key: cord-0259077-ew3q4b3u authors: Tanwettiyanont, J.; Piriyachananusorn, N.; Sangsoi, L.; Boonsong, B.; Sunpapoa, C.; Tanamatayarat, P.; Kanchanasurakit, S.; Na-Ek, N. title: The efficacy of Andrographis paniculata (Burm.f.) Wall. ex Nees crude extract in hospitalised mild COVID-19 patients: a retrospective cohort study date: 2022-01-03 journal: nan DOI: 10.1101/2022.01.01.22268609 sha: 74c8dc9a67c87c7a29026452888ceb384dae24a6 doc_id: 259077 cord_uid: ew3q4b3u Background: Andrographis paniculata (AP) crude extract has been widely used in Thailand to treat mild COVID-19 infection since early 2020; however, supporting evidence was lacking. Purpose: To evaluate the efficacy of AP compared with standard treatment among hospitalised mild COVID-19 patients. Study design: Single-centre retrospective cohort study Methods: We collected data between March 2020 and August 2021 from COVID-19 patients admitted to one hospital in Thailand. Patients whose infection was confirmed by Real-Time Polymerase Chain Reaction (RT-PCR) and had normal chest radiography were included, whereas those receiving favipiravir or had unclear chest X-rays at admission were excluded. Participants were categorised as either AP or standard of care and followed for pneumonia confirmed by chest radiography. Multiple logistic regression was used to analyse the main results controlling for age, sex, history of having diabetes, hypertension, receiving statins, and antihypertensive drugs. Results: 605 out of 1,054 patients were included in the analysis. Of these, 59 patients (9.8%) developed pneumonia during the median follow-up of 7 days. The incidence rates of pneumonia were 13.93 (95%CI 10.09, 19.23) and 12.47 (95%CI 8.21, 18.94) per 1,000 person-days in AP and standard of care group, respectively. Compared to the standard of care group, the odds ratios of having pneumonia in the AP group were 1.24 (95%CI 0.71, 2.16; unadjusted model) and 1.42 (95%CI 0.79, 2.55; fully adjusted model). All sensitivity analyses produced consistent findings with the main results. Conclusion: We do not have sufficient evidence to show the efficacy of AP in mild COVID-19 infection. Interestingly, we observed the potentially harmful signal of using AP. While waiting for insights from ongoing trials, the use of AP in this condition should be done with caution. Andrographis paniculata (Burm.f.) Wall. ex Nees (AP), also known as 'Fa Thalai Chon' or 'Fa Thalai', has been 2 widely used in Thailand for treating upper respiratory tract infections and non-infectious diarrhoea after being 3 in the National List of Herbal Medicines of Thailand in 1999 (Karbwang and Na-Bangchang, 2021) . The main 4 phytochemical constituent of the aerial parts of AP is a diterpenoid lactone compound called 'andrographolide ', 5 which has shown antiviral and immunomodulatory properties from preclinical and clinical studies (Dai et al., 6 2019). Recently, in silico study has shown the potential effect of andrographolide on SAR-CoV-2 as the 7 compound can bind and inhibit the viral protease enzyme and viral spike glycoprotein (Enmozhi et al., 2021; 8 Rajagopal et al., 2020) . Moreover, in vivo and in vitro studies consistently supported the effect of AP extract 9 on COVID-19 infections (Phumiamorn et al., 2020; Sa-ngiamsuntorn et al., 2021) . 10 In addition to preclinical studies, two small clinical trials of using a high dose of AP crude extract to treat mild 11 COVID-19 infections has shown its efficacy in terms of reducing COVID-19 symptoms (Rattanaraksa et al., 12 2021) and C-reactive protein (CRP) levels (Wanaratna et al., 2021) . However, its efficacy on an important 13 clinical outcome, such as pneumonia, is unclear. This is because the results from the trial of 57 patients had 14 shown no significant decrease in the incidence of pneumonia in the AP group, compared to placebo, after five 15 days of treatment (p-value = 0.11) (Wanaratna et al., 2021) . Currently, five ongoing trials are investigating the 16 efficacy of AP in terms of pneumonia for treating mild COVID-19 cases. 17 Although AP's efficacy on the major clinical endpoint is still ambiguous, its widespread use has been 18 encouraged. This is due to the situation in which Thailand experienced a shortage of favipiravir and vaccines at the start of a new pandemic wave in early 2020. Therefore, the data on the efficacy and safety of 20 AP from a pharmacovigilance study is necessary to support the decision of clinicians and policymakers whether 21 AP's use in COVID-19 should be further supported. 