The results of this study showed that with the Markov model approach, the MetS in its developmental process has the capability of a total of 144 bidirectional transitions in twelve states. With the Markov approach, the most important trigger for the MetS chain was dyslipidemia with overweight/obesity in the under-50 age group, and dyslipidemia with hypertension in the over-50 age group. Also, with the non-Markov approach, the most common component of the MetS initiation process was hypertension.
It is believed that the first component with the highest incidence is the main trigger of the MetS development process and this is a basis for trigger determination [
19]. In non-Markovian models, the basis for trigger determination is not longitudinal and based on the predictive power of the components, their role and importance are judged. As shown in Table
5, the central core of triggers in different Markovian studies, which is essentially a controversial issue [
21,
25,
26], is the set of Dyslipidemia, overweight/obesity and hypertension components. In all of the studies listed in Table
5, these components are present at almost all levels of study. Of course, the reason for the difference between the various studies is that its dependency on demographic characteristics such as age, sex, or other characteristics; as if, Xiaoxian [
19] also refers to this issue. Of course, there is also a marked difference between the results of the Markov and non-Markov studies, which may be due to the difference in the MetS definition criteria of the Markov and Non-Markov studies. Because the criteria used in non-Markovian studies is a global criteria such as IDF (international diabetes federation), NCEP, etc. While in three Markovian studies of four studies, the CDS local criterion was used (Table
2). In our study, in both Markovian and non-Markovian approaches, the prevalence of all components in both sexes was increasing with age and with time. In only one study with Markov’s approach, the prevalence of overweight/obesity and dyslipidemia in men didn’t increase with age. According to the available evidences, prevalence of the Mets increases with age due to endogenous sex hormone levels changes [
37,
38]. In women for example, Wu et.al [
38] indicate that if age at menopause is greater than 49 years, an increase in years since menopause confers a negative influence on glucose tolerance that is associated with central fat distribution. Also, 5α-reductase activity which is associated with adipose tissue, increased with indexes of insulin resistance [
39]. On the other hand, visceral adiposity [
40], which is intensified by the aging process [
41], has been announced to be a primary trigger for most of the pathways involved in the MetS especially for insulin resistance [
42]. It seems that changes in hormone levels in individuals, which are influenced by the aging process and trigger metabolic changes in humans, can largely explain some of these differences. In Chedraui et al.’s study [
43], an increase in IL-6 and a decrease in urokinase-type plasminogen activator levels in postmenopausal women have been introduced as the reasons for the effects of these hormonal changes on the occurrence of the MetS. Also, in relation to the more initiating effects of hypertension in the higher age group in women, the effect of endocrine system activity regress, which reflects changes in hormone levels, has been noted [
44]. Here, the key issue is the combination of the triggers and their comorbidity, which partly justifies the rising trend of the MetS due to the cumulative effects of the components over time and with age. In other words, approximately, no studies have addressed the role of an isolated component in the create conditions for subsequent occurrence of the MetS. This may be an indication that metabolic components tend to co-occur, suggesting a close association and shared roots. Many researchers believe that the four components of the MetS are overlapping in their pathogenic activities, and this occurs through similar metabolic pathways [
45]. Therefore, it is advisable to check the other components carefully and periodically for people who usually have one of the components. Therefore, it should be said that there are two important issues in relation to the triggers in different studies; one is its combination, and the other is its demographic variability, which justifies both the rising trend with age and gender difference based on hormonal differences and interplays. In general, it seems that the role of overweight/obesity and dyslipidemia components in the age group below the age of 50 and the role of hypertension in the age group over the age of 50 were be explained by hormonal changes. Of the various available studies, none has mentioned to hyperglycemia as a trigger. Dyslipidemia is said to play a mediating role in the free fatty acid (FFA) damage to the islet function process (which causes hypertension and hyperglycemia) [
19,
46]. So, at first glance, the combination of dyslipidemia with hypertension may seem justifiable in the development of the MetS. On the other hand, overweight/obesity introduced also as an important risk factor for dyslipidemia and hypertension and hyperglycemia as major components of the MetS, especially in younger people [
