Cystic fibrosis (CF) is the most frequent monogenic disease in Caucasian populations, with incidences ranging from 1/1800 to 1/25000.1 2 According to the American Cystic Fibrosis Foundation registry, there are currently more than 100 000 CF patients throughout the world. China is the world’s most populated country with a population of approximately 1.4 billion. However, up to now, only approximately 110 CF patients of Chinese origin were reported in literature,3 and most of publications were case reports, without any epidemiological data on the prevalence available. Interestingly, most of the cases have been diagnosed in the last 5 years, which suggests that the actual incidence of CF in China may be seriously underestimated. The contributing factors may include underdiagnosis, under-reporting, lack of national registries and variability in the frequency of mutation carriers. It has been identified that the profiles of CF transmembrane conductance regulator (CFTR) gene mutation spectrum vary widely among different populations based on their geographic and ethnic origins. Although CF in China is being increasingly recognised, there is still an urgent need to understand the genetic spectrum of the CFTR gene in Chinese patients, which will greatly contribute to prevention, diagnosis and even future molecular genetic management of the disease.

In this study, we describe 103 children with CF receiving care at the main referral centre in China. Meanwhile, other 99 genetically confirmed cases with Chinese origin reported worldwide from 1993 to 2022 were also summarised. To our knowledge, this is the largest study and most comprehensive analysis of genotypic features of CF in Chinese population to date.

A total of 103 patients (40 males and 63 females) from 100 Chinese families presented to Beijing Children’s Hospital were recruited into this study. All 103 patients were of Chinese origin, and none of them had a family history of intermarriage with Caucasians. Patients no. 30-1/no. 30-2, no. 31-1/no.31-2 and No. 76-1/no.7622 were siblings respectively both suffering from CF. Only two children (case no. 30-1/no. 30-2) from the same family were the products of consanguineous marriage. The mean (±SD) ages at CF diagnosis were 7.6 (±4.4) years. Eighty-four children accepted sweat conductivity analysis; among them, 78 children had positive results and six of them showed intermediate values. The mean value was 114.1 mmol/L ranging from 60.0 to 168.0 mmol/L (online supplemental e-table 1).

In addition, other 99 Chinese patients from 94 different families described in 45 references were included in this study (Mainland China (82 cases), 4 6 9–41 Taiwan area (nine cases),42–47 Hong Kong (five cases),48 Australia (one case),49 Canada (one case)50 and USA (one case)).51 Overall, a total of 202 CF patients of Chinese origin constituted the cohort for the current analysis.

The CFTR gene variants identified in all the 103 patients (100 families) presented to Beijing Children’s Hospital are listed in online supplemental e-table 1. Of note, cases 1~194 and cases 20~316 (except for patient no. 30-1 and no. 31-1) have been reported in 2016 and 2020, respectively, by the authors. The remaining cases have never been reported so far. Twenty-three novel observations were identified after sequencing (c.222delG[p.Arg75AspfsX16], c.298C>T[p.Leu100Phe], c.464C>G[p.Ala155Gly], c.940G>T[p.Gly314Trp], c.1064C>G[p.Pro355Arg], c.1219G>T[p.Glu407X], c.1265C>T[p.Ser422Phe], c.1347_1350delAGAA[p.Arg450AspfsX18], c.1368delT[p.Ala457LeufsX12], c.1393-?_1584+?del, c.1514delA[p.Asn505IlefsX22], c.1523_1534delTTGGTGTTTCCT[p.Phe508_Ser511del], c.1772T>C[p.Val591Ala], c.1810A>C[p.Thr604Pro], c.2042A>T[p.Glu681Val], c.2058_2061delTT [p.Phe687X], c.2328dupA[p.Val777SerfsX2], c.2489dupA[p.Glu831GlyfsX5], c.2909-?_3468+?del, c.3140-?_3367+?del, c.3469-12T>G, c.3469-2A>T and c.3659C>T[p.Thr1220Ile]).