22 In this study, we primarily aim to use real-world data to investigate the efficacy of AP crude extract for the 23 treatment of hospitalised mild COVID-19 patients. We also examined the course of COVID-19 and the incidence 24 of pneumonia due to COVID-19 in a country-specific context. Our ultimate goal is to make the best use of 25 available data to inform the public and improve patient care. 26 The report of this study followed the Strengthening the Reporting of Observational Studies in Epidemiology 28 (STROBE) guidance for reporting cohort study (Vandenbroucke et al., 2007) (Table S1 ). 29 This is a single-centre retrospective cohort study in which the data were collected from medical records of 31 patients diagnosed with COVID-19 infection. We used the 10 th revision of the International Classification of 32 Diseases (ICD-10) code U07.1 to identify potential participants from 1 st March 2020 to 31 st August 2021. The 33 ethical committee for clinical research of Phrae hospital approved this study (no. 70/2564). 34 The setting of our study is Phrae hospital, a 500-bed secondary hospital located in northern Thailand. Eligible 35 participants were at least 18 years old and diagnosed with COVID-19 infection by Real-Time Polymerase Chain 36 Reaction (RT-PCR) test. According to the definition of mild COVID-19 used in previous work (Zhang et al., 37 2021), we included only patients who had normal chest radiography by the time of hospital admission. In 38 contrast, individuals who did not have chest radiography results received antiviral drugs (i.e., favipiravir) or 39 received systemic corticosteroids on the first day of admission were excluded. In addition, we also excluded 40 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi. org/10.1101 org/10. /2022 those who recently received AP prior to hospital admission, had a history of allergy to AP, had elevated liver 1 enzyme, or were pregnant or breastfeeding from the analysis. Since the preliminary data suggested that AP's 2 efficacy was shown if it was given to patients as soon as they were diagnosed, we additionally excluded 3 patients who received AP after five days of admission from our analysis. (Thai Clinical Trials Registry, 2009) 4 Exposure 5 Included participants who received AP crude extract within five days of admission in addition to a supportive 6 treatment were categorised as an exposed group. AP crude extract was given as a capsule of 500 mg of crude 7 extract containing andrographolide content of approximately 4%w/w (~20 mg/capsule). According to a previous 8 trial, (Wanaratna et al., 2021) the AP product was given three capsules thrice daily after a meal to reach a total 9 dose of andrographolide 180 mg/day for five consecutive days. Song hospital, Phrae, Thailand, produced the 10 AP product used in this setting. The quality of AP product was tested and certified by the Medicinal Plant 11 Research Institute and the regional Medical Sciences Centre, Chiang Rai, Thailand (details can be found in 12 supplementary appendices). Supportive treatment, including antipyretics, mucolytics, expectorants, 13 antihistamines, oral rehydration salts, and anxiolytics, was given to patients who did not receive AP (unexposed 14 group). 15 Outcomes 16 The primary outcome was developing pneumonia based on chest radiography during hospital admission. The 17 diagnosis of pneumonia was based solely on chest X-rays (CXR) of category four or above according to the 18 Modified Rama-Co-RADS for the first CXR in confirmed COVID-19 patients (Supplementary appendices). The 19 categorisation was made mainly by infectious disease physicians or radiologists. Patients with ambiguous CXR 20 results were excluded from the analysis. For patients who did not have CXR results during follow-up and did 21 not die or refer to the intensive care unit, we assumed that they did not develop pneumonia and used discharged 22 date as the end of the follow-up. 23 In addition, we analysed the association between receiving AP and a secondary outcome, which was a 24 composite of receiving favipiravir, systemic corticosteroids, or ventilator support; having oxygen saturation drop 25 along with worsening signs and symptoms; or presenting regressive CXR findings (i.e., category three or above) 26 after admission. The CXR results, all clinical data, and relevant medications were collected from electronic 27 medical records. 28 We collected all covariates for the admission date from medical records. These covariates included age, sex, 30 weight, height, comorbidity, current medications, and laboratory parameters. According to our proposed directed 31 acyclic graphs (DAGs, Figure S1 ), age, body mass index, hypertension, type 2 diabetes (T2DM), ACEIs/ARBs, 32 statins, and COVID-19 severity were associated with both receiving AP (from discussion with a health care 33 team) and developing pneumonia (Table S5 ) (Centers for Disease Control and Prevention, 2021a, 2021b) and 34 did not lie in a causal pathway between these two variables and were thus considered confounders. Admittedly, 35 during the data collection period, there were only two patients who had received a COVID-19 vaccine before 36 diagnosis and hospital admission. Consequently, we did not include vaccination profiles in our analysis. 