Author + information
- Received December 30, 2000
- Revision received August 14, 2001
- Accepted August 29, 2001
- Published online December 1, 2001.
- Paolo Verdecchia, MD, FACC*,a (, )
- Giancarlo Carini, MDa,
- Antonio Circo, MDa,
- Emilio Dovellini, MDa,
- Ezio Giovannini, MDa,
- Michele Lombardo, MDa,
- Pasquale Solinas, MDa,
- Marco Gorini, MDa,
- Aldo Pietro Maggioni, MDa,
- the MAVI Study Group
- ↵*Reprint requests and correspondence: Dr. Paolo Verdecchia, ANMCO Research Center, Via La Marmora, 39, 50121-Firenze, Italy
This study investigated the prognostic value of left ventricular (LV) mass at echocardiography in uncomplicated subjects with essential hypertension.
Only a few single-center studies support the prognostic value of LV mass in uncomplicated hypertension.
The MAssa Ventricolare sinistra nell’Ipertensione study was a multicenter (45 centers) prospective study. The prespecified aim was to explore the prognostic value of LV mass in hypertension. Admission criteria included essential hypertension, no previous cardiovascular events, and age ≥50. There was central reading of echocardiographic tracings. Treatment was tailored to the single subject.
Overall, 1,033 subjects (396 men) were followed for 0 to 4 years (median, 3 years). Mean age at entry was 60 years, and systolic/diastolic blood pressure was 154/92 mm Hg. The rate of cardiovascular events (×100 patient-years) was 1.3 in the group with normal LV mass and 3.2 in the group (28.5% of total sample) with LV mass ≥125 g/body surface area (p = 0.005). After adjustment for age (p < 0.01), diabetes (p < 0.01), cigarette smoking (p < 0.01) and serum creatinine (p = 0.03), LV hypertrophy was associated with an increased risk of events (RR [relative risk] 2.08; 95% CI [confidence interval]: 1.22 to 3.57). For each 39 g/m2(1 SD) increase in LV mass there was an independent 40% rise in the risk of major cardiovascular events (95% CI: 14 to 72; p = 0.0013).
Our findings show a strong, continuous and independent relationship of LV mass to subsequent cardiovascular morbidity. This is the first study to extend such demonstration to a large nationwide multicenter sample of uncomplicated subjects with essential hypertension.
It is generally established that left ventricular (LV) mass determined at echocardiography is a powerful predictor of cardiovascular disease in apparently uncomplicated subjects with essential hypertension (1,2). However, only two single-center studies explored the prognostic value of LV mass (3,4)in the specific setting of initially uncomplicated subjects with essential hypertension. Evidence that serial changes in LV mass over time predict prognosis is also limited (5–7). The other outcome-based studies on the prognostic value of LV mass have been conducted in subjects with suspected coronary artery disease (8), previous myocardial infarction (MI) (9), renal failure (10,11), or in the general population (12,13). Hence, although a large body of evidence indicates the prognostic value of LV mass in different clinical settings, the specific applicability of these results to the clinical management of the vast majority of uncomplicated subjects with essential hypertension remains poorly supported.
To explore the prognostic value of LV mass at echocardiography in the specific setting of uncomplicated subjects with essential hypertension, we planned the MAssa Ventricolare sinistra nell’Ipertensione arteriosa (MAVI) study, a prospective, observational, multicenter investigation.
The MAVI study was a multicenter, prospective, observational study carried out in 58 hospital centers in Italy and endorsed by the Italian Association of Hospital Cardiologists (ANMCO, Associazione Nazionale Medici Cardiologi Ospedalieri). The protocol of MAVI had been published before the start of the study (14). The prespecified aim (14)was the assessment of the independent prognostic value of LV mass at echocardiography in subjects with essential hypertension. In brief, admission criteria to the MAVI study included clinic blood pressure (BP) in sitting position ≥140 mm Hg systolic or 90 mm Hg diastolic or current treatment for hypertension, no previous cardiovascular morbid events, age ≥50, and absence of valvular heart disease. Both genders have been included. Patients with cancer, other important diseases or serum creatinine ≥2.0 mg/dl were excluded. Patients under antihypertensive treatment were not required to withdraw their medications when they entered the study.
