Young adults with coeliac disease may be at increased risk of early atherosclerosis


Correspondence to:

Dr S. De Marchi, Divisione di Riabilitazione Vascolare, Ospedale Policlinico di Verona

37100 Verona, Italy.




Accelerated progression of atherosclerosis and increased cardiovascular risk have been described in immune-mediated disorders, but few data are available in coeliac disease.


To evaluate instrumental and biochemical signs of atherosclerosis risk in 20 adults at first diagnosis of coeliac disease and after 6–8 months of gluten-free diet with mucosal recovery.


We analysed total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol, triglycerides, homocysteine, C-reactive protein, folate and vitamin B12; ultrasound measurement of carotid intima-media thickness (IMT) and endothelium-dependent dilatation were both carried on at diagnosis and after gluten withdrawal. Twenty-two healthy members of the hospital staff served as matched controls for vascular examinations.


At baseline, mean total and HDL-cholesterol (HDL-C) were both within normal range, while mean LDL-cholesterol concentration was slightly increased; diet was associated with an increment in total and HDL-C (68.2 ± 17.4 vs. 51.4 ± 18.6 mg/dL; P < 0.001) and a significant improvement in total/HDL-C ratio (3.05 ± 0.71 vs. 3.77 ± 0.92; P < 0.02). Mean plasma homocysteine was elevated and not influenced by diet. C-reactive protein significantly decreased with diet (1.073 ± 0.51 vs. 1.92 ± 1.38 mg/dL; P < 0.05). At baseline, in coeliacs, IMT was increased (0.082 ± 0.011 vs. 0.058 ± 0.012 cm; P < 0.005), while endothelium-dependent dilatation was decreased (9.3 ± 1.3 vs. 11.2 ± 1.2%; P < 0.05). Both parameters improved after gluten abstinence.


Adults with coeliac disease seem to be at potentially increased risk of early atherosclerosis as suggested by vascular impairment and unfavourable biochemical risk pattern. Chronic inflammation might play a determining role. Gluten abstinence with mucosal normalisation reverts to normal the observed alterations.


Coeliac disease is a chronic small bowel immune-mediated enteropathy precipitated by exposure to dietary gluten in genetically predisposed people.[1] The immunological response to gluten in affected individuals causes histological abnormalities of the small intestinal mucosa with chronic inflammation, leading eventually to villous atrophy and secondary nutrients malabsorption. Symptoms of overt intestinal malabsorption may not be present and coeliac disease may be ultimately suspected after diagnosing an associated autoimmune disorder.[2] Disordered immune response in coeliac disease seems not to be limited to the bowel, but may involve different organs through unknown mechanisms.[3] As a matter of fact, a number of autoimmune diseases have been reported in association with coeliac disease. Hashimoto thyroiditis, insulin-dependent diabetes mellitus, autoimmune liver disease and connective tissue disease have all been reported as occurring in coeliac disease more frequently than in general population.[4] Recently, heart disease, both idiopathic and of immune pathogenesis, has also been reported in coeliacs.[5-7] Moreover, some patients with contemporary coeliac disease and cardiomyopathy showed improvement of heart function after gluten withdrawal.[5] In connective tissue diseases, an increased risk of developing early atherosclerosis and related complications has been reported, particularly in rheumatoid arthritis.[8-13] This risk seems to be related to systemic inflammation, suggesting potential pathogenetic implications for other chronic inflammatory diseases.[9, 12, 13] Risk of death for cardiovascular diseases is increased in coeliacs compared with the general population according to two epidemiological studies run in Scandinavian countries.[14, 15] In addition, an increased prevalence of coeliac disease has been reported in patients with low serum HDL-cholesterol (HDL-C) concentration, an acknowledged risk factor for cardiovascular disease.[16, 17] However, a recent study from the UK failed to show an increased risk for cardiovascular disease in coeliacs.[18] The measurement of intima-media thickness (IMT) at carotid artery and the calculation of endothelial-dependent dilation (EDD) at humeral artery are both considered markers of early atherosclerosis and seem to correlate with cardiovascular morbility and mortality.[19-21] IMT can show early atherosclerosis process that involves artery wall remodelling; this depends in part on endothelial dysfunction and may be partially reversible. In addition, EDD is considered an index of endothelial function and it expresses the capacity of endothelium to dilate vessels depending on a flow-mediated mechanism.[22-24]

