The Effect of Traditional Chinese Medicine on Postviral Olfactory Dysfunction: A Systematic Review and Meta-Analysis

Objective The aim of this study is to evaluate the efficacy and safety of traditional Chinese medicine (TCM) for postviral olfactory dysfunction (PVOD). Methods PubMed, EMBASE, Cochrane Central Register of Controlled Trials, China Network Knowledge Infrastructure (CNKI), Chinese Scientific Journal Database (VIP), Chinese Biomedical and Medical (CBM) Database, and Wanfang Database were electronically searched from their inception to July 25, 2022. Two authors independently performed study selection, data extraction, and quality assessment to ensure systematic quality evaluation. Randomized controlled trials (RCTs) comparing TCM with olfactory training and/or drug therapy (OTDT) were included. The outcomes were the effective rate, QOD-P, TDI score, UPSIT score, and adverse effects. Cochrane RoB was the guideline used to evaluate the methodological quality of the included trials. RevMan 5.3 software was used for statistical analysis. Results A total of 6 RCTs involving 467 patients with PVOD were selected. Compared with OTDT, TCM plus OTDT decreased QOD-P (MD = −1.73, 95% CI (−2.40, −1.06), P < 0.0001) but did not increase the effective rate (T&T) (RR = 1.28, 95% CI (0.86, 1.90), P=0.22, I2 = 61%). Compared with no treatment, TCM seemed to increase the treatment success rate (UPSIT) (RR = 3.17, 95% CI (1.78, 5.65), P < 0.0001, I2 = 0%), but there was no statistically significant difference in improving the UPSIT score (MD = 3.44, 95% CI (−1.36, 8.24), P=0.16). Compared with drug therapy, TCM plus drug therapy appeared to increase the effective rate (ΔVAS) (RR = 2.36, 95% CI (1.41, 3.94), I2 = 0%), but there was no statistically significant difference in improving the TDI score (MD = 2.10, 95% CI (−1.99, 6.19), P=0.31). No significant differences in adverse reactions were reported between TCM and OTDT. Conclusion TCM may be an effective treatment for PVOD. With a lack of high-quality RCTs, further large-scale and high-quality RCTs are still warranted.

