Nocardiosis is a rare but potentially serious infection, which is mainly considered to be involved in ISPs [5, 6]. However, recent studies showed that nocardiosis could also occur in ICPs with increasing rates [7, 8]. Delay in nocardiosis diagnosis may be due to its nonspecific and variable clinical characteristics and nonspecific radiologic findings . Thus comparing the characteristics of nocardiosis patients with different immune status could help in early diagnosis of nocardiosis.
The detailed comparison of characteristics of nocardiosis between ISPs and ICPs from tertiary general hospitals was still limited. Kim et al’s study  compared the clinical characteristics and treatment outcomes between ICPs and ISPs, however, not all Nocardia strains were identified to species levels and not all the strains were available for the antibiotic susceptibility results, mainly because these microbiological data were retrospectively collected, and some of them were unavailable. Thus a total of 23 cases of culture proven, non-repetitive nocardiosis patients from our hospital, a 1500-bed tertiary general hospital in Beijing China, during 8-year period were characterized in our study, and we performed the Nocardia species identification by DNA sequencing and antibiotic susceptibility testing by E-test method with all the available strains at the same time, instead of collecting these data retrospectively. Then these patients were divided into ISPs and ICPs groups according to the previously defined criteria [10, 11], based on their clinical characteristics. Our study showed that the 39.1% of these nocardiosis patients were ICPs and most of them recovered. Meanwhile, the overall outcome in both ICPs and ISPs were better than previous studies [7, 8].
From the data of our study, we found that most demographic features, underlying diseases, microbiological identification, and their antibiotic susceptibility were similar between the 2 groups. However, there was only 14.3% (2/14) disseminated nocardiosis patients (the detailed description of the two cases was in the notes of Table 1) in ISPs group, and it was significantly fewer than Kim et al’s study , which showed the disseminated cases accounted for 33% in ICPs and 36% in ISPs, respectively. It could be explained by the severity and sites of infection of included patients may be different between the 2 studies, and the nocardiosis patients in our study seem to be less severe and have better outcomes.
Among the underlying diseases we analyzed in this study, an interesting association was found between bronchiectasis and nocardiosis in both ISPs and ICPs, with the rates of 44.4% and 21.4%, respectively, which was similar with Woodworth’s study of 40% . Moreover, they found the increasing number of nocardiosis over time was driven by patients with non-cystic fibrosis bronchiectasis. We also analyzed the incidences of nocardiosis cases among different years in both ICPs and ISPs (Figure 1), however due to the small sample size, we found the numbers in each year went up and down with no obvious trend over time, which was different from Woodworth’s study , while the studies from Taiwan  and Spain  showed stable case numbers over time. The difference may be explained by various population and underlying diseases among different studies.
The outcome of nocardiosis patients was better in ICPs than in ISPs, in other words, their prognosis was associated with various immune statuses. We found that the 3 cases of death was attributed to severe infections, but Nocardia infections were not the only reason, as they were combined with other co-infections (e.g. co-infection with aspergillosis, virus, or bacteria was found in our study). The situation was reasonable, as assessment of attributable mortality in the context of complex conditions was often challenging and inaccurate . The cases of nocardiosis co-infection with aspergillosis in our study were more than previous study , mainly in ISPs, as these patients were more prone to mixed infection that caused by multiple pathogens, which should be notice by physicians.
The microbiological identification of Nocardia to species level showed that, N. farcinica, N. cyriacigeorgica, N. otitidiscaviarum, N. abscessus, N. brasiliensis distributed equally in ICPs. In contrast, N. farcinica and N. cyriageorgica were the predominant species in ISPs, accounting for 42.9% and 35.7%, respectively, which was different from the species distribution in United States  and Taiwan . The antimicrobial susceptibility testing of first-line drugs for Nocardia was also compared between 2 groups (Table 2), and found there was no significant difference between 2 groups. However, the resistance profiles were also different from other reports [2, 25, 26], which could be explained by the geographic variation to some extent.
Our study had several limitations. Firstly, it was the study from one tertiary hospital in Beijing, China, thus our results could not stand for other geographic regions of China. Secondly, the sample size of nocardiosis patients that met the criteria for inclusion in this study was small, compared with studies from Korea  and United States , which may indicate the low incidence of nocardiosis in our hospital. Thus we could not get the statistically significant results in some clinical characteristics, although we could see the difference between the 2 groups of patients (e.g. bronchiectasis occurred more frequently in ICPs, while chronic kidney diseases and co-infection with aspergillosis occurred more frequently in ISPs, although they did not reach statistical significance). Meanwhile, due to the limited sample size, we could not determine statistically whether the observed difference was attributed to compromised immune status of ISPs, and further studies with larger sample size are still needed. Thirdly, due to the retrospective nature of the study, sample selection bias could not be avoided.
The advantage of this study was that, due to the limited data of comparison of nocardiosis in ICPs and ISPs from China, our study collected the detailed data of clinical characteristics, and performed Noardia species identification and antibiotic susceptibility testing, then compared these data between the 2 groups. Thus our study added new value in the characterization of nocardiosis in China, and showed different characteristics from other geographic regions. Understanding the geographic distribution and characteristics of nocardiosis in both ISPs and ICPs would help physicians diagnose and treat nocardiosis more effectively.
In conclusion, our study showed that nocardiosis could occur in both ISPs and ICPs. Skin and soft tissue infection and bronchiectasis occurred more frequently in ICPs than in ISPs. Chronic kidney diseases and co-infection with aspergillosis occurred more frequently in ISPs than in ICPs, which should be noticed and considered by physicians in the diagnosis of nocardiosis.