The aim of this pilot study was to evaluate prospectively biochemical and instrumental markers of risk of atherosclerosis in a group of recently diagnosed coeliacs and the influence of gluten withdrawal on these parameters. The study was approved by our Institution Ethical Committee.


From January 2008 to April 2011, 29 consecutive patients with positive serology for coeliac disease were considered for the study. Inclusion criteria were: (i) signs or symptoms suggestive of intestinal malabsorption, (ii) plasma anti-tissue transglutaminase (tTG) and anti-endomysial antibodies (EMA) both tested positive after diagnostic workload, (iii) willingness of complying with the study procedures. Exclusion criteria were: (i) <18 age >50 year old, (ii) actual or desired pregnancy, (iii) former use of contraceptive pill, (iv) vitamin supplementation in the previous 3 months, (v) former gluten-abstinence diet, (vi) hypertension or other cardiovascular comorbidities by history, (vii) former diagnosis of inflammatory bowel disease or other immune-related diseases. Twenty patients meeting these criteria signed an informed consent form to join the study and underwent study investigations while on the waiting list to endoscopy. Appropriate investigations (i.e. colonoscopy, barium meal and follow-through, abdominal ultrasound scan) to exclude digestive diseases other than coeliac disease were performed when deemed indicated with no evidence of associated diseases. A careful biochemical search for immune disorders was also carried on including erythrocyte sedimentation rate, complete blood cell count, fasting blood glucose and serum potassium levels, alanine–aspartate transaminase, anti-nuclear antibody test, rheumatoid factor blood test and plasma thyroid-stimulating hormone. All patients underwent an evaluation of plasma lipid profile (total cholesterol, HDL and LDL cholesterol, triglycerides), C-reactive protein (CRP), plasma concentration of folic acid, vitamin B12 and homocysteine. All biochemical tests were determined by standard laboratory methodologies. Body mass index (BMI: kg/m2) was also recorded in all subjects.

In addition, all subjects underwent an hemodynamic study protocol as follows: After an overnight fast, hemodynamic measurements were carried out approximately at 9:00 AM in a thermostatic chamber (22 ± 1 °C) according to a previously published technique.[25] Briefly, after a 30-min resting interval, they underwent colour Doppler ultrasound measurement of IMT at both common carotid arteries and examination for detection of atherosclerotic lesions. Then EDD at humeral artery was measured at the left arm.

Intima-media thickness was examined by one of the authors blinded to the pertinence group of the subject tested (either coeliac disease patients or healthy controls) at both common carotid arteries by means of automatic ultrasound detection of IMT [Esaote MyLab (Florence, Italy) 30 Gold – quality intima-media thickness (QIMT)]. We used QIMT with automatic algorithm for detection of intima media previously described.[26] Common carotids were examined at standard angles bilaterally; 1 cm proximally to the bulb, a segment of 2 cm was selected with cursor of the system and automatically calculated the IMT; we took into account median IMT values for each common carotid artery. At our Laboratory, intraobserver variability was been previously reported as little as 0.5–0.9%. Media between the sites were measured and expressed as centimetres. Calibre and flow of the humeral artery at rest and after ischemia were obtained to evaluate EDD. The procedure was previously validated in our laboratory and it was carried on in accordance with Celermajer et al.[23, 24] Briefly, the probe was placed 6 cm proximally to the antecubital crease and fixed in position with stereotactic clamp. Baseline images were recorded for 1 min followed by inflation of a pneumatic tourniquet around the left forearm to a pressure exceeding 20 mmHg the systolic blood pressure for 3 min; the cuff was released and the artery was scanned for further 5 min for measurement of calibre and flow. The ultrasound images were analysed by another operator who was blinded to the identity of the volunteer and study phase.

