Circulating adhesion molecules and carotid artery structural changes in patients with noninsulin-dependent diabetes mellitus.

Hypertension and non insulin-dependent diabetes mellitus (NIDDM) are well-known risk factors for atherosclerotic disease. Intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) may exert a relevant role in the pathogenesis of atherosclerosis; their prognostic relevance has been recently demonstrated. The aim of the study was to investigate possible inter-relation between circulating adhesion molecule levels, carotid artery structure and endothelial function in 15 patients with NIDDM, as well as in 15 patients with both NIDDM and essential hypertension (NIDDM+EH) compared with 15 normal subjects (NS) and 15 euglycaemic patients with EH, matched for age, sex and body weight. All subjects were submitted to a biopsy of the gluteal subcutaneous fat. Small arteries were dissected and mounted on a micromyograph, and the media-to-lumen (M/L) ratio was then calculated. Carotid artery structure was investigated by Doppler ultrasound. Endothelial function was evaluated by investigation of the flow-mediated dilatation (FMD) of the brachial artery. ICAM-1 and VCAM-1 plasma levels were measured by ELISA. ICAM-1 and VCAM-1 plasma levels were significantly greater and FMD smaller in EH, NIDDM and NIDDM+EH than in NS, but no difference was observed among the three pathological groups. Carotid artery structural changes were more pronounced in NIDDM+EH. No significant difference was observed among NIDDM, EH and NS. The M/L ratio of subcutaneous small resistance arteries was significantly greater in NIDDM+EH than in NIDDM or EH. NS had a smaller M/L ratio than the other groups. Significant correlations were observed between ICAM-1 plasma levels and indices of carotid artery structure in diabetic patients. However, the relations were close only in NIDDM+EH. In conclusion, our data suggest that NIDDM+EH may present more pronounced vascular structural alterations than NIDDM, and that adhesion molecules plasma levels are closely inter-related with carotid artery structural alterations, at least in NIDDM+EH, but not with M/L ratio of small resistance arteries.

Endothelial dysfunction in small resistance arteries of patients with non-insulin-dependent diabetes mellitus.

OBJECTIVE: Arterial hypertension is frequently associated with the presence of endothelial dysfunction in human subcutaneous small resistance arteries, as evaluated by responses to acetylcholine or bradykinin; however it is not known whether patients with diabetes mellitus show similar alterations. Therefore, we have investigated endothelial function in subcutaneous arteries of normotensive subjects (NT), of patients with essential hypertension (EH), of patients with non-insulin-dependent diabetes mellitus (NIDDM), as well as of patients with both essential hypertension and non-insulin-dependent diabetes mellitus (NIDDM+EH). PATIENTS AND METHODS: All subjects were submitted to a biopsy of the subcutaneous fat Small arteries were dissected and mounted on a micromyograph. The media to lumen ratio (M/L) was calculated. A concentration-response curve to acetylcholine, to bradykinin as well as to the endothelium-independent vasodilator sodium nitroprusside were performed. We also evaluated the contractile response to endothelin-1. Intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) plasma levels were also measured. RESULTS: The vasodilatation to acetylcholine and bradykinin (but not to sodium nitroprusside) was significantly and similarly reduced in EH, in NIDDM, and in NIDDM+EH compared with NT. The contractile response to endothelin-1 was similarly reduced in EH, in NIDDM and in NIDDM+EH. Plasma ICAM-1 and VCAM-1 concentrations were higher in EH, NIDDM and NIDDM+EH than in NT. CONCLUSIONS: An evident endothelial dysfunction was detected in patients with NIDDM, and the simultaneous presence of EH did not seem to exert an additive effect. The contractile responses to endothelin-1 were reduced possibly as a consequence of ET(A) receptor down-regulation.

The smoothness index, but not the trough-to-peak ratio predicts changes in carotid artery wall thickness during antihypertensive treatment.

