Chronic cough is associated with reduced mental and physical health and increased health care resource use. The lack of nationwide data on chronic cough hinders our understanding of the prevalence, demographics, and impacts of chronic cough in China. The aim of this study was to estimate the lifetime and annual prevalence of chronic cough in mainland China and to describe the sociodemographic and health-related characteristics of adults with chronic cough.

This was a cross-sectional study using self-reported information from a nationally representative sample of 20,051 adults. Respondents with chronic cough (daily cough for ≥ 8 weeks) during the prior 12 months were matched to respondents without chronic cough.

We estimated a lifetime prevalence of 7.0% and an annual prevalence of 4.1% for chronic cough among adult residents of mainland China. Affected individuals had a mean age of 43.5 years, and there were no significant differences in prevalence between men and women. Compared to matched controls, the chronic cough group had more comorbidities and included a higher proportion of individuals who had been diagnosed with respiratory and sleep-related conditions. Chronic cough was also associated with significantly worse depression and anxiety symptom severity and health-related quality of life, as well as increased impairment of work and other daily activities and higher rates of all-cause health care resource use.

Chinese adults with chronic cough show a middle age, equal gender distribution. Chronic cough affects an estimated 46.4 million adults in mainland China during their lifetime and is associated with significant individual and health care system burden.

Chronic cough affects approximately 4.1% (27.2 million) of adults per year in mainland China and is associated with worse health and health-related quality of life, impairment of work productivity and other daily activities, and increased all-cause health care resource use.

Cough lasting > 8 weeks (chronic cough) is a debilitating condition associated with poor physical and mental health, reduced work productivity and quality of life, and increased health care resource use [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27]. Many cases resolve via treatment of common underlying causes such as asthma, gastroesophageal reflux disease (GERD), upper airway cough syndrome, and eosinophilic bronchitis, or cessation of prescription drugs that can cause cough as a side-effect [28,29,30]. However, other cases remain unexplained and/or refractory to treatment. The current evidence-based Chinese clinical practice guidelines for the management of cough recommend treatment of underlying medical comorbidities, consideration of multiple concurrent causes and occupational exposures such as airborne particulates, and empirical treatment (i.e., treatment selection based on the physician’s best judgment) of unexplained/refractory cough with Western and/or traditional Chinese medicines [29].

The estimation of the global and country-specific prevalence of chronic cough has been hampered by the use of differing methodologies and definitions [26, 31]. The global prevalence of chronic cough has been estimated as 9.6% [32], but other studies have estimated lower rates (3.8–5.5%) in the US and Western European countries [11, 12, 15, 33, 34]. Studies of the demographics of chronic cough in China have identified interesting potential differences compared to other national populations. A recent meta-analysis assessed 35 studies of the prevalence of chronic cough in China, conducted in different regions using different methodologies and definitions of chronic cough, and estimated an overall prevalence of 8.8% among adults ≥ 18 years of age when only studies defining chronic cough as daily cough for > 8 weeks were considered [35]. Most of the included studies had small sample sizes and a low-to-moderate grade of evidence.

The primary objective of the current study was to use a nationally representative sample of adults in mainland China to estimate the lifetime and annual prevalence of chronic cough. We also aimed to compare the sociodemographic characteristics and the health-related characteristics and outcomes of adults with and without chronic cough, providing comprehensive nationwide data on the demographics and burden of this condition in China.

This was a cross-sectional study using self-reported information from a national online survey.

The study population comprised residents of China ≥ 18 years of age who had completed the 2020 National Health and Wellness Survey (NHWS), a self-administered online questionnaire that is conducted annually in China and other countries. Studies using similar methods have described the prevalence and characteristics of chronic cough in nine other countries and regions that participate in the NHWS (Table 1) [11, 12, 15, 34, 36,37,38,39]. The NHWS uses a targeted quota sampling framework to select panelists who mirror the urban adult population of mainland China by age, binary sex, and region of residence using a demographic distribution based on data from the National Bureau of Statistics of China [40]. The annual sample size for the NHWS in China is approximately 20,000.

All NHWS respondents were asked whether they had experienced chronic cough (defined as daily cough for ≥ 8 weeks) at any time during their life, and also specifically during the prior 12 months. To compare individuals with and without chronic cough during the prior 12 months, respondents who had experienced chronic cough were matched (nearest neighbor without replacement) at a 1:3 ratio to respondents with no chronic cough who had similar (≤ 0.2 SD) chronic cough propensity scores. Propensity scores were calculated based on age, sex, interaction of marital status and household income, and a modified Charlson Comorbidity Index (CCI) score. The modified CCI score excluded chronic obstructive pulmonary disease (COPD) due to its close association with chronic cough.