The increase in humeral calibre is expressed as incremental percentage of the basal value. Beat-to-beat blood pressure was monitored continuously by plethysmography equipment at left hand (Finapress 2300, Ohmeda, Englewood, CO, USA), throughout the test.

A group of 22 healthy subjects was enrolled among hospital staff and served as control group for the hemodynamic protocol. The control group was closely matched with coeliacs in terms of sex, age, BMI and smoking habits. In addition, all controls were tested and scored negative for both plasma anti-EMA and anti-tTG to exclude subclinical coeliac disease.

Eventually, all coeliac patients underwent conventional endoscopy to obtain duodenal biopsies that were scored according to a modified Marsh classification.[27] Between four and six biopsies were taken from the descending duodenum to ensure adequate sampling.[28] The suspected diagnosis of coeliac disease was confirmed by histology in all subjects. After 6–8 months of gluten withdrawal, all patients agreed to undergo both repeated endoscopy and biochemical–hemodynamic evaluations. This repeated endoscopy was deemed necessary to confirm mucosal recovery on gluten abstinence and patients' compliance with dietetic suggestions.[28, 29]

Statistical analysis

Data are expressed as mean ± standard deviation (s.d.). Analysis of variance was performed, followed by a post hoc t-test to analyse data. Values of P < 0.05 were considered significant.


Twenty patients (11 females, age range 23–41 years) agreed to take part in the study. Reasons for the referral were: unexplained anaemia (six patients), irritable bowel syndrome unresponsive to conventional treatment (four patients), severe functional dyspepsia (four patients) and chronic diarrhoea (six patients). One coeliac patient was diagnosed as affected by Hashimoto thyroiditis, but no other immune disorders, diabetes mellitus included, were present in this group of patients. We decided to include patient/s where an immune disorder was found on diagnostic work-up to decrease potential selection bias. One patient was currently smoking approximately 20 cigarettes per day, while the others were no-smokers. Four patients reported familiarity for cardiovascular diseases, two for stroke, five for hypertension, while no vascular-related familiar risk was reported by the remaining nine patients. Blood pressure and electrocardiogram (12 derivations) were normal in all subjects. The presumed diagnosis of coeliac disease was first confirmed by histology in all subjects. Fifteen patients showed features of partial villous atrophy (Marsh IIIA), while subtotal–total villous atrophy was evident in the remaining five (Marsh IIIB-C). After gluten abstinence, repeat small bowel biopsy showed a normal duodenal histology in all patients, but one in which a Marsh I lesion was described (normal mucosal architecture and increased intraepithelial lymphocytes ≥40/100 enterocytes).

Only two patients were underweight (BMI = 17.9), while all other subjects showed BMI within normal range (mean BMI 20.51, range 17.9–22.9). BMI increased significantly on gluten withdrawal (mean BMI 21.07, range 18.9–23.2) compared with baseline (P < 0.03). At baseline evaluation, mean total plasma cholesterol and mean HDL cholesterol concentration were both within normal limits, while mean LDL cholesterol concentration was slightly out of the normal range (110.73 ± 24.65 mg/dL reference range <100 mg/dL, Table 1). Compared with baseline, gluten abstinence was associated with a significant increment in both total plasma cholesterol and HDL plasma cholesterol concentrations (P < 0.03 and P < 0.001, respectively), while LDL cholesterol concentration remained unaffected (Table 1). As a consequence, the total cholesterol/HDL ratio was significantly improved by gluten withdrawal (3.05 ± 0.71 vs. 3.77 ± 0.92 ratio, reference range >4; P < 0.02). On the contrary, plasma triglyceride concentration was within normal range in all patients and not influenced by gluten withdrawal (Table 1). CRP was also significantly decreased on gluten-free diet (P < 0.05 – Table 1). At baseline, mean plasma homocysteine was superior to normal range for our laboratory (18.94 ± 7.35 μmol/L; reference range <15 μmol/L). However, gluten abstinence was associated with a nonsignificant decrease in homocysteine concentration and both folate and vitamin B12 concentration remained unaffected as well (Table 1).