37 In this study, we included all eligible COVID-19 patients in the analysis. Therefore, sample size calculation was 39 unnecessary, and we planned to calculate statistical power afterwards. Descriptive and inferential statistics 40 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi. org/10.1101 org/10. /2022 were used to compare participants' characteristics at hospital admission according to their exposed groups. In 1 addition to the calculated incidence rate of pneumonia (per 1,000 person-days) according to exposed groups, 2 a Kaplan-Meier plot for the probability of a pneumonia-free event between groups was also created and 3 statistically compared using a log-rank test. 4 The main analysis was performed using a multivariable logistic regression based on a complete-case approach. 5 The justification for using a logistic model is that each participant had a relatively similar follow-up time: a 6 median of 7 days (interquartile range 6 to 9 days). To investigate the association between receiving AP and 7 incident pneumonia, we did serial adjustment as follows: 1) unadjusted model, 2) age-adjusted model, and 3) 8 full adjustment (i.e., adjusting for age, hypertension, T2DM, ACEIs/ARBs, and statins). Regarding BMI, we 9 further performed multiple imputations by chain equation (MICE) to impute missing values. In addition to all 10 variables in the main analysis, the following auxiliary variables were added to the imputation model: follow-up 11 time, age 2 , Nelson-Aalen cumulative hazard function, the use of proton pump inhibitors at baseline, weight, 12 height, and alkaline phosphatase (logarithmic scale). BMI was then included in a model as part of a sensitivity 13 analysis since we assumed that the missing BMI values were unlikely to be under a missing at random (MAR) 14 mechanism (i.e., there are other variables significantly affecting BMI that were not recorded and collected, such 15 as daily caloric intake, physical activity, and smoking status) and using MICE might bias the results. We 16 performed 100 imputations, and results were combined using Rubin's rule. 17 Moreover, for the sensitivity analysis, we analysed the data using Cox's proportional hazards model, in which 18 a fully adjusted model was stratified by diabetes. The Schoenfeld residuals test and log-minus-log plots were 19 checked accordingly to ensure the satisfaction of the proportional hazards assumption. The severity of COVID-20 19 was conditioned by restricting the analysis to a mild case only. Furthermore, we performed subgroup 21 analyses according to sex, age group (i.e., <60, ≥60), hypertension, T2DM, ACEIs/ARBs, and statins to 22 examine effect modifiers. Lastly, to minimise a cohort effect due to differences in admission period, since some 23 patients in an unexposed group were admitted a few months before an exposed group ( Figure S3 ) and 24 guidelines for COVID-19 treatment and coverage of immunisation can improve drastically over a short period, 25 we, therefore, excluded individuals admitted before the 1 st of July, 2021, then re-analysed the main group of 26 patients. 27 All analyses were performed using STATA version 16.1MP (StataCorp LLC, College Station, Texas) and R 28 version 3.3 with a two-sided alpha error of 5%. As we did not adjust for multiplicity, findings of the secondary 29 outcome, sensitivity analyses, and subgroup analyses should be used for exploratory purposes only. 30 Among 1,054 COVID-19 patients admitted to the hospital between March 2020 and August 2021, 605 were 32 included in the final analysis ( Figure 1 ). Of these, 351 individuals (58%) received AP within five days of 33 admission. Regarding the characteristics of included participants at hospital admission (Table 1) , the majority 34 of the participants were male (50.4%), with a mean age of 35.41 years old and a mean BMI of 24.2 kg/m 2 . 35 Only a small proportion of individuals had hypertension (7.3%), T2DM (2.2%), and cardiovascular disease 36 (0.8%). In addition, 3.8% and 2.6% of the patients received ACEIs/ARBs and statins, respectively. Comparing 37 between groups, most of the characteristics were relatively similar, except for alkaline phosphatase (ALP) 38 levels, as the levels in the AP group were slightly higher than in the standard of care group. However, all 39 laboratory parameters were within the normal range (Table 1) . 