Prognostic implications of left ventricular hypertrophy defined by the thresholds from the international and Chinese guidelines

Abstract To compare the predictive value of mortality between left ventricular hypertrophy (LVH) defined by Chinese thresholds and defined by international guidelines in hypertension individuals and investigate better indexation methods for LVH in Chinese population. We included 2454 community hypertensive patients with Left ventricular mass (LVM) and relative wall thickness. LVM was indexed to body surface area (BSA), height2 7 and height 1 7. The outcomes were all‐cause and cardiovascular mortality. Cox proportional hazards models were used to explore the association between LVH and the outcomes. C‐statistics and time‐dependent receiver operating characteristic curve (ROC) was used to evaluate the value of those indicators. During a median follow‐up of 49 months (interquartile range 2–54 months), 174 participants (7.1%) died from any cause (n = 174), with 71 died of cardiovascular disease. LVM/BSA defined by the Chinese thresholds was significantly associated with cardiovascular mortality (HR: 1.63; 95%CI: 1.00‐2.64). LVM/BSA was significantly associated with all‐cause mortality using Chinese thresholds (HR: 1.56; 95%CI: 1.14‐2.14) and using Guideline thresholds (HR: 1.52; 95%CI: 1.08‐2.15). LVM/Height1.7 was significantly associated with all‐cause mortality using Chinese thresholds (HR: 1.60; 95%CI: 1.17‐2.20) and using Guideline thresholds (HR: 1.54; 95%CI: 1.04‐2.27). LVM/Height2.7 was not significantly associated with all‐cause mortality. C‐statistics indicated that LVM/BSA and LVM/Height1.7 by Chinese thresholds had better predictive ability for mortality. Time‐ROC indicated that only LVM/Height1.7 defined by Chinese threshold had incremental value for predicting mortality. We found that in community hypertensive populations, race‐specific thresholds should be used to classify LV hypertrophy related to mortality risk stratification. LVM/BSA and LVM/Height1.7 are acceptable normalization method in Chinese hypertension.


INTRODUCTION
Hypertension was an important risk factor for cardiac morbidity and mortality. 1 Left ventricular hypertrophy (LVH), mainly pathologically increased left ventricular mass, independent contributed to cardiovascular events in diverse populations, including hypertension. 2 Although the American Society of Echocardiography (ASE) and the European Association of Cardiovascular Imaging (EACVI) guideline has recommended LVM and relative wall thickness (RWT) reference value, normal values of these echocardiographic parameters differ significantly among different racial and ethnic groups. 8 The Echocardiographic Measurements in Normal Chinese Adults (EMINCA) study, performed in China, revealed that most echocardiographic normal values for the health Chinese were smaller than those recommended by the ASE/EACVI guideline. 9 Recently, the same investigators found that using the ASE/EACVI guideline thresholds would overestimate the proportion of abnormal LV geometry in the Chinese hypertensive patients.
They subsequently proposed the Chinese thresholds to define LVH and LV geometry patterns. 10 These findings once again support the notion that ethnic-specific normal reference values to defined LV geometric patterns are needed.
LVH is also determined by body size, apart from age, sex and race.
One method for overcoming the limitation has been indexing LVM to body size. Most commonly this is done by dividing LVM by body surface area, height 1.7 or height 2.7 . The prognostic value of LVH defined by different body size should be certified in large population-based cohorts so that the best indexing methods can be ascertained in Chinese hypertension. Therefore, the aims of the current study were to (a) quantify and compare the prognostic value of LVH defined by the Chinese thresholds and the ASE/EACVI guideline and (b) the prognostic value of LVH indexed by different methods.

Study population
This was a community-based cohort study conducted in Liao Bu County, Dongguan, Guangdong Province, China, as previously reported. 11 Center. Written informed consent was obtained before enrolment.  14 and eGFR <60 mL/min/1.73 m 2 was defined as chronic kidney disease (CKD).

Blood pressure measurement
Participants were required to sit in a quiet room for 5 min before blood pressure (BP) measurements. According to the Chinese hypertension guideline, 15 the individuals kept the upper arm at the heart level and BP was measured using the Omron HEM-7051 device (Omron Health-Care, Kyoto, Japan). Two BP measurements should be taken with 1 min interval and the average value was used. An additional measurement was required if the first two readings differ by >5 mmHg, and the mean value of the three readings was recorded.

Definitions of LVH
The echocardiographic values of LVM were indexed to BSA and height 2.7 based on the ASE/EACVI recommendations 16,17 and the European Society of Cardiology (ESC) and European Society of Hypertension (ESH) recommendations. 18 We also used another indexation method by Chirinos and colleagues 19 who proposed to index LVM to H 1.7 . Chinese thresholds were generated from recently published research. 10 LVH thresholds were showed in Table S1.

Outcome
The outcome was all-cause mortality and cardiovascular mortality.

