This nationwide study identified 70,517 patients with extracranial vascular malformations in Korea from 2008 to 2021, with an overall annual incidence of 9.85 cases per 100,000 population. In the subgroup analysis, the annual incidences of VMs, CMs, AVMs, and LMs during the same period were 1.48, 2.31, 0.24, and 5.82 cases per 100,000 population, respectively. Some studies have reported the incidence of VMs, CMs, and LMs; however, the exact incidence from a comprehensive perspective remains unidentified to date. To the best of our knowledge, the incidence of AVMs has not been reported owing to the rarity of extracranial vascular malformations30. In the present study, the incidence of all VMs was higher in female patients, and the malformation subtype with the highest female dominance was AVMs. According to a previous study, CM was found to be the most common vascular malformation8; however, in our nationwide study, LM was the most common malformation (5.82 cases per 100,000 population). This difference may be attributed to national or racial differences between the two studies. Furthermore, in our study, the incidence of LMs increased with age; however, relatively older patients may have been underestimated in other studies because they were not well diagnosed given that extracranial vascular malformations are congenital anomalies. In our study, the second most common type of malformation was CM (231 cases per 100,000 population). Further, VMs are known to be the most common type of extracranial vascular malformations for which patients are referred to multidisciplinary vascular anomalies centers30,31. However, in the present study, the incidence of VMs was approximately 3.93 times lower than that of LMs, indicating that several patients with LMs are not referred to multidisciplinary vascular anomalies centers. Regarding the age and sex distribution, the incidence of VMs was the highest among female patients aged 0–4 years, indicating that VMs (congenital anomalies) are well diagnosed immediately after birth. Notably, the incidence of CMs peaked immediately after birth, and it was two times higher than that at birth. Moreover, a previous study reported that CMs are present at birth and grow proportionately with the individual9. It is possible that the lesion, which was initially small and undiagnosed, may have gradually grown and become diagnosed with age. AVMs are rare and fast-flow malformations, and their exact incidence is currently unknown; however, in the present study, the incidence of AVMs was 0.24 cases per 100,000 population. Among patients with AVMs, male and female patients were most frequently diagnosed in the ages of 20–24 and 30–34 years, respectively, suggesting that AVMs are expressed relatively slower in female patients. Moreover, the history of normal LMs indicates a slow progressive increase in size and complications, including pain, bleeding, and oozing32. In the present study, the incidence of LMs tended to increase with age, and this finding is consistent with the known history. Regarding LMs, our nationwide study reported that this diagnosis should not be neglected even in older adults.
Although all pathological mechanisms of extracranial vascular malformations cannot be explained by single gene mutations, VMs, CMs, AVMs, and LMs are known to be related to the Tie-2, GNAQ, MAP2K1, and PIK3CA genes, respectively1. VMs are known to be caused by mutations in the Tie-2 gene, which are easily recognizable due to their bluish color and are inferred to be diagnosed at the earliest age among all subtypes. The incidence of CMs and AVMs was highest in women in their 20 s and 30 s, suggesting that related genes may be stimulated by hormones. The activation of the MAP2K1 gene and subsequent production of the MEK1 protein can be triggered by various extracellular signals, including hormones. The binding of a hormone to its specific receptor on the cell surface initiates a cascade of intracellular events that ultimately lead to the activation of the mitogen-activated protein kinase (MAPK) pathway. The PIK3CA gene which is associated with LMs is one of the most commonly mutated genes in human cancers, and mutations in this gene can lead to the dysregulation of the phosphatidylinositol 3-kinase (PI3K) pathway. These mutations can result in increased and sustained activation of the pathway, promoting cell growth and survival and contributing to tumorigenesis and cancer progression. Considering the gradually progressive nature of tumors and cancers, the increasing incidence of LMs with age can be explained.
Extracranial vascular malformations are known as congenital malformations, and the incidence of congenital malformations is generally expressed as the number per 1000 live births33,34. Therefore, we initially attempted to express the incidence of these malformations as the number per 1000 births. However, as reported earlier, VMs are mostly diagnosed immediately after birth, but CMs, AVMs, and LMs are not well diagnosed immediately after birth. In particular, in cases of LMs, it might be meaningless to express the number per births because its incidence increases with age. Therefore, we calculated the incidence of extracranial vascular malformations as number per 100,000 of the general population. We believe that there is no definitive approach for classifying types of epidemiological studies, and different classification schemes may be useful for different purposes35.
