A Controlled Descriptive Study to Determine the Nutritional Status and Biochemical Parameters of Multiple Sclerosis Patients
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.293462
Source of Support: None, Conflict of Interest: None
Keywords: Biochemical parameters, decision tree, malnutrition, multiple sclerosis, nutritional status
Multiple sclerosis (MS) is a slowly progressive central nerve system disorder that causes demyelinating regions of the brain and spinal cord. Affecting 2.5 million people worldwide, MS is one of the most common neurological disorders that can cause disruption in young adults.
Etiology of MS is not fully known. It is thought that etiological factors may include genetic factors, disorders in the immune system, and environmental factors. Obesity and dietary components are among the environmental factors affecting the disease. Based on the epidemiological results showing an increase in the incidence of MS in populations with high saturated fat intake and low vitamin D intake, a hypothesis was established that a diet could affect the course and incidence of disease. Smoking is among the risk factors affecting the disease development and progression.
Since the cause of MS is not precisely defined, there is no definitive treatment for the disease. There are preventive measures to improve quality of life and treatments of attack episodes. Changes in serum 25(OH)D levels are associated with inflammatory diseases such as MS, and patients with MS have lower levels of 25(OH)D. High levels of 25(OH)D have protective effect against MS. Calcium is a mineral that acts as a membrane stabilizer for the communication of nerve impulses. This mineral is required for the control of nerve conduction.
It is believed that the presence of MS may vary according to the nutritional status and biochemical information of the individual. This paper aims to investigate such association to reveal certain parameters that differ in MS patients. Hence, nutritional precautions might be taken to tonify the immune system.
This study was designed as a controlled descriptive study. The output was the presence of MS. Two groups, one including the patients having the outcome of interest (MS) and one not having, were compared in terms of nutritional status and biochemical information. Socioeconomic and demographic characteristics and nutritional status of the participants were determined based on the answers collected from the surveys conducted using “face to face interview” method. Biochemical information was obtained from the blood samples of the participants.
The study group consisted of patients who applied to a state hospital in Trabzon, Turkey between the dates of June 28, 2016 – January 5, 2017. All individuals carrying the following characteristics were included in the study group: 1) He/she was eager to participate, 2) His/her age was between 19 and 65 years, 3) He/she did not have any disease other than MS, and 4) He/she did not change his/her eating habits after diagnosed with MS. There were 196 patients satisfying the requirements. However, for the validity of the biochemical analyses, three more criteria were added: 5) He/she has not used disease modifying drugs for at least 1 year, 6) He/she has not received any type of mineral supplements, and 7) He/she has not received any type of corticosteroids for the previous three months. The number of patients who satisfied these constraints was 56 (42 female, 14 male). In order to constitute the control group, a list of patient names who applied to the internal medicine or the endocrinology outpatient clinics of the hospital within the aforementioned time period was recorded. Among these patients, individuals who were between 19 and 65 years of age and did not have any chronic health problems were determined and asked to be a volunteer for the current study. 196 of the patients who accepted to be a volunteer were randomly selected. To be able to carry out sex adjusted analyses, 42 female and 14 male participants were again randomly selected from these 196 patients. The resulting participants were requested to give their blood samples when they did not have any health problem. Thus, the control group was established.
General characteristics, health status, and eating habits of participants were determined. Through a well-organized survey, participants were asked to report their average consumption of most frequently consumed nutrients on an average day. The average 24-hour retrospective food consumption records of the individuals participating in the study were evaluated using the software Ebispro for Windows. The energy and nutrient intakes of all individuals were calculated. Information regarding to the daily average food consumption was collected from 33 food and drinks related questions. Participants were asked how often they would consume specific food and drinks in a regular day. Based on the responses, average daily energy intake, and macro and micronutrient intakes were computed.
Based on the laboratory standards of the hospital, the reference values for 25(OH)D were taken as 20–120 ng/ml and 10–60 ng/ml for the summer and winter seasons, respectively. All parameters except serum 25(OH)D were measured with “Beckman Coulter AU-5800” device. Serum 25(OH)D level was analyzed on a “Beckman Coulter Dxl-800” by using “chemiluminescence immunoassay” method. Serum Ca, Mg, and inorganic phosphate values were determined by “spectrophotometry” method, whereas “ion selective electrode” method was used to measure serum K levels. Lipid profile values were calculated using “spectrophotometric” methods. The serum levels of alanine amino transferase (ALT), aspartate aminotransferase (AST) and alkaline phosphatase (ALP) were analyzed using “kinetic spectrophotometric” method, whereas to assess C-reactive protein (CRP), “immunoturbidimetric” methods were used.
