DISCUSSION
In this study, we verified the prevalence of OPLL and OLF in Koreans using PET-CT. Our results revealed the following prevalence rates: OPLL, 8.9%; OLF, 6.5%; C-OPLL, 8.2%; T-OPLL, 1.8%; L-OPLL, 0.6%; C-OLF, 0.1%; T-OLF, 6.3%; and L-OLF, 0.2%.
The prevalence of OPLL, which was previously studied using plain radiographs, was relatively lower than that in studies using recently developed imaging techniques, such as CT and MRI. In studies based on plain radiographs in East Asians, the prevalence of C-OPLL varied from 0.60% to 4.1% [
6-
8], that of T-OPLL varied from 0.56% to 0.8% [
9,
10], and that of T-OLF varied from 4.5% to 6.2% [
10,
11]. Recently, with the development of imaging modalities, such as CT and MRI, the sensitivity in detecting ossification of the spinal ligament has increased. Sohn et al. [
3] used thyroid CT and reported that the prevalence of C-OPLL was 5.7% in the Korean population. Moon et al. [
4] used MRI and reported that the prevalence of T-OLF was 16.9% in the Korean population. Kim et al. [
5] used chest CT and reported a T-OLF prevalence of 21.8% in the Korean population. In the Japanese population, the following prevalence rates were found based on PET-CT: C-OPLL, 6.3%; T-OPLL, 1.6%; L-OPLL, 0.7%; T-OLF, 12%; and L-OLF, 0.3% [
1]. In the Chinese population, the following prevalence rates were found based on PET-CT: C-OPLL, 4.1%; T-OPLL, 2.25%; L-OPLL, 0.8%; T-OLF, 37.65%; and L-OLF, 1.45% [
2]. In the Indian population, the following prevalence rates were found based on whole-body CT: C-OPLL, 5.12%; T-OPLL, 0.56%; and T-OLF, 9.90% (
Table 4) [
12]. The prevalence of spinal ligament ossification in this study was consistently similar to that in previous studies based on CT and MRI. However, the prevalence of T-OLF in our study was relatively lower than that in other studies involving the Korean population and other Asian populations. Moon et al. [
4] investigated patients complaining of back pain, and Kim et al. [
5] included patients suffering from pulmonary disease; therefore, it is thought that there was selective bias in the patient groups. Moreover, there is a difference in the age distribution of the population. The average ages were 56 years and 60.9 years in the studies by Moon et al. [
4] and Kim et al. [
5], respectively, which were higher than that of this study. In both studies, the age distribution of the population was high; thus, the proportion of the elderly population would have been relatively higher than in this study. For this reason, the T-OLF prevalence seemed relatively low in this study. Although Fujimori et al. [
1] did not report the age distribution in detail, the average age and standard deviation in their study were higher than in this study, suggesting that the prevalence was higher. It is unclear why Liang et al. [
2] have a higher prevalence compared to this study. It is believed that there may be a genetic or lifestyle-related cause for the difference in the prevalence rate in each country [
13,
14]; however, more detailed research is needed.
In this study, the prevalence of C-OPLL was approximately twice as high in men than in women, and this finding corresponded to that in previous studies [
1,
3,
12]. The prevalence of T-OPLL was higher in women than in men, which was consistent with the findings of previous studies [
1,
2,
12,
15]. However, the difference in the prevalence of T-OLF among men and women was inconsistent with that in previous studies. In some studies, the prevalence of T-OLF was higher in women than in men [
2,
4,
16], while in others, it was higher in men than in women [
1,
5,
12,
17]. In this study, the prevalence of T-OLF was approximately twice as high in men than in women, and this difference was significant (men, 6.9%; women, 2.8%; p < 0.008). It remains unclear whether T-OLF is more prevalent in men or women.
The prevalence of C-OPLL was the highest in C4 (19.2%), followed by C5 (18.7%), C3 (15.4%), and C6 (15.4%). In other studies, the overall prevalence of C-OPLL was the highest in C5, followed by C6 and C4 [
1-
3]. It is presumed that the prevalence of OPLL was high in C3 in this study compared to other studies due to the relatively high number of male patients. Sohn et al. [
3] reported that the prevalence of OPLL was significantly higher at the C3, C4, C5, and C6 levels in men. Therefore, in this study, the prevalence at the C3 level was relatively high compared with in previous studies.