Clinical data were stored on a computer using an ad hoc software for hypertension laboratories provided by ANMCO and sent to the study headquarters in Florence. Prespecified primary outcome events included fatal and nonfatal MI, sudden cardiac death, fatal and nonfatal stroke, other cardiovascular deaths, all-cause death, severe heart failure requiring hospitalization and severe renal failure requiring dialysis. Secondary events included new-onset angina (typical chest pain symptoms plus objective evidence of ischemic changes at electrocardiography [ECG]), transient ischemic attack, and peripheral arteries occlusive disease verified at angiography.
Original source documents of patients who suffered an end-point event were collected by centers and sent to the study headquarters. Standard international criteria were used for definition of outcome events. Myocardial infarction was diagnosed on the basis of at least two of three standard criteria (typical chest pain, QRS changes at ECG, transient elevation of myocardial enzymes by more than twofold the upper normal laboratory limits). New-onset angina was defined by chest pain accompanied by typical ischemic changes on the ECG. Sudden death was defined as a witnessed death that occurred within an hour after the onset of acute symptoms, with no history that violence or accident played any role in the fatal outcome. Stroke was diagnosed on the basis of rapid onset of localizing and persistent neurological deficit in the absence of any other disease process explaining the symptoms. Transient ischemic attack was defined by the diagnosis, made by a physician, of any sudden focal neurological deficit that cleared completely in less than 24 h.
The original ECG tracings were sent to a central laboratory for reading. The ECG reading was manual and was performed by one expert reader. Readers for echocardiographic and ECG tracings were different, did not work together and were not aware of the clinical characteristics of patients. The ECGs were recorded at 25 mm/s and 1 mV/cm calibration. Subjects with complete bundle branch block, previous MI, Wolff-Parkinson-White syndrome and atrial fibrillation were excluded from analysis. The LV hypertrophy at ECG was diagnosed using the Perugia score (15,16), which requires positivity of at least one of the following three criteria: SV3+ RaVL>2.4 mV in men or >2.0 mV in women, a typical strain pattern, or a Romhilt-Estes point score ≥5.
A two-dimensional-targeted M-mode echocardiographic study was carried out at the beginning of the study in the context of a complete diagnostic workup including laboratory examinations, ECG and clinical examination. Working meetings among involved investigators were held periodically. Echocardiograms were recorded on tape and sent to a central laboratory for reading. Three expert readers examined the tracings, and the allocation of tapes to readers was randomized. All measurements were made on the screen using calipers. A long-axis parasternal approach was first examined to check perpendicularity of the ultrasonic beam with respect to the septum. Then, the short-axis approach was used to take LV diastolic and systolic measurements (the average of three consecutive cycles on the best single reading set was considered). The M-mode study was performed under 2D control using commercially available instruments. End-diastolic and end-systolic measurements were taken with the patient in partial left lateral decubitus according to the American Society of Echocardiography recommendations (17,18). Frames with optimal visualization of interfaces and showing simultaneous visualization of septum, LV internal diameter and posterior wall were used for reading. The LV mass was calculated using the following formula introduced by Devereux et al. (19)on the basis of necropsy validation studies: For definition of LV hypertrophy we considered an LV mass ≥125 g/body surface area (BSA) [m2], a partition point supported by ample prognostic evidence (3–7,17), and an LV mass >51.0 g/height [m2.7], in order to provide a more stringent allowance for obesity (20).
On the basis of previous data (3,4), we predicted a prevalence of echocardiographic LV hypertrophy of 30% and a rate of events of 2 per 100 person-years in the absence versus 4 per 100 person-years in the presence of LV hypertrophy (14). On this basis, a sample size of 1,811 subjects with an average follow-up time of two years per patient was estimated to detect a significant difference between the groups (two-tailed test) with a type I error of 5% and a type II error of 10%. Survival curves were determined using the Kaplan-Meier product-limit method (21)and compared by the Mantel (log-rank) test (22). The effect of prognostic factors on survival was evaluated by the stepwise Cox model (23). The tested variables were age (years), gender (men, women), systolic and diastolic BP at entry and after a median of three years of follow-up, evidence of coronary heart disease in the father at age <55 or in the mother at age <65 (yes, no), body mass index (weight[kg]/height[m2]), current cigarette smoking (yes, no), diabetes (yes, no), serum cholesterol (mg/dl), serum creatinine (mg/dl), LV mass (g/BSA[m2]) and antihypertensive treatment (no drugs; diuretics and beta-blockers alone or combined; different antihypertensive regimens). For the subjects who experienced multiple events, survival analysis was restricted to the first event. In two-sided tests, p values <0.05 were considered significant.
