional medicine, largely vis-à-vis modalities of complementary and alternative medicine (CAM), provided therapeutic care beyond the recommended supportive management (Kumar et al., 2022) of rest, isolation and over-the-counter medications for symptom management.These treatments ameliorated impacts of inflammatory cytokines throughout disease course (Lim et al., 2021) and strengthened physical and mental resilience (Seifert et al., 2020).Numerous countries have implemented traditional medicine supplementation in biomedical protocols for COVID-19 cases (Kumar et al., 2022;Lyu et al., 2021;Wang et al., 2021;Panda et al., 2022;Zhang et al., 2023).A review assessing 32 randomized control trials, totaling 3177 COVID-19 patients treated with herbal intervention as adjuvant therapy, showed significantly greater improvement in clinical outcomes compared to conventional Euroamerican biomedicine alone (Kumar et al., 2022).
In the absence of treatment during the first year of the pandemic, Tibetan communities in North America, as well as those familiar with such resources, relied on Tibetan medicine for their healthcare support.As such, we had an opportunity to conduct an observational study of patients under Tibetan medicine standard of care administered by ambulatory clinics across North America.Patient cases observed in the study relied on Tibetan medicine exclusively.
Tibetan medicine approaches patient care through comprehensive paradigms that integrate dietary, behavioral, and mental health guidance; as well as herbal formulas and external therapies-massage, moxibustion, needle therapies, compresses, and medicinal bath (Hofer, 2014).Herbal formulas comprise botanical and mineral species from high altitude Himalayan regions and South Asian lowlands, developed over long exposure histories to epidemics.
Our purpose is to describe symptom presence and recovery in a population that uses traditional Tibetan medicine.We conducted the study prior to vaccine availability and were able to collect detailed information about symptom presentation and time to recovery.

Study design and participants
The study team invited physicians of Tibetan medical clinics in North America, recognized by national medical licensing bodies in Tibet, China and India (Craig and Gerke, 2016;Pordié and Blaikie, 2014), to attend online meetings explaining the study.Twenty physicians attended recruitment meetings: fifteen enrolled.Reasons cited for study non-participation included burden of time for documentation and patient record deidentification limitations.
Patients contacting the Tibetan outpatient clinic from March 12, 2020 to May 5, 2021 with suspected COVID-19 infection had data collected using a standard reporting template (Supplementary Table 1) (N = 145).Patients with incomplete symptom resolution information (N = 28), hospital admission (N = 3, including 1 incomplete data), and treatment start after the first two weeks of symptoms (N = 2) were excluded.Data on use of Western medications (over-the-counter drugs like acetaminophen) were collected, and participants using both Tibetan and over-the-counter medications simultaneously were also excluded (N = 3, including 2 of the 3 hospitalized patients).See Supplementary Figure 1 for a flow diagram of inclusion and exclusion criteria.
Confirmed presence of SARS-CoV-2 was desired but testing site limitations during this period led to a decision to accept suspected cases for potential later confirmation by antibody testing.However, time and resources required for antibody tests prohibited confirmations and none of our suspected cases were subsequently confirmed.Only cases confirmed using polymerase chain reaction (PCR) or equivalent laboratory test (e.g., rapid antigen test) were included.All patients were unvaccinated.Follow up on September 15, 2021 assessed recurrent or persistent symptom reports.
The study followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) requirements.

Informed consent
Patients consented at the outpatient clinic to have their data deidentified and analyzed for research purposes.Physicians consented with study staff for data submission of confirmed cases.Institutional Review Boards of University of Wisconsin-Madison, University of Minnesota, and Dartmouth College approved the protocol.All procedures followed ethical standards of the Helsinki Declaration of the World Medical Association.