As one of the most common causes of OD, PVOD is an infection caused by several respiratory viruses, including parainfuenza virus, human coronavirus, and rhinovirus [4][5][6]. Te varying severity of PVOD is proposed to refect the degree of epithelial destruction, viral load, and viral serotype [6]. PVOD has a long-term impact on the quality of life of infected patients, which is associated with depressive symptoms [7], malnutrition [8], cognitive decline [9], and mortality [10].
Before the emergence of SARS-CoV-2, 18% to 42% of patients with OD had been associated with a preceding viral upper respiratory infection [11]. Since the outbreak of the COVID-19 pandemic, PVOD is becoming more apparent in this population, often in otherwise asymptomatic patients [12]. Change in the sense of smell and taste is highly prevalent in patients with COVID-19, with 40-50% of people on average reporting these symptoms globally [13,14], and up to 98% showing olfactory dysfunction when tested objectively [15]. A meta-analysis has shown that an estimated 74%, 86%, 90%, and 96% of patients have selfreported smell recovery at 30, 60, 90, and 180 days, respectively. Meanwhile, 5% of patients with olfactory dysfunction have been lasting for more than six months after COVID-19 infection [16]. With more than 550 million people worldwide confrmed as having COVID-19 as of July 2022, of whom about 50% report smell or taste dysfunction, 5.6% of patients with persistent smell dysfunction that translates to more than 15 million patients with long-term smell dysfunctions [14].
Te continuing spread of SARS-CoV-2 has caused a large burden of the disease across the world [17]. Infected patients have been facing long-term sequelae, repeated infections, repetitive positive, and prolonged or intermittent incubation periods [18]. PVOD usually follows the onset of respiratory symptoms and is associated with the infammatory change in the respiratory mucosa and mucous discharge [11,19,20]. Te current understanding and sequence of olfactory dysfunction following a viral upper respiratory infection (URI) begin as nasal mucosal infammation, disrupting natural airway conduction within the nasal cavity and inhibiting the delivery of odorants to the olfactory epithelium. Te persistence of olfactory dysfunction following recovery from the URI is likely explained by direct damage to the olfactory epithelium and olfactory bulb by the virus itself [6].
Tere are a few research studies to explore the treatment of PVOD in the COVID-19 event. Olfactory training is recommended as a frst-line therapy for the treatment of PVOD, with topical corticosteroids, sodium citrate, oral vitamin A, and traditional Chinese medicine considered optional therapies for appropriately selected patients [21][22][23]. However, no data are available on the efcacy of treating postviral olfactory disorders.
Traditional Chinese medicine (TCM) has been used as a healing technique in China for almost 2000 years. Olfactory disorders have been associated with lung-related pathogenesis in the traditional Chinese classic "Te Yellow Emperor's Internal Canon of Medicine" [21]. TCM has accumulated rich literature and case studies, among which treatment methods mainly include TCM decoction, acupuncture, acupoint injection, and the combination of various means [24]. Acupuncture is a form of traditional Chinese medicine that has minimal side efects, is cost-efective, can be easily administered, and may serve as a useful nonpharmaceutical therapy for certain conditions. Recent acupuncture studies have shown favorable results for a variety of conditions within otolaryngology, including allergic rhinitis, chronic rhinosinusitis, tinnitus, sudden sensorineural hearing loss, postviral olfactory dysfunction, dysphonia, and posttonsillectomy pain [25]. Acupoint injection is a common clinical therapy of acupuncture, and its role is widely known and recognized in the treatment of rhinitis. Te acupuncture method of acupoint injection is similar to common acupuncture, but the syringe needle is replaced by the acupuncture needle. Te process of acupoint injection is that the syringe needle is injected vertically without blood refuxing, which generates needle feeling; then, the syringe needle slowly injects the solution under the skin of bilateral acupoints. As a promising method for olfactory dysfunction, TCM has the advantages of internal and external treatment, defnite clinical curative efects, and no obvious adverse reactions [19,[26][27][28][29][30][31].
Few studies have been collected and evaluated on the treatment of postviral olfactory dysfunction. Terefore, this meta-analysis is used to analyze the results of relevant clinical trials and evaluate the efect of TCM on PVOD, which would provide more reliable evidence-based medical evidence for clinical practice.

Types of Participants.
Tis study considered patients of any age with olfactory dysfunction attributed to viral illness without restrictions on race or gender. Te diagnosis of olfactory disorder includes the collection of the patient's medical history, the examination of nasal endoscopy, and the examination of olfactory function.
Studies reporting PVOD with a history of an upper respiratory tract infection without a gap period were included, while olfactory dysfunction must be accompanied by virus infection.
Studies on all other conditions for olfactory loss attributed to previous surgery of the nose and the paranasal sinuses, exposure to toxic substances, congenital diseases, occupying lesions, acute or chronic infammatory nasal disease, and psychiatric factors were excluded.

Types of Intervention
(1) Treatment Interventions. We focused on treatment of traditional Chinese medicine combined with or without conventional treatment. We included the studies that evaluated any type of traditional Chinese medicine, such as traditional Chinese medicine decoction, acupuncture, moxibustion, massage, cupping, and acupoint injection.
(2) Control Interventions. We included studies recruiting patients receiving conventional treatment (western medicine and/or olfactory training, OTDT), no treatment, or placebo control to prevent postviral olfactory dysfunction as control.