Table 1. Metabolic biochemical parameters and significance levels in 20 coeliacs at diagnosis (baseline) and after 6–8 months of gluten withdrawal (means ± s.d.)
 Baseline6–8 months of gluten abstinence P
  1. RR, reference range.

Total cholesterol (mg/dL)

(RR <200 mg/dL)

185.4 ± 37.77204.42 ± 35.2<0.003

HDL cholesterol (mg/dL)

(RR >40 mg/dL)

51.4 ± 18.668.2 ± 17.4<0.001

LDL cholesterol (mg/dL)

(RR <100 mg/dL)

110.73 ± 26.65107.91 ± 22.8NS

Total cholesterol/HDL ratio

(RR >4 ratio)

3.77 ± 0.923.05 ± 0.71<0.02

Triglycerides (mg/dL)

(RR <150 mg/dL)

85.9 ± 46.187.4 ± 45.9NS

CRP (mg/L)

(RR <5 mg/L)

1.92 ± 1.381.073 ± 0.51<0.05

Homocysteine (μmol/L)

(RR <15 μmol/L)

18.6 ± 8.916.6 ± 5.8NS

Folate (ng/mL)

(RR 5–16 mg/dL)

9.1 ± 4.410.9 ± 4.6NS

Vitamin B12 (pg/mL)

(RR 200–900 pg/mL)

346 ± 233520 ± 175NS

Blood pressure and heart rate did not change throughout the study and no difference was found between groups (Table 2). QIMT was significantly increased in the coeliac group compared with the controls (0.082 ± 0.011 vs. 0.058 ± 0.012 cm; P < 0.005). Gluten-free diet was associated with a significant decrease in QIMT in coeliacs (0.064 ± 0.010 vs. 0.082 ± 0.011 cm; P < 0.03). EDD was reduced in coeliac patients compared with the controls (9.3 ± 1.3 vs. 11.2 ± 1.2%; P < 0.05) and it recovered after dietetic intervention (12.1 ± 2.3 vs. 9.3 ± 1.3%; P = 0.05 – Figure 1).

Table 2. Blood pressure and heart rate in 20 coeliacs at disease diagnosis (baseline) and after 6–8 months of gluten-free diet and in 22 controls at the same interval (means ± s.d.)
  Baseline6–8 months
  1. SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate.

SBP (mmHg)Coeliacs118.7 ± 4.6117.8 ± 6.3
Controls123.2 ± 6.4122.6 ± 5.2
DBP (mmHg)Coeliacs71.2 ± 674.6 ± 8
Controls75.1 ± 874.7 ± 5
HR (bpm)Coeliacs69.6 ± 668.7 ± 5
Controls71.5 ± 770.9 ± 8
Figure 1.

Intima-media thickness in 20 coeliacs at disease diagnosis (CD baseline) and in 22 controls; additional testing was performed in coeliacs after 6–8 months of gluten-free diet (*< 0.005 coeliacs vs. controls; **< 0.03 gluten-free coeliacs vs. baseline coeliacs).