Treatments
Patients contacted clinics based on infection status.Attending physicians administered a telemedicine consultation to assess additional diagnostics from the Tibetan medical perspective, including psychophysiological constitutional characteristics (Cameron et al., 2012;Luo and Li, 2021).Physicians applied personalized medicine treatment approaches endemic to Tibetan medicine that: (1) specifies formulas to constitution, symptom presentation, and disease course; and (2) provides personalized guidance on diet, thermoregulation, nurturing social interactions (though remote), and methods of maintaining positive mental attitudes.Several patients applied physician-guided external therapies (medicinal bath, warm medicinal oil compresses, herbal vapor inhalations).See Supplementary Figure 2 for a schematic diagram detailing treatment by disease phase with Tibetan herbal formulas among participating study physicians.

Statistical analysis
The primary outcome of interest was time to recovery, defined as number of days between reported symptom onset and symptom remission (last date all symptoms were no longer reported).Additional variables of interest included baseline participant characteristics, namely, age, race/ethnicity, presence of comorbidities, and presence of specific symptoms.
Comparison of characteristics between those eligible for analyses (confirmed COVID-19 infection) to unconfirmed and excluded cases was performed using Analysis of Covariance for continuous covariates (e.g., age) and Chi-Square test (Mantel-Haenszel test of trend (Mantel and Haenszel, 1959)) for categorical covariates (e.g., age group, gender, symptom presence).Time to recovery was log-transformed and geometric mean and 95% confidence intervals were computed.Analysis of Variance was used to compare time to recovery between levels of covariates, including specific symptom presence.SAS (v9.4,SAS Instituted, Cary, NC) was used for all analyses.We also conducted exploratory cluster and factor/principal components analyses to assess symptom groupings based on explained variation and similarity metrics, where symptom patterns in time to recovery were assessed (e.g., if someone reports cough, they often also report fever).Symptom clusters were defined to replicate symptom groups reported in prior work (Lane et al., 2021): upper respiratory (sore throat, rhinorrhea); lower respiratory (cough, shortness of breath); gastrointestinal (nausea, diarrhea); neurologic (myalgia, fatigue, headache).

Demographic data and comorbidities
Of the 145 patient cases assessed, the included cohort comprised patients.Table 1 presents the demographic data.For the included cohort, mean (SD) age was 44.7 (12.7) years and females comprised 49.3% of the sample.In terms of race and ethnicity, patients were predominantly Tibetan (80.6%), followed by White non-Hispanic (14.5%).

Clinical outcomes
We found cough and fever to be the most common presenting symptoms at disease onset among included cases (Table 1).Other common presenting symptoms included fatigue, headache, sore throat, and reduced appetite; and less commonly, insomnia, chest pain, digestive difficulties, sweats, rhinorrhea, and brain fog.
Exploratory cluster analysis identified headache and joint pain as symptoms often reported concurrently.Cough and fever grouped distinctly from the cluster of symptoms most highly associated with COVID-19, namely, shortness of breath and loss of smell and taste.The latter cluster also included reduced appetite and nausea.

Time to recovery
Patients contacted an outpatient clinic an average of 1.0 (IQR 0-5.0)T.L. Tidwell et al. day after confirmed COVID-19 test, and 3.0 days (IQR 0-4.0) after symptom onset.Symptom resolution occurred a median of 11.7 (IQR 10.1-13.5)days from symptom onset, shown in Table 2A.Approximately 64.2% of patients recovered by the first or second weeks of their illness, and a cumulative 94.0% reached recovery by the third week.Out of the 4 patients that did not reach recovery by the third week, 2 had shortness of breath that took approximately 90 days to resolve; the other 2 expressed intermittent experiences of anxiety and fatigue at follow up.This may represent long COVID, defined as symptoms lasting 60 days or more (Nasserie et al., 2021).

Table 1
Demographic characteristics, comorbidities, and initial symptoms.

Table 2
Recovery time and covariates by analysis group.