In previous studies based on plain radiographs, it was reported that T-OPLL occurred most frequently in T6 [
10,
18]. However, in recent studies based on CT images, T-OPLL occurred most frequently in T1 [
1,
2,
12]. On plain radiographs, many bony structures, such as the shoulders and ribs, were found to mask upper thoracic OPLL; therefore, OPLL was not detected at the upper level of the thoracic spine [
19]. In this study, the prevalence of T-OPLL was the highest in T6, followed by T1, and there was little difference in the values: the prevalence was 2.4% in T6 and 2.0% in T1. These results are somewhat consistent with those of recent studies on T-OPLL. The distribution pattern of OPLL was roughly consistent with that in previous studies [
1,
2].
The distribution pattern of OLF in this study was similar to that in previous studies. The distribution of OLF exhibited a bimodal pattern. OLF occurs most frequently in T10 and T11, followed by T5 and T4. The true ribs transition to the floating ribs at the T10 and T11 level, and the kyphotic curve changes to the lordotic curve; therefore, stability is lost at this level. It is considered that this characteristic causes repetitive and high tensile stress, thereby leading to ossification [
20,
21]. The second peak in the segmental distribution of OLF was consistent with that in previously published reports. However, there was a difference in the level of the second peak, and most previous studies reported a second peak at the T3 level [
1,
4,
5]. Liang et al. [
2] reported a second peak at T4, and Mori et al. [
17] reported a second peak at T4/5. The presumed reason for the occurrence of the second peak is that the kyphotic curve angle is high at T5, and thus, the mechanical stress is also high. Kim et al. [
5] reported that T-OLF patients had more pronounced thoracic kyphosis. The reason for the occurrence of the second peak at the upper level of the thoracic spine is still unclear, and further research is needed on this subject.
The prevalence of OPLL and OLF increases with age. The number of lesions with OPLL and OLF also tends to increase with age. We found a significant relationship between age and the prevalence of OPLL. This finding was consistent with the results of previous reports [
1-
3]. The prevalence of OLF tended to increase with age. On the other hand, there was no significant relationship between age and the prevalence of OLF in this study. This tendency was consistent with that in previous studies, but the statistical significance was different. Kim et al. [
5] reported that the relationship between age and the prevalence of T-OLF was not significant, whereas Moon et al. [
4] reported that there was a positive correlation between age and T-OLF prevalence. Further data and studies are warranted for investigating the statistical correlation between age and the prevalence of OLF.
This study shows that the coexistence of spinal ligament ossification at different levels was high. Several past studies have reported that patients with C-OPLL also have ossification in other areas of the spine [
22,
23]. Fujimori et al. [
1] reported that T-OPLL patients have a systemic tendency to develop ossification. In this study, even in healthy people, the presence of OPLL and OLF in all the locations was correlated with the presence or absence of spinal ossification in other locations. Prior to treating patients with locally observed OPLL or OLF, it should be determined whether ossification in other areas may become symptomatic. Additionally, we recommend performing an evaluation through whole-spine CT to determine whether ossification is present at any other level.
This study has several limitations. First, PET-CT performed at this institute did not provide data related to the bone setting view. The resolution of the PET-CT fusion image was relatively lower than that of the bone setting view; therefore, we could only check for the presence or absence of ligament ossification. It is also possible that we were not able to identify cases with subtle ossification. Therefore, there is a possibility that the reported prevalence rate is lower than the actual prevalence rate. Second, the male to female ratio was not uniform. There were many more men than women in the subject group. In addition, since treatment agreements were made between heavy industry and electric companies and this hospital, bias resulting from occupations must be considered. Despite these limitations, this is the first study of the prevalence of ossification in healthy Koreans, and we believe that the results of this study sufficiently reflect the prevalence of OPLL and OLF in the general Korean population.