The MAVI study began in October 1995 in 58 hospital centers within Italy. Overall, 1,857 subjects were enrolled up to December 1998 and follow-up was concluded on December 31, 1999. Incomplete follow-up, predefined by follow-up information available in <50% of subjects, was reason for exclusion from the study for 13 centers and 197 subjects. The remaining 45 centers enrolled 1,660 subjects. Of these, 627 (38%) were rejected because of echocardiographic tracings of suboptimal technical quality on judgment of the reading center, thus leaving 1,033 subjects for the final analysis.
Table 1shows some baseline characteristics of these subjects. Distribution of antihypertensive regimens at entry and on follow-up is shown in Figure 1. Distribution of the six stages of the Joint National Committee VI classification (optimal BP, normal BP, high-normal BP, stage I, stage II, stage III) (1)was 0.3%, 2.1%, 8.9%, 45.2%, 33.0% and 10.4%, respectively. Thirty-eight percent of the subjects did not show any concomitant traditional risk factors (1,2)in addition to hypertension, and the remaining 62% showed at least one risk factor. Diabetes was present in 7.3% of subjects, family history of premature coronary artery disease in 8.3%, cigarette smoking in 13.5%, body mass index >30 kg/height [m2] in 21.1%, a serum cholesterol >250 mg/dl or a total cholesterol/high density lipoprotein cholesterol ratio ≥6.0 in 28.1% of subjects.
Prevalence of LV hypertrophy at echocardiography was 28.5% using the cutoff of 125 g/BSA [m2], and 47.5% using 51.0 g/height [m2.7]. Prevalence of LV hypertrophy at ECG was 19.7%.
After a median of three years of follow-up, average BP was 146/86 mm Hg (SD 15/8). Only 32.5% of subjects achieved adequate BP control (BP <140 mm Hg systolic and <90 mm Hg diastolic). The distribution of antihypertensive regimens at entry and on follow-up is reported in Figure 1.
The subjects who developed a first cardiovascular event during follow-up numbered 83. Fifty-five of these were in the group (n = 1,033) who had good-quality echocardiographic tracings (Table 2). Thirty of the 55 cardiovascular events were primary events. Total event rate (per 100 person-years) did not differ between the subjects with suboptimal tracings and the other group (1.6 vs 1.8, respectively; p = 0.54 [log-rank test]). In the subset with good-quality echocardiographic tracings, event rate was 1.8 in the total population, 1.3 in the subset with normal LV mass (<125 g/BSA [m2]) and 3.2 in that with increased LV mass (log-rank test: p = 0.0005). Incidence of total cardiovascular events in the two groups is shown in Figure 2.
Overall, there were 4 subjects with sudden cardiac death, 9 with MI, 13 with stroke, 9 with transient ischemic attack, 3 with heart failure, 11 with new-onset unstable angina, 5 with arterial occlusive disease and 1 with severe progressive renal failure requiring dialysis.
As shown in Table 3, after adjustment for the significant influence of age, cigarette smoking and diabetes, there was an independent 37% increase in the risk of primary events for any 39 g (1 SD) increase in LV mass (95% CI: 5 to 80; p = 0.020). Furthermore, after adjustment for age, cigarette smoking, serum creatinine concentration and diabetes, for any 39 g (1 SD) increase in LV mass there was a 40% rise in the risk of total cardiovascular events (95% CI: 14 to 73; p = 0.0013). None of the other variables yielded statistical significance, including LV hypertrophy at ECG. When LV mass entered the Cox analysis as a binary variable (below vs. above 125 g/m2), the risk associated with LV hypertrophy remained significant for total cardiovascular events (RR = 2.08; 95% CI: 1.22 to 3.57; p = 0.007) and primary events (RR = 1.51; 95% CI: 1.19 to 1.92; p = 0.008). Systolic and diastolic BP at entry and on follow-up did not achieve significance.
