A Comparative Study to Evaluate the Interplay of Lipoprotein (a) With Traditional Lipid Parameters in Overt and Subclinical Hypothyroidism

study. and the literature searches. Authors AD and SR carried out the analyses of the study performed and the spectroscopy analysis. Authors AD, SR and UB managed experimental process. ABSTRACT Aims: In conjunction with triglyceride (TG) and HDLc, changes in the Lp (a) level in hypothyroidism have shown variable results. In the present study we made an effort to evaluate the role of Lp(a) as cardiovascular risk factor in both subclinical(SH) and overt hypothyroidism(OH) along with its dependence with dyslipidemic changes found in both groups. Study Design: It was a cross sectional, observational, non interventional, hospital based study Place and Duration of the Study: The study period was one year spanning a duration from February 2014 to January 2015 in the Dept. of Biochemistry, Calcutta National Medical College, Kolkata. Methodology: We evaluated the changes in Lp(a) TG, HDLc and fT4 levels in 30 overt and 34 subclinical hypothyroid patients and compared them with 34 control subjects in a hospital based cross-sectional study. Data were compared for difference between mean values and obtaining dependence of Lp(a) on lipid parameters. Results: Mean values of Lp(a), TG, TC and LDLc were found to be increased most in the OH group followed by that in the SH patients, the difference between two groups being significant statistically (p < 0.001). In contrast, fT4 and HDLc showed decreased levels in both SH and OH groups with a significant difference between them. Results of multiple linear regression analysis revealed that changes in the Lp(a) levels showed significant positive and negative dependence on the TG ( β = 0.377 for OH and 0.296 for SH), and fT4 ( β = -0.699 for OH and -0.380 for SH) and HDLc ( β = -0.341 for OH and -0.393 for SH) respectively. Conclusion: Dyslipidemic features are evident in patients with SH as well as in the OH group that play also an important predictive role on the changes in Lp(a), indicating that in addition to traditional dyslipidemia, nontraditional risk factors like Lp (a) play a major role in initiating cardiovascular adverse events even in the early stages of hypothyroidism (SH).


INTRODUCTION
Depending on the TSH and thyroxine levels, overt (OH) and subclinical hypothyroidism (SH) are two broad spectrums of primary hypothyroidism. OH is defined as an elevated serum TSH with thyroxine level below normal reference interval, whereas SH, the milder spectrum of the disease, is defined as elevated serum TSH with blood thyroxine within the normal reference interval. The prevalence of SH has been stated to be about 4-10 percent worldwide [1] among general population and about 7 -10 percent among the elderly [2]. Although, being denoted as compensated hypothyroidism, SH has been found to be closely associated with several clinical and metabolic abnormalities like memory loss and depression that was improved significantly after levothyroxine therapy [3], increased coronary heart disease (CHD) [4] and elevated cardiovascular risk factors like increased LDL cholesterol and dyslipidemia [5]. Furthermore, one of the studies projected that 36.7% and 11.2% of SH patients showed persistently elevated TSH and ultimate progression to OH respectively during a course of 5 year period [6]. Keeping in mind this high frequency of SH and the related cardiovascular risks, it is becoming imprudent to consider it as truly compensated and hence to deem it as a state of mild thyroid failure with its potential adverse consequences related to metabolic abnormalities of thyroid deficiency.
Normal level of thyroid hormones is supposed to play a major role in the regulation of cellular metabolism including that of catabolic energy production by the cellular lipids [7]. The resulting hypercholesterolemia and hypertriglyceridemia along with a decreased HDL-cholesterol level are particular features of hypothyroidism [8,9]. Increased level of the oxidized LDL along with an elevated triglyceride (TG) initiate the generation of foam cells resulting into atherosclerotic plaques that ultimately leads to major cardiovascular complications. On the other hand, a persistent decrease in HDL-cholesterol level has been observed in hypothyroidism that hampers the reverse cholesterol transport from peripheral tissues back to the liver, promotes oxidized LDL formation, and potentiates mediated damage to the endothelial walls [10]. Thyroxine treatment has been found to increase the expression of Apo A-1 levels both at the transcription and post transcriptional levels by increasing its mRNA transcription and stabilizing its precursor form in the nucleus respectively, in both the liver and the intestine, thus explaining partly the reduction of HDLc fraction in hypothyroidism [11][12][13]. In contrary to this study, another one carried out by Tan et al. [14] described an increase in the HDL-2 fraction in the hypothyroid patients indicating an incongruity in observations regarding changes in HDLc fractions in hypothyroid patients. Furthermore, although the dyslipidemic features along with endothelial dysfunction, coagulation defects, metabolic, hormonal and hemodynamic changes pose pertinent risk factors for cardiovascular diseases and early death in OH patients [15], findings related to the SH patients have been not consistent so far. Some studies showed no changes in cholesterol level [16], some reported significantly increased values [17] whereas in some studies a non significant trend of increase was observed [9]. These inconsistent findings potentiate the need for further research for delineating the lipid abnormalities in SH patients regarding both conventional and unconventional markers like Lp(a) for the cardiovascular risk.
Lp(a), a non-traditional marker for cardiovascular risk, is a carbohydrate rich protein with structural homology to the plasminogen and is linked to the LDL particles [18][19][20]. Being structural homologue of the plasminogen, it inhibits the natural thrombolytic activity of the plasminogen and therefore, acts as a pro-atherogenic molecule that potentiates the cardiovascular risks significantly in subjects with its increased level [21,22]. Studies have indicated genetic predispositions [23] and several systemic disorders causing a significant elevation in its blood level. Although, thyroid hormones, growth hormones and sex steroids have been found to put an effect on the circulating level of Lp(a) [20,24] the role of thyroid hormone on its metabolism has shown conflicting results till date. Different studies have indicated both significant changes [25] as well as absence of any alteration in its level in the hypothyroid patients [17,26,27]. Moreover, in a cross sectional study, the plasma Lp(a) level was found to be significantly increased in OH patients that was normalized after effective therapy by levothyroxine [25]. Keeping in mind its potential elevation in hypothyroid disorders, further studies are needed to explore any possible alteration in its activity regarding to different stages of hypothyroidism, particularly the SH.
Relationship between the Lp (a) and HDL also remain controversial and is without any common agreement in different pathogenic conditions. Sharma et al. [28] reported a positive correlation between the overweight people and their HDLc, thus negating the Lp(a) as a potential cardiovascular risk factor in overweight people. On the other hand, recent studies have indicated a significant inverse correlation between the HDLc and Lp (a) in OH patients that underscores the importance of Lp(a) as an independent risk factor in this disorder [29].
Keeping these factors in mind, we hypothesized that dyslipidemia and Lp (a) levels may be affected differently between the OH and its milder form SH groups with changes in Lp(a) being more dependent on the biochemical parameters indicating dyslipidemia. Accordingly, the present study was designed and was carried out to find out the validity of this hypothesis and thereby its clinical application.

