Thirty years with LIFE—a randomized clinical trial with more than 200 published articles on clinical aspects of left ventricular hypertrophy

Losartan Intervention for Endpoint Reduction in hypertension (LIFE) is the only large clinical trial of antihypertensive treatment ever performed in hypertensive patients with left ventricular hypertrophy (LVH). Patients 55 years of age and older (average 67 years) with hypertension and electrocardiographic (ECG) LVH were randomized to the angiotensin receptor blocker losartan or the beta-1 selective adrenergic receptor blocker atenolol in this prospective, double blinded clinical outcome trial. The protocol was written in 1993, and the main outcome results published in 2002 [1]. At the same level of achieved systolic blood pressure (SBP), 144 mmHg, losartan lowered the primary composite endpoint of cardiovascular (CV) death, myocardial infarction and stroke by 13% more than atenolol. Losartan also had greater effects than atenolol on regression of ECG- and echocardiographic measured LVH, and on prevention of incident atrial fibrillation (AF) and type 2 diabetes mellitus.

While these interesting results of LIFE clearly influenced practice, LIFE has also left us with some important unresolved issues. For example, a major unresolved problem is how far BP should be lowered in elderly patients with LVH. A subsequent analysis of LIFE data showed that mortality increased in participants if in-treatment systolic BP was <130 mmHg [2]. This is important since the most recent American hypertension guidelines [3] recommend a target SBP < 130 mmHg in all hypertensive patients, including in the elderly. The outcomes of a trial of intensive BP control in elderly patients with hypertension (The STEP Study) [4] recently supported the American guidelines: 8511 patients were enrolled in the trial, 4243 were randomly assigned to the intensive treatment, and 4268 to the standard treatment group. At 1 year of follow-up, the mean SBP was 127.5 mmHg in the intensive-treatment and 135.3 mmHg in the standard treatment group. During a median follow-up period of 3.34 years, primary outcome events occurred in 147 patients (3.5%) in the intensive treatment group, as compared with 196 patients (4.6%) in the standard treatment group (HR = 0.74; 95% CIs 0.60–0.92; p = 0.007). However, the question of increased mortality with in-treatment SBP < 130 mmHg in elderly patients with LVH, as seen in LIFE [3], remains unresolved.

In the process of publishing, more than 200 manuscripts containing LIFE data, most of which have appeared in major medical journals, some draft articles have been side lined, and forgotten and not finalized. Another complication in the process is that an article may have been returned to the responsible authors following a thorough journal review with multiple requests for additional analyses. The authors may either have not had time or found the requests too complicated and put the work aside for later, only to forget about the task or never regain the energy needed to finish the analyses and resubmit the revised manuscript. Such untoward events occurred with some drafted LIFE manuscripts. Compared to the large body of published LIFE articles, these unfinished manuscripts are small in number (eight articles detected). Importantly, the LIFE Study investigators made a major cleanup of unpublished articles in the past year, and substantial amounts of important data came out in the current round of yet unpublished articles [5–11].

First, we refer to data from the main LIFE study. Incident AF and heart failure (HF) were pre-specified secondary endpoints in LIFE, and both of these CV outcomes are strongly related to hypertension, particularly in patients with LVH. A recently published analysis [5] investigated the incidence and severity of AF and HF in 8,702 hypertensive patients with LVH receiving antihypertensive treatment. Incident AF occurred in 679 patients (7.8%) and HF in 246 patients (2.8%) during an average of 4.7 years of follow-up. Incident AF was associated with a >four-fold increased risk of developing subsequent HF (p< 0.001) in multivariable Cox analyses adjusting for age, sex, race, randomized treatment, standard CV risk factors, and incident myocardial infarction. The development of HF as a time-dependent variable was associated with a multivariable-adjusted three-fold increase of the primary study endpoint (p < 0.001), a composite of myocardial infarction, stroke or CV death. Incident HF was associated with a >three-fold increased risk of developing subsequent AF (p < 0.001). Development of AF was associated with greater than a two-fold increase of the composite primary study endpoint in multivariable Cox analysis (p = 0.028). Thus, treated hypertension patients with LVH and either AF or HF are at increased risk of developing the alternative complication. High-risk hypertensive patients who subsequently develop both AF and HF are at a particularly high risk of a composite outcome of myocardial infarction, stroke, or CV death.