There were only 13 fatal events and their rate (×100 person-years) was 0.23 in the subset with normal LV mass (<125 g/BSA [m2]) and 0.99 in that with increased LV mass (log-rank test: p = 0.007). The low number of fatal events precluded their multivariate assessment.
Left ventricular mass was also corrected by height2.7to improve allowance for obesity. Event rate (×100 person-years) was 1.36 in the subset with LV mass ≤51.0 g/height2.7(20), and 2.41 in that with greater LV mass (p = 0.034; log-rank test). When gender-adjusted cutoff values were used for definition of LV hypertrophy (104 g/m2in women and 116 g/m2in men) (24,25), prevalence of LV hypertrophy was 47.1% and event rate (×100 person-years) was 1.46 in the subset with normal LV mass and 2.45 in that with LV hypertrophy (log-rank test: p = 0.025). The list of independent predictors of outcome in the Cox analysis remained unchanged.
The present study is the first to extend demonstration of the prognostic value of LV mass to a nationwide multicenter sample, the largest available so far, of uncomplicated subjects with essential hypertension. Qualifying features of the present study were its prospective design, with prespecified definition of primary and total outcome events, and the central blind reading of ECG and echocardiographic tracings. These characteristics may have reduced the possibility of technical biases potentially inherent to single-center studies.
In several observational studies, conducted in a variety of clinical conditions (3–13), the predictive power of LV mass remained significant after adjustment for the confounding effect of age, BP and other risk markers including diabetes, smoking, serum lipids and coronary lesions, whose number and prognostic impact differed across these studies. However, one could question to what extent the conclusions of these studies can be applied to the majority of subjects with uncomplicated essential hypertension. In fact, only a few of these studies have been specifically conducted in uncomplicated hypertensive subjects (3,4). Most studies have been done in quite different settings including subjects undergoing cardiac catheterization for presumed coronary artery disease (8), survivors of MI (9), subjects with renal failure (10,11)or in the general population (12,13).
It is out of the question that this large body of evidence strongly supports, in general, the prognostic value of LV mass. This, however, might lead to a partially unsupported feeling of scientific and clinical confidence in the use of LV mass for cardiovascular risk stratification in the specific setting of uncomplicated subjects with essential hypertension, which could become a target for excessive extrapolation of conclusions not specifically obtained in these subjects.
The LV mass and prognosis
In the present study, LV hypertrophy at echocardiography doubled the risk of both primary and total events, and for any increase in LV mass by 1 ± SD (39 g/m2) there was an independent 40% rise in the risk of events. The relation between LV mass and subsequent cardiovascular risk persisted after adjustment for the influence of several traditional risk factors, while other factors including baseline BP and antihypertensive treatment did not achieve significance. The overall event rate (1.8 × 100 patient-years) was somewhat lower than expected in the year 1995 when the study was designed (2.5 × 100 patient-years), but the absolute excess risk associated with LV hypertrophy remained virtually as expected (1.9 events vs. 2.0 events predicted). Also, the observed prevalence of LV hypertrophy (28.5%) was comparable to that expected (30.0%).
The apparently high prevalence of LV hypertrophy at ECG (19.7%) resulted from the use of the Perugia score (15,16), a diagnostic score that is positive in the presence of typical LV strain, a Romhilt-Estes score of 5 points or more, or Cornell voltage exceeding 2.0 mV in women or 2.4 mV in men (15,16). In a previous study (15), the combination of three specific criteria (Romhilt-Estes, LV strain and Cornell) in this score produced a rise in sensitivity to about 34% without excessive deterioration of specificity (93%), with a prevalence of LV hypertrophy at ECG of about 18%. A recent analysis of the Losartan Intervention For Endpoint (LIFE) study showed that accuracy of the Perugia score for diagnosis of LV hypertrophy is independent of overweight or obesity (26). In the present study, LV hypertrophy at ECG did not achieve significance as a predictor of cardiovascular risk after controlling for LV mass.