Types of Outcomes.
Te primary outcome was the efective rate of postviral olfactory dysfunction in the treatment: (1) Efective rate (T&T): Te T&T olfactometer test was conducted to measure the sense of smell subjectively with fve favors such as fecal odor, fruit, corrupt, coke, and fragrance. Efciency � heal + obviously efective + efective. (2) Treatment success rate (UPSIT): Olfactory function was evaluated using the University of Pennsylvania Smell Identifcation Test (UPSIT) before and after treatment. Te sum of the test results was used as a measure of olfactory function, which allows the grouping of participants as patients (>15 scores), hyposmic (15-35 scores), and normosmic (>35 scores). Treatment success was defned as a score increased by at least four points. (3) Efective rate (ΔVAS): Olfactory function was evaluated using the Visual Analogue Scale (VAS). ΔVAS score � the VAS score before the treatment-the VAS score after the treatment. Heal: the VAS score reached 0∼1 score; obviously efective: the ΔVAS score ≥4 scores; improvement: 2 scores ≤ the ΔVAS score ≤ 3 scores; invalid: the ΔVAS score ≤1 score. Efciency � heal + obviously efective + efective.
Te secondary outcomes included the questionnaires of olfactory disorders (QOD-P), Snifn' Sticks test (TDI score), and the UPSIT score. All side efects and adverse events reported were included as safety outcomes.
(1) QOD-P: the questionnaires of olfactory disorders (parosmia statements) classifed olfactory function objectively into four degrees: P1, P2, P3, and P4 (agree, partly agree, partly disagree, and disagree) according to the score (3, 2, 1, and 0). (2) TDI score: Te TDI score evaluated olfactory function by Snifn's Sticks Test before and after treatment. Odorants were presented in felt-tipped pens; for odor presentation, the cap was removed by the investigator, and the pen's tip was placed in front of the subject's nostrils for approximately 15 seconds. Tis test battery assessed olfactory function bilaterally and involved subtests for odor threshold (T), discrimination (D), and identifcation (I). Te TDI score was used as a measure of olfactory function, which allowed grouping of patients into anosmic (TDI score <15), hyposmic (15 < TDI score < 30), and normosmic (30 < TDI score). (3) UPSIT score: Olfactory function was evaluated using the University of Pennsylvania Smell Identifcation Test (UPSIT) before and after treatment. Te sum score of the test results was used as a measure of olfactory function.

Selection of Studies.
All the authors were trained regarding the purpose and process of the review. Two authors independently performed the study selection by screening the titles and abstracts of all retrieved studies and then by reading through the full text independently to decide eligibility. If any disagreement existed throughout the process, the third author made the fnal decision. Te selection process of selecting the eligible study is shown in a Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) fow diagram ( Figure 1).

Data Extraction and Management.
Two independent reviewers conducted data extraction based on a predesigned data extraction form. Te third author settled any disagreement during the process. According to the recommendation of the Cochrane Handbook, all extracted data and information management were recorded in an Excel extraction form. Te following data were extracted: (1) Basic information of the study: title, frst author's name, year of publication, country, and journal. (2) Participants' characteristics: age, sex, number of participants, disease, inclusion criteria, exclusion criteria, and baseline situation. (3) Interventions: details of TCM (such as treatment methods, sessions, and frequency), treatment duration, study design, randomization, allocation concealment, and blinding methods. (4) Comparators: western medicine or/and olfactory training.
Evidence-Based Complementary and Alternative Medicine (5) Outcomes: measures and primary and secondary outcomes.

Assessment of Risk of Bias.
Te Cochrane Collaboration's 'risk of bias" assessment tool was used to assess the potential source of bias in the included studies. Two authors frst evaluated the risk of bias in eligible studies separately and then cross-checked their fndings. Tis quality assessment was based on random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. We graded the risk of bias in the included trials and classifed it into 3 levels: "high risk of bias," "low risk of bias," and "unclear risk of bias." Te third author resolved any disagreement and made a fnal decision.

Measures of Treatment Efects.
Te available data on treatment outcomes were extracted and meta-analyzed.
Te weighted mean diference (MD) or standardized mean diference (SMD) with 95% CIs was used for continuous data. Te risk ratio (RR) with its 95% CI was used for dichotomous data.

Dealing with Missing Data.
If the presented data of the study were inconsistent or missing, we tried to contact the corresponding author or the relevant author for the required data via email. Otherwise, the study was excluded.
2.4.6. Data Synthesis. RevMan software (V.5.3) was used to complete the data analysis and synthesis by the Cochrane Collaboration. If little heterogeneity existed among the trials, a fxed-efect model was established, while a random-efect model was carried out for signifcant heterogeneity. Dichotomous data were analyzed by the risk ratio (RR) with 95% CIs, and continuous data were analyzed by the mean diference (MD) or standard MD (SMD) with 95% CIs.     2.6. Subgroup Analysis. If obvious heterogeneity existed in a single meta-analysis, subgroup analysis was conducted to analyze the heterogeneity of available data according to the change in characteristics of trial participants, type of traditional Chinese medicine, and type of conventional treatment.