40 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint During a median follow-up time of 7 days (IQR 6, 9 days) and a median hospital stay of 8 days (IQR 6, 10 1 days), 59 out of 605 participants (9.8%) developed pneumonia -an overall incidence rate of 13. Figure S2 ). According to 7 Table S3 -S4, it is worth noting that, regardless of exposure group, 1) the incidence rate of pneumonia before 8 seven days of follow-up was higher than that afterwards, and 2) the incidence rate of pneumonia among 9 patients aged over 60 years was drastically higher than that among younger individuals. 10 According to Table 2 , compared to a standard of care, receiving AP was associated with increased but not 11 statistically significant odds of having pneumonia: odds ratio ( unadjusted-, age-adjusted-, and fully adjusted-model, respectively. Additionally, receiving AP was also 16 associated with a slight but not significant increase in the odds of worsening symptoms. Further adjusting for 17 BMI did not change the direction of the association (Table S6) . 18 Interestingly, excluding participants admitted before the 1 st of July, 2021 (most were from the standard of care 19 group) further strengthened the association of receiving AP with the increased odds of having outcomes. The 20 OR of having pneumonia in an unadjusted-, age-adjusted-, fully adjusted-model, and a model additionally 21 adjusting for BMI was 1.83 (95% CI 0.93, 3.61), 1.94 (95% CI 0.97, 3.92), 1.88 (95% CI 0.92, 3.81), and 1.72 22 (95% CI 0.78, 3.79), respectively (Table S6) . 23 Results from subgroup analyses are shown in Figure 2 . It can be observed that sex was not an effect modifier 24 of the association between receiving AP and pneumonia. However, the association seems stronger among the 25 elderly (i.e., >60 years). Although AP might be related to the increased risk of pneumonia in overall populations 26 and all p-values for interaction >0.05, we found the opposite direction of the associations among individuals 27 with hypertension, receiving ACEIs/ARBs, and receiving statins. 28 In this retrospective cohort study of 605 hospitalised COVID-19 patients who had normal chest radiography at 31 the time of admission, 9.8% of them developed pneumonia after a median follow-up time of 7 days. However, 32 we did not have sufficient evidence to show the efficacy of AP crude extract in decreasing the risk of pneumonia 33 or worsening clinical symptoms. Interestingly, individuals, mainly the elderly, receiving AP were associated with 34 an increased, but not statistically significant, risk of pneumonia and worsening clinical symptoms. Moreover, all 35 sensitivity analyses provided consistent findings, ensuring the robustness of the main results. 36 So far, the clinical evidence of using AP to treat COVID-19 is still lacking. After performing a systematic search 38 on three databases (i.e., PubMed, Google Scholar, and Thai Clinical Trial Registry), we found only two complete 39 trials (Rattanaraksa et al., 2021; Wanaratna et al., 2021) and five ongoing trials relevant to this subject, with 40 the largest trial of 736 patients expected to end in October 2022 (Table S2 ). One trial investigated the efficacy 41 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint of AP in improving clinical symptoms and duration of disease in 62 mild COVID-19 patients (Rattanaraksa et 1 al., 2021) . It showed that all COVID-19 symptoms in the AP group had disappeared by day 7 (i.e., two days 2 after completing an AP course). Compared with our observation, the median length of hospital stays before 3 being discharged alive in an AP group and a standard of care group was eight days (IQR 6, 10 days) and 4 seven days (IQR 6, 9 days), respectively. Therefore, the course of the disease in ours was comparable to the 5 previous one. Another trial reported the incidence of pneumonia in the AP group (0%) and the placebo group 6 (10.7%) after five days of treatment (Wanaratna et al., 2021) . The figure was similar to the incidence of 7 pneumonia in our study's standard of care group (8.7%), confirming the validity of our collected data. 8 Furthermore, we found that increased age, having hypertension and diabetes, and receiving ACEIs/ARBs and 9 statins were associated with an increased risk of pneumonia (Table S5 ). This is consistent with previous reports 10 (Centers for Disease Control and Prevention, 2021b, 2021a) and can further ensure the validity of the data 11 used in our analyses. 12 To the best of our knowledge, this is the first cohort study of AP's use in treating mild 14 Thailand was confronted with the favipiravir and vaccine shortages at the beginning of the second wave of the 15 pandemic crisis, leading to the unproven AP's use for this condition. Consequently, a pharmacovigilance study 16 is required since the real-world data from using AP has been already available so that its efficacy and safety 17 can be clinically ensured. 