In addition to the expected concomitant effect of age, diabetes and cigarette smoking, our study showed an independent association between serum creatinine and subsequent risk of total cardiovascular events. Previous studies in hypertensive subjects showed an association between elevated creatinine levels and subsequent cardiovascular events (27–29). Also, high-normal levels of serum creatinine appear to be predictive of future cardiovascular events in subjects with essential hypertension (30). Because renal function was normal in all subjects at entry into the study, our data suggest that serum creatinine levels within the normal range may work as a marker of long-term exposure to high BP values, and therefore predict outcome also in nonrenal target organs.
Our epidemiological findings are unable to define the basic mechanisms underlying the prognostic value of LV mass in subjects with essential hypertension. The LV mass is generally considered a biological assay, which reflects and integrates the long-term cumulative level of activity of several risk factors for cardiovascular disease (31). Increased LV mass is a powerful predictor not only of ischemic cardiac events, an association partially expected by the unfavorable effects of LV hypertrophy on the balance between myocardial oxygen requirement and supply (32), but also of stroke (33).
Because our study was conducted among white subjects, results may not be extended to different ethnic groups. Another limitation was the lack of assessment of the prognostic value of serial changes in BP and LV mass over time (5–7). The large percentage of subjects with inadequate echocardiographic tracings reflects the use of stringent criteria in the central reading process. Central reading of echocardiographic tracings in clinical trials frequently results in a large percentage of rejections due to suboptimal quality (34,35). In the Framingham Heart Study, 20.7% of echocardiographic tracings from men and 21.1% from women were rejected because of inadequate technical quality (13).
Our study is the first to demonstrate a powerful association between LV mass determined at echocardiography and subsequent occurrence of cardiovascular disease in a nationwide representative multicenter sample of uncomplicated subjects with essential hypertension. These findings support the view that both prevention of LV hypertrophy development and regression of hypertrophy once established are key targets in the management of asymptomatic patients with essential hypertension.
Steering Committee: E. Giovannini (chairman), G.C. Carini, A. Circo, E.V. Dovellini, M. Lombardo, P. Solinas, P. Verdecchia.
Scientific and Organizing Secretariat: S. Ghione, M. Gorini, D. Lucci, A.P. Maggioni.
Echocardiogram and ECG Central Reading: C. Borgioni, A. Ciucci, G. Gozzellino, A. Milletich.
Participating clinical centers:Avola (E. Mossuti, G. Canonico, G. Romano), Bari Policlinico (I. De Luca, N. Ciriello), Bazzano (A. Baldini, G. Castelli), Belluno (G. Catania, L. Tarantini), Bologna Osp. S. Orsola-M. Malpighi (E. Ambrosioni, F. Marchetta), Brindisi Osp. Di Summa (G. Ignone, D. Zuffianò, A. Storelli), Cagliari Osp. Binaghi (G. Isaia, E. Muscas), Camerino (R. Amici, B. Coderoni), Capua (E. Mercaldo, R. Esposito), Casale Monferrato (M. Ivaldi, G. Gozzelino, A. Capilli), Casarano (G. Pettinati, M. Ieva, A. Marzo), Caserta (G. Corsini, S. Romano, A. Martone), Catanzaro Policlinico (F. Perticone, C. Cosco), Chieti