Article Text
Abstract
Objectives: To investigate the association of left ventricular outflow tract (LVOT) obstruction with blood coagulation, platelet activity and inflammatory response in patients with hypertrophic cardiomyopathy (HCM) and sinus rhythm.
Patients and main outcome measures: In 42 patients with HCM with sinus rhythm, including 16 patients with resting LVOT obstruction (gradient ⩾30 mm Hg) and 29 age- and sex-matched controls, markers of thrombin generation (thrombin–antithrombin complex (TAT), prothrombin fragment 1+2 (F1+2)), platelet activation (soluble CD40 ligand (sCD40L), β-thromboglobulin (β-TG), P-selectin) and inflammation (C-reactive protein (CRP), interleukin (IL)6, tumour necrosis factor-α (TNFα)) were determined.
Results: Thrombin, platelet and inflammatory markers were higher in the entire HCM group than in controls (p<0.005 for all compared parameters). Compared with non-obstructive HCM, obstructive HCM was associated with increased thrombin formation (TAT, F1+2), platelet activation (sCD40L, β-TG, P-selectin) and both CRP and IL6 levels. Only the level of TNFα was similar in both forms of HCM. In contrast, a comparison of non-obstructive HCM with controls showed that all these variables (except for P-selectin) were similar; P-selectin was higher in non-obstructive HCM. The LVOT gradient correlated positively with all the raised blood markers (r from 0.39 to 0.73; p<0.05), except for TNFα. In multiple regression analysis models, the LVOT gradient was the only independent predictor of TAT (R2 = 0.61; p<0.001), sCD40L (R2 = 0.59; p<0.001), F1+2 (R2 = 0.55; p = 0.002), P-selectin (R2 = 0.49; p = 0.004) and β-TG (R2 = 0.38; p = 0.005) in patients with HCM.
Conclusions: LVOT obstruction is independently associated with enhanced thrombin generation and platelet activity in patients with HCM with sinus rhythm.
Statistics from Altmetric.com
In hypertrophic cardiomyopathy (HCM) acute thromboembolic vascular events such as stroke and peripheral arterial embolisation are not rare, particularly in patients aged >50 years.1 2 During the 7-year follow-up,2 stroke and peripheral vascular events occurred in 6% of the patients with HCM and these thromboembolic events were associated with atrial fibrillation.2 However, 35% of patients with HCM with atrial fibrillation are free of stroke, with a benign outcome during the follow-up.3 Therefore, atrial fibrillation cannot be the only determinant of stroke in patients with HCM and other potential contributors include advanced age, left atrial dilatation, left ventricular outflow tract (LVOT) obstruction and female gender.1 4 It might be hypothesised that HCM leads to the prothrombotic state. The only report of this subject5 demonstrated activation of blood coagulation, reflected by increased levels of thrombin generation markers. However, the number of patients was small (13 patients), the LVOT gradient was not measured and inflammatory markers were not determined.5 Accumulating evidence indicates that the hypercoagulable state is linked to the inflammation process.6 Interestingly, it has been reported that patients with HCM had higher interleukin 6 (IL6) levels than normal subjects, while tumour necrosis factor α (TNFα) levels were similar in both groups.7
LVOT obstruction is related to unfavourable clinical outcome in HCM,8 9 including the occurrence of stroke.8 However, it has not been studied to date whether LVOT obstruction affects haemostatic and inflammatory markers. Our hypothesis of an association between LVOT obstruction and levels of these markers is based on the observation that accelerated and turbulent flow increases the shear stress,10 which predisposes to the prothrombotic state.11
PATIENTS AND METHODS
Study group
We studied 42 patients (24 men and 18 women, mean (SD) age 47 (15) years) diagnosed with HCM on the basis of typical clinical, echocardiographic and haemodynamic features.12 The severity and distribution of left ventricular (LV) hypertrophy were assessed by echocardiography and maximal LV wall thickness was determined at the septum in all patients. A resting LVOT gradient >30 mm Hg (obstructive form) was found in 16 patients. Exclusion criteria were acute illness, arterial hypertension, cancer, paroxysmal or persistent atrial fibrillation, autoimmune disorders, history of bleeding or venous thromboembolism, endocarditis, previous cerebrovascular event, myocardial infarction, renal or hepatic dysfunction, and treatment with oral anticoagulants or antiplatelet drugs. Patients who were subjected to alcohol septal ablation were also excluded from the study. Based on echocardiographic data, subjects with LV cavity dilatation and depressed LV contractility were also excluded from the study. Patients were treated with verapamil (n = 12) or β-blockers (n = 13). In these patients, it was considered unsafe to discontinue drug treatment. Seventeen patients who were referred for the initial evaluation received no drug treatment.