BACKGROUND: It has recently been demonstrated that the smoothness index (SI) (the ratio between the average of the blood pressure changes computed for each hour of the recording and its standard deviation), a new and reproducible measure of the homogeneity of blood pressure reduction by antihypertensive treatment, has evident advantages over trough-to-peak ratio (T/P) in the prediction of the regression of left ventricular hypertrophy. Therefore we considered it to be worthwhile to compare the ability of SI and T/P to predict changes of the carotid artery intima-media thickness (IMT) during pharmacological treatment in patients with essential hypertension. METHODS: In 100 patients with essential hypertension, 24 h ambulatory blood pressure and carotid artery IMT were measured after 3 weeks of therapeutic wash-out and after 12 months of antihypertensive treatment (calcium antagonists, diuretics, angiotensin converting enzyme (ACE) inhibitors or beta-blockers). The homogeneity of the effect of treatment over blood pressure was evaluated by computing T/P and SI. RESULTS: Twenty-four hour blood pressure was significantly reduced by therapy, while, on average, a small but significant increase in indices of carotid artery wall thickness was observed. However, IMT was clearly reduced in patients with high SI. Statistically significant correlations were observed between changes in indices of carotid artery IMT during therapy and SI. No significant correlation was observed between indices of carotid artery morphology and T/P, basal 24 h blood pressure or changes in blood pressure during therapy. CONCLUSIONS: SI, but not T/P is the predictor of changes in carotid artery wall thickness. The information provided by SI is independent from basal blood pressure values. For carotid artery morphology, the smoothness of blood pressure reduction is even more important than its absolute change.

Flow-mediated dilatation of the brachial artery and left ventricular geometry in hypertensive patients.

OBJECTIVES: In arterial hypertension, the spectrum of geometric patterns in the left ventricle may parallel the structural alterations detected in the carotid arteries and in subcutaneous small arteries. It has been also reported that hypertensive left ventricular hypertrophy (LVH) may be associated with endothelial dysfunction, as evaluated by the response of coronary or forearm vasculature to acetylcholine infusion. The aim of this study was to evaluate the flow-mediated vasodilatation (FMD) of the brachial artery, non-invasive estimate of endothelium-dependent vasodilatation according to left ventricular geometric adaptations in hypertensive patients. METHODS AND RESULTS: In 16 normotensive (nine males, seven females, aged 40-68 years) and in 78 hypertensive subjects (50 males, 28 females, aged 42-67 years), we performed an echocardiographic study for the measurement of left ventricular mass index (LVMI) and relative wall thickness (RWT); we measured to a high resolution the brachial artery diameter at rest, during reactive hyperaemia (5 min of brachial artery occlusion) and after sublingual glyceril trinitrate (GTN); brachial artery flow velocity was measured by pulsed Doppler. Twenty-six hypertensive patients had a normal LVMI (LVMI < 51 g/ m2.7) and geometry (RWT < 0.44), five had concentric remodelling (RWT > or = 0.44), and concentric and eccentric LVH were observed in 19 and 28 patients, respectively. FMD was reduced in hypertensive patients as compared with normotensive subjects (P< 0.01). No correlation was found between FMD and LVMI (r= -0.078) or RWT (r = 0.049); in addition, no difference in FMD was found among the left ventricular geometric patterns in hypertensive patients. CONCLUSIONS: In hypertensives, the presence of endothelial dysfunction is not associated with the LVH or with different left ventricular geometric patterns, suggesting that different and independent mechanisms may be responsible for the presence of LVH and of endothelial dysfunction.

Cardiovascular status of carriers of the apolipoprotein A-I(Milano) mutant: the Limone sul Garda study.

BACKGROUND: Carriers of the apolipoprotein A-I(Milano) (apoA-I(M)) mutant present with very low plasma HDL cholesterol and moderate hypertriglyceridemia, apparently not leading to premature coronary heart disease. The objective of this study was to establish whether this high-risk lipid/lipoprotein profile is associated with structural changes in the carotid arteries and heart, indicative of preclinical atherosclerosis. METHODS AND RESULTS: Twenty-one A-I(M) carriers were compared with age- and sex-matched control subjects from the same kindred and with 2 series of matched subjects with primary hypoalphalipoproteinemia (HA). Structural changes in the carotid arteries were defined as the intima-media thickness (IMT) measured by B-mode ultrasound. HA subjects, both recruited among patients attending our Lipid Clinic and blood donors, showed significant thickening of the carotids (average IMT, 0.86+/-0.25 and 0.88+/-0.29 mm, respectively) compared with control subjects (average IMT, 0.64+/-0.12 mm); the apoA-I(M) carriers instead showed normal arterial thickness (average IMT, 0.63+/-0.10 mm). Moreover, a significantly higher prevalence of atherosclerotic plaques was found in patients and blood donors with HA (both 57%) compared with apoA-I(M) carriers (33%) and control subjects (21%). Echocardiographic findings and maximal treadmill ECG did not differ significantly between apoA-I(M) carriers and control subjects, apart from a slight increase in left ventricular end-diastolic dimension in the carriers. CONCLUSIONS: Despite severe HA, carriers of the apoA-I(M) mutant do not show structural changes in the arteries and heart, in contrast to HA subjects, who are characterized by a marked increase in carotid IMT and increased prevalence of atherosclerotic plaques.

