Introduction

Drusen are extracellular debris deposits that accumulate beneath the retinal pigment epithelium with age1. These deposits serve as diagnostic indicators for various diseases, including age-related macular degeneration (AMD)2, cardiovascular disorders (CVD)3, Alzheimer’s disease (AD)4, and dense deposit disease5. While drusen may be associated with these and other diseases, the mechanisms underlying drusen development remain poorly understood.

Studies have shown that individuals with moderate to large drusen have an estimated 18% probability of progressing to late-stage AMD within five years6. Previous research has extensively explored the relationship between AMD and survival outcomes, but the results have been inconsistent7,8,9,10. Most studies have focused on the correlation between AMD and mortality specifically related to cardiovascular disease. Soft drusen, an early indicator of high risk for AMD, has been associated with various systemic health conditions; however, its implications for overall mortality are not well understood. A comprehensive understanding of the link between soft drusen and specific causes of death could provide insights into the pathophysiology of AMD.

Therefore, we aimed to investigate the association between soft drusen and mortality using the National Health and Nutrition Examination Survey (NHANES).

Methods

Design and participants

The NHANES is a comprehensive initiative designed to evaluate the health and dietary conditions of individuals in the United States. It operates through a biennial survey that employs a nationally representative sample design to gather pertinent data11. The NHANES adheres to the ethical guidelines of the National Center for Health Statistics research ethics review board, aligning with the principles of the Helsinki Declaration, and secured written consent from all participants12.

For our analysis, we utilized data from 2005 to 2008 NHANES cycles, focusing on individuals aged 40 years and older, totaling 7081 persons. We excluded 2584 individuals due to ungraded retinal images, missing drusen data, missing covariate data, or missing death data, resulting in a final sample of 4497 persons (Fig. 1).

Fig. 1
figure 1

Schematic showing inclusion criteria for study participants.

Full size image

Retinal photography and soft drusen grading

During the 2005–2008 period of NHANES, retinal photographs were collected from individuals aged 40 and above. When retinal images were available for both eyes, the data from the worse eye were used. Soft drusen are classified into three categories based on their characteristics: soft distinct drusen, soft indistinct drusen, and the presence of both types of soft drusen. The diameter of soft drusen also plays a role in their classification. Soft drusen are categorized into three types based on their diameter: <125 μm, 125–500 μm, and ≥ 500 μm.

Mortality data

The mortality data was extracted from the 2019 Public Access Link Mortality dataset which linked with the National Death Index by a specific ID. During the study period, all participants were followed until December 31, 201913. Any participant not matched to a death record was considered alive for the entire follow-up period. Causes of death encompassed any fatalities arising from any source. Utilizing the Tenth Revision of the International Classification of Diseases-10 codes, cardiovascular deaths were identified within the categories I00-I09, I11, I13, I20-I51, and I60-I69. Cancer deaths were classified under C00-C97, and all remaining cases were attributed to other causes of mortality. The survival of the participants was computed by measuring the interval from the time of the interview to either the occurrence of death or the endpoint of December 31, 2019.

Covariates data

Factors such as age, gender, race, educational level, marital status, family income, body mass index (BMI), alcohol consumption, and smoking habit were collected through household interviews14,15. Age groups were set as 40 to 60 years and above 60 years. Ethnicity was classified into non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, or other races encompassing Multiracial. Educational level was bifurcated into less than a high school diploma and a high school diploma or higher qualification. Marital status was either married or cohabiting with a partner, or unmarried or in any other relationship status. The poverty income ratio (PIR) was assessed as below the poverty line (PIR < 1) or at or above it (PIR ≥ 1). Alcohol consumption was classified into non-drinkers, those who consume 1–5 drinks monthly, 5–10 drinks monthly, and over 10 drinks monthly. Smoking patterns were classified as non-smokers, former smokers, or current smokers.