Coronary angiography was performed in 12 patients and normal coronary arteries were detected.
The control group included 29 subjects matched for age, sex, diabetes mellitus and smoking status, who were recruited from hospital staff. Exclusion criteria were identical to those for patients with HCM. Echocardiographic evaluation disclosed no abnormalities of cardiac structure and function in the control group.
All subjects gave their informed consent to participate in the study. The Jagiellonian University Ethical Committee approved the study.
Echocardiographic study
Transthoracic echocardiographic examinations were performed using a Vivid 7 ultrasound machine. In each patient M-mode and two-dimensional echocardiograms were obtained, followed by pulsed and continuous-wave Doppler recordings. Conventional techniques were used to measure the left ventricle size. The LV contractility was assessed by fractional shortening as recommended in a previous study.13 The LVOT gradient was measured using Doppler recordings and a value of >30 mm Hg was considered significant. As in the study of Maron et al,14 care was taken to report only those gradients derived from Doppler velocity profiles typical of subaortic obstruction, which allows contamination by the mitral regurgitation jet to be avoided. Morphological assessment of the narrowed LVOT was performed by measurement of the minimal distance between the mitral valve and ventricular septum during systole in the parasternal long-axis view with M-mode echocardiography.15
Laboratory variables
Blood samples were taken with minimal stasis between 08:00 and 09:00 after an overnight fast within 10 minutes following echocardiography with Doppler measurement of the LVOT gradient. Blood morphology, glucose, creatinine, aminotransferases and creatine kinase were assayed by routine laboratory techniques. Plasma samples were centrifuged at 2000 g for 20 minutes and stored in aliquots at −80°C until analysis. Using commercially available assays according to the manufacturers’ instructions, technicians unaware of all subject data determined plasma markers of:
thrombin generation: thrombin–antithrombin complexes (TAT; Dade Behring, Marburg, Germany), interassay and intra-assay coefficients of variation, 5.6% and 5.2%, respectively; and prothrombin fragment 1+2 (F1+2; Dade Behring), interassay and intra-assay coefficients of variation, 4.0% and 3.8%, respectively;
platelet activation: soluble CD40 ligand (sCD40L; R&D Systems, Abingdon, UK), interassay and intra-assay coefficients of variation, 7.8% and 8.1%, respectively; β-thromboglobulin (β-TG; Diagnostica Stago, Asnieres, France), interassay and intra-assay coefficients of variation, 7.0% and 5.7%, respectively; P-selectin (R&D Systems), interassay and intra-assay coefficients of variation, 6.7% and 6.9%, respectively;
inflammation: C-reactive protein (CRP; Dade Behring), interassay and intra-assay coefficients of variation, 8.3% and 8.0%, respectively; IL6 (R&D Systems), interassay and intra-assay coefficients of variation, 7.2% and 5.6%, respectively; tumour necrosis factor-α (TNFα; R&D Systems), interassay and intra-assay coefficients of variation, 8.8% and 6.0%, respectively.
Statistical analysis
Data are expressed as means (SD) or number and percentage. The Kolmogorov–Smirnov test was used to determine normal distribution. Differences between the groups were tested for statistical significance, using the analysis of variance test for continuous variables, and χ2 test (or Fisher exact test) for categorical variables. Linear regression analysis was performed to correlate the variables. Multiple linear regression analysis was performed to determine independent predictors for all biochemical variables using age and echocardiographic measures (LVOT obstruction, mitral–septal distance, LV end-systolic and LV end-diastolic diameter, septal thickness at end diastole, fractional shortening, left atrial diameter). The data presented represent the best-fit model as determined by the R2 value. A p value <0.05 was considered significant.
RESULTS
Study participants
Table 1 shows the demographic and clinical characteristics of patients with HCM and controls. Both groups had similar age, sex distribution, body mass index, hypercholesterolaemia, smoking and diabetes. In the HCM group 12 patients had a history of syncope, 20 had non-sustained ventricular tachycardia in Holter monitoring, and 23 patients had a family history of HCM. Results of routine blood tests were normal (data not shown).
The LVOT gradient (from 30 to 104 mm Hg) was found in 16 patients and this group with obstructive HCM had similar demographic and clinical features to those of the non-obstructive HCM group (26 patients) (data not shown).
Echocardiographic examination of the entire HCM group showed a non-dilated LV cavity (LV end-diastolic diameter, 44.3(5.5) mm, LV end-systolic diameter, 24.9 (5.5) mm) with normal contractility (fractional shortening, 44.1 (8.4)%). Septal thickness was increased to 25 (3.5) mm and the left atrium was dilated to 44.8 (5.4) mm despite sinus rhythm. The mean value of minimal distance between mitral valve and ventricular septum at systole was 9.6 (6.2) mm. In six patients the LVOT obstruction was severe and expressed by a complete mitral–septal contact during systole.