All NHWS participants read an informed consent statement and provided electronic confirmation of their consent to participate. Respondents who provided consent were assigned a unique code that can only be linked to personal information by a team within Oracle Life Sciences that was not involved in the current study. The NHWS protocols and questionnaires were reviewed and granted an exemption by the Pearl Institutional Review Board (Protocol #20-KANT-228).

The NHWS collects information on respondents’ sociodemographic and health-related characteristics, all-cause health care resource use (e.g., number of visits to general and specialty health care providers from a list provided) and other health-related outcomes, and diseases experienced and diagnosed. It also includes several validated patient-reported outcome instruments. Depression severity was assessed via the Patient Health Questionnaire 9-item scale (PHQ-9) [41], and anxiety symptoms using the Generalized Anxiety Disorder 7-item scale (GAD-7) [42]. Health-related quality of life was assessed using 3 instruments: the Medical Outcomes Study 12-Item Short Form Survey Version 2 (SF-12v2; QualityMetric Inc., Lincoln, RI, USA) [43,44,45,46]; the SF-6D Health Utility Score [45, 46]; and the EuroQol 5-Dimension Health Questionnaire (EQ-5D-5 L^™; EuroQol Research Foundation) [47]. Finally, the Work Productivity and Activity Impairment Questionnaire assessed health-related absenteeism (percent work hours missed), presenteeism (percent impairment experienced while at work), and overall work productivity loss (combined absenteeism and presenteeism score) as well as activity impairment (percent impairment of daily non-work activities) during the prior 7 days [48].

The unweighted lifetime and annual prevalence of chronic cough were calculated as the number of respondents indicating they had ever experienced daily cough for ≥ 8 weeks (or had experienced daily cough for ≥ 8 weeks within the last 12 months, respectively), divided by the total number of NHWS respondents. Lifetime and annual prevalence calculations were also performed stratified by age category, binary sex, smoking status, and sex stratified by smoking status. To adjust for potential sampling biases, weighted lifetime and annual prevalence estimates were calculated by projecting the unweighted prevalence rates from the study sample to the full urban adult population of China using data from the National Bureau of Statistics in China [40]. Sampling weights were calculated using rake weighting / random iterative methods and were based on age, sex, region, and rurality [49].

For comparisons between respondents who had experienced chronic cough during the prior 12 months and matched respondents who had not (matching strategy described in ‘Study population’ section above), the matched groups were first compared using kernel density plots to visually compare the propensity score groupings and ensure that their characteristics were balanced (Supplementary Fig. 1). Respondent characteristics and outcomes for each group were then analyzed using descriptive statistics including percentages, means, SDs, medians, and ranges. Characteristics were compared between the matched groups using bivariate analyses (t-tests for continuous variables and chi-square tests for categorical variables). Differences with two-tailed p-values of < 0.05 were considered statistically significant.

A total of 20,051 adult residents of China completed the 2020 NHWS, of whom 1,389 (6.9%) reported having experienced chronic cough (daily cough for ≥ 8 weeks) during their lifetime and 812 (4.0%) during the prior 12 months (Table 2). These proportions were weighted to the 665.8 million total adult national population to estimate a lifetime prevalence of 7.0% (~ 46.4 million individuals) and an annual prevalence of 4.1% (~ 27.2 million individuals) for chronic cough in China.

Current or former smokers of cigarettes or other tobacco products had a significantly higher weighted annual prevalence of chronic cough compared to respondents who had never smoked (7.0% versus 3.1%; P < 0.001). The weighted annual prevalence of chronic cough was also statistically significantly different between age categories (P < 0.001). The distribution was bimodal, with the highest proportions observed for individuals ≥ 75 years of age (7.4%) and 30–39 years of age (5.0%). A similar trend was found with different age groupings for 50 years and above ( 50–59, 60–69, ≥ 70 years) (data not shown). The weighted annual prevalence of chronic cough was similar for male versus female respondents (4.2% versus 3.9%). Stratified by smoking status, the unweighted annual prevalence of chronic cough was higher among female than that among male current/formal smokers (10.2% versus 6.3%), however among never smokers, no sex difference was found for the annual chronic cough prevalence (Supplementary Table 3).