Another topic of major interest in LIFE relates to whether incident left bundle branch block (LBBB) is associated with increased CV morbidity and mortality in treated hypertensive patients with LVH [6]. Annual follow-up ECGs identified 295 patients (3.4%) with incident LBBB associated with male gender, older age, higher Cornell voltage and history of diabetes, isolated systolic hypertension, and prevalent CV disease. When adjusted for the history of previous CV disease, diabetes, isolated systolic hypertension, Framingham risk score, ECG-LVH, and randomized study treatment, Cox regression models showed that incident LBBB predicted higher risk of the composite endpoint of CV death, myocardial infarction, stroke, HF, and all-cause mortality (p < 0.001 for all). These data suggest that among hypertensive patients with ECG-LVH receiving aggressive antihypertensive therapy, incident LBBB independently predicts increased risk of subsequent CV events, including HF and CV and all-cause mortality.

Because losartan reduces serum uric acid (SUA) levels, the LIFE Study also investigated the relationship of SUA with incident AF [7]. Time-varying SUA was associated with incident AF defined by Minnesota code (p < 0.0001), independent of losartan treatment (HR =0.75, p = 0.007), older age (p < 0.0001), male sex (p = 0.010), and higher Cornell voltage-duration product on ECG (HR = 1.10 per 1023 msec·mm, p = 0.034). Similar results were obtained in Cox models with SUA levels partitioned according to baseline quartiles and in which AF was defined by physician reports or by both Minnesota coding and physician reports. Thus, in-treatment SUA is a strong predictor for new-onset AF in hypertensive patients, independent of effects of antihypertensive treatment, age, sex, and ECG-LVH. Further research needs to clarify how increased uric acid levels may provoke AF.

LIFE also assessed easily available patient characteristics, including urinary albumin excretion, as predictors of the primary composite outcome of myocardial infarction, cerebral stroke, and CV death, in patients with hypertension and LVH, and developed risk algorithms for outcomes [8]. A Cox proportional hazards model identified baseline variables that had a significant impact on the occurrence of the composite endpoint in the 9193 elderly patients with LVH. The investigators developed a risk point table that assigned points to various levels of the risk factors, including the important age-gender interaction term, by adapting the point system algorithm developed from Framingham data. The Cox model identified 12 predictors of risk for the endpoint. Points for age were calculated separately for women and men due to the interaction. Risk estimates defined by point totals were compared to risk estimates determined by the Cox model; the weighted kappa for the agreement was 0.89. Thus, a number of patient characteristics predicted CV events in patients with hypertension and LVH, with age, gender, and urinary albumin excretion being the most important. The risk point table represents an intuitive method to compare the relative contributions of the various CV risk predictors and explain a complicated statistical model, including an interaction term, to patients and their physicians.

Important results also came from the LIFE echocardiographic sub-study. Available nomograms to predict aortic root (AoR) diameter from body surface area have limitations. The purpose of the LIFE sub-study was to evaluate the use of a new multivariate predictive model to identify AoR dilatation in hypertensive patients with LVH [9]. A total 943 of the 961 participants in LIFE had the baseline characteristics and echocardiographic 2 D measurements of AoR size needed to be included in the sub-study. Predicted AoR (Sinus of Valsalva) diameter was: 1.519+(age[years] × 0.010)+(height[cm] × 0.010) – (gender[1 = M, 2 = F] × 0.247), and a measured AoR diameter > the 97.5-percentile of this estimate was considered dilated. Measured AoR diameter was larger in men than in women (3.75 vs. 3.48 cm, p < 0.001) and AoR diameter predicted by the model was larger than predicted by the existing nomogram (3.52 vs. 3.28 cm, p < 0.001). Using the multivariate model to identify patients with AoR dilatation, the prevalence was 13.7% in men and 12.3% in women (p = 0.537). There was consensus of AoR phenotype (normal/dilated) between model and nomogram in 92.8% of the patients. In multivariate logistic regression, AoR dilatation by model definition was predicted by the presence of aortic regurgitation (OR 2.67, p < 0.001) and the SD increase in age (OR 0.75, p = 0.023), pulse pressure (OR 0.64, p < 0.001), left ventricular (LV) mass index (OR 1.36, p = 0.08) and stroke volume (OR 1.45, p = 0.002), but not by body weight. Using the proposed model, the prevalence of AoR dilatation was equal in men and women, and the model seems to address the effects of gender, age and body size on AoR size.