An accelerated progression of atherosclerosis has been reported in immune disorders.[30] This is likely to contribute to the increased mortality due to cardiovascular disease observed in some of them (i.e. rheumatoid arthritis).[8] Systemic inflammation seems to be a relevant pathogenetic factor, but chronic immune inflammation limited to the bowel may act differently.[12, 13, 30] Bernstein et al. found an overall increase in the prevalence of ischaemic heart disease and cerebrovascular disease, although the latter was limited to Crohn's disease patients.[31] However, a recent meta-analysis provided indirect evidence against an association between inflammatory bowel disease and increased risk of cardiovascular disease.[32] Coeliac disease is an autoimmune disorder triggered by gluten ingestion.[28] The small bowel is the privileged immune target, but a number of associated immune disorders have been described, suggesting a diffuse immune activation.[2, 4] Severe heart disease of potentially immune pathogenesis has been described in coeliac disease.[7] Cardiac performances seem to improve on a gluten-free diet in some first diagnosed coeliacs with associated cardiomyopathy.[5] Studies on morbidity and mortality in coeliac disease have traditionally focused on the increased risk of malignancy.[28] Few data are available on mortality from other causes in coeliac disease with conflicting results. Two large population studies run in Scandinavian countries have both shown a significant increment in mortality rates secondary to cardiovascular disease in coeliac patients when compared with the general population.[14, 15] More in depth, Peters et al. reported vascular disease as the most important single cause of mortality accounting for 40% of all deaths in coeliac disease.[15] However, UK-based population studies have given a somewhat different scenario. Whorwell et al. found a 40% reduction in ischemic disease mortality in coeliac disease.[33] On the contrary, data from a Scottish community-based cohort study suggested that cardiovascular risk seems to be increased in coeliac disease.[34] However, a recent large population study run on 3790 coeliac adults from the General Practice Research Database of Nottingham, UK, failed to show any increment in mortality rates secondary to cardiovascular disease.[18] The authors concluded that the effect of a gluten-free diet on cardiovascular risk factors should be determined before entertaining screening programmes in coeliac disease.[18] As far as our knowledge goes, only biochemical signs of atherosclerosis have been previously studied in coeliac disease. Low plasma total cholesterol has been traditionally considered to correlate with fat malabsorption in coeliac disease.[35] However, a recent study has failed to confirm this feature of coeliac disease,[36] most likely because subclinical diagnosis of coeliac disease is becoming increasingly common.[37] On the contrary, a low-serum HDL-C concentration has been described as a potential sign of coeliac disease.[16] Disordered lipid absorption and/or decreased Apo-A1 secretion may be responsible for the observed alteration.[16] As a matter of fact, gluten withdrawal seems to have a beneficial effect on lipid profile, increasing plasma HDL-C concentration in coeliac disease.[16, 36] In addition, an increased plasma homocysteine level has been described in coeliac disease most likely due to a poor vitamin status.[38] Both low-plasma HDL-C and high plasma homocysteine are considered relevant risk factors for cardiovascular disease.[39] Our data confirm the beneficial effect on lipid profile of a gluten-free diet in coeliac disease with a secondary increase in HDL-C plasma concentration.[36] However, we provide additional evidence to correlate biochemical improvement with recovered duodenal histology that was lacking in previous studies. As guidelines on repeated biopsy in coeliac disease are controversial, we choose to biopsy coeliac patients after 6–8 months of gluten-free diet according to our previous experience.[28, 29]

Hyperhomocysteinemia was evident in most of our patients, but gluten abstinence failed to improve it. We have no clear explanation for this finding. In this regard, a potential limitation of our study is our failure to evaluate red blood cell folate levels and the methylenetetrahydrofolate reductase genetic variant that impairs the ability to process folate in affected subjects.[40] However, additional time to observe a beneficial influence on homocysteinaemia by gluten withdrawal might be needed. Moreover, homocysteinaemia may persist elevated even in dietetically compliant coeliacs.[38]