18 However, there are some limitations worth noticing. First, we cannot avoid residual confounders embedded in 19 an observational study design. For instance, smoking status and mental disorders (e.g., depression) were 20 suggested to be risk factors for developing severe COVID-19 (Centers for Disease Control and Prevention, 21 2021b, 2021a), and this can be prevalent in people in their 30s and 40s. In addition, patients receiving AP may 22 have a higher risk of developing pneumonia than those who do not (i.e., confounding by indication). Therefore, 23 the observed association might result from residual confounders instead, and the causality cannot be inferred. 24 However, baseline characteristics between groups were mostly similar in terms of statistical (i.e., p-values) and 25 nonstatistical (i.e., eyeballing) consideration. Furthermore, since our study populations were relatively young, 26 many medical conditions that can increase the risk of severe COVID-19, such as cancer, chronic lung disease, 27 and chronic kidney disease, were rare and should not be major concerns. Additionally, the results were less 28 likely to be confounded by favipiravir as proportions of patients receiving favipiravir during admission were 29 similar between groups (i.e., 9.7% in a standard of care group versus 10.6% in an AP group). 30 Second, our results suffered from being underpowered. With the sample size of 605, we had only 11% of power 31 to detect the difference in the incidence of pneumonia between exposed (10.5%) and unexposed groups (8.7%). 32 A total of 9,000 participants would be required to achieve at least 80% of power to detect such a slight 33 difference. However, when one carefully examines the effect size and the lower and upper limit of the 34 confidence interval (e.g., OR 1.42 [95% CI 0.79, 2.55]), increasing the sample size is prone to strengthen the 35 harmful signal of the association. 36 Third, our primary outcome was evaluated based solely on chest radiography by each physician leading to 37 potential misclassification bias. Nevertheless, it is likely to be a non-differential misclassification, since the same 38 data source was used to evaluate the outcome throughout, and this would bias our results toward the null. 39 Accordingly, the actual association should be more potent than our observation. Nevertheless, the results from 40 the secondary outcome were consistent with the primary one, suggesting the robustness of our findings. 41 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10. 1101 /2022 Lastly, the data on viral strains was lacking, which might affect the external validity of our study. The most 1 prevalent variants initially found in Thailand were B.1.1.7 (Alpha), B.1.351 (Beta), and P.1 (Gamma), whereas 2 the first report of B.1.617.2 (Delta) in Thailand was back in May 2021. Hence, all four strains can be found in 3 study patients at data collection. Nonetheless, since the incidence of pneumonia in the standard of care group 4 in our study was similar to a previous trial (Wanaratna et al., 2021) and no death occurred, it can be assumed 5 that the viral strains in our study were comparable to the previous trial. Additionally, our findings may limit the 6 generalisability to unvaccinated patients. However, since the COVID-19 vaccine's efficacy in the reduction of 7 the severity of symptoms and pneumonia has been proved and widely accepted (Jara et al., 2021), the role of 8 AP in COVID-19 may, unfortunately, become less prominent over time. 9 Implications 10 For the clinical implications, while waiting for the results from ongoing trials together with improved availability 11 of favipiravir and the COVID-19 vaccine, we suggested that physicians should suspend the use of AP to treat 12 COVID-19. This is because we observed potentially harmful effects without the proof of efficacy, even though 13 causality cannot be established. For the research implications, collaborated multicentre is required to achieve 14 a sufficient sample size and confirm our findings. In addition, the safety parameters of using AP were rarely 15 monitored. We noticed that less than one-fourth of patients receiving AP underwent liver function (e.g., AST, 16 ALT, ALP) and renal function test (e.g., Scr and eGFR) at baseline and rarely measured afterwards. Although 17 a previous study has shown the safety of AP used in other indications (Worakunphanich et al., 2021) , the 18 safety of using such a high dose of AP in COVID-19 is still unclear and needs further investigation. 