Night time blood pressure and cardiovascular structure in a middle-aged general population in northern Italy: the Vobarno Study.

The aim was to determine, in a cross-sectional study, the relation between structural alterations in the heart and carotid arteries, and blood pressure (BP) changes from day to night time, measured by ambulatory BP (ABP). In 225 untreated subjects (107 F, 118 M, age range 48-64 years) and 59 treated subjects (24 M, 35 F, age range 50-64), living in a small town of northern Italy (Vobarno, Brescia) carotid intima media thickness as well as the occurrence of plaque, were evaluated by ultrasound. Echocardiographic left ventricular (LV) mass was measured according to the Penn Convention. BP was determined by clinic measurement and by 24-h non-invasive ABP monitoring. Subjects were divided in two groups, according to the decrease of night time systolic BP (SBP) "dippers" (SBP decreased by at least 10% during night time) and "non-dippers" (decrease of night time SBP <10%). The intima-media thickness in the common carotid, in the carotid bifurcation, in the internal carotid artery and average intima-media thickness were significantly greater in untreated non-dippers as compared with dipper subjects (ANOVA P < 0.05). A significantly higher prevalence of plaque was observed in untreated non-dippers as compared with dippers (P = 0.002). After adjusting for age, sex, 24-h SBP, and smoking, IMT in the carotid bifurcation and average intima-media thickness remained significantly greater in non-dipper subjects (P < 0.05 for all comparisons). No significant differences in LV mass were observed between dippers and non-dipper subjects. In conclusion, in a general population of unselected middle-aged subjects, night time BP values, among other risk factors, seem to represent an important determinant of carotid wall structure.

Changes in midwall systolic performance and cardiac hypertrophy reduction in hypertensive patients.

OBJECTIVE: To investigate changes in left ventricular (LV) performance, as evaluated by measurement of midwall LV fractional shortening (FS), after reduction of cardiac hypertrophy. DESIGN AND METHODS: Echocardiographic evaluation of LV anatomy and function was performed by M-mode echocardiography at baseline, after long-term antihypertensive therapy, and after treatment withdrawal in 68 asymptomatic hypertensive patients (50 males, 18 females, age range 22-62 years). Patients were divided according to the presence of LV hypertrophy (LVH) at baseline (LV mass index, LVMI, > or = 51 g/m(2.7)). RESULTS: At baseline patients with concentric (relative wall thickness > 0.44) LV hypertrophy (n = 38) or remodelling (n = 7) had reduced midwall shortening with respect to patients with normal LV geometry (n = 4) or eccentric LVH (n = 19); no differences were observed for endocardial FS. After long-term treatment (average 15 months), in 11 patients LV mass remained within normal limits, in 45 patients LVH reduction was obtained, while in 12 patients LV mass remained persistently elevated. Midwall FS was significantly increased in patients with reduction of LVH both during treatment and after withdrawal of treatment, while it remained significantly lower in patients with persistently elevated LV mass. Changes in midwall fractional shortening were independently associated with modifications in relative wall thickness (P < 0.00001), with changes in end-diastolic dimensions (P < 0.0001) and those of LVMI (P< 0.02) as shown by multivariate analysis. CONCLUSION: LV midwall systolic performance significantly improved after reduction of LVH, even in the presence of high blood pressure values. Modifications in relative wall thickness are more independently associated with changes, in LV diastolic dimensions and mass, to midwall improvement

Effect of treatment on flow-dependent vasodilation of the brachial artery in essential hypertension.