47,
48]. Even elsewhere, it, especially abdominal obesity, has been independently introduced as the most important cause of the MetS [
49]. Advocators of this hypothesis that point to the higher value of this component than other components at the onset of the MetS, using the same argument, namely temporal priority of the overweight/obesity component over others and suggests that this component is, in fact, an important risk factor for the subsequent occurrence of other components [
47]. However, the causal link between dyslipidemia and overweight/obesity remains a controversial issue [
16]. Also, in this study, TP values from no component to MetS were higher in all Markovian studies in men than in women; With the exception of one study in which the transition from no component to overweight/obesity and low HDL states and the TP from overweight/obesity to the MetS after the age of 50 was higher in women. The latter finding, namely the higher TP from no component to overweight/obesity and low HDL states, as well as the possibility of transition from overweight/obesity to the MetS in women, is largely attributable to hormonal differences between men and women after 50 years old. However, with age increasing, the gap between men and women gradually decreased except for the recent exception noted previously. Also, in the Markov model, the probabilities of reverse transitions from each of the states to the no component state were lower in men than in women; However, these probabilities decreased significantly with age. These findings primarily indicate that men at higher risk (in terms of components) are more likely to develop the MetS than women, and are therefore more resistant to return to the normal state. As mentioned earlier, the central triggers in this study included the components of dyslipidemia, overweight/obesity and hypertension. On the other hand, although its mechanism is not yet known, dyslipidemia is claimed to cause overweight/obesity and hypertension in men more than women [
19,
50]. The some of the higher developmental trend of the MetS in men may be due to this issue. Interestingly, most studies [
51‐
55] have been reported a higher prevalence of the MetS in women than men but in all studies (Markovian studies) on the natural history of the MetS, it has been said that men have a higher rate of progression to the MetS than women. Various studies have shown that weight gain, hypertension, and hyperglycemia are important risk factors for dyslipidemia [
56,
57]. On the other hand, the prevalence of all these factors has been reported higher in women than men [
58]. Thus, the speed of the MetS development is expected to be higher (as the prevalence of the MetS) in women than men, but despite the evidence available for faster growth of the MetS in women than men, the accumulation of evidence is in favor of the higher speed of the MetS development in men than women. The prevalence and clustering speed of the components appear to be higher in men than in women [
59], and this may be partly responsible for the faster occurrence of the disorder in men. One of the individual studies used in this review [
19] has also shown that after hyperglycemia, men are more likely to develop hypertension and, as a result, different forms of multi-component, which is in line with the recent justification. But it may be possible to attribute the higher rate of the MetS development in men than women to the important problem of androgen deficiency in men in the best way [
60]. Men with the MetS relative to women are said to be at higher risk for androgen deficiency and Late-onset hypogonadism or Testosterone Deficiency Syndrome. The pathophysiology of these disorders is multifactorial and contains defective inflammatory, enzymatic, and endocrine mechanisms that provide the basis for further development of the MetS in patients so that by treating these mechanisms, the parameters of the MetS are improved [
60‐
63]. Overall, the most common transitions (with a high TP) with the Markovian approach were primarily due to the transition from dyslipidemia to no component, and second, the transition from all-states to 2-component. Regarding the high transition from dyslipidemia to no component, it should probably be said that, in the first place, much of this, is due to the success of dyslipidemia treatment, which response better to treatment compared to other components (whether medications or lifestyle changes). However, in the articles used in this review, the treatment status is unclear, so this argument is only a speculation. Without any treatment, in the long-term, it seems that natural changes in lifestyle and related events can lead to lower lipid profile and normalization to other components of the MetS. Secondly, it must be said that in this transition, there may be several biological and physiological mechanisms, such as the effect of thyroid hormones [
64,
65], female hormone interactions in women [
66] and the effect of preventive actions on other components on lipid profile [
67,
68] that due to the interdependence of the components and the pathophysiological mechanisms between them or unknown self-therapies, they play a role that is not mentioned in the articles used.