Patients with HCM versus controls
Baseline comparison
At the first step of the analysis the entire HCM group was compared with controls. Both thrombin generation markers—that is, TAT and F1+2, were raised in patients with HCM as compared with controls (table 1). All three platelet activation markers—that is, sCD40L, P-selectin and β-TG, were also raised in patients with HCM (table 1). Comparative analysis of inflammatory markers showed that the inflammatory response was significantly enhanced in the HCM group compared with controls (table 1).
Correlations between laboratory variables
In the entire HCM group, TAT and F1+2 were closely correlated (table 2). Like thrombin markers, platelet markers were significantly correlated (table 2). Moreover, thrombin generation markers showed positive correlations with platelet activation indicators. Similar associations were found for CRP and IL6 (table 2). Only TNFα was not related to all the remaining parameters. In controls there were significant correlations among all the markers studied (r between 0.4 and 0.8; p<0.05) with the exception of the lack of any association between F1+2 and P-selectin (r = 0.29; p>0.1).
Obstructive versus non-obstructive HCM
In patients with an obstructive form of HCM, we observed significantly higher levels of all markers (apart from TNFα) than those found in patients with non-obstructive HCM (table 3). An additional analysis was performed to compare non-obstructive HCM and controls (table 3). Patients with non-obstructive HCM differed from controls only with respect to plasma P-selectin levels which were higher by almost 30% in the former group (table 3). The levels of all remaining blood parameters were similar in these groups (table 3).
Correlations between LVOT obstruction and blood markers
Correlations between the LVOT gradient and blood markers were studied both in obstructive HCM and in the whole group of patients with HCM as LVOT obstruction (even with a gradient <30 mm Hg) has a continuous spectrum. We found that the magnitude of the LVOT gradient (both in obstructive HCM and in the entire HCM group) correlated positively with markers of thrombin generation and platelet activation, and also with IL6 and CRP (table 4). TAT, F1+2 and sCD40L showed the strongest associations with the LVOT gradient (table 4). Only TNFα was not correlated with the LVOT gradient.
The correlations between the minimal mitral–septal distance and these variables were weaker, but significant, with that of TAT being strongest (table 4). The associations of TAT, sCD40L, P-selectin, and β-TG with the left atrial diameter were weaker, but significant. However, CRP, IL6 and TNFα did not correlate with the left atrial size (table 4). There were no associations between laboratory variables and the LV cavity dimension (systolic and diastolic), fractional shortening and LV diastolic filling parameters (E wave and E/A ratio). Septal thickness showed no correlation with any of the markers studied (data not shown). Age, body mass index or sex showed no associations with markers of thrombin formation, platelet activation or inflammation (data not shown). There were no differences in these variables related to the treatment used in patients with HCM (data not shown).
In regression analysis models after adjusting for potential confounding factors, including age and echocardiographic parameters (LVOT obstruction, mitral–septal distance, LV end-systolic and LV end-diastolic diameter, septal thickness at end diastole, fractional shortening, left atrial diameter), the LVOT gradient was the only independent predictor of TAT (R2 = 0.61; p<0.001), sCD40L (R2 = 0.59; p<0.001), F1+2 (R2 = 0.55; p = 0.002), P-selectin (R2 = 0.49; p = 0.004), β-TG (R2 = 0.38; p = 0.005) in patients with HCM.
DISCUSSION
Shear stress abnormalities are known to impair haemostasis. We decided to study patients with HCM without atrial fibrillation. This can be justified by the fact that atrial fibrillation might lead to additional shear abnormalities and bias in the study.
As far as we know, our study is the first to demonstrate that patients with obstructive HCM and sinus rhythm are characterised by enhanced thrombin generation, platelet activity and inflammation as compared with patients with non-obstructive HCM and controls. Our findings confirm and extend the previous study by Yamamoto et al5 by showing that the new determinant of coagulation activation in HCM is the LVOT obstruction, and its impact is stronger than that found for the left atrial dilation.
Interactions between coagulation and inflammation
In our study thrombin formation is closely linked to augmented platelet activation and inflammatory response mediated by IL6 both in patients with HCM and controls. The regression analysis models in patients with HCM showed that both thrombin generation and platelet activation (but not the inflammatory response) are predicted independently by the LVOT obstruction, suggesting that increased inflammatory markers are probably a consequence of activation of blood coagulation in this disease. According to Esmon,6 the interactions between inflammation and coagulation are bidirectional. Enhanced inflammation can augment blood coagulation which, in turn, can stimulate the inflammatory response, whereas the failure of natural anticoagulant mechanisms to control blood clotting will enhance the inflammatory process.