Sensitivity
Analysis. Sensitivity analysis was needed to evaluate the robustness and reliability of the result when suffcient data existed. We conducted the sensitivity analysis in two ways: (1) exclude any of the studies; (2) change the efect model to verify the synthesized result. When a low-quality study was identifed and excluded, the meta-analysis pertained to low heterogeneity. A certain result could determine whether a lowquality study should be included or not. Te fnal result depended on the sample size, missing data, risk of bias, and quality of methods of each study.

Certainty Assessment.
Te Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) system was used to assess the quality of evidence. Two investigators performed the assessment independently and gave a summary of the fnding of the table together. Te third person would be necessary when there was any disagreement between the two investigators.

Ethics and Dissemination.
Since we performed a secondary analysis of the published article, ethical approval was not required. Te results will be published in a peer-reviewed journal and presented at a relevant conference.

Literature Search Results.
A total of 1638 studies were retrieved in the initial search. After removing 316 duplicates, 1322 studies were identifed for further analysis. Trough screening the title and abstract, 1287 studies were excluded because they were literature reviews, case reports, letters, duplicates, or irrelevant studies. Of the remaining 35 studies, 6 studies met our inclusion criteria by reading the full text, as shown in Figure 1.

Characteristics of the Included Studies.
Te basic characteristics of the included studies are listed in Table 1 [32][33][34][35][36][37]. By integrating 6 included studies, a total of 467 cases were reported, with 233 cases in the treatment group and 234 cases in the control group. Te publication period was from 2010 to 2020. Two studies were published in English, and the other 4 studies were published in Chinese. All of the included patients were concentrated in China with a large diference in ages (18-80 years); meanwhile, the overall patients in the 2 groups were mainly balanced in terms of age, mean outcome measures, and the distribution of disease baseline.

Risk of Bias and Assessment of the Quality.
According to the Cochrane Handbook for Systematic Reviews of Interventions, we assessed the risk of bias in the included literature. Te details of the risk of bias (ROB) assessment are provided in Figures 2 and 3. Six RCTs described the appropriate random sequence generation method in detail. Te other 2 RCTs were historical cohort studies in which random sequence generation was assessed to be at high ROB. Tree studies reported allocation with low ROB, and 5 studies did not report the allocation concealment with unclear ROB. Due to the nature of traditional Chinese medicines like acupuncture and acupoint injection not blinding participants and personnel, all studies were assessed to be at high ROB. All of the included studies reported the complete outcome data, and we considered them to be at low ROB. Tree studies reported the details of adverse efects or published protocols with low ROB, while the other 5 studies did not report them with unclear ROB. Six studies reported the patients' baseline characteristics with low ROB based on other sources of bias. Te other 2 studies were judged to be at unclear ROB due to lack of reporting details. Te overall quality of the trials was assessed as a moderate risk of bias.
(3) UPSIT Score. One study [37] showed no signifcant diference in the TDI score between TCM and no treatment (n � 50, MD � 3.44, 95%CI (−1. 36, 8.24), P � 0.16), as shown in Figure 5(c). (4) Adverse Event. Tree studies [34,36,37] investigated adverse events, whereas the other 4 studies did not report any adverse event. Two studies [36,37] showed that no harm or adverse events were experienced during and after the treatment between the two groups. One study [34] reported a total of 9 mild adverse events in 90 cases including dizziness (n � 2) and pain (n � 2) in the treatment group and stomach discomfort (n � 5) in the control group treated with sleeping pills, and adverse events in the two groups were improved after symptomatic treatment. Because of the lack of adequate studies, there exist no signifcant diferences reported between TCM treatment and olfactory training and/or drug therapy on PVOD.

Sensitivity
Analysis. Sensitivity analysis was not available as there were inadequate data in the meta-analysis.