To determine whether parametric or nonparametric methods will be used to compare the control (healthy) group with the MS group, Shapiro–Wilk normality tests were carried out to see if the interested variables follow a normal distribution. Depending on whether the Shapiro–Wilk test was significant or not, Mann–Whitney's U test or Student's t test was conducted for the main analysis. Pearson's Chi-square test was used to determine whether there is a relation between variables. A decision tree model was built to demonstrate the differences in MS patients and healthy people in terms of the socioeconomic and demographic characteristics, nutritional status, and biochemical information. For all tests, the statistical significance level was taken as 0.05. The decision tree model was constructed using Rapid Miner Studio version 8.2. All the statistical calculations were done in IBM SPSS statistics for windows, version 22.0.
For both control and MS groups, frequency analyses were conducted in terms of socioeconomic and demographic characteristics [Table 1]. There were equal number of men and women in each group. Men made up 25% of total participants. 67 (59.8%) of total participants were of the age range 31–50 years, 31 of them were of the age range 18–30 years, and 14 of them were 51 years old and above. The number of participants regards to the age categories were almost the same across the groups. 35.7% of the total participants had an education level higher than high school. In the control group, most of the participants (55.4%) graduated from a university, whereas in the MS group, the education level was less than high school for the majority of the participants (57.1%). 80.3% of the total participants had monthly average income above the subsistence level. With regard to marital status, 86 (76.8%) of the individuals were married. Most of the respondents (75.9%) were living in provinces, only 8.9% of the participants were from villages.
Diagnosis dates for the MS patients were as follows: 36 (64.3%) of them were diagnosed less than one year ago, 8 (14.3%) of them 1–5 years ago, and 12 (21.4%) of them more than 5 years ago. Measurements for 46 (82.1%) of the patients in MS group and 47 (83.9%) of the participants in the control group were taken in summer. Rest of the measurements were carried out in winter.
Descriptive statistics and group comparisons
In this study, group statistics were expressed as the mean (x–) ± standard deviation (SD). Daily average energy and macro and micronutrient intakes were compared to see whether there were statistically significant differences among the participants of the control and MS groups. Descriptive statistics along with the comparisons regarding to the nutritional status of participants were presented in [Table 2]. The percent lipid and carbohydrate (CHO) intakes, protein, cholesterol, calcium, phosphorus, vitamin B2, vitamin B12, and zinc intakes were not statistically significantly different between the control and MS groups (P values were 0.383, 0.694, 0.219, 0.132, 0.119, 0.056, 0.123, 0.650, 0.548, respectively). However, strong statistical differences were observed among the calculated intakes of CHO, lipid, potassium, vitamin B1 and B6 of control and MS groups. In general, the control group was found to have more energy and nutrient intakes than the MS group.
Biochemical information of the participants was collected to investigate possible differences between the control and MS groups. Descriptive statistics along with the comparisons regarding to the biochemical information of participants were displayed in [Table 3]. As can be seen from the table, there were statistically significant differences between the groups in terms of the serum levels of Ca, Mg, K, vitamin D, and LDL. Especially for the serum levels of 25(OH)D and potassium, the difference was obvious. The serum 25(OH)D level of MS group (8.9 ± 4.0 ng/ml) was significantly lower than the control group (18.9 ± 11.0 ng/ml). Similarly, the mean serum potassium level of MS group was significantly lower than the control group (4.11 ± 0.41 mEq/l v. 4.45 ± 0.41 mEq/l respectively, P < 0.001).
Existence of relation between the presence of MS and the state of wearing head and body-covering clothes
Vitamin D is known as the sunshine vitamin. Deficiency of vitamin D is common among especially dark-skinned individuals who migrate to high northern or low southern latitudes. Nevertheless, this effect could not be observed in the current study since the participants were all white skinned. Recently, vitamin D deficiency has been associated with the pathogenesis and/or progression of MS. Based on the fact that the vitamin D deficiency is a modifiable risk factor for MS, it is predicted that for females, there might be a relation between the presence of MS and the state of wearing head and body-covering clothes. To test this hypothesis, the contingency table shown in [Table 4] was created based on the female participants of this study. Chi-square test result showed that the association between the presence of MS and the state of wearing head and body-covering clothes were statistically significant (P = 0.009).