Eleven respondents with chronic cough during the prior 12 months had no qualifying matches, 9 had 1 match, 4 had 2 matches, and the remainder had 3 matches. The matched analyses therefore compared 801 respondents with chronic cough to 2,381 respondents without chronic cough.

Respondents who had experienced chronic cough within the prior 12 months had a mean (SD) age of 43.5 (15.9) years and a median (range) age of 41.0 (18.0–112.0) years, and 51.3% were male (Table 3). Continuous age was used as a matching criterion, but the age group distribution was different between the matched groups (P < 0.001). A majority of respondents with chronic cough had a medium or high income (68.5%) and were married or cohabiting (79.3%). Significantly higher proportions of respondents with chronic cough had ≥ 2 sources of health insurance compared to the control group (78.2% versus 61.4%) and a primary insurance type of Urban Resident Basic Medical Insurance (39.8% versus 31.9%), while a significantly lower proportion had a primary insurance type of Urban Employee Basic Medical Insurance (39.0% versus 44.1%; P < 0.001 for all comparisons).

A significantly greater proportion of respondents with chronic cough had ≥ 1 comorbidity compared to matched controls (37.2% versus 24.7%; P < 0.001; Table 4); this group also had a significantly higher mean (SD) unmodified CCI (0.56 [0.94] versus 0.37 [0.79]; P < 0.001; Fig. 1). Experiencing chronic cough was also statistically significantly associated with being a current or former smoker (42.3% of respondents with chronic cough versus 27.3% of respondents with no chronic cough) and consuming alcohol (≤ 3 times per week, 60.0% versus 45.7%; ≥4 times per week, 8.7% versus 6.0%; P < 0.001 for all comparisons; Table 4).

Health-related characteristics of respondents with chronic cough, compared to a matched control group. CCI, Charlson Comorbidity Index; COPD, chronic obstructive pulmonary disorder; GAD-7, Generalized Anxiety Disorder symptom scale; PHQ-9, Patient Health Questionnaire depression symptom scale. Respondents who reported having chronic cough in the last 12 months were matched 1:3 by chronic cough propensity score (based on age, gender, interaction of marital status and income, and modified CCI, excluding COPD) with respondents with no chronic cough in the last 12 months. Values are displayed as the mean of the continuous scores for each measure. P-values were calculated using the two-tailed t-test. All values were statistically significantly different (P < 0.001) between the chronic cough (N = 801) and the no chronic cough (N = 2,381) groups, except modified CCI (p = 0.134)

Full size image

Individuals with chronic cough had significantly worse mental health than did matched controls. Among the chronic cough group, the mean (SD) score on the PHQ-9 scale of depression symptom severity was 7.48 (5.04) and the mean (SD) score on the GAD-7 scale of anxiety symptom severity was 5.80 (3.99), versus 4.49 (4.66) and 3.42 (3.63), respectively, for the matched control group (P < 0.001 for both comparisons; Fig. 1). Significantly more respondents with chronic cough also reported experiencing moderate-to-severe symptoms of depression (30.2% versus 13.2%; P < 0.001) and anxiety (18.2% versus 7.5%; P < 0.001; Table 4).

Respondents were also asked whether they had ever experienced or been diagnosed with specific medical conditions; they were not asked whether these conditions occurred before or after they experienced chronic cough. Significantly greater proportions of the chronic cough group had ever been diagnosed with each of the respiratory conditions on the list provided (P = 0.002 for emphysema and P < 0.001 for allergies, asthma, hay fever, chronic bronchitis, and COPD) as well as heartburn (P < 0.001) and other conditions (Supplementary Table 1). Significantly more respondents with chronic cough also reported having ever experienced or being diagnosed with each of the sleep-related conditions on the list provided, i.e., insomnia, narcolepsy, sleep apnea, and other sleep difficulties (P = 0.031 for the percent diagnosed with narcolepsy and P < 0.001 for all other comparisons; Fig. 2). Chronic cough was also significantly associated with regularly experiencing each of the specific sleep problems on the list provided (P = 0.005 for difficulty staying awake and P < 0.001 for all other listed problems; Supplementary Table 2).