The high transition from all states to the 2-component state can also be mainly due to the natural development of isolated components toward complete the MetS establishment that discussed earlier. However, in our study, also the TP from any of the 2-component states to the MetS, which is a continuation of the natural history of the MetS, was also high. Also, the least TP occurred from hypertension, dyslipidemia, and hyperglycemia to each of the model states except for 2-component and MetS states and from the MetS state to all states except hypertension and 2-component states that all of them, in other words, indicate a high rate of progress toward the establishment of the MetS. Regarding the reverse transition from MetS to 2-component state in this process, it is possible to point to a similar mechanism as that described for the reverse transition to no component state. However, the dynamics mentioned in the natural process of the MetS development suggest that this process is influenced by various factors, that many of which are still unknown, and their discovery requires detailed studies of the natural history of the MetS in the other general populations.
Regarding the most common combinations with the Markovian approach, the combination of dyslipidemia with overweight/obesity should be noted and the combination of hyperglycemia and overweight/obesity with the non-Markovian approach. The commonality of both approaches is the overweight/obesity component. Combining dyslipidemia with overweight/obesity is a common combination, because obesity increase triglyceride and LDL, and decrease HDL as components of dyslipidemia, as well as elevated blood glucose and insulin levels through various pathways such as hepatic overproduction of VLDL and decreased circulating TG lipolysis [
69]. However, one should not overlook the role of age; however, with Markov’s approach, the probability of combining dyslipidemia with overweight/obesity was higher in the under-50 age group, and in the over-50 age group the probability of combining dyslipidemia with hypertension was higher [
19]. Regarding the non-Markovian approach, it should be said that this approach is mainly due to the use of predictive and regression-based statistical models and, of course, it has a serious difference with the Markovian models approach that looks at longitudinal and more realistic to evolution of a phenomenon over time and also the possibility of various errors, there is no evidence-based judgment regarding the combination of hyperglycemia and overweight/obesity. However, similar to the argument used in the Markovian approach, the role of the overweight/obesity component for providing insulin resistance and hyperglycemia in relation to other components may be mentioned that provides the higher linkage between the two components and is discussed elsewhere [
70]. Finally, with the Markovian approach, the most important components, which in the simulation process, increasingly were moving toward a full establishment of the MetS, were 2-component states and hyperglycemia and, the most important predictor of the MetS occurrence with non-Markovian was overweight/obesity. The importance of 2-component states in the development and establishment of the MetS in individuals reflects the cumulative effects of the components together. Although it is shown that in isolated component states, the probability of developing the MetS over the next 10 years is higher than in no component states that of course, with the Markovian approach, as shown, hyperglycemia is progressively more likely to lead to the MetS among the isolated components. Also, it is shown that in 2-component states, the highest probability of establishing the MetS appears in the first 5 years [
4,
71]. Regarding hyperglycemia, also it should be said the effect of insulin resistance on the subsequent development of the MetS is a well-defined pathologic process and is often associated with the occurrence of other components of the MetS [
49,
72]. Regarding non-Markovian approaches, as stated earlier, regression models have been used for prediction in most of the studies and due to the different nature of the statistical analyses and approaches to prediction in this model compared to the Markovian model that using the simulation process, no evidence-based judgment can be made. In spite of this, new evidence is in favor of the parallel development of the obesity epidemic with the MetS, especially in children and adolescents, and obesity is routinely used and evaluated in studies with non-Markovian approaches in these areas. On the other hand, the role of obesity in the pathogenesis of insulin resistance has been established as a key factor in the development of the MetS, and it is obvious that the overweight/obesity in a well-designed regression model provides a good estimate of the MetS.
The strength of our study is that it is for the first time that the process of the natural history of the MetS, especially summarize of studies conducted with the Markov model, in the form of a comprehensive review is discussed. On the other hand, the lack of primary studies in this field which reduced the possibility of pros and cons discussions with available evidences, also, difficultness of understanding of inferences made from Markov model processes for non-specialists were of the most important limitations of this study which of course, it was tried to help to better understand the concepts as much as possible by simplifying the explanations and interpretations.