The spectrum of CFTR variants detected in all Chinese patients with CF (194 families, 388 alleles) was summarised in online supplemental e-table 2. As a result, 373 mutated alleles were detected (detection rate: 96.1%). The identified variants included 56 missense, 31 nonsense, 27 frameshift, 21 splicing, 11 large insertion/deletion, 8 sequence variation and 4 in-frame insertion/deletion. Overall, 158 different variants were identified after sequencing. The majority of CFTR variants (82.3%, 130/158) in Chinese have been observed only once or twice. Approximately half of the variants (43.7%, 69/158) were only identified in patients of Chinese origin thus far, and to our knowledge, they have never been reported in Caucasians (online supplemental e-table 2). The c.2909G>A(p.Gly970Asp) was found to be the most frequent variant among Chinese CF patients with the highest allele frequency of 12.1% (47/388), followed by c.1766+5G>T (5.4% (21/388)) and c.1657C>T[p.Arg553X] (3.6% (14/388)), which were the second and third most common variants, respectively (online supplemental e-table 2, figure 1). They also showed significant differences in geographical distribution, that is, the c.2909G>A(p.Gly970Asp) variant was found in the Northern and Eastern China, while the c.1766+5G>T and the c.1657C>T(p.Arg553X) variants were most common observed in the Southern and Eastern coasts (figure 2). Meanwhile, c.1521_1523delCTT(p.F508del) was observed in six patients of pure Chinese origin, with an allele frequency of 1.8% (7/388). Interestingly, two de novo variants (c.960dupA[p.Ser321IlefsX43] and c.2491-2A>G) and two deep-intronic variants (c.3718-2477C>T and c.3874-4522A>G) were identified, which were also quite rare among Chinese.

Frequency (>1%) of CFTR variants observed among 202 Chinese CF patients. CF, cystic fibrosis; CFTR, CF transmembrane conductance regulator.

Frequency distribution of the top three most frequent variants among 202 Chinese patients with CF. CF, cystic fibrosis.

Herein, to the best of our knowledge, we report results of the most comprehensive analysis to date of genotypic features associated with CF patients originating from across China. The first CF patient of Chinese origin was reported in 1975, which was diagnosed by sweat test.52 Then, it was not until 1993 that there was the first genetically confirmed case.46 From 1993 to 2022, a total of 202 CF patients of Chinese origin have been diagnosed with definite CFTR variants. In addition, 158 different variants of CFTR gene were identified, including 23 novel observations (current report). Of these, only 45 variants are known to be CF causing in CFTR2.53 In addition, 50 loss-of-function variants (nonsense, frameshift and large insertion/deletion) are likely to be CF causing. However, the pathogenic significance for the remaining 63 variants is unknown, and further functional elucidation is necessary.

The variant spectrum of CFTR among Caucasians in Western countries has been well established. c.1521_1523delCTT(p.F508del) is the most frequent variant in Caucasians, accounting for approximately 70% of mutated alleles in general.54 However, it is quite rarely seen in Asia, especially East Asia.55 The c.1521_1523delCTT (p.F508del) was only observed in six patients of Chinese origin (one in homozygosity and five in compound heterozygosity), with an allele frequency of only 1.8%. No cases have ever been reported in other East Asian countries so far (except for mixed Asian-Caucasian parentage). By contrast, among Chinese population, the majority of CFTR variants (82.3%, 130/158) have been observed only once or twice. Approximately half of the variants (43.7%, 69/158) were only identified in patients of Chinese origin thus far. The c.2909G>A(p.Gly970Asp) and c.1766+5G>T variants were the most predominant observations, occurring in 12.1% and 5.4% of the alleles among all the reported Chinese patients, respectively. Notably, both variants show significant Chinese ethnic tendency, because to our knowledge, most cases with them were reported among Chinese. The c.1657C>T(p.Arg553X) was the third most commonly observed variants, with an allele frequency of 3.6%. Interestingly, the c.1657C>T(p.Arg553X) was also present in the panel of 23 mutations proposed by the American College of Medical Genetics and Genomics.56 It has been associated with Central European-derived populations, and the clinical consequence of this variant is known to be CF causing.57 In terms of geographic distribution, the c.2909G>A(p.Gly970Asp) variant was found in the Northern and Eastern China, while the c.1766+5G>T and the c.1657C>T(p.Arg553X) variants were most common observed in the Southern and Eastern coasts (figure 2). We presume that the different distribution of these variants may be due to the different ethnic groups in different parts of China. Eastern coast is considered the most prosperous region of China’s economy and trade, with more frequent population migrations from both Northern and Southern areas. That may be a possible explanation that the top three most frequent variants were all found in the Eastern area. The first description of c.1657C>T(p.Arg553X) among Chinese was made in homozygosity genotype in a native Taiwanese boy diagnosed with CF. Chen et al 45 proposed that the occurrence of c.1657C>T(p.Arg553X) variant in Taiwan area may correspond to the colonisation by the Dutch and Spanish 300 years ago. Interestingly, in this study, half of the c.1657C>T(p.Arg553X) variants were found in Zhejiang Province, which located on the southeastern coast of China. Historically, since China established diplomatic relations with European countries in the 1950s, Zhejiang people immigrated to Europe through Macao and Hong Kong. At present, Zhejiang migrants live throughout Europe. Based on this, we speculate that the origin of the c.1657C>T(p.Arg553X) variant most likely came from Europeans, which may be due to the intermarriage between Taiwanese and Europeans during the colonial period, or the wave of European economic immigration after the founding of new China.