Study Design and Settings
The present study was conducted in the Dept. of Biochemistry of a tertiary care Medical College and Hospital, West Bengal, India as a cross sectional, observational, non interventional, hospital based study. The study period was 1 year that spanned the duration from February 2014 to January 2015.

Inclusion criteria for cases into OH and SH groups
Patients diagnosed to be suffering from primary hypothyroidism on the basis of serum TSH and fT4 values were selected on convenience basis attending the thyroid clinic of Biochemistry department during the stipulated study period. Sixty four cases aged between 20 to 40 years were selected for the final study. No gender preference was made during the case selection. Age and sex matched 34 healthy control subjects were selected from the same population group belonging to similar socioeconomic and nutritional status.
On the basis of fT4 values the cases were divided into two groups; i) SH group with raised TSH but fT4 within reference range, and ii) OH group with raised TSH with fT4 below the lower reference limit. Following these criteria thirty four (34) patients were included in the SH group whereas thirty (30) patients were put into the OH group.

Exclusion criteria for case subjects
Subjects having history of chronic alcohol ingestion, regular smoking habit, addiction to any drug, any malignant disorder, chronic inflammatory diseases and any other metabolic or endocrinological disorders other than primary hypothyroidism were excluded from the study. Same exclusion criteria were applied for selection of normal healthy population as the control group.

Ethical considerations
Informed consents were obtained from both the case and control groups as per protocol. Institutional ethical clearance was obtained before start of the study. Total study strictly adhered to the guidelines of the Helsinki Declaration 1975, revised in 2000. The study was started after getting written permission from the properly constituted ethical clearance from the institutional ethical committee (No. 90, dated 28.12.13 of Calcutta National Medical College, Kolkata).

Estimation of Serum Hormones
Serum TSH and fT4 were estimated by competitive and non competitive ELISA respectively by the kits obtained from Accubind, USA. Methods for TSH and fT 4 were reported to have pre-recorded regression coefficients of 0.995 and 0.920 when compared with a reference immunochemiluminescence assay and RIA respectively. The coefficients of variation (CV) were found to be 6 and 8 percent respectively for the TSH and fT4 in our lab during the study period.