In the present study, all patients with extracranial vascular malformations, except for those with VMs, had higher mortality than the matched controls. Among them, the LM subgroup had the highest mortality (17.05 per 1000 person-years). In addition, based on the Kaplan–Meier curve, patients with LMs had the lowest survival rate as the age of several patients at the diagnosis of LMs was higher than that of patients with other extracranial vascular malformations. Interestingly, the survival rates of patients with AVMs differed in terms of sex. The survival rate of male patients with AVMs was lower than that of the matched controls; moreover, the rate was lower than that of patients with LMs until approximately 3 years following diagnosis. Furthermore, as the incidence of LM in male patients increased with age, the survival rate of male patients with AVM was considered relatively low. In contrast, the survival rate of female patients with AVMs was similar to that of patients with CMs; however, it was higher than that of patients with LMs. Additionally, the survival rate of female patients with AVMs was slightly lower than that of patients with CMs until 3 years following diagnosis, and this rate increased after 3 years. Briefly, the mortality of patients with AVMs was relatively high up to 3 years of diagnosis. Further, the survival rate of patients with VMs was higher than that of matched controls because the incidence of VMs was the highest in individuals aged 0–4-years, which may be attributed to their relatively young age. Age- and sex-adjusted HRs could be obtained in the Cox proportional hazards model, which was used to determine whether the subgroups of extracranial vascular malformations were risk factors for mortality. We found that VM was not a statistically significant risk factor for mortality; however, the adjusted HRs of CMs, AVMs, and LMs were 1.94, 2.06, and 1.38, respectively, indicating that they were risk factors for mortality. Among these, it was confirmed that the HR of AVMs was the highest.
Based on the clinical examination, the Schobinger stage classification categorizes AVMs into stages I–IV according to the severity of symptoms. In stage I, physical examination reveals a warm pink-blue mass, which is occasionally confused with CMs and infantile hemangioma. When this lesion worsens, it enlarges and is accompanied by pulsatility, thrill, and bruit, and this is considered stage II. If the artery-to-vein shunting continues, oxygen diffusion to the capillary does not occur, resulting in an ischemic state, thereby causing secondary pain, tissue ulceration, and bleeding. Moreover, the skin undergoes dystrophic changes, and this is considered stage III. Notably, AVMs allow direct blood flow through the nidus from the high-resistance high-pressure arterial system to the low-resistance low-pressure venous system without capillary perfusion, leading to venous hypertension. When this becomes severe, it leads to high-output cardiac failure corresponding to Schobinger stage IV (Table 5)36. The low survival rate and the highest adjusted HR of AVMs among the vascular malformation subgroups, despite the incidence of AVMs being the highest in the 20 s and 30 s, can be attributed to the fact that they can cause cardiovascular disease. Kim et al. suggested that patients with AVM had to bear higher hospital costs owing to the complexity of their condition, which requires a higher level of inpatient care37. This report can be evidence of the high mortality of AVMs.
Regarding the mortality of LMs, because they are associated with mutations in the PIK3CA gene, and the PIK3CA gene is associated with several cancers, patients with LMs may be more likely to develop cancer. In such a case, it is possible to explain the higher mortality rate of LMs patients than control cohort.
This study has some limitations. First, as the ICD-10 code was used to identify the diagnostic classification, the presence of combined types of extracranial vascular malformations (including lymphatico-venous malformations and CM-AVM) was impossible. Second, there was no information on the findings of imaging tests, including MRI to diagnose VM or LM and computed tomography angiogram to diagnose AVM, or photographs derived from these data. Thus, there may be a bias in the classification of diagnoses. However, the Korean NHIS provides fairly accurate diagnostic data. Before claiming the medical fee, the ICD-10 code and fees for surgery and intervention were verified by the insurance review team of each general hospital. Subsequently, the Korean NHIS data were stored after reverification by the Health Insurance Review and Assessment service—a Korean government agency25. Further, we evaluated the diagnostic accuracy of diagnostic classification, which revealed high sensitivity and specificity. Therefore, the risk of bias for misdiagnosis in the present study was low. Third, regarding mortality, it would have been more useful to determine the causes of death in patients with extracranial vascular malformations and in matched controls. In NHIS-NHID–based studies, the database did not contain data about the causes of death. In contrast, Statics Korea has information on the causes of deaths, and we have not conducted a study that linked the NHIS-NHID and Statics Korea. However, as data of all patients with extracranial vascular malformations were extracted and those of matched controls were extracted by age and sex from the general populations, no selection bias was noted between patients with extracranial vascular malformations and matched controls in this study. Therefore, the only difference in the cause of death between the two groups was attributed to extracranial vascular malformations.
This study also has several strengths. First, to the best of our knowledge, this is the first study to investigate the incidence of extracranial vascular malformations and related mortality in a nationally representative Asian population. Previous studies investigated hospital-based data or used small sample sizes, posing difficulties in the identification of the incidence of rare extracranial vascular malformations, including AVMs. Second, by evaluating the incidence in terms of age and sex, were identified the age and sex that demonstrated the highest incidence of each vascular malformation subtype. Finally, the survival and mortality of the patients in each subgroup with extracranial vascular malformations were identified by comparison with the nationally matched control population.