Variants occurring de novo in the CFTR gene are extremely rare, with approximately 10 cases of de novo CFTR variants published to date.58 There was an interesting finding that two de novo variants (c.960dupA[p.Ser321IlefsX43] and c.2491-2A>G) were identified among Chinese, which were confirmed after paternity test as well as CFTR gene screening for the biological parents. Furthermore, both de novo variants were found in Beijing Children’s Hospital. The insertion c.960dupA(p.Ser321IlefsX43) has been reported in 2016,4 while the splicing variant c.2491-2A>G was a recent observation. Among previous reports, the only description of the c.2491-2A>G was made in an Irish CF patient, without any additional clinical data available.59 In the present study, c.2491-2A>G was the first time observed as a de novo variant in a 11-year-old Chinese boy (case no. 68) with severe sinopulmonary manifestations, liver disease and pancreatic insufficiency (PI), who bore the c.2491-2A>G/c.3196C>T genotype disease. Compared with inherited variants, de novo variants are probably more deleterious because they have been subjected to less stringent evolutionary selection.60 Most often, the de novo variants appeared on the paternal chromosome, and the same observation was also shown for c.24912A>G in our patient. Casals et al 61 proposed that this tendency may reflect a higher mutation rate in paternal gametes. Nevertheless, the c.960dupA (p.Ser321IlefsX43) earlier found in our study was located on the maternal chromosome. The reason for this is unknown.

The CFTR genotype remains incomplete in 1% of CF cases, deep-intronic variants are putative candidates to fill this gap.62 A collection of variants in non-coding regions of the CFTR gene could help to assess their potential role as genetic factors that modify the phenotype. To our knowledge, only seven deep-intronic disease-causing variants have been identified in the CFTR gene until now.62 Interestingly, two deep-intronic variants (c.3718-2477C>T[c.3717+12 191C>T] and c.3874-4522A>G) were the first time identified among Chinese patients. The c.3718-2477C>T has been associated with multiethnicity-derived populations, including Ashkenazi-Jewish, Southern European, Middle Eastern, Iranian and Indian.57 In addition, it belongs to class V mutations in the CFTR2 database, which result in reduced synthesis of CFTR protein with normal function at the epithelial cell membrane.57 63 In our report, the c.3718-2477C>T variant was found in two Chinese CF children (case no. 44 and case no. 83) with typical pulmonary features but without PI. The results of sweat conductivity test were intermediate (64 mmol/L) and weak positive (87 mmol/L), respectively, which were consistent with previous reports and may be considered milder CF phenotype. As far as c.3874-4522A>G was concerned, the ethnic origins were reported to be France, Iran and Laos.62 Moreover, this variant was considered CF causing associated with a large phenotypic spectrum, including CFTR-related disorders and typical CF.62 In the present study, c.3874-4522A>G was found in two Chinese children (case no. 39 and case no. 57) with severe sinopulmonary diseases from early childhood, including progressive bronchiectasis, recurrent airway Pseudomonas aeruginosa, lung function defect and sinusitis (both cases), and allergic bronchopulmonary aspergillosis (only for case no. 39). Due to the limited sequencing methods at the first time they presented to us, we only detected the c.2936A>C(p.Asp979Ala) and the c.1368delT(p.Ala457LeufsX12) variant on one allele, respectively. Three years later, we found the c.3874-4522A>G variant deeply located in intron 23 of the other allele and eventually supplemented the genetic data for both children.

In terms of the novel observations, 12 variants (c.222delG[p.Arg75AspfsX16], c.1219G>T[p.Glu407X], c.1347_1350delAGAA[p.Arg450AspfsX18], c.1368delT[p.Ala457LeufsX12], c.1393-?_1584+?del, c.1514delA[p.Asn505IlefsX22], c.1523_1534delTTGGTGTTTCCT[p.Phe508_1521Ser511del], c.2058_2061delTT[p.Phe687X], c.2328dupA[p.Val777SerfsX2], c.2489dupA[p.Glu831GlyfsX5], c.2909-?_3468+?del and c.3140-?_3367+?del) are likely to be CF causing, and the clinical consequences of remaining 11 variants are uncertain. Clinical characterisation includes the presence of high frequency of sinopulmonary diseases and relatively low frequency of PI compared with Caucasians, which were consistent with previous reports on phenotype in individuals with CF of Chinese origin (online supplemental e-table 3). The geographical distribution of the novel variants showed no significant discrepancies between North and South in China.

The genetic spectrum of CF in Chinese is unique and quite different from that observed in Caucasians. Therefore, the Caucasian CFTR common mutation-screening panel is not applicable for Chinese patients. We recommend using the extensive CFTR gene sequencing (including all CFTR exons, their intronic boundaries and selected regions of deep introns) followed by MLPA analysis for effective diagnosis, although it is relatively expensive. If patients with CF are still incompletely genotyped, whole genome sequencing for the identification of unknown deep-intronic variants is advocated.

To date, 2110 variants of CFTR have been identified worldwide,64 but the disease liability of only 401 variants has been ascertained in CFTR2.53 Based on the nature of the molecular defects in CFTR, small molecules (CFTR modulators) can successfully restore activity to the mutant protein, thereby ameliorating disease manifestations.55 Unfortunately, few studies on the molecular consequences of Chinese-specific CFTR variants have been reported. The most frequent CFTR mutation in Chinese, c.2909G>A(p.Gly970Asp), was predicted to be a gating mutation with partial trafficking defect.65 Furthermore, lumacaftor/ivacaftor therapy was proven efficacy in both ex vivo65 and in vivo66 studies of a patient with c.1521_1523delCTT(p.F508del)/c.2909G>A(p.Gly970Asp) genotype. These findings have significant implications for CFTR modulator use in or availability to the Chinese population. Future studies on functional consequences analysis of Chinese ethnicity-specific variants would be beneficial to take the first step to research on CFTR modulator therapies in China.

The limitation of the present study would be that the sweat chloride concentration test, which has been the gold standard for diagnosis of CF, was not available in China. However, the sweat conductivity measurement has been shown to have excellent correlation with the sweat chloride concentration.67 68 It is accurate, simple to perform and economical. Thus, the sweat conductivity measurement is used as an assistant diagnostic test for CF in China. As the patients in this study were enrolled over a period of decades, this could have cause unintended bias such as improvement in laboratory technology and lifestyle changes. First, sweat conductivity analysis is not readily available in all the paediatric centres in China, except Beijing Children’s Hospital, which is considered the main referral centre for patients with CF from all over the country. Besides this, even in our centre, sweat conductivity analysis was not performed on all the patients enrolled because the testing facility was only available from 2014. Finally, MLPA analysis was only adopted in patients recruited from 2016 onwards also due to inaccessibility of this technique.

In conclusion, results presented herein describing the genetic spectrum of CF in Chinese is unique and quite different from that observed in Caucasians, consistent with our previous statement.4 The c.2909G>A(p.Gly970Asp), c.1766+5G>T and c.1657C>T(p.Arg553X) are the most frequent variants among Chinese CF patients studied. The geographical distributions of the most frequent variants were reported for the first time. These data demonstrate that it is very urgent and necessary to establish a national CF registry in China, which would be beneficial to compare genetic data intranationally and internationally from an epidemiological perspective, to evaluate the phenotype–genotype association in China and to enhance our understanding of CF pathogenic processes for improvement of disease management and prognosis.