Interobserver variability
To test the reproducibility of echocardiographic measurements, three key parameters, including IVSTd, LVEDD and LVPWTd, were remeasured in 200 randomly selected subjects from the study participants.
Interobserver variability was assessed between two investigators (Z.D. and X.Q.). Intraobserver variability was assessed by Z.D. Bland-Altman analysis was used to assess interobserver and intraobserver variability and interclass correlation coefficients were calculated ( Figure S1).

Statistical analysis
Continuous variables were presented as mean ± standard deviation if normal distribution, otherwise were presented as median and interquartile range (IQR). Categorical variables were presented as number (proportion). For categorical variables, the chi-square test or Fisher exact test were used. When the continuous data were normal distribution, two groups were compared using the paired t-test.
Non-parametric tests were used for analyzing data that were skew distribution. Univariable and multivariable Cox proportional hazards models were used to explore the association between LVH and out-

The association of LVH with cardiovascular mortality
Analysis was performed with indexing LVM to BSA, height 1.7 and height 2.7 , respectively using different thresholds. Kaplan-Meier curve showed that hypertensive patients with LVH, which using the LVM/BSA and LVM/Height 2.7 , defined by the international guideline and the Chinese thresholds had higher mortality risk (Figure 2A,B,E,F; Log-rank ratio, all p<.05). LVH defined by LVM/Height 1.7 using Chinese thresholds, not guideline threshold, had higher risk of mortality ( Figure 2C,D).
After adjusting for covariates, only LVH, which using the LVM/BSA, defined by the Chinese thresholds was associated with cardiovascular mortality (HR: 1.63, 95%CI: 1.00-2.64) ( Table 2). LVH defined by LVM/BSA using Chinese threshold had the lowest AIC.

The association of LVH and all-cause mortality
Kaplan-Meier curve showed that hypertensive patients with LVH, which using the LVM/BSA, LVM/Height 2.7 and LVM/Height 1.7 , defined by the international guideline and the Chinese thresholds had higher mortality risk, except LVM/Height 1.7 defined by international guideline ( Figure 3A,B,D,E,F; Log-rank ratio, all p<.05). After adjusting for covariates in Table 3, LVH, which using the LVM/BSA, defined by the Chinese thresholds was associated with all-cause mortality (HR: 1.56, 95%CI: 1.14-2.14), while LVH defined by the international guide was associ-TA B L E 2 Comparison of the effect of LVH defined by Chinese thresholds or Guideline thresholds on cardiovascular mortality.

Predictive value of those LVH indicators for mortality
The predictive value of those LVH indicators was showed in Table 4.
There was independent predictive value for most indicators except LVM/Height 2.7 and LVM/Height 1.7 defined by Guideline thresholds.
Among those indicators, LVH defined by LVM/BSA using Chinese threshold was better associated with all-cause mortality (C-statistics 0.578) and cardiovascular mortality (C-statistics 0.587). LVH defined by LVM/Height 1.7 using Chinese thresholds was also better associated TA B L E 3 Comparison of the effect of LVH defined by Chinese thresholds or Guideline thresholds on all-cause mortality. with all-cause mortality (C-statistics 0.576) and cardiovascular mortality (C statistics 0.578). In Table 5

DISCUSSION
We revealed that LVH defined by the Chinese thresholds was more specific to identify LVH-related mortality. LVM/BSA and using Chinese threshold was consistently associated with all-cause mortality, however, failed to predict cardiovascular mortality. Previous study showed normalization for height 1.7 fully eliminates the ethnic-independent, sex-independent relationship between LVM and height. 19 The incremental value of LVM/Height 1.7 beyond clinical model also been revealed with using Chinese thresholds, not Guideline thresholds.
There are some limitations of our study. First, the Chinese threshold from EMINCA study was limited by the fact that it was obtained by age-sex matching hypertensive patients. This threshold may not be generalizable to an otherwise general population. Second, we collected complete data on mortality but no other cardiovascular events.
There were less cardiovascular mortality and less power to detect statistical significance. Although the number for cardiovascular mortality was small, LVM/BSA defined by Chinese thresholds have shown independent prognostic value for cardiovascular mortality, not guideline thresholds. Third, no probing biomarkers explain the potential underlying mechanisms involved in the relationship between LVH and mortality. Fourth, currently, our study did not have sufficient power to explore sex differences and further follow-up was needed.
In conclusion, in community hypertensive populations, current international guideline thresholds may lead to significant overestimate of LVH and the mortality risk attributable to LVH that is identified with the Chinese thresholds is different from the international guideline thresholds. The potential value of this paper is that race-specific thresholds should be used to classify LV hypertrophy related to mortality risk stratification. The ability to properly account for ethnic body size difference will allow for more accurate quantification of LVM in Chinese threshold.

ACKNOWLEDGMENTS
This study is a joint effort of many investigators and staff members whose contribution is gratefully acknowledged. This research was supported by the Key Area R&D Program of Guangdong Province