Decision tree model for comparative purposes
Based on the nutritional status and biochemical information obtained from the participants, a decision tree model was built to compare the threshold values of certain parameters in healthy people and MS patients. Since there were many features obtained from the participants, the constructed tree would be too complicated to be examined. Therefore, a forward selection method was performed to find out the optimal subset of features that will make the tree simpler. Serum levels of vitamin D, magnesium, calcium, and intakes of potassium and CHO were found to be the five features that constituted the decision tree, which achieved the best classification performance. The constructed decision tree was presented in [Figure 1]. The interpretation of the tree can be done by following the arrows from top to bottom and paying attention to the limit values of the features. For example, in this study, there were 29 participants who had serum 25(OH)D level below 15.40 ng/ml, serum magnesium level below 2.29 mg, serum calcium level below 9.56 mg/dl, and potassium intake below 1798.05 mg. Being a descriptive tool, the constructed tree serves perfectly to compare the parameter differences in the control and MS groups. Descriptive findings of the tree might help physicians to take nutritional precautions accordingly.
On the other hand, when utilized as a classification tool, it is obvious that the model can lead to erroneous predictions. The statistical classification performance of the constructed decision tree model was evaluated based on 10-fold cross validation. The confusion matrix obtained after validating the model with 10-fold cross validation i's shown in [Table 5]. The accuracy of the decision tree model was calculated as 86.52%, whereas the precision, recall, and F1 score were computed as 86.71%, 86.61%, and 86.66%, respectively. According to the table, out of 56 MS patients, 6 of them were misestimated as healthy, and out of 56 healthy people, 9 of them were mispredicted as having MS.
Summary of main findings
Unlike many studies in literature that compared mainly the serum 25(OH)D levels of MS patients and healthy people, the current study focused on constructing a decision tree model to reveal the threshold values of certain parameters in healthy people and MS patients. Critical values for serum levels of 25(OH)D, magnesium, calcium, and intakes of potassium and CHO were determined using this decision tree.
As stated in the study of Jena et al., MS has a spectrum of heterogeneity, as seen in western and eastern hemispheres, in the clinical features, topography of involvement, and differences in natural history. It must be noted that the current study was designed as a controlled descriptive study. That is, the main goal is to compare the control and MS groups descriptively in terms of the nutritional status and biochemical information obtained from the participants. As a major disadvantage, this type of design lacks establishing causality. Just by looking at the descriptive differences, drawing conclusions about the causes of MS is not possible, but the selected parameters of the decision tree suggest that the associated variables are the most distant ones between the compared two groups. That is, these parameters might provide a baseline for a well-organized cohort study, where the causality can be searched broadly.
Due to the lack of the studies in literature on critical value determination for MS patients, only the mean value comparisons could be performed in the current study. The comparison of the findings of the MS based studies was presented in [Table 6]. The literature research revealed that the parameters of the constructed decision tree have commonly been investigated as the possible causes of MS.
Serum 25(OH)D level has been the most frequently analyzed parameter that is suspected to be the main variable, differs in healthy people and MS patients. Knowing that the 25(OH)D levels are known to fluctuate according to season, as mentioned in “Measures” section, for winter and summer seasons, different reference values for 25(OH)D were used in this study. For both summer and winter seasons, the 25(OH)D levels of the control and MS groups differed significantly (P values were 0.029 and <0.001, respectively). Since both groups included almost equal number of patients whose measurements were taken in winter (or summer), the comparison was also performed without seasonal effect. Serum 25(OH)D levels of the MS group (8.87 ± 4.02 ng/ml) were found to be significantly lower than the control group (18.86 ± 11.04 ng/ml). Similar results were obtained in the studies of Ozgocmen et al. and Hatamian et al. In many studies, 25(OH)D deficiency was associated with the increase in MS risk. Moreover, an inverse correlation between MS prevalence and sunlight exposure, which is thought to be the main source of vitamin D, was observed in three studies: one performed on US veterans, one conducted in Australia, and one carried out on immigrants to Israel.
Seasonal fluctuations of 25(OH)D levels may differ based on skin tone. That is, a white individual, who is more likely to be able to absorb ultraviolet radiation than an African American individual, may be more likely to experience fluctuations in 25(OH)D levels based on the season. However, this situation was not a concern in the current study since the participants were all white-skinned.
Magnesium plays an essential role in nerve transmission and neuromuscular conduction. It also functions in a protective role against excessive excitation that can lead to neuronal cell death and has been implicated in multiple neurological disorders. Magnesium is required for the axon stabilization and influences the release of neurotransmitters. When Mg level is low, nerve conduction velocity is increased, and the threshold for axon stimulation is reduced. Due to these important functions within the nervous system, magnesium is a mineral of intense interest for the potential prevention and treatment of neurological disorders. Hence, it seems reasonable to expect differences between the serum Mg levels of the control and MS groups. Indeed, in the current study, the serum Mg levels of the individuals in the MS group (1.97 ± 0.18 mg/dl) were found to be significantly lower than the individuals in the control group (2.43 ± 2.82 mg/dl, P = 0.037). Most of the studies in literature,, also support these findings.