Time to recovery by covariates
The distribution of time to recovery by covariates does not show distinct differences between males or females, nor by age.The only individual symptoms to show significant relationships with time to recovery are sore throat and muscle cramps, which, when present at baseline, predict faster recovery times (Table 2B).See Supplementary Figure 3 for summarized diagram of recovery time by baseline symptom versus no symptom for confirmed cases.Some symptom clusters were associated with faster time to recovery.Headache and joint pain as a symptom pair, identified from cluster analysis, had a median 10.5 days to recovery with a 95% CI [9.2, 11.9, p = 0.010].The fever cluster (as Lane et al., 2021 defined, comprising one or more symptoms of fever, joint pain, chills and/or cramps) also demonstrated significant (p = 0.04) reduction in time to recovery of 11.1 days with a 95% CI [9.9, 12.4].

Discussion
We found the median time to resolution of symptoms for COVID-19 in the outpatient setting, under care management by Tibetan medicine practitioners administering Tibetan herbal formulas, was 11.7 (IQR 10.1-13.5)days.Time to recovery following public health recommendations for supportive management was previously reported as 15-21 days (Lane et al., 2021;Blair et al., 2021aBlair et al., , 2021b;;Pettrone et al., 2021), during a similar span of time as reported here.Though a statistical comparison with these prior studies is not possible, they report a total time to recovery greater than the upper bound of the 95% confidence interval.This suggests that patients infected with similar COVID-19 variants (based on calendar periods) who used Tibetan medicine may have had faster symptom resolution than patients using other methods of supportive management.
Importantly, patient characteristics among our sample largely match those of prior studies, except the predominant Tibetan ethnicity among our participants.Only Lane et al. (2021) and Blair et al. (2021aBlair et al. ( , 2021b) ) provided similar details on patient characteristics, and so are employed as comparison samples.Our study had comparable distributions of age and sex, relative to those of the comparison studies, and all 3 studies had similar dominant comorbidities.However, our sample did have a high incidence of liver disease (10.1%), reflecting higher rates of hepatitis infection in the Tibetan community (Sangmo et al., 2007), which was not present in the comparison samples; whereas the comparison samples had over 40% and 27%, respectively, of participants with current or previous smoking history, of which our sample had few.Thus, we cannot exclude the possibility that ethnicity, in addition to the Tibetan herbal formulas, contributed to the recovery time reported in our sample.Indeed, we did find a slightly more rapid recovery (though not statistically significant) in Tibetan compared to non-Tibetan patients in our sample (Tables 2B) -11.3 days to symptom resolution in Tibetan and 13.5 days in non-Tibetan patients (i.e., predominantly White non-Hispanic) (p = 0.19).Socioeconomic status was not reported in any of the 3 studies.
Cough and fever, our most commonly reported symptoms and with a distinct relationship in the cluster analysis, presented with similar frequency to that of other studies, which also identified cough and fever as the predominant symptoms among first ambulatory COVID-19 cases recorded in China, North America, and globally (WHO, 2020;Goyal et al., 2020;Guan et al., 2020).The relationships identified by cluster analysis that grouped symptoms most highly associated with COVID-19, namely shortness of breath and loss of smell and taste, with reduced appetite and nausea might identify symptoms underreported in a condition widely framed as a respiratory infection, though with known gastrointestinal presentation.It might also reflect physiological responses to taste/smell loss.
The Lane et al. (2021) study was the only study to report illness duration differences by specific symptoms.While this study found longer symptom durations when lower respiratory symptoms were present at disease onset, our study did not replicate this result.However, sore throat and muscle cramps at baseline did predict faster recovery times among our sample.This might be due to the lack of good standard of care treatment options for lower respiratory symptoms among conventional, over-the-counter options, whereas these symptoms are well-addressed by Tibetan herbal formulas.Likewise, sore throat and muscle cramps represent two different presentations of virulent infectious disease in Tibetan medical nosology-one presenting with symptoms more isolated to the upper body such as classic upper respiratory infection; and the other, more systemic in its presentation affecting various muscle groups, and impeding liver and kidney function.
The significant relationship present in the exploratory cluster analysis for the factor grouping of headache and joint pain illustrates a symptom pairing recognized in Tibetan medical nosology for virulent respiratory infections of this type (Gönpo, 2008).Though this symptom pair is often reported to present with more intense illness experiences, it also was found to be associated with swifter recovery times in our study, which may result from stronger antibody responses and/or more rapid symptom relief consequent to Tibetan formulas.The Tibetan medical understanding of such virulent infections links central nervous system inflammation to that in interstitial and synovial fluid spaces.Several formulas target these inflammatory pathways specifically and thus might account for swifter recovery times.Conversely, those who do not gain treatment access early with this symptom constellation would be predictably more susceptible to long COVID due to persistent central nervous system inflammation.
Grouping symptoms, either based on previously published groupings or on our cluster analysis, identified differences in time to recovery not present when considering symptoms individually, particularly swifter recovery among patients with both headache and joint pain, as well as those who had one or more fever cluster symptoms, including joint pain, chills and cramps in addition to fever.This may be because Tibetan medicine targets presentation of symptoms groups, rather than diagnosis, and thus will be highlighted in time to recovery analyses.
During the period in which our study occurred, development of severe symptoms requiring hospitalization was more likely among individuals with hypertension and/or high BMI.The prevalence of hypertension among our cases is consistent with that reported by other outpatient studies as the most common comorbidity among ambulatory cases (Lane et al., 2021;Ramasamy et al., 2021;Arons et al., 2020;Blair et al., 2021aBlair et al., , 2021b;;Kirtana et al., 2020;Li et al., 2020).Although our study could not assess BMI, 65.2% of our included cases had one or more other risk factor known to predict more severe disease such as diabetes, chronic liver disease, dyspnea, male sex and older age (Cheng et al., 2021;Dinnes et al., 2021).Despite the high prevalence of risk for severe disease, very few patients in our sample experienced long COVID, which is more common among those with severe disease (Hedberg et al., 2023).This further emphasizes the value of the Tibetan medicine approach beyond rapid time to recovery.