Increased IMT is a sign of initial atherosclerotic damage.[21] It is related to a series of cardiovascular risk factors such as hypertension, dyslipidaemia, diabetes, autoimmune diseases.[11, 21, 41, 42] Enhanced immune and inflammatory cascade is a condition of endothelial damage, as it causes an endothelial activation with over-expression of adhesion molecules and production of reactive oxygen species.[30] Cytokines are also involved in vascular damage produced by several risk factors through activation of leucocytes and inflammation cascade (e.g. in plaque formation); cytokine-mediated mechanisms are involved both in autoimmune disorders and in cardiovascular damage.[9, 10, 14] These conditions cause modifications of intima-media structure with augmented collagen synthesis, increased cellular population and lipid deposition.[19] Endothelium modifies its function with depressed synthesis of nitric oxide and/or increased destruction, thus producing a reduced vasodilating capacity.[11] Our data show that IMT is significantly increased in young adults with coeliac disease compared with healthy volunteers. Additional information might have been given by repeated ultrasound investigation in healthy volunteers. However, we decided not to perform a double hemodynamic evaluation in the healthy control group for the expected BMI increase in coeliac disease patients when on a gluten-free diet. In addition, the variability in repeated hemodynamic evaluation in healthy volunteers by the techniques employed has been reported as little as 1%.[24] Normalisation of the small bowel mucosa secondary to gluten abstinence is associated with reversal to normal of this parameter of early atherosclerosis. An analogue beneficial trend was evident on EDD (Figure 2), another parameter of early atherosclerosis.[23] Endothelial-dependent dilatation was significantly decreased in coeliacs compared with healthy volunteers and significantly improved by gluten withdrawal. Our data are consistent with a decreased inflammatory response in coeliac patients secondary to gluten abstinence as confirmed by duodenal histology and by a significant decrement in CRP levels. We may speculate that this decrement may have a beneficial effect on early atherosclerotic lesions in coeliac disease in accordance with the theory of an inflammatory genesis of atherosclerosis in immune disorders.[30] Additional benefit may be given by the observed improvement in lipid profile secondary to gluten abstinence. Regression of early and advanced atherosclerotic lesions secondary to appropriate treatment has been previously reported.[42]

Figure 2.

Endothelium-dependent dilatation in 20 coeliacs at disease diagnosis (CD baseline) and in 22 controls; additional testing was performed in coeliacs after 6–8 months of gluten-free diet (*< 0.05 coeliacs vs. controls; **< 0.05 gluten-free coeliacs vs. baseline coeliacs).

However, there are a number of limitations in our preliminary study. The main limitation is given by the small sample size and the lack of a power calculation at the beginning of the study. However, considering the invasive nature of repeated endoscopy, we choose to rely on a sample size similar to that reported as meaningful in previous pilot studies evaluating non-invasive signs of cardiovascular risk in coeliac, metabolic and connective tissue diseases.[16, 43-45] The small sample size has also hampered the search for a potential correlation between bowel inflammation and atherosclerosis risk. However, a recent study has failed to provide evidence of correlation between clinical presentation and extent of bowel involvement in coeliac disease.[46] Physical activity was neither analysed nor matched between coeliac patients and healthy volunteers. Exercise and regular physical activity are established means to prevent early atherosclerosis.[47] However, we expected some of our patients to be affected by chronic anaemia, leading physical activity to a difficult parameter to control for. Nonetheless, the rapid improvement of signs suggestive of early atherosclerosis in our patients stands against a relevant contribution of a likely increased physical activity secondary to an improved nutritional status in our patients.


In conclusion, our pilot study raises concerns on potentially increased cardiovascular risk in coeliac disease secondary to accelerated atherosclerosis and unfavourable biochemical profile. Normalisation of the small bowel mucosa due to gluten withdrawal seems to be associated with a beneficial effect, but further larger studies are warranted to better define cardiovascular risk factors in coeliac disease and potential positive modification induced by gluten abstinence.


Guarantor of the article: Sergio De Marchi.

Author contributions: De Marchi S designed the protocol, analyzed the data and wrote the paper. Chiarioni G designed the protocol, collected the data and wrote the paper. Prior M collected and analyzed the data. Arosio E designed the protocol and provided data interpretation. All authors approved the final version of the article, including the authorship list.


Declaration of personal and funding interests: None.