Percent of respondents with chronic cough A) ever experiencing sleep-related conditions during the prior 12-month period and B) ever diagnosed with sleep-related conditions, compared to a matched control group. Respondents who reported having chronic cough in the last 12 months were matched 1:3 by chronic cough propensity score (based on age, gender, interaction of marital status and income, and modified Charlson Comorbidity Index, excluding chronic obstructive pulmonary disease) with respondents with no chronic cough in the last 12 months. P-values were calculated using the two-tailed chi-square test. All values were statistically significantly different between the chronic cough (N = 801) and the no chronic cough (N = 2,381) groups (p = 0.031 for the percent diagnosed with narcolepsy; P < 0.001 for all other comparisons)

Full size image

Chronic cough was associated with significantly worse health-related quality of life. Among respondents with chronic cough, the mean (SD) score on the SF-12v2 scale was 46.75 (7.53) for the physical component and 44.89 (7.86) on the mental component (a clinically meaningful difference compared to the accepted population benchmark mean of 50), versus 51.11 (7.31) and 49.11 (8.14), respectively, among matched controls (P < 0.001 for both comparisons; Table 5). The mean (SD) SF-6D health utility score was significantly lower in the chronic cough group, at 0.64 (0.09) compared to 0.73 (0.13) for the control group (P < 0.001)—again, a clinically meaningful difference on this scale. Respondents with chronic cough also had significantly lower mean (SD) scores on the EQ-5D index (0.84 [0.19] versus 0.93 [0.13]; P < 0.001) and the EQ-5D VAS (74.85 [18.89] versus 83.20 [17.19]; P < 0.001) than did matched controls, indicating a worse health state. The chronic cough group also reported a significantly greater extent of impairment of daily activities, as measured by the mean (SD) percent work absenteeism (10.47 [15.25] among those with chronic cough versus 4.77 [12.34] among those without), work presenteeism (36.70 [27.22] versus 21.27 [25.92]), overall work impairment (40.87 [28.83] versus 23.33 [27.28]), and impairment of non-work-related daily activities (34.89 [26.04] versus 21.36 [24.88]; P < 0.001 for all comparisons).

Compared to matched controls, significantly greater proportions of the chronic cough group had used in the prior 6 months, for any reason, each of the health care resources on the list (P < 0.01 for emergency room visits, hospitalizations, and all specialty and non-specialty health care provider visits; Fig. 3). Chronic cough was also associated with significantly more all-cause visits per person to each of the health care resources on the list provided, except for neurologist visits (P < 0.001 for all comparisons).

Health care resource use within the prior 6 months by respondents with chronic cough, compared to a matched control group. ER, emergency room; HCP, health care provider. Respondents who reported having chronic cough in the last 12 months were matched 1:3 by chronic cough propensity score (based on age, gender, interaction of marital status and income, and modified Charlson Comorbidity Index, excluding chronic obstructive pulmonary disease) with respondents with no chronic cough in the last 12 months. P-values were calculated using the two-tailed chi-square test. All values were statistically significantly different between the chronic cough (N = 801) and the no chronic cough (N = 2,381) groups (p = 0.004 for the percent visiting a psychiatrist during the last 6 months; P < 0.001 for all other comparisons). Inset shows health care resource use by specialty

Full size image

The nationally representative nature of the study population in the present manuscript allowed us to more accurately estimate the prevalence of chronic cough and characterize the demographics of chronic cough among Chinese adults. This study estimated that 4.1% of adults in mainland China (~ 27.2 million people) have experienced chronic cough during the prior 12 months, and 7.0% (~ 46.4 million) during their lifetime. The prevalence of chronic cough was similar among men and women. Compared to matched controls, significantly higher proportions of adults reporting chronic cough in the prior 12 months had experienced respiratory, sleep-related, and other conditions; the chronic cough group also had significantly more comorbidities. Respondents with chronic cough also had a lower health-related quality of life, worse depression and anxiety, more impairment of work and other activities, and higher rates of all-cause health care resource use.

The estimated 4.1% weighted annual prevalence of chronic cough was lower than the estimated overall prevalence reported by a recent meta-analysis, which was 8.8% among studies that defined chronic cough as daily cough for > 8 weeks [35]. However, most of the included studies had small sample sizes and a low-to-moderate grade of evidence [35]. The current study population was larger and more nationally representative than that of most of the studies included in the prior analysis, and prevalence was weighted to the national population to minimize sampling bias. Further, the estimated weighted annual prevalence was lower than the estimated three-year average probability (8.9%) reported from an administrative claims database study among Chinese adults from nine cities in 2015, 2016, and 2017 [50]. However, the prevalence estimated herein is also broadly similar to the findings of equivalent studies that analyzed NHWS data from other countries, which reported a weighted annual prevalence that ranged from 4.3% in Japan and South Korea to 6.3% in Italy and a lifetime prevalence that ranged from 6.2% in South Korea and the UK to 9.2% in Italy (Table 1) [11, 12, 15, 34, 36,37,38,39].

