Effects of low-density lipoprotein cholesterol on cardiovascular disease and all-cause mortality in elderly patients (≥75 years old)

Although increased low-density lipoprotein cholesterol (LDL-C) is one of the major risk factors for the cardiovascular disease (CVD), the associations of LDL-C with CVD and all-cause mortality are unclear in elderly (≥75 years) individuals. A total of 3674 individuals aged 75 or older underwent medical examinations at the Kailuan Group from 2006 to 2007, including 3478 males (94.67%) and 196 females (5.33%). Participants were divided into three groups based on the LDL-C level: the ideal level (LDL-C < 100 mg/dl), appropriate level (100 mg/dl ≤ LDL-C < 130 mg/dl) and elevated level (LDL-C ≥ 130 mg/dl) groups. CVD and all-cause mortality events were recorded during the follow-up period. The Cox proportional hazards regression model was applied to evaluate the effects of LDL-C on CVD and all-cause mortality events. The average follow-up time was 9.87 ± 3.60 years. After adjustment for confounding factors, the multivariate Cox proportional hazards regression model showed that the CVD risk in the elevated level group was 1.45 (95% CI, 1.08–1.95), acute myocardial infarction risk was 1.96 (95% CI, 1.19–3.24) and all-cause mortality risk was 1.18 (95% CI, 1.02–1.37) compared with those in the ideal level group. For every standard deviation increase in LDL-C levels, the CVD risk increased by 10%, acute myocardial infarction risk increased by 21% and all-cause mortality event risk increased by 4%. No association was observed between elevated LDL-C levels and the risk of stroke. In the sample of older Chinese individuals investigated in the present study, elevated LDL-C levels (≥130 mg/dl) are a risk factor for CVD and all-cause mortality.


Introduction
Cardiovascular disease (CVD) is the leading cause of mortality in the global population [1], and animal experiments and clinical studies have confirmed that an increased low-density lipoprotein cholesterol (LDL-C) level is one of the major risk factors for the occurrence and development of atherosclerosis [2,3]. A number of randomised controlled trials using lipid-lowering drug interventions, such as a meta-analysis of 27 randomised controlled trials on the effectiveness and safety of treatments reducing LDL-C levels, have shown that lowering LDL-C levels reduces the risk of CVD in the future. For every 39 mg/dl decrease in LDL-C levels, the risk of major coronary events decreased by 24%, and the risk of all-cause mortality decreased by 9% [4]. However, in randomised controlled trials on lipidlowering therapy, only some randomised controlled trials included some patients ≥75 years of age. The PROSPER study provided the first evidence of a benefit from statins in the primary prevention of coronary heart disease in the elderly population (70-82 years), showing a significant 15% reduction in the composite cardiovascular endpoint [5]. None of the trials were specifically conducted in elderly individuals ≥75 years of age. Therefore, the various guidelines based on the results of randomised controlled trials were applicable only to adults aged <75 years.
At present, with the improvement of living standards and medical standards, elderly individuals are healthier than before; therefore, the life expectancy of elderly individuals is significantly longer than before, which delays the onset of disease [6]. The proportion of the global elderly population that is over 75 years old is estimated to exceed 20% for the first time in 2046 and that their number will reach 410 million [7]. Currently, approximately 73 million people over the age of 75 live in China. With the advent of an ageing society, approximately 123 million people over the age of 75 are estimated to be living in China by 2030. With population growth and ageing, CVD events are expected to increase by 50% annually, and they mainly occur in the elderly population [8]. Therefore, an urgent need is to address the issue of abnormal lipid metabolism in individuals aged 75 years and older, especially regarding the effect of LDL-C on cardiovascular events and whether intervention is needed. Therefore, this article prospectively analysed the associations of LDL-C levels with CVD and all-cause mortality in elderly individuals ≥75 years of age in the Kailuan study cohort.

