Prevalence and characteristics of thoracic diffuse idiopathic skeletal hyperostosis in 3299 black patients

The purpose of this study was to examine the prevalence and characteristics of thoracic diffuse idiopathic skeletal hyperostosis (T-DISH) in the Black patients using the computed tomography (CT) analysis. This study is a cross-sectional study. All patients who underwent chest CT for the trauma screening and whose race was categorized as “Black” on the questionnaire were recruited in the study from Mar 2019 to Mar 2020. Demographic data, including age, sex, body mass index (BMI), and presence of diabetes mellitus (DM), were recorded. A total of 3299 Black patients (1507 women and 1792 men) were included for the analysis. The prevalence of T-DISH was 7.7% (255 patients), with 8.6% for females and 7.0% for males. The highest prevalence was observed in patients at the age of 70 years (11.7%), followed by the age of 80 years (10.5%). The highest prevalence level of T-DISH segment was at T8, followed by T9, and T7. The most frequent number of contiguous vertebrae was seven (21%). BMI was not associated with T-DISH. The presence of DM was significantly higher in male patients with T-DISH than those without T-DISH (P = 0.02).

). The overall prevalence of T-DISH was 7.7%, with 8.6% for females and 7.0% for males. (Fig. 1). No difference was found in the prevalence of T-DISH between female and male patients (8.6% vs 7.0%, P = 0.1). The mean age of patients with T-DISH was significantly higher than that without T-DISH (62.1 vs 57.2, P < 0.001).
With regard to the prevalence of T-DISH among each 10-year age group, the highest prevalence was observed in patients at the age of 70 years (11.7%), followed by the age of 80 years (10.5%), and the age of 60 years (8.1%) (Fig. 2). The distribution of T-DISH in the thoracic segments is shown in Fig. 3. The highest prevalence level of T-DISH segment was at T8 (94%), followed by T9 (93%), and T7 (89%). The number of contiguous vertebrae of T-DISH is shown in Fig. 4. The most frequent number of contiguous vertebrae was seven (21%), followed by four (18%), and five (18%).
Association between T-DISH and BMI/DM. There was no difference in the mean BMI between patients with and without T-DISH (27.5 vs 27.8, P = 0.76) ( Table 2). No difference was found in the mean BMI between female patients with and without T-DISH (27.6 vs 27.5, P = 0.60), and between male patients with and without T-DISH (27.3 vs 28.0, P = 0.37). There was no difference in the presence of DM between patients with and without T-DISH (36.9% vs 33.7%, P = 0.34) ( Table 2). No difference was found in the the presence of DM between female patients with and without T-DISH (26.4% vs 29.8%, P = 0.48). The presence of DM was significantly higher in male patients with T-DISH than those without T-DISH (47.6% vs 37.0%, P = 0.02).