The nationally representative nature of the study population also allowed us to characterize the demographics of chronic cough among Chinese adults. Globally, chronic cough is generally associated with older age, with a peak prevalence at ≥ 50 years of age [51, 52]. In the current study, the numerically highest prevalence of chronic cough was observed among individuals 30–39 and ≥ 75 years of age, and the mean age of 43.5 years among individuals with chronic cough was lower than in studies that used NHWS data from other countries (US, 50.1 years; Germany, 52.1 years) [12, 15]. Studies of the demographics of chronic cough in China have identified some interesting potential differences compared to other national populations. Our findings are consistent with previous reports that Chinese populations with chronic cough are younger than equivalent populations from other countries, and that there may be a peak prevalence of chronic cough during middle age in China [8, 22, 51, 52]. However, in a study that surveyed a nationally representative sample of approximately 51,000 Chinese adults, Huang et al. reported a mean age of 52.0 years among a group of 1,985 individuals with chronic cough, which is more consistent with the demographics of the condition in Western nations [5]. The use of different definitions of chronic cough (≥ 8 weeks in the current study versus ≥ 3 months in Huang et al.; the latter definition was based on that of chronic bronchitis) may explain this discrepancy, highlighting the importance of consistent definitions of chronic cough.

Further, studies in other countries have consistently found that chronic cough is strongly associated with female sex, potentially due to a more sensitive cough reflex among women [11, 12, 15, 26, 31, 51,52,53,54,55,56,57]. However, as comprehensively reviewed by Bai et al. [51], almost all studies that have reported the sex distribution of chronic cough in Chinese populations have identified either a weaker female preponderance than is seen in most studies from other countries, a roughly equal prevalence among men and women, or even a slight male preponderance [5, 10, 17,18,19, 22, 24, 26, 27, 58]. Our finding that men and women had a similar prevalence of chronic cough may be explained by the higher prevalence of smoking among men compared to women. In an additional analysis, we found that the proportion of current and former smokers was five times higher among men than women. However, the prevalence of chronic cough among male smokers was lower than that among female smokers. Among never smokers, we found no difference in the prevalence between men and women (data not shown). Smoking could be an effect modifier which at least partially contributed to the observed sex differences in China compared to the global pattern of a strong female preponderance [5, 8, 10,11,12, 17,18,19, 22, 24, 39, 51, 52, 58]. The smoking distribution in additional analyses varied by age and sex, suggesting that the overall demographics of chronic cough in China may be due to a higher prevalence of the condition among younger men than is seen elsewhere in the world [8, 22, 51]. Our analysis thus contributes further evidence that the sex distribution of chronic cough in China differs from that observed in many Western nations.

Our study confirms previous reports from China and other countries that chronic cough is more common among individuals with a history of smoking cigarettes [5, 11, 12, 15, 20, 22, 23]. However, more than half (57.7%) of the current study population who reported experiencing chronic cough had never smoked, which is consistent with other reports from China [5, 8, 22]. It has previously been suggested that the distinct demographics of chronic cough in China may be driven at least partially by cigarette smoking, since Chinese men (and particularly young men) have one of the highest rates of cigarette smoking in the world while Chinese women have one of the lowest [8, 22, 51, 59, 60]. However, our findings are consistent with previous reports suggesting that other factors are also involved. For example, in the current study, chronic cough was significantly associated with lower-coverage forms of health insurance (i.e., Urban Resident rather than Urban Employee insurance), which may reflect lower socioeconomic status and less access to health care [61]. Air pollution and occupational or residential exposures, which were not assessed in the current study, have also been suggested as potential risk factors for chronic cough in China, which has one of the highest levels of air pollution in the world; however, the relationship between these variables and chronic cough remains poorly understood [16, 21, 29, 51, 62,63,64].