Serum
Lp (a) was measured by immunoturbidimetric method obtained from ERBA, Transasia that reported a correlation coefficient r of 0.989 with standardized ELISA method. Lab CV for this assay was 5 percent.

Estimation of Serum Lipids
Serum lipids were estimated by standard spectrophotometric techniques in the autoanalyzer, XL 600 from ERBA, Transasia. Cholesterol was estimated by the CHOD-PAP reagent, whereas the LDLc and HDLc were measured by direct methods. All reagents were obtained from ERBA, Transasia. The CV for these assays remained within 6 percent throughout the assay process.

Statistical Analysis
Data obtained were analyzed for difference among the mean vaules of SH, OH and the control group by post hoc ANOVA with Bonferroni correction for correcting multiple testing. Dependence and predictive values of the study parameters in the case group were analyzed with the help of multiple linear regression study. For all studies, the p value was considered to be significant at a level of 0.05 or less with a confidence interval of 95%. The statistical analysis were carried out by using SPSS software for Windows, version 17.0.

Sample Characteristics
A Shapiro-Wilk's test (p>0.05, Table 1) suggested that values of TSH, fT4, TG, LDLc, HDLc and Lp(a) were approximately normally distributed. Furthermore, the ratio of the skewness and kurtosis to their corresponding standard errors were between -1.96 and +1.96 for all study parameters (Table 1) suggesting that the data were approximately normally distributed for all study groups. Graphical analysis with the help of Q-Q plots and histograms also indicated that the study parameters follow approximately the normal distribution pattern. Based on these observations, ANOVA and multiple linear regression analyses were performed to determine the significance of difference between the mean values and the strength of predictive values of thyroid and lipid parameters on Lp(a) in OH, SH and control groups.
The Table 2 shows the results of one way ANOVA performed on OH, SH and normal control subjects. From the p values (p<0.001 for all parameters), it is evident that there was significant difference between these three groups. However, to ascertain the degree of difference between the individual groups of OH, SH and control groups post hoc ANOVA with Bonferroni correction was performed and the results were shown in the Table 3. In the Table 3, the post hoc ANOVA revealed the significance of difference between the individual groups separately while Bonferroni correction compensated for the multiple comparisons. Mean values of serum TSH, TG, TC, LDLc and Lp(a) showed significant increase (p <0.001) in the OH groups compared to the SH groups, that in turn, showed a significant higher values than the control group (p < 0.001). On the other hand, values of fT4 and HDLc showed significant trend in opposite direction showing marked reduction in the OH group in comparison to the SH counterparts which was again lower than the normal controls. Results of ANOVA also revealed that there was no significant difference between the age and BMI distribution indicating that all three groups were matched for age and body weight. Chi square test indicated that these three groups were sex matched as well (χ2 = 0.09, p = 0.95, data not shown in Table).  Similarly, Table 5 showed the predictive values of the above parameters on the rise in serum Lp (a) level in the SH group. Results showed same trends as for the OH group (Table 4)