BMI was divided into underweight (less than 18.5), normal to overweight (18.5–30.0), and obese (equal to or exceeding 30.0). Diabetes was confirmed through a glycosylated hemoglobin level, a fasting plasma glucose level, a self-reported doctor’s diagnosis, or insulin usage. Hypertension was identified if an individual had a past history of hypertension reported, or a systolic blood pressure reading of 130 mmHg or more and/or a diastolic blood pressure of 80 mmHg or higher, based on the average of four measurements. CVD encompassed congestive heart failure, coronary heart disease, angina, myocardial infarction, and stroke. Walking impairment was self-reported as difficulty in walking or dependence on special aids for assistance. Health status was classified into two groups: poor to fair and good to excellent. Ocular coexisting conditions were defined as the presence of cataracts, glaucoma, or retinopathy.

Statistical analysis

Statistical evaluations were conducted considering the intricate and stratified structure of NHANES, adhering to the NHANES guidelines for analysis and reporting. Descriptive statistics are depicted as mean ± standard deviation for continuous data and as counts and proportions for categorical data. To contrast the baseline attributes between individuals with and without soft drusen, the Wilcoxon rank-sum test and a chi-square test, corrected by Rao & Scott’s second-order method, were employed. Only statistically significant covariates were incorporated into the multiple Cox regression model for adjustment purposes.

The analysis of differences in mortality age between the group with soft drusen and the one without involved employing the Wilcoxon rank-sum test, along with assessing the time elapsed to death from the initial examination. The comparison of mortality rates between these two groups was carried out using Fisher’s exact test.

Survival probabilities and their corresponding 95% confidence interval (CI) were assessed using standardized survey-weighted Cox proportional hazards models adjusting for age, gender, racial background, family income, educational attainment, hypertension, cardiovascular conditions, walking difficulties, self-reported health status, history of cancer, and ocular co-morbidities, which calculated hazard ratio (HR)14. Separate models, also survey-weighted and were employed to evaluate the impact of various soft drusen conditions and diameters on survival.

Data analysis was conducted using R software, specifically version 4.3.1, released on June 16, 2023, with the ucrt package. A two-tailed P-value of less than 0.05 was considered statistically significant for all tests.

Results

In the 2005–2008 NHANES, there were 7081 participants aged 40 years or older. This study excluded 2584 NHANES participants with ungraded retinal images, missing drusen data, missing covariate data, and missing death data, resulting in a final sample of 4497 participants for analysis.

Of the enrolled population, 767 (17.1%) had soft drusen at the baseline examination, whereas 3730 (82.9%) did not have soft drusen. Table 1 presents the demographic attributes, health-related behaviors, and prevalent health comorbidities of the total participant cohort, as well as those with and without soft drusen. Participants with soft drusen were older than those without soft drusen (62.7 ± 0.5 vs. 54.5 ± 0.2 years), and a higher proportion were male (52.0% vs. 47.0%). More participants with soft drusen had hypertension (72.0% vs. 60.0%), CVD (18.0% vs. 9.3%), and walking disorders (9.6% vs. 6.5%). They were also more likely to report poor to fair health status (20.0% vs. 16.0%), more likely to have cancer (15.0% vs. 11.0%), and more likely to have eye complications (31.0% vs. 18.0%) (Table 1).

Table 1 Demographic, health behaviours, and general health characteristics of participants with and without drusen.
Full size table

Mortality characteristics

After a median follow-up of 12.33 (11.33, 12.58) years, 1014 (22.5%) participants died from all causes. The all-cause mortality rate in the soft drusen group was 38.0%, significantly higher than the 19.0% in the non-soft drusen group (P < 0.05). Analysis focusing on specific causes revealed a higher mortality risk in individuals with soft drusen compared to those without, irrespective of whether the cause was cardiovascular disease (CVD) or any other non-CVD or non-cancer cause (Table 2).

Table 2 Mortality characteristics overall and by soft drusen.
Full size table

Stratified analyses were performed according to soft drusen status. Of the 1,014 participants who died, 723 (71.3%) had no soft drusen, 170 (16.8%) had soft distinct drusen, 25 (2.5%) had soft indistinct drusen, and 96 (9.5%) had both types of soft drusen. The all-cause mortality rates were 33.0%, 34.0%, and 53.0%, respectively, all higher than the 19.0% in the non-soft drusen group (P < 0.05). The age at death in the soft distinct drusen and both soft drusen groups was higher than in the non-soft drusen group (P < 0.05). No substantial variation was observed in the time from baseline assessment to mortality among the three categories of soft drusen groups and the group without soft drusen (P > 0.05) (Table S1).

