Objective:To summarize the clinical characteristics, common images, pathogens, and gene mutation types of chronic granulomatosis disease (CGD) in 19 children.Methods:The clinical manifestations, laboratory findings, treatment, and prognosis of 19 patients diagnosed with CGD in Hong Kong University-Shenzhen Hospital from December 2012 to December 2018 were analyzed.Results:The 19 patients were all males and confirmed as CGD by the dihydrorhodamine test and gene sequencing.The age of the first infection was mostly 1 month after birth(13 cases), and the age of clinical diagnosis ranged from 2 months to 10 years.Sixteen mothers were carriers.The patients presented with pulmonary fungal infection (19/19 cases), Bacillus Calmette Guerin (BCG)-osis (14/19 cases), lymphadenitis (14/19 cases), perianal abscess (9/19 cases), skin abscess (5/19 cases) and ulcerative colitis (2/19 cases). There were 59 positive cultures.Pathogens included fungi (9 cases), Klebsiella pneumonia (8 cases), mycobacteria (7 cases), Streptococcus Viridans (5 cases), Escherichia coli (3 cases), gram-positive bacteria (3 cases), Staphylococcus aureus (3 cases), and Burkholderia cenocepacia (2 cases). Gene mutations were found in all 19 patients, including 17 cases of CYBB, 1 case of CYBA and 1 case of NCF2.The type of mutations included nonsense mutations (6 cases), deletion mutations (5 cases, including 2 large fragment deletions), splice mutations (3 cases) and missense mutations (5 cases). Five mutations were novel.Splice mutations in 3 cases often led to skin abscess, perianal abscess and lymphadenitis.Two patients with large deletion mutations had more serious infection than other patients. Conclusions:In China, CGD is characterized with pulmonary infection and disseminated BCG-osis.Mycobacteria are common pathogens of CGD, and fungi are dominant pathogens of CGD.The most common infection is respiratory infection. Klebsiella pneumonia and Escherichia coli often lead to perianal abscess.The relationship between gene mutation types and clinical phenotypes requires further verification by big data.