Villa Pini D’Abruzzo (C. Ciglia, P. Di Giovanni), Citta’ Della Pieve (G. Benemio, G. Schillaci, N. Sacchi), Cosenza INRCA (E. Feraco, A. M. Nicoletti), Desio (V. Baldini, M. Cristofari), Firenze Osp. Careggi (P. F. Fazzini, D. Antoniucci, E. V. Dovellini, E. Taddeucci), Firenze Osp. Camerata (F. Marchi, A. Buzzigoli), Foligno (M. Massi Benedetti, C. Pagnotta, R. Liberati), Gallarate (R. Canziani, G. Cozzi, M. Alberio), Gioia Del Colle (F. Barba), Gorizia (A. Fontanelli, R. Marini, L. Scarpino), Grottaglie (V. Portulano, G. Sportelli), Matera (L. Veglia, T. Scandiffio), Mesoraca (F. Schipani, C. Tangari), Messina Osp. Papardo (L. Cavallaro, G. Sergi, S. Mangano), Messina Policlinico Universitario (S. Coglitore, C. De Gregorio, D. Cento), Milano Osp. Niguarda (A. Pezzano, M. Lombardo), Monfalcone (T. Morgera, G. Zilio, D. Chersevani), Napoli Osp. Cardarelli (L. D’Aniello, F. D’Isanto), Oliveto Citra (G. D’angelo, V. Iuliano, P. Bottiglieri), Palermo Osp. Ingrassia (P. Di Pasquale, F. Clemenza, S. Cannizzaro, G. Caramanno), Perugia (C. Porcellati, P. Verdecchia, A. Ciucci, C. Borgioni), Pieve Di Cadore (J. Dalle Mule, M. Mazzella), Polistena (R. M. Polimeni, F. Catananti, F. Terranova), Pontedera (G. Squarcini, S. Giaconi), Pordenone (G. B. Cignacco, G. Zanata, D. Pavan), Potenza (A. Lopizzo, M. Chiaffitelli, M. Faruolo), Prato (A. Petrella, M. Paoletti), Roma CTO (M. Uguccioni, D. Mocini), Roma Osp. Forlanini (A. Majid Tamiz, A. Avallone), Roma Osp. San Camillo II Divisione (E. Giovannini, A. Chiantera, F. De Santis), Roma Osp. San Camillo Servizio (P. Tanzi, L. Boccardi, D. Colecchia), Rovigo (P. Zonzin, A. Bortolazzi, S. Aggio), San Giovanni Rotondo (R. Fanelli, A. Russo, A. De Vita), San Pietro Vernotico (L. Vergallo, S. Pede, A. Renna), Sarzana (G. Filorizzo, D. Bertoli, M. Corradeghini), Scorrano (E. De Lorenzi, A. Bergamo, A. Colizzi), Sesto San Giovanni (R. Melloni, E. Choicca), Sondalo (G. Occhi, G. Frisullo), Sondrio (S. Giustiniani, M. Marieni, M. L. Ghirimoldi), Soveria Mannelli (G. Bellieni, A. Marotta, A. Andricciola), Termoli (D. Staniscia, T. Alfieri), Udine (P. Fioretti, L. Pilotto), Vasto (G. Di Marco, G. Levantesi, L. Cavasinni), Viareggio (A. Pesola, G. Marracci, M. Pardini), Viterbo (R. Guerra, A. Achilli, S. De Spirito, R. Castellani).
☆ This study was supported in part by Pfizer Italia, S.p.A.
- Associazione Nazionale Medici Cardiologi Ospedalieri
- blood pressure
- body surface area
- left ventricular
- Massa Ventricolare Sinistra nell’Ipertensione
- myocardial infarction
- Received December 30, 2000.
- Revision received August 14, 2001.
- Accepted August 29, 2001.
- American College of Cardiology
- Yurenev A.P.,
- Dyakonova H.G.,
- Novikov I.D.,
- Vitols A.,
- Pahl L.,
- Haynemann G.
- Verdecchia P.,
- Schillaci G.,
- Borgioni C.,
- et al.
- Verdecchia P.,
- Schillaci G.,
- Borgioni C.,
- et al.
- Committee on M-mode Standardization of the American Society of Echocardiography,
- Sahn D.J.,
- DeMaria A.,
- Kisslo J.,
- Weyman A.
- de Simone G.,
- Daniels S.R.,
- Devereux R.B.,
- et al.
- Cox D.R.
- Hypertension Detection and Follow-up Program Cooperative Group,
- Shulman N.B.,
- Ford C.E.,
- Hall D.,
- et al.
- Gueyffier F.,
- Boissel J.-P.,
- Pocock S.,
- et al.
- Devereux R.B.,
- Roman M.J.
- Houghton J.L.,
- Frank M.J.,
- Carr A.A.,
- von Dohlen T.W.,
- Prisant L.M.
- Gottdiener J.S.,
- Reda D.J.,
- Massie B.M.,
- Materson B.J.,
- Williams D.W.,
- Anderson R.J.