DISCUSSION
The present study was undertaken to assess the metabolic effects of both subclinical and overt hypothyroidism on both traditional and nontraditional lipid parameters including the pro-atherogenic Lp(a). Hypothyroid patients were divided into the SH and OH categories according to their fT4 levels and the effects of thyroxine and its trophic hormone TSH were noted separately in each group. Post hoc ANOVA in the Table 3 indicated that serum values of TSH and fT4 were significantly higher and lower respectively in both OH and SH groups in comparison to controls, with the changes being more significant in the OH group than that in the SH category (p < 0.001). This indicated that in spite of a significant compensatory elevation, in the advanced stage of the disease process, TSH could not maintain the thryoxine levels within the normal range in the OH patients. Similar trends were observed for the increase in serum TC, LDLc, and TG values and for the decrease in serum fT4 and HDLc levels in both OH and SH groups ( Table 3). Hypercholesterolemia in OH is a consistent feature [8] and is supposed to be due to lack of T3 and T4 mediated LDL receptor upregulation and cellular uptake of LDL particles [30] along with the absence of stimulatory effect of thyroxine on the gene of LDL receptor [31]. Our results regarding the higher values of LDLc in hypothyroid cases are in close agreement with these explanations. Almost similarly, elevated TG levels in hypothyroidism may result from the absence of the stimulatory effect of thyroxine hormones on the lipoprotein lipase activity that is essential for degradation of the TG molecules into VLDL and chylomicron remnants [32]. These observations proposed potential elevations in the atherogenic lipid fractions in the mild thyroid failure found in SH. Results of our present study validate these cues and strengthen the notion of a mild thyroid failure in SH resulting in significant metabolic derangements in the patients before they proceed into the OH state.
Regarding the non-traditional cardiovascular risk parameters, significant increase in the Lp (a) level in both OH and SH groups was found in our study that may be explained by an impaired catabolism of Lp (a) in hypothyroidism [25]. Higher values of Lp (a) in hypothyroidism is attributed to its impaired catabolism in the disease due to a functional deficiency of thyroxine hormone [33]. Furthermore, higher values in the OH group in comparison to the SH group (p < 0.001) in our study groups indicated a proportionate increase of this plasminogen inhibitor according to the severity of the disease.
This observation was reiterated by the fact that there was a significant negative dependence of Lp (a) on the circulating level of fT4 in both OH (β = -0.699, p = 0.015, Table 5) and SH (β = -0.347, p = 0.027, Table 5) groups i.e. the rise in serum Lp (a) levels showed a graded response with decrease in the fT4 values in our study subjects. From these results we stress on the fact that the thyroxine is the major negative predictor for increase of this non traditional cardiovascular risk factor as well. Furthermore, as Lp (a) levels show significantly elevated levels in the SH patients in comparison to the normal controls, we suggest that inspite of fT4 levels within the reference range, derangement in the biochemical functions start at the cellular levels at this stage.   (Tables 4 and 5). This dependence is furthermore important regarding the fact that along with a decrease in thyroxine levels, increased levels of TG and a decrease in HDL-c become significant predictors for an elevation in Lp(a). As these two parameters define dyslipidemia and are major criteria of the MS, SH patients become vividly linked to the cardiovascular risks of MS also significantly. However, in spite of showing significant dependence on the HDLc and TG levels, Lp (a) did not show any such relationship on the LDLc levels in either group (Tables 4 and 5). These findings keep in track with the findings that the catabolic regulation of the Lp(a) occurs in an independent way from that of the LDL and its receptor [35]. Lack of this mutual dependence have been earlier potentiated by the experiments involving the mutated LDL receptors and the inhibitors of the HMG CoA [36].
Importantly, the levels of HDL cholesterol on the other hand, showed significant decreases in both the OH and SH groups, the decrease being significantly more in the OH group (Table 3). An increase in the expression of Apo A-1 levels both at the transcriptional and post transcriptional level in the liver and the intestine suggest that lack of the stimulatory effects of the thyroxine in hypothyroid patients are one of the major causes for reduction in the HDLc levels in this disorder [11][12][13]. When considered with the elevated TG levels, their significant reciprocal alteration strongly suggests that the dyslipidemic changes involving decreased HDLc and increased TG start as early as the subclinical stages of hypothyroidism that further strengthens the cellular deficiency of thyroxine signifying a mild thyroid failure in SH. Furthermore, as raised TG level with decreased HDLc are typical characteristic features of MS [37], their cooccurrence in both SH and OH patients in our study population indicated a close link between the MS and hypothyroidism which are in accordance with the association of dyslipidemia of MS with SH as suggested in some recent studies also [38].
Considering the fact, that dyslipidemic features are evident in patients with subclinical hypothyroidism and they possess significant predictive effects for altering the non-traditional cardiovascular risk factors like Lp (a) towards an adverse directions, we can conclude that SH patients are equally prone to develop myocardial infarction and atherosclerosis as their OH counterparts. These findings are in well congruence with the report from Rotterdam study which concluded that patients with mild thyroid failure have a significantly increased prevalence of aortic atherosclerosis and myocardial infarctions [2]. In addition to this, the present study, not only heralds the importance of Lp (a) as an independent potential risk factor for atherogenesis at the subclinical stages of hypothyroidism, but also suggests its close link with increased TG and reduced HDLc, the major dyslipidemic features of the MS. Clinically it is important to recognize this association, and keeping this factor in mind, we recommend for an early diagnosis of the SH patients by improved biochemical assays followed by initiation of an early replacement therapy. Furthermore, we suggest for monitoring of these patients by screening for early changes in the Lp (a) levels along with the traditional markers of dyslipidemia for assessment of future cardiovascular risks. Hence, along with an early replacement therapy with thyroxine, proper management of hyperlipidemia and dyslipidemia by dietary or therapeutic measures are needed to reduce the risks for atherogenesis and cardiovascular deaths.

CONCLUSION
However, our observations and interpretations need to be judged in the context of inherent limitations of the case control non interventional studies, limited sample size and lack of genetic evidences affecting any population variation. In addition to above conclusions we thus further propose the need for further research considering larger sample sizes in different cohorts involving both biochemical and genetic factors.

CONSENT
Informed consents were obtained from both cases and control subjects in appropriately prepared consent forms as per protocol of the institutional ethical committee and the Helsinki guidelines.