Participants were also classified according to soft drusen diameter. Of the 1,014 people who died from all causes, 723 (71.3%) had no soft drusen, 97 (9.6%) had < 125 μm soft drusen, 95 (9.4%) had 125 to 500 μm soft drusen, and 97 (9.6%) had ≥ 500 μm soft drusen. The all-cause mortality rates were 29.0%, 36.0%, 60.0%, respectively, all higher than the 19.0% in the non-soft drusen group (P < 0.05). The average age of death from all causes was 79.0 ± 1.1, 80.0 ± 0.9, and 83.0 ± 0.9 years, respectively, showing a gradual upward trend. Each soft drusen group was older at the time of death than the non-soft drusen group (76.0 ± 0.4 years) (P < 0.05). There was no statistically significant difference (P > 0.05) in the duration from initial assessment to all-cause mortality between groups with varying degrees of soft drusen and those without soft drusen (Table S2).

Table S3 presents the correlation analysis between baseline variables and survival outcomes. Males exhibited an increased likelihood of all-cause mortality (HR 1.42; 95% CI 1.25 to 1.62; P < 0.05). After adjusting for age and sex in Cox proportional hazards regression models, several factors were significantly associated with heightened all-cause mortality risks, including race, PIR, educational level, BMI, marital status, alcohol consumption, smoking status, hypertension, diabetes, CVD, walking disability, self-rated health, cancer, and the presence of comorbid ocular diseases (all P values < 0.05).

Cause-specific mortality

Among the 1014 deaths, 235 (23.2%) were due to CVD, 250 (24.7%) to cancer, and 529 (52.2%) to causes other than CVD or cancer. Several adjusted survey-weighted Cox proportional hazards models revealed an increased mortality risk among individuals with soft drusen compared to those without (HR 1.41; 95% CI 1.22 to 1.64; P < 0.05). When examining cause-specific deaths, the presence of soft drusen was significantly associated with a higher likelihood of death from cardiovascular disease (HR 1.53; 95% CI 1.15 to 2.04; P < 0.05) and non-cardiovascular or non-cancer-related causes (HR 1.48; 95% CI 1.20 to 1.83; P < 0.05). However, no significant association was observed with cancer-related mortality (HR 1.19; 95% CI 0.77 to 1.83; P > 0.05).

Stratified analysis by the presence of soft drusen status revealed a heightened mortality risk among individuals with soft distinct drusen (HR 1.26; 95% CI 1.07 to 1.49; P < 0.05) and those with both soft drusen (HR 1.81; 95% CI 1.44 to 2.27; P < 0.05). Both soft distinct drusen and the presence of both types of soft drusen were linked to increased risks of death due to cardiovascular disease (HR for soft distinct drusen: 1.48; 95% CI 1.08 to 2.02; HR for both types of soft drusen: 1.76; 95% CI 1.05 to 2.95) as well as other causes (HR for soft distinct drusen: 1.26; 95% CI 1.04 to 1.53; HR for both types of soft drusen: 2.13; 95% CI 1.55 to 2.94).

Analysis stratified by the size of soft drusen revealed that participants with a soft drusen diameter of 500 μm or more exhibited an elevated mortality risk of all cause (HR 1.90; 95% CI 1.44 to 2.50; P < 0.05), CVD cause (HR 1.82; 95% CI 1.03 to 3.24), and other causes (HR 2.34; 95% CI 1.55 to 3.53) (Table 3)

Table 3 Cox proportional hazards models for all-cause mortality and specific-cause mortality.
Full size table

Sensitivity analysis

The sensitivity analysis revealed that all subgroup HR estimates exceeded 1, aligning with the primary findings of this investigation. Despite the HR for individuals with a BMI under 18.5 (HR 0.43, 95% CI 0.09 to 1.99, P > 0.05) and those below the poverty line (HR 0.96, 95% CI 0.57 to 1.62, P > 0.05) being less than unity, the non-significant P-values indicate the robustness of our model (Fig. S1).