Limitations
Although physicians in the current observational study gave patients guidance on adaptive mindsets to cultivate, and maladaptive stress responses to avoid, herbal formulas comprised the primary therapy, targeting the virulent infectious disease phases recognized by Tibetan medicine (Tidwell and Gyamtso, 2021).See Supplementary Figure 2 for a schematic diagram detailing treatment by disease phase with Tibetan herbal formulas.Nevertheless, we recognize the potential therapeutic significance of the patient-physician relationship.Additional limitations of our study are multifold.To address a few, the naturalistic study design prohibits the ability to make causal claims.Reliance on patient self-reports has high susceptibility to recall bias and potential cultural differences in reporting certain symptom types.In addition, our sample does not have a comparator control population for the same study period following standard of care and it represents limited diversity and thus limits generalizability.Finally, testing site limitations for confirming cases led to a relatively small study population.

Conclusion
Despite the limitations, this is the first study to characterize the duration of COVID-19 symptoms in an outpatient setting for patients using Tibetan medicinal formulas.With median time to symptom resolution for COVID-19 under Tibetan medicine care management in the outpatient setting as 11.7 days and time to recovery following public health recommendations for supportive management previously reported as 15-21 days (Lane et al., 2021;Blair et al., 2021aBlair et al., , 2021b;;Pettrone et al., 2021), our results suggest patients with similar COVID-19 variants may have had faster symptom resolution under Tibetan medicine care than patients using other supportive management methods.An increasing number of SARS-CoV-2 infections in those first evaluated as outpatients has increased and variants of concern demonstrating differential transmission dynamics and disease course progressions continue to emerge (CDC and NCIRD, 2022).A better understanding of symptom duration among outpatients with COVID-19, particularly those treated by traditional, complementary and alternative medicine such as Tibetan medicine, can help direct care, inform transmission reduction, tailor public health messaging, and boost recognition of CAM modalities that may ameliorate disease severity and reduce recovery time.Furthermore, this study attempts to address concerns of populations susceptible to epidemiological invisibility (Gurung et al., 2021) by contributing one of the few assessments of outpatient experiences among such marginalized communities, particularly those drawing upon cultural resources for therapeutic care in one of the greatest global health crises of the current era.