19 In summary, there is insufficient evidence to show the effectiveness of AP crude extract for the treatment of 21 mild COVID-19 in our study. Moreover, we observed the signal that AP might potentially harm. Results from 22 ongoing randomised controlled trials should provide insight into this issue. In the meantime, using AP in this 23 condition should be cautious or suspended. 24 This study was financially supported by the Thailand Science Research and Innovation Fund and the University 26 of Phayao (Grant No. UoE62010). However, the funding body did not involve a study design, data collection, 27 data analysis, and study's interpretation. Additionally, we thank members of staff at Phrae hospital for facilitating 28 the data collection process of this study. 29 Authors' contribution 30 JT, SK, and NN conceptualised the study objectives and designed and collected the data. JT, SK, and NN 31 contributed to the literature reviewing, data cleaning, data analyses, and interpretation of the findings. JT 32 prepared an initial manuscript, and NN further developed subsequent manuscripts. SK, PT, NP, LS, BB, and 33 CS critically revised the initial manuscript, and all authors participated in further revisions. The final manuscript 34 was read and approved by all authors before submission. 35 The funding body and authors' affiliations bear no responsibility for analysing and interpreting this study. 37 38 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. Vandenbroucke, J.P., von Elm, E., Altman, D.G., Gøtzsche, P.C., Mulrow, C.D., Pocock, S.J., Poole, C., 4 Schlesselman, J.J., Egger, M., Blettner, M., Boffetta, P., Brenner, H., Chêne, G., Cooper, C., Davey-5 Smith, G., Gagnon, F., Greenland, P., Greenland, S., Infante-Rivard, C., Ioannidis, J., James, A., 6 Jones is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint person-days (95% confidence interval), ‡ Analysis using a Cox's proportional hazards model in which the fully 4 adjusted model was additionally stratified by diabetes, ¶ Worsening symptoms were the composite of receiving 5 antiviral drugs, systemic corticosteroids, or ventilator support; having oxygen saturation drop along with 6 worsening signs and symptoms; or presenting regressive chest X-rays findings (i.e., category three or above). 7 Abbreviations: AP; Andrographis paniculata, CI; confidence interval 8 9 . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. Andrographis paniculata (exposed) group (n=351) . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint This is supplement to 'the efficacy of Andrographis paniculata (Burm.f.) Wall. ex Nees crude extract in hospitalised mild COVID-19 patients: a retrospective cohort study' Quality assurance of Andrographis paniculata product ( is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi. org/10.1101 org/10. /2022 Describe any efforts to address potential sources of bias 5-6 Study size 10 Explain how the study size was arrived at 5-6 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why 5-6 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding (b) Describe any methods used to examine subgroups and interactions (c) Explain how missing data were addressed (d) If applicable, explain how loss to follow-up was addressed (e) Describe any sensitivity analyses 5-6 Participants 13* (a) Report numbers of individuals at each stage of study-eg numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analysed is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. Table 2 Other analyses 17 Report other analyses done-eg analyses of subgroups and interactions, and sensitivity analyses Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and magnitude of any potential bias 8-9 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence 7-9 Generalisability 21 Discuss the generalisability (external validity) of the study results 9 Other information Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based 10 *Give such information separately for cases and controls in case-control studies, and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint Quality assurance of Andrographis paniculata product (exposure) used in Phrae hospital Certificate of analysis: The analysis and quality control reported from the Medicinal Plant Research Institute and the regional Medical Sciences Center 1/1, Chiang Rai, Thailand is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 Source (Suwatanapongched et al., n.d.) Categories 1-6 and C are for the initial chest X-rays (CXR) in a new patient or the first CXR in a home isolation or community isolation patient. In this regard, whenever the patient has an old CXR before having COVID-19, the newly performed CXR after confirmed COVID-19 should be interpreted using categories 1, 2, 3, 4, 5, 6, or C. Details of each category were as follows: § Category 