In the current study, the serum calcium levels of the MS group (9.14 ± 0.63 mg/dl) were statistically significantly lower than the control group (9.33 ± 0.92 mg/dl, P = 0.004). Soilu–Hanninen et al. also found similar results in their study. There might be a few reasons for that finding. Glucocorticoid, a hormone produced by the adrenal glands that has anti-inflammatory effects, is known to be used to manage acute exacerbations or relapses in people with MS. Since glucocorticoids decrease intestinal calcium absorption and increase renal calcium excretion, patients with MS may have suffered from Ca deficiency. Moreover, under conditions of vitamin D insufficiency or deficiency, intestinal Ca absorption decreases, resulting in a subtle reduction in serum Ca. Low levels of vitamin D in MS patients may have caused a decrease in calcium absorption in our study. Furthermore, increased calcium losses due to menopause in women may have affected calcium consumption.
Potassium intake was one of the parameters that differed significantly in the control and MS group of the current study. As shown in [Table 2], higher intakes of potassium were observed in healthy individuals. WHO has revealed that increased potassium intake reduces systolic and diastolic blood pressure and risk of incident stroke in adults. Moreover, the studies have shown that there is no significant association between potassium intake and incident cardiovascular disease or coronary heart disease. On the other hand, there is insufficient literature on studies looking at a possible relation with potassium intake and MS risk. In their case control study, Ghadirian et al. found out that the higher intakes of potassium were negatively associated with the MS risk.
Diets lower in carbohydrate are positively correlated with improvements in ambulation and physical activity in people with MS. Diets high in CHO can contribute to impairment of lipid and glucose metabolism in MS. In the current study, CHO intakes of MS patients were significantly lower than healthy subjects (P = 0.001). In general, it is assumed that MS patients tend to have a healthy diet and reduce the consumption of carbohydrates or tend to eat fast foods because of their physical activity limits. This circumstance might seem as an obligation rather than a conscious behavior. In literature, there are studies, in accordance with our findings, arguing that the MS patients are determined to consume more fat and protein and less CHO comparing to normal population. Moreover, some studies,, compared male and female MS patients in terms of their CHO intakes.
Vitamin B12 and folic acid deficiency may adversely affect the methylation reactions, which are essential for the metabolism of components of the myelin sheath of nerve cells as well as for synthesis of neurotransmitters. In MS, the immune system of the patient can attack the central nervous system (CNS), and the myelin sheath might be damaged. Considering that folic acid and vitamin B12 may have roles in the prevention of disorders of CNS development, lower intakes of folic acid and vitamin B12 might be expected in MS patients. In the current study, although strong statistical differences were observed between the control and MS groups in terms of the folic acid intake, vitamin B12 intakes did not differ significantly between the groups. In literature, research on this subject has revealed contradictory results.,,,
The analyses and the results of this paper aimed to make contributions to the literature in terms of the recognition of MS by comparing nutritional status and biochemical information of people with and without MS. By supplying a decision tree model, a new research area was tried to be created on the determination of the critical values for MS patients. Being a descriptive tool, the constructed tree serves perfectly to compare the parameter differences in the control and MS groups. It indicates that the main parameters that differ between an MS patient and a healthy person are as follows: serum levels of vitamin D, magnesium, calcium, intakes of potassium and CHO. It is vital to note that the parameters that constructed the decision tree should not be thought as the causes of MS. The aim in this study was not to seek for causality. As a controlled descriptive study, the differences between the MS patients and healthy individuals were tried to be observed in terms of certain parameters. The results reveal that especially the vitamin D, magnesium, and calcium levels in blood are expected to be lower for an MS patient.
The study could not be designed at MS onset. All the measurements were taken after the diagnosis of MS in the patient population. Effect of sunlight exposure on vitamin D levels could not be observed. Since all the patients were from the same nation, there is still ambiguity on whether or not the ethnicity has an effect on the presence of MS. The ratio of “male patients to female patients” in the MS group of this study was one third, which may have resulted from the fact that MS is more prevalent in women than men. For further studies, the sample size might be increased, and the study can be carried out on different ethnic groups. Moreover, since the current study was designed as a controlled descriptive study, causality between the presence of MS and the associated variables could not be searched. However, descriptive findings of this study can be used to construct a well-organized cohort study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]