Quality of Evidence.
Te certainty of evidence for the outcomes of the meta-analysis was assessed by the Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) method, and it showed that the quality of evidence for the results of the meta-analysis ranged from "very low" to "low." Since the nature of traditional Chinese medicine as acupuncture and acupoint injection could not meet blinding of participants and personnel, heterogeneity existed in the intervention. Te sample size was small, so some studies showed no signifcant diference between TCM and no treatment. Te main reasons for the downgrade were the small sample size and the unclear risk of bias in the selected studies, as shown in Figure 6.

Discussion
Te pathophysiology of PVOD is not clear, and its methods of treatment consist of corticosteroids, supplementation with vitamin B complex, olfactory training, and others [36]. Studies showed that the human olfactory nervous system Evidence-Based Complementary and Alternative Medicine could be reshaped. Olfactory sensory neurons in the olfactory mucosa had the characteristics of sustainable regeneration lifetime in the nervous system of all vertebrates. Olfactory training could repeatedly stimulate the olfactory epithelium and the olfactory pathway through various olfactory elements so that the damaged olfactory function could be improved or restored [38][39][40][41]. Te systemic glucocorticoid and the Ginkgo biloba extract were recommended, but the curative efect was uncertain [22]. One study showed that the total efective rate of PVOD patients treated with nasal pneumatic spray aerosol-inhaled budesonide suspension was 90%, but most of the patients (65%) did not return to normal [42]. Tere were few efective ways of western medicine to treat PVOD.
Tere are a large number of records about the treatment of "no smell of the nose" documented in traditional Chinese medical classics. A wealth of treatment information has been accumulated, not only on internal clothing but also on various external treatments such as the stufng nose, the blowing nose, nasal irrigation, acupuncture, and massage. Acupuncture and moxibustion theory have been proved efective in postviral olfactory dysfunction [43], and they could improve the olfactory sensitivity of healthy people [44]. Te decoction of TCM, like "Danggui Shaoyao powder" and "Ginseng Yangrong decoction," has been proved more efective than nasal glucocorticoids. 43% of patients treated with Danggui Shaoyao powder and 36% of patients treated with Ginseng Yangrong decoction, respectively, improved olfactory disorders after viral infection [45].
Te aim of this study was to evaluate the efectiveness and safety of TCM in patients with PVOD. Based on the included studies, we pooled the data from 6 studies involving 467 patients for further analysis. Our pooled analysis indicated that, compared with conventional treatment (western medicine or/and olfactory training), TCM seemed to improve postviral olfactory dysfunction measured by  QOD-P and increase the efective rate of PVOD with mild adverse events. As treatment success was defned as a score increase by at least four points, the UPSIT score [37] (RR � 3.44, 95% CI (−1. 36, 8.24), P � 0.16) or the TDI score [43] (RR � 2.10, 95% CI (−1.99, 6.19), P � 0.31) in this study could not meet the standard of treatment success. Due to the small number of included studies, further research needs to enlarge the sample size.
Te common acupoints of the acupuncture for treatment of PVOD in the included studies were LI20, EX-HN8, EX-HN3, LG20, GV23, BL7, LI4, Liv3, SP10, ST36, LG16, LU7, LU9, KD3, BL2, and BiQiu (EX-HN). As extra nerve points, BiQiu (EX-HN) was located at the anterior end of the middle turbinate in the lateral wall of the nasal cavity that preserves the nerve, and sphenopalatine nerve branches were sensitive parts of the nasal cavity and the target organs of allergic rhinitis. Te needle frequency of the treatment of PVOD was 1 time a day for 3 weeks or 3 times a week for 3 months. Te common acupoints of the acupoint injection for treatment of PVOD in this study were LI20, RN22, EX-HN8, and BiQiu (EX-HN). Needle frequency was 3 times a week. 10 times was 1 course with a treatment interval of 3 to 5 days. Te total treatment duration was 3 months.
In our review, there were two common methods of TCM, acupuncture and acupoint injection treating PVOD, which had signifcant improvement in smell and taste function. It is believed that TCM treatment could reestablish an equilibrium for a diseased body [36]. According to Chinese acupuncture guidelines, therapy should be individualized to restore physical balance and to bring yin and yang into equilibrium.
Te injection solution of acupoint injection in this study was 0.1∼0.2 ml lidocaine mixed with 1 ml vitamin B12. Te injection was 0.5 ml per side. Needle frequency was 3 times a week. 10 times was 1 course with a treatment interval of 3∼5 days. Te total treatment duration was 3 months. Acupoint injection therapy is a method of injecting an appropriate amount of liquid medicine into specifc acupoints, combining the function of acupoints and the mechanical efect of acupuncture with the pharmacological efect of drugs to prevent and treat diseases. It has the characteristics of simple operation, safety, small dosage, rapid action, and good acceptance by patients [46].
Te quality of evidence ranged from low to high based on the Cochrane Bias Evaluation Tools and the Jadad scale. Among the 6 studies included, 1 study was of high quality, 3 studies were of moderate quality, and 2 studies were of low  Evidence-Based Complementary and Alternative Medicine 9 quality. Te quality of the included studies was generally moderate. Te inappropriate random method, allocation concealment, and lack of blinding in most studies exaggerated the result of the outcome measures.
No statistical heterogeneity was found in the efective rate of the efectiveness of TCM treating PVOD, which may be because all the included studies explored the acupointstimulating therapy without traditional Chinese herbs. Missing data on Chinese herbal medicine, despite its widespread clinical use, reveal an evidence gap.
Furthermore, these studies had problems such as low quality, the faw in test design, small sample sizes, or the variety of the intervention and control groups. To a certain extent, these indicated that the results of this meta-analysis were afected by the risk of bias.
Tere were several limitations in this systematic review and meta-analysis. Te methodological quality of most included studies was relatively low, which resulted in the potential risk of bias. To some extent, it also weakened the credibility and reliability of the evidence of TCM for the treatment of PVOD in this systematic review and meta-analysis. For example, due to few control groups being placebo control or false TCM, it was difcult to eliminate the placebo efect. Since the nature of traditional Chinese medicine like acupuncture and acupoint injection could not meet the blinding of participants and personnel, it might lead to potential implementation bias and selection bias.
Traditional Chinese medicine has little or no side efects, and it is cost-efective and easily administered. Regardless of the potential bias and limitations of this review, all of the included studies showed that TCM intervention seemed to have signifcant efects in improving PVOD. As an efective adjunct treatment, TCM intervention has turned out to be a superior complementary and alternative theory, which has potential efcacy and caused few adverse efects.