It remains unknown whether left atrial systolic force (LASF), a measure of left atrial function, predicts incident AF. Furthermore, the influence of treatment with atenolol and losartan on LASF is unclear. To address this question, 758 patients without AF at baseline were enrolled from the LIFE echocardiography sub-study [10]. Mean follow-up was 59 months. LASF was calculated using average mitral orifice area and mitral peak A velocity obtained by Doppler echocardiography. At baseline, 25% of patients had a LASF ≤ 10.3 kdyn. This quartile had a higher proportion of men, lower heart rate, body mass index and age than the other three quartiles. After controlling for these variables, patients in the first quartile had lower stroke volume compared to other quartiles. Incident AF occurred in 29 (8.1/1000 patients-years of follow-up) patients. In multivariable Cox regression analyses with backwards elimination, increasing LASF was associated with lower risk of incident AF (HR = 0.90, p = 0.001). Integrated discrimination improvement was 0.054 (p = 0.004) and there was a borderline significant net reclassification improvement of 19.2% (p = 0.075). Over time LASF decreased more in the atenolol-based than the losartan-based treatment group (p < 0.001). Low LASF was associated with higher risk of incident AF. Thus, losartan-based treatment was associated with better preservation of LASF compared to atenolol-based treatment.

While it is commonly thought that LV systolic function tends to deteriorate over time in hypertensive patients, few prospective data are available to support this notion. The LIFE investigators evaluated 680 hypertensive patients (66 years on average, 45% women) with ECG-LVH enrolled in the LIFE echocardiographic sub-study who was free of prevalent CV disease and with baseline ejection fraction (EF)≥55% [11]. Echocardiographic examination was performed annually for 5 years during anti-hypertensive treatment. Development of reduced systolic function was defined as incident EF < 50%. During a mean follow up of 4.8 years, 37 patients (5.4%) developed reduced EF without an inter-current myocardial infarction (5.4%). In analysis of covariance, patients who developed reduced EF were more often men, had greater baseline LV diameter and LV mass, lower mean EF (all p < 0.05), and similar diastolic function indices to those who maintained their EF. At the last available examination before EF reduction, independently of covariates, patients with reduced EF showed a significant increase in LA size, LV diameter, end-systolic stress, and mitral E/A ratio, as compared to those who did not develop reduced EF (all p < 0.05). In time-varying Cox regression analysis, also controlling for baseline EF, predictors of developing reduced EF were higher in-treatment LV diameter (HR = 5.19/cm) and higher in-treatment mitral E/A ratio (HR = 2.37 per unit, both p < 0.0001). Thus, in treated hypertensive patients with ECG-LVH at baseline, incident reduced EF occurred in only a few patients (n = 37), and was associated with the development of LV chamber dilation and signs of increased LV filling pressure.

In summary, these seven new publications provide novel information on incident AF, HF and LBBB in elderly hypertensive patients with LVH, important effects of uric acid for CV disease, and a risk point table representing an intuitive method to compare the relative contributions of the CV risk predictors. Further, a new model for predicting AoR diameter did better than previous nomograms. LASF was shown to predict incident AF and to be influenced by type of antihypertensive treatment, while incident reduction in systolic function occurred in few patients and was related to dilate LV chamber size and increased LV filling pressure. Since we believe that it is extremely important that written publications with novel data and analyses are available to the public with easy access beyond the scientific meetings where the data originally appeared, all of these new articles are available as open access publications [5–11]. We are uncertain whether additional analyses of LIFE data will be carried out, resulting in publication of another generation of LIFE articles in the future.

Disclosures statement

SEK & SO were member of the Steering Committee of the LIFE Study. None of the authors has otherwise any conflicts relevant to this editorial.