Discussion
Our study revealed that the prevalence of T-DISH in the Black patients was 7.7%, who underwent chest CT for the trauma screening. The prevalence of T-DISH has been reported to be 2.6-17% [5][6][7]9,10,14 (Table 3). However, it is difficult to compare with the prevalence of T-DISH in the previous epidemiological literature due to the differences in the diagnostic modality, diagnosis criteria, and the study population.
Traditionally, the evaluation of DISH was conducted by the plain chest or abdominal radiographs. However, Hirasawa et al. 12 implied that it seems too ambiguous to judge the presence of spinal column continuity only based on the radiographs and rely on the reliability and accuracy as an interpretation. They compared the prevalence of DISH assessed by reconstructed CT of the chest to pelvis with that evaluated by plain radiographs of the chest and abdomen. The results revealed that the prevalence of DISH based on CT was 27.1% and that based on radiographs was 17.6% 12 . This fact confirms that CT scan is a better diagnostic modality for the evaluation of DISH.    17 proposed the modified Resnick and Niwayama criteria for the diagnosis of DISH on chest CT (Table 1). They evaluated inter-observer agreement related to the diagnosis of DISH on chest CT and found inter-observer agreement was fair or poor by kappa analysis using the Resnick and Niwayana criteria 17 . This indicates that the reliability of the original criteria on chest CT may be disputable. In contrast, they also evaluated the modified Resnick and Niwayama criteria and demonstrated that the modified criteria to identify DISH on CT scan led to moderate to excellent agreement between the observers with different levels of experience and expertise 17 . In our study, the results of kappa coefficient of intra-and inter-observer agreement also demonstrated a substantial and excellent agreement (0.90 and 0.91). Thus, the modified Resnick and Niwayama criteria is more reasonable than the original criteria to evaluate DISH on CT scan images.
The range of age in the study population also affects the prevalence of DISH. Most previous reports only included people who are older than 40 years and the average age is around 65 years 6,7,9,10,12 . In contrast, our study includes the range of age from 5 to 105 and the average age is 57.6 years. We believe our study population reflect more accurate prevalence of T-DISH compared with the previous reports which only included people older than 40 years.
The prevalence of previous reports showed a considerable difference by race. Weinfeld et al. 7 showed that the prevalence was different between ethnic groups. They reported that DISH is less common in the Black, Native-American and Asian populations, although the sample size of their study was small 7 . Our study revealed the prevalence of T-DISH in the Black population seems to be not as low based on the CT analysis. Regarding the prevalence of T-DISH in the Black population, Cassim et al. 6 reported it was 3.9% in 1500 African in South Africa, using conventional chest radiographs; however, there is no study that reported the prevalence of T-DISH in the Black population using the CT scan. Our study first reported the prevalence of T-DISH in the Black population using the CT scan and modified Resnick and Niwayama criteria. This study demonstrated that the prevalence of T-DISH was 8.6% in females and 7.0% in males. This is an interesting finding because most of the previous studies showed the prevalence of DISH was higher in male patients compared with female patients. One study showed the prevalence of DISH was higher in female patients. Cassim et al. 6 found the prevalence of T-DISH in the African population was 4.2% in females and 3.9% in males. These results may suggest that female dominance in the prevalence of T-DISH is specific to the Black population.
The prevalence of T-DISH was the highest at the age of 70 years (11.7%), followed by the age of 80 years (10.5%). This is consistent with other previous reports 12,14,15 . Mori et al. 14 found that T-DISH was observed after the age of 40 years and the prevalence of T-DISH rose rapidly with age until its peak distribution, the age of 70 years. Interestingly, in our study, the prevalence of T-DISH below the age of 40 years was 4.3% (25/577). Recent studies that included all age groups and analyzed CT or PETCT revealed there was no or minimum DISH patients below the age of 40 years in Japanese 14,15 . We do not have a clear explanation for this. Patients with genetic diseases resulting in abnormalities of the parathyroid hormone-calcium-phosphate pathway may develop ossification of spinal ligaments and/or ectopic calcifications in young individuals 18 , although we were unable to obtain such information of those patients. Or this finding may be also specific to the Black population.
In our study, the highest prevalence level of T-DISH segment was at T8, followed by T9, and T7. Hirasawa et al. 12 reported that T-DISH most commonly occurred at T7/8, T8/9 and T9/10. Hiyama et al. 15 performed a whole-spine CT study and found that most of the ossification occurred in the middle and lower thoracic spine: T8 (88%), T9 (91%), and T10 (85%). They described that such result may suggest that an anatomical effect which these vertebrae are vulnerable to compressive mechanical stress because T8 is located almost at the peak of the physiological spinal kyphosis 15 . The most frequent number of contiguous vertebrae was seven (21%), followed by four (18%), and five (18%). In Mori's study, the most frequent number of contiguous vertebrae was eight in Japanese 14 .
Previous studies reported that high BMI and the presence of DM are the risk factors for DISH 13,14,19 . In Japanese patients, Fujimori et al. 13 reported that high BMI was associated with DISH in female subjects; in contrast, Mori et al. 14 found a significant association between T-DISH and high BMI in male subjects. However, our study showed no association between T-DISH and BMI in the Black patients. A proposed mechanism of DM for DISH is the prolonged and high levels of insulin or insulin-like growth factors occurring in DM patients, stimulating new bone growth 19 . Our study revealed that the presence of DM was significantly higher in male patients with T-DISH than those without T-DISH (P = 0.02). Similarly, Fujimori et al. 13 found that high glucose level was associated with DISH in Japanese males. However, our study showed no association between T-DISH and DM in female patients.
Our study is limited by several limitations. First, this study population may not represent the general population and may have a selection bias due to the limited area (New York City). However, as it is unethical to take CT scans of normal volunteers, we consider that our study participants may represent the best possible sampling for the general population. Second, we could not evaluate the whole spine or sacroiliac joint. Third, we were unable to differentiate Type 1 and Type 2 DM. Type 2 DM has been reported to be associated with DISH, whereas Type 1 DM is not 20 . Fourth, we were unable to obtain genetic disease information such as abnormalities of the parathyroid hormone-calcium-phosphate pathway, particularly in young patients. Previous studies reported the correlation between DISH and abnormalities of that pathway 21,22 . Fifth, the study lacks sufficient metabolic data, including both anthropometric (waist circumference for metabolic syndrome) and chemical blood measurements. Despite these limitations, we believe that our data are invaluable because this was the first study to examine the prevalence of T-DISH using CT analysis in the Black population.

Conclusion
The prevalence of T-DISH was 7.7% in the 3299 Black patients who underwent chest CT for the trauma screening, with 8.6% for females and 7.0% for males. The highest prevalence was observed in patients at the age of 70 years (11.7%), followed by the age of 80 years (10.5%). The highest prevalence level of T-DISH segment was at T8, followed by T9, and T7. The most frequent number of contiguous vertebrae was seven (21%). BMI was not associated with T-DISH. The presence of DM was significantly higher in male patients with T-DISH than those without T-DISH (P = 0.02).