We provide evidence for increased platelet activation in patients with HCM. Platelet activation also contributes to the inflammatory response—for example, through release of large amounts of the proinflammatory mediator, CD40L, which increases production of inflammatory cytokines such as IL6. Among platelet indices, sCD40L appears to reflect changes in platelet stimulation typical of obstructive HCM better than β-TG or P-selectin. In the current study there were significant correlations between platelet and inflammatory markers in patients with HCM and controls. Associations between platelet and inflammatory markers support a role of platelets as inflammatory blood components.16 Interestingly, patients with non-obstructive HCM differed from a well-matched control group only with regard to plasma P-selectin levels.
An intriguing finding is the association between LVOT obstruction and increased IL6 levels with the subsequent raised CRP that is produced predominantly in response to this proinflammatory cytokine. Previous studies on IL6 levels in patients with HCM which did not clearly identify those with the obstructive HCM yielded inconsistent results. In 19 patients in whom the presence or absence of the LVOT obstruction was not determined, IL6 levels were significantly raised,7 while in another study17 performed in patients with non-obstructive HCM, plasma IL6 concentrations were not increased. Our study showed that in non-obstructive HCM, IL6 levels are similar to those found in controls.
LVOT obstruction as a potential cause of the prothrombotic state
Recently, it has been documented that HCM is a predominantly obstructive disease in which LV outflow gradients are evident in most patients (ie, 70%).14 Therefore activation of blood coagulation and platelet activation in association with the LVOT gradient seem to be of importance.
There are several possible mechanisms by which LVOT obstruction may affect blood coagulation. Accelerated and turbulent blood flow due to the LVOT obstruction (with systolic anterior motion of the mitral leaflet) are likely to stimulate platelets. Both anatomic (systolic anterior motion–septal contact) and haemodynamic (pressure gradient) abnormalities18–21 generating turbulent flow in patients with obstructive HCM might play a part in the prothrombotic state reported here. Importantly, from the hydrodynamic point of view (Venturi and drag forces) the LVOT obstruction increases the shear stress,10 which has been convincingly shown to be associated with the prothrombotic state.11
Study limitations
Our study has limitations. First, the number of the patients enrolled in this study was small, mostly owing to numerous exclusion criteria. This resulted in a quite homogeneous population sample, which together with a well-matched control group, mitigate against significant recruitment bias. Second, our analysis was based on a determination of each variable at a single time point. Third, in our study we were unable to examine temporal relationships between the variables and clinical outcome, especially thromboembolic events. This relationship may be difficult to demonstrate because, for example, of silent (asymptomatic) cerebral infarcts. In a study of Kario et al,22 this asymptomatic (therefore frequently undetected) complication is not rare in hypertensive patients and it has been shown to be associated with a 30% increase in TAT and F1+2 levels. Similar percentage increases in these prothrombotic measures were seen in patients with obstructive HCM in the current study. Fourth, the percentage of patients with LVOT obstruction (38%) in HCM was higher than the percentage (about 25%) reported by other investigators.23 This difference may be related to the fact that 40% of our patients did not receive any drug treatment.
Fifth, as in the study of Thaman et al,13 only patients with a typical chest pain or risk factors for coronary artery disease underwent coronary arteriography. Consequently, not all patients, but only selected subjects, underwent coronary arteriography, which yielded normal results. Therefore, it is unlikely that coronary artery disease was an important confounding factor in the analyses performed in patients with HCM. A similar diagnostic approach was used in several previous studies.13 24
Finally, it is not known whether reduction in the LVOT gradient results in attenuated thrombin formation and platelet activation. Moreover, it is not known whether provocation of the LVOT gradient results in enhanced blood coagulation or platelet activation. A study evaluating the effect of an exercise-induced LVOT gradient on thrombin generation and platelet activation measured simultaneously is needed to elucidate this issue. Results of such a study may be important because an LVOT gradient develops frequently during exercise.14
CONCLUSIONS
LVOT obstruction is associated with enhanced thrombin generation, platelet activity and inflammation activity in patients with HCM with sinus rhythm. The strongest independent association was found between thrombin formation and the LVOT gradient. A large study is needed to assess the prognostic importance of increased coagulation, platelet and inflammatory markers in patients with HCM.
REFERENCES
Footnotes
Competing interests: None declared.
Ethics approval: The Jagiellonian University Ethical Committee approved the study.