Study design and population
The Kailuan study is a prospective cohort study based on the Kailuan community in the Chinese industrial city of Tangshan [9]. The Kailuan community is a functional and comprehensive community owned and managed by the Kailuan Group. Eleven hospitals are responsible for providing health care to the community. From June 2006 to October 2007, 101 510 participants (81 110 men and 20 400 women) had health records in the 11 hospitals, and their health status was updated every 2 years according to the follow-up protocol. All participants underwent a questionnaire assessment, clinical examination and laboratory assessment.
We selected elderly individuals ≥75 years old who participated in physical examinations for the first time between 2006 and 2007 as the observation subjects. Notably, 4060 individuals met the following inclusion criteria: persons aged ≥75 years old who participated in a Kailuan Group Medical Examination for the first time between 2006 and 2007, persons who could complete the questionnaire, and persons who agreed to participate in the study and signed an informed consent form. Three hundred eighty-six individuals were excluded due to a history of CVD or cancer. Thus, 3674 subjects were included in our analysis. This research complied with the Declaration of Helsinki and was approved by the ethics committee of Kailuan General Hospital (batch number: 200608).

LDL-C measurement and grouping method
Blood samples were collected from the antecubital vein in the morning after an overnight fast and transfused into vacuum tubes containing EDTA. The samples were centrifuged within 4 h, and routine blood examinations were conducted to measure serum total cholesterol (CHOD method), HDL cholesterol (direct test method-selective inhibition method), LDL cholesterol (direct test methodsurfactant removal method) and triglyceride levels (GPO method) and then analyses were carried out (interassay coefficient of variation<10%; Mind Bioengineering Co. Ltd., Shanghai, China). All blood parameters were measured using an autoanalyser (Hitachi 7600; Hitachi, Japan) at the central laboratory of the Kailuan General Hospital.
International guidelines for blood lipid management mostly set LDL-C treatment targets based on the overall risk level of CVD. Due to the lack of a model for risk stratification of CVDs in elderly individuals (≥75), the population in this study was divided into groups according to the appropriate blood lipid level and abnormal stratification criteria for primary prevention of atherosclerotic CVD in 2016 Chinese guidelines for the management of dyslipidaemia in adults [10]. The participants were divided into three groups: an ideal level group (LDL-C < 100 mg/dl), an appropriate level group (100 mg/dl ≤ LDL-C < 130 mg/dl) and an elevated level group (LDL-C ≥ 130 mg/dl).

Data collection
Information on demographic, socioeconomic and other variables was obtained using standardised questionnaires at the baseline interview between 2006 and 2007, including age, sex, education, physical activity, cigarette smoking, alcohol consumption and medical history. Diabetes was defined as a fasting blood glucose level ≥7.0 mmol/l, selfreport of a physician's diagnosis, or self-reported use of antidiabetic medication. Hypertension was defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg or systolic blood pressure <140 mmHg and diastolic blood pressure <90 mmHg along with a history of clearly diagnosed hypertension or use of antihypertensive drugs. Smoking was defined as smoking at least 1 cigarette a day on average over the past year for at least 1 year. Drinking was defined as drinking liquor (alcohol content >50%) ≥100 ml/d on average over the past year for a duration >1 year. Regular physical exercise refers to physical exercise ≥3 times/week with a duration ≥30 min/ session and body mass index (BMI) was defined as body weight (kg)/height 2 (m 2 ).

Outcome
The first physical examination of the Kailuan Group between 2006 and 2007 was considered the starting point for the follow-up, and CVD and all-cause mortality were considered the endpoints. CVDs included acute myocardial infarction (AMI) and stroke. Stroke included ischaemic stroke and haemorrhagic stroke, which included subarachnoid haemorrhage and cerebral haemorrhage.
For participants with two or more CVDs, the first occurrence of CVD was the endpoint. For those individuals who died but did not have CVD, the time of death was the endpoint of follow-up. For those without CVD and those who survived, the date of the last follow-up was 31 December 2019.
During the follow-up period, trained medical staff verified the information of the observation subjects at the hospitals of the Kailuan Group and the designated hospitals of the municipal medical insurance every year, determined the diagnosis of CVD according to the international disease classification code ICD-10 and recorded the endpoint events. All diagnoses were confirmed by professional physicians according to the patients' hospitalisation records. Every year, information on deaths was obtained by consulting the Kailuan Social Security Information System.