off aim of our study was to evaluate the effect of antihypertensive treatment on flow-mediated dilation (FMD)of a large artery, a noninvasive estimate of endothelial function, in hypertensive patients. In 78 consecutive hypertensive patients (40%men; age range, 42 to 67 years) we measured by a high-resolution ultrasound system the changes of brachial artery diameter during reactive hyperemia and after sublingual glyceryl trinitrate (400 microg); brachial artery flow velocity was measured by pulsed Doppler. The results of 2 studies are reported. In the first study, this procedure was repeated in 58 patients after 6 and 12 months of treatment with a combination of antihypertensive drugs; in a second study, the FMD was assessed in 20 patients after 2 months of monotherapy with either nifedipine or hydrochlorothiazide. In the first study, FMD was significantly increased after treatment compared with baseline (from 3.1+/-3% at baseline to 6.5+/-4.5% at 6 months and to 8.12+/-4. 6% at 12 months; P<0.001 by ANOVA), concomitant with blood pressure reduction (from 162+/-24/102+/-13 mm Hg to 141+/-12/89+/-6 mm Hg and to 141+/-9/89+/-6 mm Hg; P<0.001 by ANOVA); significant changes of endothelium-independent dilation were also observed, but only after 12 months of treatment (from 14.2+/-4.8 at baseline to 15.5+/-4.7 at 6 months and 16.8+/-5.9% at 12 months; P=0.03 by ANOVA). In the second study, FMD was significantly increased during nifedipine treatment as compared with baseline (from 5+/-6.18% at baseline to 9. 45+/-3.94%, P<0.001), while it did not change in patients receiving hydrochlorothiazide (from 5.15+/-5.28% at baseline to 4.69+/-4.34%, NS). No significant changes of endothelium-independent dilation were observed with both drugs (from 17.10+/-2.4% to 18.14+/-3.76% and from 18.73+/-4.07% to 17.46+/-4.27% during nifedipine and hydrochlorothiazide, respectively, NS). Thus, in essential hypertensive patients an improvement of the impaired FMD of the brachial artery, evaluated by noninvasive ultrasound, may be observed after long-term, effective blood pressure reduction, suggesting a beneficial effect of antihypertensive treatment on endothelial function. It seems that beyond blood pressure control, a calcium antagonist may be more effective than a diuretic in this respect.

Structural association between the carotid artery and the left ventricle in a general population in Northern Italy: the Vobarno study.

OBJECTIVE: To evaluate the interrelationships between structural alterations in the carotid arteries and left ventricular geometric patterns in a middle-aged general population. DESIGN AND METHODS: We studied 223 untreated subjects (106 females, 117 males; aged 48-64 years) living in a small town in Northern Italy (Vobarno, Brescia), using a cross-sectional design. The left ventricular mass index was assessed by echocardiography, while the intima-media thickness and the occurrence of plaque were evaluated by ultrasound. Blood pressure was determined by clinic measurements and by 24 h noninvasive ambulatory blood pressure monitoring. RESULTS: According to the presence of left ventricular hypertrophy (left ventricular mass index > 50 g/m2.7 in males and > 47 g/m2.7 in females) or concentric left ventricular remodelling (normal left ventricular mass index and relative wall thickness > 0.42), the subjects were divided into four groups: 124 subjects with normal left ventricular geometry, 73 subjects with left ventricular hypertrophy (55 with eccentric and 18 with concentric hypertrophy) and 26 subjects with concentric remodelling. The common carotid intima-media thickness and cross-sectional area were significantly greater in the subjects with concentric left ventricular hypertrophy (analysis of variance, P< 0.05) than in those with normal left ventricular geometry. A significantly higher number of plaques was observed in subjects with concentric and eccentric left ventricular hypertrophy. CONCLUSIONS: In a general population of unselected middle-aged subjects, the presence of concentric left ventricular hypertrophy was associated with an increase in intima-media thickness and with the presence of plaque in the carotid arteries, possibly contributing to the worse prognosis observed in this group of patients.

Various ways of calculating echocardiographic left ventricular mass and their relative prognostic values.

OBJECTIVE: To compare the calculations of left ventricular mass according to thick-wall [American Society of Echocardiography (ASE) and Penn convention] and thin-wall (Wikstrand formula) models. METHODS: We have reexamined the data from the cross-sectional study on the general population sample of Vobarno and from a prospective longitudinal study of hypertensive patients assessing the prognostic significance of changes in left ventricular mass during a follow-up period of 10 years on average (Brescia population). RESULTS: For the Vobarno and Brescia populations, we found a close relationship between values of left ventricular mass calculated by using a thin-wall ellipsoidal model (Wikstrand formula) and those calculated using a thick-wall model with Penn convention or ASE left ventricle measurements (r = 0.99, for both the Vobarno and Brescia populations). Highest values of Penn left ventricle mass were slightly underestimated by use of the thin-wall formula. The numbers of nonfatal cardiovascular events and the relative risks, evaluated by Cox proportional hazard models for 151 patients seen at follow-up did not differ for patients with persistence of left ventricular hypertrophy (LVH), those with regression of LVH, and those with normal left ventricle mass, both at baseline and at follow-up, when these different ways of measuring left ventricle mass and partition values for LVH were used. CONCLUSIONS: The calculation of left ventricle mass according to the ASE recommendations or to the Penn convention, both of which are based on the assumption that the left ventricle can be represented by a prolate ellipsoid with both the internal and external long axes twice the short axis, produces results similar to those obtained using an alternate formula for the calculation of left ventricle mass, considering wall thickness constant around the ellipsoidal cavity. The cardiovascular risk stratification, in relation both to baseline left ventricular mass and to its change during long-term antihypertensive treatment, does not differ significantly among the results of these three different calculations.