Conclusion
In this systematic review and meta-analysis of RCTs, the efectiveness of traditional Chinese medicine in the treatment of postviral olfactory dysfunction is positive. TCM has the potential to become a more efective therapy than conventional treatment (western medicine or/and olfactory training). Because of the lack of adequate studies, there exist no signifcant diferences in the adverse reactions reported between TCM treatment and conventional treatment of POVD.
In order to improve the credibility of the research and better guide clinical practice, well-designed, rigorous, large sample, and multicenter prospective RCTs are needed. Te long-term efectiveness of TCM intervention in PVOD must also be evaluated.

OD:
Olfactory dysfunction URI: Upper respiratory infection PVOD: Postviral olfactory dysfunction TCM: Traditional Chinese medicine CNKI: China Network Knowledge Infrastructure VIP: Chinese Scientifc Journal Database CBM: Chinese Biomedical and Medical Database RCTs: Randomized controlled trials VAS: Visual Analogue Scale QOD: Questionnaires of olfactory disorders OTDT: Olfactory training and drug therapy UPSIT: University of Pennsylvania Smell Identifcation Test TDI score: Snifn's Sticks Test assessed olfactory function bilaterally and involved subtests for odor threshold (T), discrimination (D), and identifcation (I).

Disclosure
Fangfang Ma and Hewei Zhang are the co-frst authors.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
FFM, HWZ, and XMW provided the concept and design. FFM and HWZ searched the literature databases. BXL and PYC selected eligible studies and extracted the data. FFM and HWZ evaluated the risk of bias. FFM drafted the frst version of the manuscript. XMW and MWY were responsible for the preliminary revision of the manuscript. FFM and XMW supervised the project. All the authors reviewed the manuscript.