Statistical analysis
SAS (version 9.4; SAS Institute, Cary, NC) was used for statistical analyses. Measurement data conforming to the normal distribution are reported as x ± s, and comparisons between multiple groups were performed using singlefactor analysis of variance. Measurement data with a skewed distribution are reported as the medians (P25, P75), and comparisons between groups were performed using the Kruskal-Wallis rank sum test. The count data are reported as n (%), and the χ 2 test was applied. The Kaplan-Meier method was used to calculate the cumulative incidence of endpoint events, and the log-rank test was performed; the Cox proportional hazards regression model was used to analyse the hazard ratio (HR) and 95% confidence interval (CI) of endpoint events in different LDL-C groups. Each additional standard deviation (SD) in LDL-C levels was used as an independent variable in the Cox proportional hazards regression model to observe any dose-response relationship. Since the rate of all-cause mortality in this study population was as high as 54.49%, a death competition risk model analysis was performed when analysing the effect of LDL-C levels on CVD events. Because subjects with high LDL-C levels had higher risks of CVD and allcause mortality within 2 years after the start of follow-up, this population was excluded from the sensitivity analysis. Missing values for LDL-C levels were input with the median. Missing values of covariates were added using multiple imputation methods. A two-sided test was used, and P < 0.05 was considered statistically significant.

Results
We analysed 3674 subjects, including 3478 males (94.67%) and 196 females (5.33%), who participated in health checkups between 2006 and 2007 and were ≥75 years old, with an average age of 79.40 ± 3.67 years old. Sixty-five missing LDL-C levels were input as the median. Of these participants, 2610, 726 and 338 were in the ideal level group, the appropriate level group and the elevated level group, respectively. The median LDL-C levels of the three groups were 72.78, 110.23 and 147.88 mg/dl, respectively. Compared with the ideal level group, the elevated level group had lower rates of middle school or college education and higher levels of high-sensitivity C-reactive protein (hs-CRP) and systolic blood pressure ( Table 1).

Incidence of CVD and all-cause mortality in groups with different LDL-C levels
The average follow-up time of the 3674 subjects was 9.87 ± 3.60 years. Twenty-four subjects experienced two CVD events, of whom 10 had both AMI and stroke and 14 had both ischaemic stroke and haemorrhagic stroke. Considering the first occurrence of CVD as the endpoint, 439 cases were observed, with 297, 88, and 54 cases in the ideal level group, appropriate level group and elevated group, respectively. The incidence density was 12.01 per thousand person-years, 13.31 per thousand person-years and 17.97 per thousand person-years, respectively. We identified 122 cases of AMI events, with 79, 23, and 20 cases in the three groups, and the incidence density was 3.09 per thousand person-years, 3.32 per thousand person-years and 6.35 per thousand person-years, respectively. Notably, 2002 cases of all-cause mortality, and the three groups had 1376, 418 and 208 cases; the incidence density was 52.97 per thousand person-years, 59.07 per thousand person-years and 64.87 per thousand person-years, respectively ( Table 2).

Cox proportional hazards regression analysis of the effect of different LDL-C levels on CVD and all-cause mortality events
The dependent variables were whether CVD, AMI, stroke, ischaemic stroke, haemorrhagic stroke and all-cause mortality occurred during the follow-up period. The ideal level group was used as the control group for the Cox proportional hazard regression analysis. After adjusting for age, sex, smoking, drinking, physical exercise, education level, hypertension, diabetes, atrial fibrillation history, hs-CRP level, BMI, use of lipid-lowering drugs and use of antihypertensive drugs, the risk of CVD in the appropriate level group and the elevated group was 1.05 (95% CI, 0.82-1.33) and 1.45 (95% CI, 1.08-1.95), the risk of AMI was 1.00 (95% CI, 0.63-1.59) and 1.96 (95% CI, 1.19-3.24) and the risk of all-cause mortality was 1.10 (95% CI, 0.99-1.23) and 1.18 (95% CI, 1.02-1.37), respectively. For every additional SD (36 mg/dl) in LDL-C levels, the risk of CVD was 1.10 (95% CI, 1.01-1.20), and the risk of AMI was 1.21 (95% CI, 1.05-1.40). The risk of death was 1.04 (95% CI, 1.00-1.08). After considering the competing risk of Values are presented as n (%), mean ± SD or median (P25, P75) SBP systolic blood pressure, DBP diastolic blood pressure, MAP mean arterial pressure, LDL-C low-density lipoprotein cholesterol, HDL-C high-density lipoprotein cholesterol, TC total cholesterol, TG triglycerides, FBG fasting blood glucose, hs-CRP high-sensitivity C-reactive protein, BMI body mass index Subjects with CVD or all-cause mortality within 2 years after the start of follow-up were excluded death to CVD, the results remained the same. Subjects with CVD or all-cause mortality within 2 years after the start of follow-up were excluded and a sensitivity analysis was performed. The following results were obtained: the CVD risk in the elevated level group was 1.41 (95% CI, 1.01-1.98), the AMI risk was 1.81 (95% CI, 1.00-3.28) and the all-cause mortality risk was 1.14 (95% CI, 0.98-1.34). We did not observe an association between an elevated LDL-C level and the risk of stroke (Table 2).