Reduction of cardiovascular structural changes by nifedipine GITS in essential hypertensive patients.

The aim of this study was to evaluate the effect of the calcium antagonist Nifedipine GITS in a double-blind, randomized comparison with the diuretic hydrochlorothiazide (HCTZ) on reduction of left ventricular (LV) mass and minimal vascular resistance in a group of essential hypertensives with left ventricular hypertrophy (LVH). The effects on blood pressure and on echocardiographic LV functional parameters were also analysed. After two months of randomized treatment with Nifedipine GITS or HCTZ, if diastolic blood pressure was > 90 mmHg, a combination of the two drugs was given and was continued for 24 weeks. M-mode, 2D-guided echocardiography was used to measure LV mass index (LVMI) according to the "Penn convention". Minimal vascular resistance was measured in the forearm, from arterial pressure and maximal blood flow, using a strain gauge plethysmography. All examinations were performed before and after 8 and 24 weeks of treatment. Changes in LVMI were analysed at 8 weeks and at 24 weeks in patients receiving monotherapy ("according to protocol" analysis), and also at the end of treatment in patients taking Nifedipine or HCTZ monotherapy or the combination of the two drugs ("intention to treat" analysis). Both Nifedipine and HCTZ significantly reduced systolic and diastolic blood pressure (p < 0.001), without any significant difference between the two drug treatments. Heart rate was not significantly modified by either treatment. A progressive decrease in LVMI was observed after 8 and 24 weeks of treatment with Nifedipine monotherapy (ANOVA, p = 0.03), while the decrease in LVMI during HCTZ treatment did not progress further at 24 weeks (ANOVA, p = 0.49). A significant reduction of minimal vascular resistance was observed in patients treated with Nifedipine GITS monotherapy (ANOVA, p = 0.001), but not in the HCTZ group (ANOVA, p = 0.06). Comparison of changes of forearm minimal vascular resistance, considering baseline values, could demonstrate a greater effect during Nifedipine monotherapy as compared to HCTZ monotherapy. In conclusion, in a group of hypertensive patients with LVH, treatment for 24 weeks with Nifedipine GITS alone or in combination with HCTZ induced a significant reduction in LVMI and of forearm vascular structural changes, as evaluated by minimal vascular resistance. The decrease of minimal vascular resistance was significantly greater in patients treated with Nifedipine monotherapy, as compared to those given HCTZ.

Relations between cardiac and vascular structure in patients with primary and secondary hypertension.

BACKGROUND: Data on cardiac and vascular structure in secondary hypertension are generally scarce, and no data on the interrelations between cardiac mass and structural characteristics of the vessel wall, both in large and in small resistance arteries, are presently available. OBJECTIVES: The aim of this study was to investigate the relation between structural changes in subcutaneous small arteries, left ventricular mass and wall thickness of the common carotid artery in patients with primary and secondary hypertension. METHODS: Seventy-four subjects were included in the study: 11 patients with pheochromocytoma, 14 with primary aldosteronism (PA), 19 with renovascular hypertension (RVH), 18 with essential hypertension (EH) and 12 normotensive (NT) control subjects. All subjects were submitted to a biopsy of subcutaneous fat. Morphologic characteristics of subcutaneous small resistance arteries (relaxed diameter <300 microm) were directly evaluated using a micromyographic technique. All subjects were submitted to calculation of left ventricular mass index (LVMI) and common carotid artery intima-media thickness (CCIMT), using ultrasound technique. RESULTS: The correlation coefficients between the media to lumen ratio in subcutaneous small arteries (M/L) and LVMI or between M/L and CCIMT were closer in RVH than in pheochromocytoma, EH or NT; in PA the correlation coefficients were slightly less close than those in RVH. An excess prevalence of carotid plaques in RVH was observed. CONCLUSIONS: A close relation between small resistance artery morphology and cardiac or carotid artery structure may be observed in those hypertensive patients in whom the renin-angiotensin-aldosterone system is activated. In constrast, in NT, EH and pheochromocytoma no significant correlation between M/L and LVMI or CCIMT was observed.