Our study comprehensively assessed the health-related characteristics and outcomes of adults with chronic cough. Adults with chronic cough had significantly worse health and health-related outcomes than matched controls, and higher proportions of this group had experienced health-related impairment of sleep, work, and other daily activities. They also reported more frequent use of a range of specialty and non-specialty health care resources. These findings add to the body of evidence that chronic cough represents a serious health burden in China in terms of physical and mental health, quality of life, work productivity, and health care resource use [5, 9, 10, 22, 29]. Our results are supported by a retrospective observational study of a nationwide claims database in China, which concluded that chronic cough is common among medical insurance users, leading to substantial utilization of medical resources and posing a significant burden on healthcare in China [50].

This study used validated instruments to comprehensively survey a large, nationally representative sample of the adult population of mainland China. Our matching strategy rigorously compared adults who had experienced chronic cough within the prior 12 months to a control group of survey respondents who had not. The assessment of a general population sample rather than the specialist samples (e.g., cough clinic attendees) included in many previous Chinese studies allowed us to generate a broader analysis of the characteristics of people living with chronic cough in China and of the national health-related burden of the disease for individuals and the health care system. Our findings bolster the existing literature reporting a distinct demographic pattern in China that merits further study.

However, this study is also subject to some known limitations. Although we used a sampling and weighting strategy designed to minimize sampling bias, some degree of sampling bias may remain, which may affect the accuracy of the weighted prevalence estimates for the full national population. The NHWS likely under-represents people without access to or comfort with online surveys, including less healthy older adults, institutionalized individuals, and individuals with severe health conditions or disabilities. Study measures were self-reported and may be subject to recall and other biases; to mitigate these issues, the survey questions focused on recent timeframes (e.g., prior week or month) when possible and the list of answers to potentially sensitive questions included “don’t know” and “prefer not to answer” options. Further, the data presented in this study are derived from the 2020 fielding of the NHWS, which was administered during the earliest stages of the COVID-19 pandemic. As research on the SARS-CoV-2 virus continues to evolve, its impact on the prevalence of chronic cough remains uncertain [65, 66]. However, we believe the impact of the pandemic was minimal as strict lockdown measures implemented across the country likely limited exposure to respiratory infections and environmental triggers.

CCI:

Charlson Comorbidity Index

COPD:

Chronic obstructive pulmonary disease

EQ-5D-5L:

EuroQol 5-Dimension Health Questionnaire

GAD-7:

General Anxiety Disorder 7-item scale

GERD:

Gastroesophageal reflux disease

NWHS:

National Health and Wellness Survey

PHQ-9:

Patient Health Questionnaire 9-item scale

SF-12v2:

Medical Outcomes Study 12-item Short Form Survey Version 2

WPAI:

Work Productivity and Activity Impairment

SF-6D:

Short-Form 6-Dimension

Download references

Conception, design or planning of the study: E.F.Interpretation of the results: K.L., V.W.L., L.C., K.X., A.M., N.A.W., and H.D. Drafting of the manuscript: H.D.Critically reviewing or revising the manuscript for important intellectual content: K.L., V.W.L., L.C., K.X., A.M., N.A.W., H.D., and E.F. All authors approve the final version to be published and agree to be accountable for all aspects of the work.

Correspondence to Helen Ding.

All NHWS participants read an informed consent statement and provided electronic confirmation of their consent to participate. Respondents who provided consent were assigned a unique code that can only be linked to personal information by a team within Oracle Life Sciences that was not involved in the current study. The NHWS protocols and questionnaires were reviewed and granted an exemption by the Pearl Institutional Review Board (Protocol #20-KANT-228. KN-NHWS-NC20 The 2020 China National Health and Wellness Survey Patient Centered Research Program).

Not applicable.

Kefang Lai has received honoraria from AstraZeneca, Chiesi, Circassia, Daiichi Sankyo (China), GSK, Merck & Co., Inc., Shionogi Inc., and Novartis; he has received research funding and other support from AstraZeneca, Chiesi, Circassia, Daiichi Sankyo (China), GSK, Merck & Co., Inc., and Novartis. Vicky W Li and Nate A Way are employees of Oracle Life Sciences, Seattle, WA, USA which received funding to conduct this study from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co, Inc., Rahway, NJ, USA. Ashley Martin was an employee of Oracle Life Sciences, Seattle, WA, USA which received funding to conduct this study from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co, Inc., Rahway, NJ, USA, at the time of the study. Helen Ding and Eileen Fonseca are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA and shareholders in Merck & Co., Inc., Rahway, NJ, USA. Lin Chen and Kai Xia are employees of MSD China and shareholders in Merck & Co., Inc., Rahway, NJ, USA.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions