Clinical and Genetic Characteristics of BCG Disease in Chinese Children: a Retrospective Study

Summarize the characteristics of a large cohort of BCG disease and compare differences in clinical characteristics and outcomes among different genotypes and between primary immunodeficiency disease (PID) and patients without identified genetic etiology. We collected information on patients with BCG disease in our center from January 2015 to December 2020 and divided them into four groups: chronic granulomatous disease (CGD), Mendelian susceptibility to mycobacterial disease (MSMD), severe combined immunodeficiency disease (SCID), and gene negative group. A total of 134 patients were reviewed, and most of them had PID. A total of 111 (82.8%) patients had 18 different types of pathogenic gene mutations, most of whom (91.0%) were classified with CGD, MSMD, and SCID. CYBB was the most common gene mutation (52/111). BCG disease behaves differently in individuals with different PIDs. Significant differences in sex (P < 0.001), age at diagnosis (P = 0.013), frequency of recurrent fever (P = 0.007), and vaccination-homolateral axillary lymph node enlargement (P = 0.039) and infection severity (P = 0.006) were noted among the four groups. The CGD group had the highest rate of males and the oldest age at diagnosis. The MSMD group had the highest probability of disseminated infection (48.3%). The course of anti-tuberculosis treatment and the survival time between patients with PID and without identified genetic etiology were similar. Greater than 80% of BCG patients have PID; accordingly, gene sequencing should be performed in patients with BCG disease for early diagnosis. BCG disease behaves differently in patients with different types of PID. Patients without identified genetic etiology had similar outcomes to PID patients, which hints that they may have pathogenic gene mutations that need to be discovered.


Introduction
Bacille Calmette-Guérin (BCG)-a live, attenuated strain of Mycobacterium bovis-is the only recommended vaccine by the World Health Organization (WHO) against tuberculosis, and neonates born in areas with high tuberculosis incidences, such as China, should receive this vaccine [1]. From 1950, China began to vaccinate children with BCG for free, requiring them to be vaccinated at birth, no later than 3 months old [2]. Danish strain is commonly used in China. BCG is generally safe, helps reduce the risk of severe tuberculosis, and prevents transmission [3,4]; however, some children tend to suffer from BCG disease with an incidence of approximately 1:10,000-1:1,000,000 or even disseminated BCG disease, which is a rare and serious adverse reaction [5] . In 2015, Aishwarya Venkataraman et al. observed sixty children who presented with adverse reactions and found that two-thirds (65%) presented with BCG lymphadenitis, one-third (30%) presented with injection site complications, and 54% had received anti-tuberculous therapy and/or a procedure [6].
Yuyuan Zeng and Wenjing Ying contributed equally to this work.
With the increasing use of gene sequencing, some pathogenic genes have been discovered in some BCG disease [7]. Rina Yue Ling Ong et al. reported that 10 patients likely had underlying PID-four with SCID, three with MSMD, one with anhidrotic ectodermal dysplasia with PID (EDA-ID), one with combined immunodeficiency (CID), and one with a STAT1 gain-of-function mutation [8] . A previous study enrolled 74 confirmed cases of BCGosis/BCGitis in our center in 2014. Thirty-two patients (43.2%) had definitive PID, and CGD was the most common PID (n = 23, accounting for 71.9% of all PID patients) [7] . In 2014, Beatriz E Marciano et al. found that 51% of 349 BCG vaccinated SCID patients from 28 centers in 17 countries developed BCG complications [9] .
Limited updated clinical data on large-sized samples of BCG disease have been reported, and differences among different genotypes and between patients with PID and without identified genetic etiology have not been described. We wanted to determine whether patients without identified genetic etiology had similar characteristics and prognoses to patients with PID. Therefore, this study aims to analyze the clinical and genetic characteristics of the largest cohort of BCG disease in China and compare differences in clinical manifestations, infectious severity, and outcomes among different genotypes and between patients with PID and without identified genetic etiology.

Methods
The study was approved by the ethics committee of the Children's Hospital of Fudan University.

Patients
Initially, all patients who were diagnosed with BCG disease were enrolled. We collected data on year of birth, age at onset, age at diagnosis, gender, residence, clinical manifestation, family history, body mass index (BMI), site of BCG disease, pathogen of coinfection, genetic results, treatment, and survival time. Patients who met the enrollment criteria were included (Table 1). According to the classification criteria reported previously, the infection severity of the enrolled patients was divided into three grades (Table 1) [10] . PID was diagnosed based on a virtual panel including 269 genes published by the International Union of Immunological Societies associated including all known inborn error of immunity (IEI), and if no genetic diagnosis was established, whole exon sequencing (WES) was performed. Panel lists and result of functional verification were provided as Appendix (see Supplementary materials 2 and 3). According to different types of gene mutations, we divided the recruited patients into four groups: the CGD group, MSMD group, SCID group, and gene negative group. Here, we had 3 patients with AR IFNGR1 deficiency and 1 AR IFNGR2 deficiency; unfortunately, we did not distinguish between complete and partial forms because functional experiments were not performed. The other genetic etiologies lacked homogeneity. Thus, we did not compare these patients with the above four groups in our study.

Statistical Analysis
SPSS 26.0 (SPSS Inc., Chicago, IL) was used to perform data analysis, and GraphPad Prism 5.0 was used Table 1 The enrollment criteria and classifications for Bacillus Calmette-Guérin (BCG) disease to generate figures. The Kolmogorov-Smirnov test was used to test the normality of the data, and numerical data are presented as the mean and SD or median. Categorical variable data are reported as frequencies and percentages. ANOVA and Kruskal-Wallis test were used to compare the numerical variables. The chi-square test and Fisher's exact test were used to compare the categorical variables depending on applicable conditions. The Kaplan-Meier method and life table analysis were applied to describe and calculate probabilities of survival among different groups. A two-tailed P < 0.05 was considered statistically significant.

General Features of All Patients
From January 1, 2015, to December 31, 2020, a total of 134 patients who were diagnosed with BCG disease were enrolled. All of the patients had a clear history of BCG inoculation before the onset of the disease, and scars were visible at the vaccination site. Most of these patients were vaccinated within 3 days after birth (Table 2). Among all of the patients, 101 were male (75.4%). Most patients were from East China (n = 88 65.7%) around Shanghai followed by South China (n = 22 16.4%) and Southwest China (n = 12 9.0%). Almost all patients were lean, and the average BMI was 16.3 ± 1.8 (kg/m 2 ). The median age at onset of BCG disease of all patients was 3.0 months old (IQR 1.67-4.0). The median age at diagnosis of BCG disease of all patients was 5.5 months old (IQR 4.5-8.6). Thus, a delay in BCG disease diagnosis was noted.

Clinical Manifestation in Different Groups
Ten other genetic mutations were lacked of homogeneity; in this study, we did not compare these patients with others. According to different types of gene mutations, we divided the remaining 124 patients into CGD group (54 cases), MSMD group (29 cases), SCID group (18 cases), and gene negative group (23 cases). The sex composition ratio of the four groups was significantly different (P < 0.001). Almost all patients in the CGD group were males. The median ages at onset in the CGD group, MSMD group, SCID group, and gene negative group were 3.0 months old (IQR 0.5-4.0), 2.5 months old (IQR 1.5-3.0), 3.3 months old (IQR 3.0-4.0), and 3.3 months old (IQR 2.0-4.8), respectively. No significant difference existed among them (P > 0.05). The median age at diagnosis of the four groups was 7.0 months old (IQR 5.0-10.3), 4.5 months old (IQR 3.9-8.1), 6.0 months old (IQR 4.5-8.0), and 5.5 months old (IQR 4.0-6.5). Significant difference was noted among the four groups, and the CGD group had a significantly older age at diagnosis than the other groups (P = 0.013) ( Table 3). Every group experienced an evident delay in the diagnosis of BCG disease, especially the CGD group.
The most common clinical manifestations of the four groups were recurrent fever, abnormal vaccination site, and vaccination-homolateral axillary lymphadenitis (Fig. 2). Patients also presented with cough, diarrhea, skin abscess, night sweats, abdominal pain, and hematochezia. Symptoms varied among the four groups. The CGD group and SCID group had a significantly higher rate of recurrent fever than the other groups (P = 0.007). Ratio of vaccination-homolateral axillary lymph node enlargement was significantly higher in the MSMD group (P = 0.039). There was no significant difference in the frequency of abnormal vaccination sites among the four groups (P > 0.05) ( Table 4). Patients in the MSMD group manifested abnormal gait (2/28), which was due to BCG bone infection.
A significant difference in infection severity was noted among the four groups (P = 0.006). In the MSMD group, 12/29 (41.4%) were regional infection, 3/29 (10.3%) were At birth Vaccination-homolateral axillary lymph node Regional disease 10 At birth Vaccination-homolateral axillary lymph node Regional disease 11 At birth Vaccination-homolateral axillary lymph node, respiratory system, bone marrow Disseminated disease 12 At birth Site of inoculation Regional disease 13 At birth Site of inoculation Regional disease 14 At birth Vaccination-homolateral axillary lymph node Regional disease 15 At birth Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 16 At birth Site of inoculation Regional disease 17 At birth Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 18 At birth Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 19 At birth Site of inoculation and vaccination-homolateral axillary lymph node, respiratory system Extraregional localized disease 20 Two months old Vaccination-homolateral axillary lymph node Regional disease 21 At birth Vaccination-homolateral axillary lymph node Regional disease 22 At birth Vaccination-homolateral axillary lymph node Regional disease 23 At birth Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 24 At birth Vaccination-homolateral axillary lymph node Regional disease 25 At birth Vaccination-homolateral axillary lymph node Regional disease 26 Four months (premature delivery) Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 27 At birth Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 28 At birth Vaccination-homolateral axillary lymph node Regional disease 29 At birth Site of inoculation Regional disease 30 At birth Site of inoculation Regional disease 31 At birth Vaccination-homolateral axillary lymph node, digestive system Extraregional localized disease 32 At birth Vaccination-homolateral axillary lymph node Regional disease 33 At birth Vaccination-homolateral axillary lymph node, digestive system Extraregional localized disease 34 At birth Site of inoculation, skin and soft tissue Extraregional localized disease 35 At birth Vaccination-homolateral axillary lymph node Regional disease 36 At birth Site of inoculation and vaccination-homolateral axillary lymph node, digestive system Extraregional localized disease 37 At birth Site of inoculation Regional disease 38 At birth Vaccination-homolateral axillary lymph node Regional disease At birth Vaccination-homolateral axillary lymph node Regional disease 42 At birth Vaccination-homolateral axillary lymph node Regional disease 43 At birth Vaccination-homolateral axillary lymph node Regional disease 44 At birth Vaccination-homolateral axillary lymph node Regional disease 45 At birth Vaccination-homolateral axillary lymph node Regional disease 46 At birth Site of inoculation and vaccination-homolateral axillary lymph node, skin and soft tissue  At birth Vaccination-homolateral axillary lymph node Regional disease 85 At birth Vaccination-homolateral axillary lymph node Regional disease 86 Fifteen months old (born in Greece) Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 87 At birth Vaccination-homolateral axillary lymph node Regional disease 88 At birth Vaccination-homolateral axillary lymph node Regional disease 89 At birth Site of inoculation Regional disease 90 At birth Site of inoculation Regional disease 91 At birth Site of inoculation, skin and soft tissue, bone marrow Disseminated disease 92 At birth Vaccination-homolateral axillary lymph node Regional disease 93 At birth Site of inoculation Regional disease 94 At birth Site of inoculation Regional disease 95 At birth Site of inoculation Regional disease 96 At birth Vaccination-homolateral axillary lymph node Regional disease 97 At birth Site of inoculation and vaccination-homolateral axillary lymph node Site of inoculation and vaccination-homolateral axillary lymph node Regional disease 114 At birth Vaccination-homolateral axillary lymph node Regional disease extraregional localized infection, and 14/29 (48.3%) were disseminated infection. In the CGD group, SCID group, and gene negative group, the proportion of regional infection was 12/29 (41.4%), 14/18 (77.8%), and 13/23 (56.5%), respectively. We found that regional infection was the most common type. However, the incidence of disseminated infection Site of inoculation and vaccination-homolateral axillary lymph node Regional disease Fig. 1 Genotype of all PID patients and classification of PID according to the genotype was significantly increased in the MSMD group, with 6/54 (11.1%) in the CGD group, 3/18 (16.7%) in the SCID group, and 5/23 (21.7%) in the gene negative group. It is worth noting that the probability of developing disseminated infection in the gene negative group which was second only to that in the MSMD group (Table 5).

Coinfections in Different Groups
Apart from BCG infection, these patients often experienced recurrent infections caused by various pathogens, even in patients without identified genetic etiology. These infectious agents were speculated to lead to disease quite likely according to patients' clinical manifestation and treatment effect. Tested samples were obtained from nasal mucosa, sputum, bronchoalveolar lavage fluid (BALF), pus of skin, stool, urine, gastric juice, ascites, pleural fluid, blood steam, and bone marrow. The common coinfections included fungi (43.5%); viruses, such as Cytomegalovirus (17.9%), Rotavirus (11.3%), human parainfluenza virus (10.5%), and Epstein-Barr virus (7.3%); bacteria, such as Klebsiella (16.1%), Staphylococcus (7.5%), Enterococcus   Fungal infections were very common in the four groups. The patients in the CGD and MSMD groups were mostly accompanied by bacterial infection, and the patients in the SCID and gene negative groups were mostly accompanied by viral infection (Table 6).

Treatment and Prognosis in Different Groups
The median course of anti-tuberculosis treatment for all children was 7.0 months old (IQR 0.6-15). The MSMD group had the longest anti-tuberculosis treatment course at 34.5 months (IQR 9.8-51.8). The CGD group, SCID group, and gene negative group had treatment course durations of 8.0 months old (IQR 0.0-15.0), 5.5 months old (IQR 0.0-11.0), and 3.0 months old (IQR 3.0-6.5), respectively. No significant difference in anti-tuberculosis treatment courses was noted among the four groups (P = 0.09). Twentytwo cases in the CGD group had completed hematopoietic stem cell transplantation (HSCT), of which 1 case was lost to follow-up, 18 cases were successfully transplanted (81.8%), and 3 cases were unsuccessful. In addition, 14 cases in the SCID group had completed HSCT, and 12 cases were successfully transplanted (85.7%). There was no significant difference in the success rate of transplantation between the two groups (P = 1.000) ( We compared HSCT − patients in the gene negative group with HSCT − patients in the other three groups, respectively, in the survival rate. One-year survival rate of the four groups was 95.0%, 100.0%, 100.0%, and   (Fig. 4).

Discussion
BCG is used to prevent severe tuberculosis, and the WHO recommends that newborns born in areas with high tuberculosis incidence should be vaccinated [11][12][13]. The incidence of severe tuberculosis has been reduced after BCG vaccination, but BCG disease still occurs occasionally [7] and can even be fatal [14][15][16]. In recent years, there have been few large-sample updated clinical data about BCG disease, and no studies have focused on differences in BCG disease in different types of PID and between patients with PID and without identified genetic etiology. This retrospective study presents and statistically analyzes the clinical and genetic features from the largest BCG disease cohort to date in China. We reported a total of 134 cases of BCG disease from January 2015 to December 2020 and divided the recruited patients into the four groups because CGD, MSMD, and SCID accounted for the most patients and were representative.
According to previous reports, patients suffering from BCG disease often have PIDs, such as SCID, MSMD, CID, and CGD [17][18][19]. In Iran in 2017, 64.0% of patients with disseminated disease also had PID [19]. Of these patients, 62.5% were classified as CGD, 25.0% were SCID, and 12.5% were MSMD. A 15-year retrospective review in Singapore showed that all disseminated BCG diseases had PID, including SCID, MSMD, anhidrotic ectodermal dysplasia with PID (EDA-ID), and CID [8]. In our study, 82.8% of the patients has a definite genetic diagnosis, and 48.6% of them were CGD. Compared with the report from our center in 2014, the rate of PID diagnosis significantly increased from 43.2 to 82.8% [17] which is related to the gradual popularization of gene sequencing technology in recent years. It also suggested that if patients can be screened for these diseases before BCG inoculation, a large proportion of serious BCG disease will be prevented.
The median age at onset was reported previously as 3.8 months old in Singapore [8] and 3.6 months old in our center previously [7]. In this study, however, the age of onset was advanced to 3.0 months old. This finding may be attributed to the fact that the previous study recruited patients with CGD, while the present study included more patients with MSMD and SCID. However, there was still a delay in diagnosis, especially in the CGD group. The median age at diagnosis was 7.0 months old in the CGD group, which was significantly later than that of the other group. Physicians should further raise awareness of this issue. A previous study showed that no case had only local infection, 49.4% of patients had a regional infection, 26.6% of patients had a distant infection, and 17.7% of patients had a disseminated infection [2]. Most of the patients in the study had regional disease, which was similar to previous results. The MSMD group had the highest probability of disseminated infection, suggesting that we may adopt more aggressive anti-tuberculosis treatment for MSMD patients.
Previous reports have shown that axillary lymphadenitis is the most common clinical manifestation [1] . Overall, 85-90% of patients have regional lymphadenitis, and 10.2% of patients have BCG-related osteomyelitis. Recurrent fever, axillary lymphadenitis, and abnormal vaccination sites were the main clinical manifestations in each group, which were consistent with previous reports. If the patient has a clear history of BCG vaccination and presents with the above three manifestations, vigilance for BCG disease should be raised.
It is worth noting that the survival time of gene negative group was not significantly different from that of the other three groups, and this group had a high probability of disseminated infection. We hypothesize that there may be undiscovered gene mutations in these patients, which is worth further research.
Overall, we reviewed the largest cohort of BCG disease in China and found that greater than 80% of patients with BCG disease had PID. Thus, gene sequencing should be performed in patients with BCG disease to facilitate early diagnosis, which will help us distinguish most patients with PID. BCG disease behaves differently in patients with different types of PID, and this information could help us manage patients more carefully. Patients without identified genetic etiology may have pathogenic gene mutations that deserve further study.