Discussion
We first confirmed that elevated LDL-C levels were a risk factor for CVD and all-cause mortality in Chinese individuals ≥75 years old, and the increased risk of CVD caused by high LDL-C levels was mainly observed for AMI, without an increase in the risk of stroke; moreover, the risk of haemorrhagic stroke and ischaemic stroke did not increase. A total of 3674 subjects aged ≥75 years in the Kailuan study were followed for approximately 10 years. After adjustment for possible confounding factors, the risks of CVD and AMI events in the group with elevated LDL-C levels were 1.45 times and 1.96 times those of the ideal level group. The risk of all-cause mortality was 1.18 times that of the ideal level group. A dose-response relationship may exist between LDL-C levels and the risks of CVD and all-cause mortality. For every SD (36 mg/dl) increase in the LDL-C level, the risk of CVD increased by 10%, the risk of AMI increased by 21% and the risk of death increased by 4%. The abovementioned associations persisted after the onset of CVD was adjusted for the competing risk of death. Very few patients took lipid-lowering therapy, and even fewer were treated with this medication in the group of patients with higher LDL-C levels. Therefore, the long exposure to high LDL-C levels might have affected the higher incidences of CVD and all-cause mortality. Subjects with CVD or all-cause mortality within 2 years after the start of follow-up were excluded, and a sensitivity analysis was performed. The association between LDL-C levels and all-cause mortality was weakened in the elevated level group, but the trend of an increased risk was still observed.
The association between elevated LDL-C levels and CVD in the adult population has been confirmed [11,12], but controversy still exists regarding this association in the elderly population ≥75 years old. The findings from the National Institutes of Health pooled cohort indicated that elevated LDL-C levels are not associated with the risk of CVD in the elderly population aged ≥75 years [13]. The results of a primary prevention cohort study in a 70-to 100-year-old population showed that an elevated LDL-C level was associated with the risk of AMI and CVD. Every 39 mg/dl increase in the LDL-C level increased the risk of AMI and CVD by 25% and 12%, respectively [14]. The present study obtained similar results. For every increase in the LDL-C level (36 mg/dl), the risks of AMI and CVD increased by 21% and 10%, respectively. Our results show that a high LDL-C level is not a risk factor for stroke. A recent Korean national longitudinal study showed that a high LDL-C level is a protective factor against ischaemic stroke among individuals ≥65 years of age. Compared with individuals in the first quartile, the risk of ischaemic stroke for subjects in the fourth quartile of LDL-C levels was reduced by 20% [15].
The relationship between LDL-C levels in the elderly and all-cause mortality is also controversial. The results of clinical studies examining a 75-year-old population by Nilsson et al. showed no correlation between LDL-C levels and allcause mortality [16]. Meanwhile, a study by Liang et al. on the association between serum total cholesterol levels and risks of cardiovascular and non-cardiovascular mortality in elderly individuals showed an inverse association between a high total cholesterol level and reduced all-cause mortality in older adults that is primarily due to non-cardiovascular mortality, especially among those who are not treated with cholesterol-lowering medications [17]. Another study assessing the relationship between lipoprotein cholesterol levels and mortality reported a negative correlation between LDL-C levels and all-cause mortality in an elderly population ≥70 years old [18]. However, a number of current interventional trials of lipid-lowering drugs have confirmed that LDL-C-lowering therapy in elderly individuals ≥75 years old significantly reduces the risk of cardiovascular mortality or all-cause mortality [19][20][21][22][23][24]. Our results are consistent with the results of the intervention study.
Due to the different life expectancies of elderly people, significant functional heterogeneity and possible weakness, comorbidities and prescriptions of multiple drugs, evidence for the management of blood lipid levels in the elderly is insufficient [25,26]. Therefore, the need for lipid-lowering therapy and the target value of lipid-lowering therapy in the elderly are highly controversial issues. Among our observation subjects, the average LDL-C level in the appropriate level group was 110 mg/dl, which was higher than the target LDL-C level of 100 mg/dl for primary prevention in adults, but the CVD risk in the appropriate level group did not increase and the risk of death increased by only 10% (P = 0.082). Therefore, a target value of 100 mg/dl for primary prevention among elderly patients may not be appropriate. According to our research, the target value can increase to 130 mg/dl. However, since the Kailuan study was a prospective cohort study based on the physical examination data from Kailuan Group employees, most of whom were male, and the male proportion in our study population was 94.67%, more data are needed to test whether the study results can be generalised to the elderly population.
Old age is an unchangeable risk factor, and elderly individuals often have multiple diseases. Therefore, randomised controlled trials often exclude elderly individuals. Even if elderly subjects are included, many conditions are set, and the results are not universal. The results of this research are derived from real-world data and have the value of promotion. Based on our research results, we support some guidelines, such as the 2018 Guideline for US Blood Lipid Management [11] and the 2019 Guideline for the Management of Dyslipidaemia in European Society of Cardiology/European Atherosclerosis Association [27], which are recommended for longer life expectancy (more than 1 year) and recommend that elderly patients ≥75 years old with elevated LDL-C levels should be administered lipid-lowering treatment.
In our observation population, the risk of myocardial infarction was similar to all-cause mortality caused by high LDL-C levels, but the absolute number of all-cause deaths was much greater than the number of myocardial infarctions. Therefore, the greatest benefit of lipid-lowering interventions may be to reduce all-cause mortality. A study on the use of statins and all-cause mortality among veterans aged 75 years and older in the United States showed that the risk of all-cause mortality was reduced by 25% in those who took statins compared with those who did not take statins [28]. Moreover, moderate-intensity statin treatment reduces LDL-C levels by 25-50% [9]. According to this rough calculation, if the elevated group used statin drugs, LDL-C levels would decrease to 74-111 mg/dl, which is close to the appropriate level. Therefore, adverse events and all-cause mortality might theoretically be reduced.
Our results may underestimate the effects of high LDL-C levels on CVD and all-cause mortality in elderly people due to the deviation of healthy survival. Previous studies have found that the presence of high LDL-C levels at a young age is the main cause of early-onset CVD and early death. The average age in our observation population was 79 years. These surviving individuals were relatively healthy compared to those with early-onset CVD or early death, and the LDL-C level of elevated group was close to the marginal increase of 130 mg/dl in the primary prevention population of CVD in China. Therefore, we did not observe an effect of higher LDL-C levels on CVD and all-cause mortality in the elderly population, and this result may underestimate the effect of LDL-C levels on CVD and allcause mortality in elderly people.

Strengths and limitations of the study
Advantages of research: our data were obtained from a relatively large and stable cohort. The penetration rate of statins in China was not high in 2006, and this study was not affected by the use of lipid-lowering drugs.
Our research also has certain limitations. No cause of death was provided in the data. The subjects were retirees using public health care, and the proportion of males among the subjects was high; therefore, the results of this study may not be applicable to other populations. Point estimation of LDL-C levels was carried out, and the effects of changes in LDL-C levels on CVD and all-cause mortality risks were not observed.

Conclusions
Briefly, in the sample of older Chinese individuals investigated in the present study, LDL-C levels higher than 130 mg/dl are a risk factor for CVD and all-cause mortality, and thus the primary prevention target value may be slightly higher than that of adults. Meanwhile, guidelines should be strengthened to recommend lipid-lowering interventions for people older than 75 years with LDL-C levels greater than 130 mg/dl.

Data availability
Data that support the findings of this study are available from the corresponding author upon reasonable request and approval.

Code availability
Codes of this study are available from the corresponding author upon reasonable request and approval.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.