Effects of long-term antihypertensive treatment with lisinopril on resistance arteries in hypertensive patients with left ventricular hypertrophy.

OBJECTIVE: To evaluate the effects of long-term antihypertensive therapy with the angiotensin converting enzyme inhibitor lisinopril on structural alterations and the endothelial function of small resistance arteries in hypertensive patients with left ventricular hypertrophy. METHODS: Fourteen patients with left ventricular hypertrophy were treated for 3 years with a lisinopril-based regimen. Patients underwent an echocardiographic evaluation of left ventricular mass index at baseline, during the first and third years of treatment. At the end of the treatment period, subcutaneous small resistance arteries (obtained by biopsy of the subcutaneous fat from the gluteal region) were dissected and mounted on a micromyograph (Mulvany's technique); the media : lumen ratio was then calculated. Data obtained were compared with those observed for 14 untreated essential hypertensive patients and 14 normotensive subjects, age- and sex-matched. RESULTS: In the present study, a significantly lower media : lumen ratio was observed in treated compared with untreated hypertensive patients, although it remained significantly higher than that in normotensive subjects. In treated hypertensive patients a significant reduction in clinic blood pressure was observed. However, their blood pressure remained significantly higher than that in normotensive subjects. Significant correlations between the media : lumen ratio and blood pressure, left ventricular mass index or changes in left ventricular mass index during treatment were observed. The response to acetylcholine administration was reduced in untreated hypertensives compared with that in normotensives. In patients treated with lisinopril, the vasodilatation obtained with the two higher doses of acetylcholine was greater than that in untreated hypertensives, thus suggesting an improvement of endothelial function. CONCLUSIONS: Long-term therapy based on lisinopril was associated with a smaller media : lumen ratio in the subcutaneous small resistance arteries of hypertensive patients with left ventricular hypertrophy. Our retrospective study confirms previous findings obtained in prospective studies with other angiotensin converting enzyme inhibitors. Endothelial function was probably improved by lisinopril therapy.

Persistence of left ventricular hypertrophy is a stronger indicator of cardiovascular events than baseline left ventricular mass or systolic performance: 10 years of follow-up.

OBJECTIVE: Left ventricular hypertrophy (LVH) and depressed left ventricular performance have been shown to be associated to an adverse prognosis in hypertensive patients. It has not been established, however, whether, during chronic antihypertensive treatment, the increased cardiovascular risk is more strictly related to the presence of LVH or of a low left ventricular performance. DESIGN AND METHODS: A total of 215 patients with uncomplicated hypertension (129 males, 86 females; age range 18-70 years, mean +/- SD 45 +/- 11) underwent an echocardiographic evaluation of left ventricular anatomy and function. In 151 patients (87 males, 64 females; age range 18-70 years, mean 45 +/- 10.4) the echocardiogram was repeated on average 10 +/- 1 years after the initial study. The presence of LVH (left ventricular mass index > 134 g/m2 in males and 110 g/m2 in females) and the midwall left ventricular shortening/end-systolic stress relationship were prospectively analysed as predictors of cardiovascular non-fatal events (n = 23) in patients who were seen at follow-up. RESULTS: The incidence of non-fatal cardiovascular events was greater in patients with LVH (n = 17, P < 0.0001) and in those with a lower midwall performance (n = 14, P < 0.01) at baseline. At follow-up, the incidence of non-fatal cardiovascular events was significantly greater in patients without a reduction in the left ventricular mass index, after adjusting for traditional risk factors (relative risk 3.52 versus 1.38 in patients with persistence and regression of LVH, respectively; P < 0.01). The baseline midwall fractional shortening was lower in patients with both persistence or regression of LVH (analysis of variance, P < 0.0001) than in patients with a normal left ventricular mass index. In logistic analysis, the left ventricular mass index at follow-up and age were independent determinants of non-fatal cardiovascular events (P < 0.001); without the left ventricular mass index at follow-up, this analysis showed that age, systolic blood pressure at follow-up and baseline midwall fractional shortening were independent determinants of non-fatal cardiovascular events. CONCLUSIONS: Our results suggest that lack of regression of LVH is a stronger indicator of cardiovascular risk than a depressed baseline midwall left ventricular performance.