1: Normal chest X-ray or no abnormality detected § Category 2: Presence of minor abnormalities unrelated to COVID-19 (e.g., mild cardiomegaly, aortic atherosclerosis, scoliosis, old fractures) § Category C: Low likelihood or atypical for COVID-19 pneumonia, but with other clinically significant diseases (e.g., bacterial pneumonia, active TB, CHF, pneumothorax, pleural effusion, malignancy) unrelated to COVID-19 § Category 3: Equivocal/indeterminate opacities, which may be due to acute or residual/post-COVID-19 pneumonia or pseudolesions § Category 4: Single or multifocal poorly defined ground-glass opacities or consolidations in one lung, suspicious for early/mild acute or post-COVID-19 pneumonia ± fibrosis-like changes § Category 5: Multifocal, peripheral, poorly defined ground-glass opacities or consolidations with or without rounded morphology involving any zones of both lungs, typical for moderate/severe acute or post-COVID-19 pneumonia ± fibrosis-like changes § Category 6: Acute or post-COVID-19 pneumonia with its related conditions or complications (e.g., atelectasis, PE, pulmonary infarction, OP, AFOP, secondary infection, pneumothorax, pneumomediastinum) . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ("andrographis" OR "Andrographis paniculata" OR "andrographolide") AND ("sars-cov2" OR "sars cov 2" OR "covid-19" OR "covid") restricted to the range between 2019 and 2021 Searching on 15/12/2021 without language restriction yielded 1,460 results: found 2 relevant studies (Rattanaraksa et al., 2021; Wanaratna et al., 2021) . There are seven currently registered RCTs. Two are completed RCTs: One small pilot study with no published results (n=6), another one can be found at https://doi.org/10.1101/2021.07.08.21259912 (n=57) (Wanaratna et al., 2021) . Additionally, there are five ongoing RCTs (Table S2) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint 19 Notes: *Analysis using multiple imputation by chain equation (MICE), † Adjusting for age, diabetes, hypertension, receiving statins, and ACEIs/ARBs, § Incidence rate of pneumonia per 1,000 person-days (95% confidence interval), ‡ Cox's proportional hazards model, in which fully adjusted model was additionally stratified by diabetes, and effect sizes were reported as hazard ratio (95% CI). † † Additional adjusting model was a fully adjusting model with further adjusting for body mass index. ¶ Worsening symptoms were the composite of receiving antiviral drugs, systemic corticosteroids, or ventilator support; having oxygen saturation drop along with worsening signs and symptoms; or presenting regressive chest X-rays findings (i.e., category three or above). . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint Figure S1 Directed acyclic graphs (DAGs) of the association between Andrographis paniculata and incidence of pneumonia Abbreviations: ACEIs/ARBs; angiotensin converting enzyme inhibitors/ angiotensin receptor blockers, BMI; body mass index, T2DM; type 2 diabetes mellitus, Sources:(Centers for Disease Control and Prevention, 2021a, 2021b) . CC-BY-NC 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 3, 2022. ; https://doi.org/10.1101/2022.01.01.22268609 doi: medRxiv preprint Science brief: evidence used to update the list of underlying medical conditions that increase a person's risk of severe illness from COVID-19 Underlying medical conditions associated with higher risk for severe COVID-19: information for healthcare providers The efficacy and safety of Andrographis paniculata extract for treatment of COVID-19 patients with mild symptoms, Nakhonpathom hospital. Reg 4-5 Modified Rama Co-RADS user guideline Identifier TCTR20211022002 -Efficacy and safety of an Andrographis capsule compared with standard therapy for the treatment of asymptomatic or mild COVID-19 without pneumonia : a double-blinded randomized-control trial Identifier TCTR20210514003 -Andrographolide as a medical tool for reduction of hospitalization in mild or asymptomatic COVID-19 patients: a randomized double-blind placebo controlled trial Identifier TCTR20210609001 -Efficacy of Andrographis paniculata extract combined with favipiravir treatment to prevent disease progression in mild to moderate COVID-19 patients: a double-blind, randomized, placebo-controlled study Identifier TCTR20210809004 -Comparison efficacy and safety of Andrographis paniculata extract capsules and placebo in COVID-19 patients: double blind randomized control trial Identifier TCTR20210906002 -Prospective study of Andrographolide and favipiravir versus favipiravir monotherapy to prevent severe pulmonary involvement in patients with COVID-19 Efficacy and safety of Andrographis paniculata extract in patients with mild COVID-19: a randomized controlled trial