Discussion

In a comprehensive national study involving 4497 American adults aged 40 and above, we revealed significant findings regarding mortality risk linked to soft drusen. Individuals presenting with soft drusen exhibited higher all-cause mortality, CVD mortality, and non-CVD or cancer-related mortality. Further analysis segmented by soft drusen status disclosed an augmented risk of all-cause mortality among participants with distinct soft drusen as well as those exhibiting both types of soft drusen. Notably, distinct soft drusen and both types of soft drusen were found to be connected with CVD mortality and non-CVD or cancer-related mortality. Additionally, our investigation stratified by the diameter of soft drusen indicated that subjects with soft drusen measuring 500 μm or more had a heightened risk of all-cause mortality compared to those without such drusen. Importantly, a soft drusen diameter of 500 μm or larger was also linked to fatalities due to CVD and non-CVD or cancer-related causes.

With the progression of age, soft drusen form as accumulations of extracellular material underneath the retinal pigment epithelium16. Subjects with moderate and large drusen have an estimated probability of 18% progression to late AMD over five years6. Numerous investigations consistently reveal a higher likelihood of all-cause mortality in individuals with late AMD, echoing the findings of our study8,17,18. Additionally, a meta-analysis has pointed out that AMD, particularly its late stage, is linked to an augmented risk of CVD mortality19. To explore the correlation between soft drusen and mortality, we conducted a similar investigation using NHANES. Our results indicate a higher prevalence of CVD and hypertension among individuals with soft drusen. Previous research has proposed a connection between CVD, hypertension, and the formation of macular drusen20. The occurrence of drusen is closely related to abnormal lipid metabolism, and some studies have found that drusen are associated with higher levels of high-density lipoprotein21. However, abnormal lipid metabolism is also an important risk factor for CVD, suggesting a heightened likelihood of CVD occurrence in individuals presenting with soft drusen.

Histological studies revealed that soft drusen contain lipoprotein particles of apolipoproteins B and E16. Therefore, investigations have employed the positioning of cholesterol and apolipoprotein B within drusen to examine the supposition that AMD and cardiovascular disease linked to atherosclerosis share common elements and processes connected to lipid deposition outside cells22,23. Additional research has revealed an association between the presence of drusen and heightened risks from the complement factor H gene, as well as two lipid-risk linked allele genes, APOE2 and CETP24. The association of these genes with cardiovascular disease supports a possible link between these two diseases. Our findings are consistent with the above theoretical basis, indicating that soft drusen are highly correlated with CVD, and thus participants with soft drusen have higher CVD mortality.

In this study, participants with soft drusen were older than participants without soft drusen. The reason for the association of soft drusen with CVD mortality and non-CVD or cancer-related mortality is not entirely clear. It is possible that soft drusen reflect a systemic comorbidity associated with frailty and aging, a hypothesis confirmed by several previous studies. A study showed that drusen are associated with aging25. Previous research has established a connection between the presence of drusen and cognitive deterioration26. The work of Chiho Shoda et al. revealed that individuals exhibiting greater levels of cerebral amyloid-β deposition exhibited increased soft drusen areas. This indicates a correlation between soft drusen and Alzheimer’s disease27. Nevertheless, in our investigation, we were unable to delve into this association due to the restricted number of Alzheimer’s disease-related deaths. Chronic kidney disease (CKD) ranked as the primary cause of mortality excluding CVD and cancer. Supporting this link, investigations have revealed a connection where CKD is associated with the presence of peripheral retinal drusen in Korean individuals aged 50 and above, further solidifying the association between drusen and CKD28. However, owing to the restricted number of fatalities resulting from CKD, further examination was impractical in our research.

In addition, it has been hypothesized that the presence of soft drusen with a diameter > 63 μm can cause difficulty in driving at night, difficulty in near work, and glare29. Our findings revealed an increased occurrence of walking disability among patients with soft drusen. Such visual impairments can give rise to various functional complications and psychological issues, including an increased risk of falls30, unintentional injuries31, and a decline in autonomy32. Regrettably, the available data on mortality linked to accidental injuries were insufficient to substantiate this proposition.

This research boasts several strengths, such as the extensive sample size of the elderly cohort, standardized procedures for evaluating soft drusen, and access to extensive demographic data, health indicators, coexisting conditions, and exhaustive mortality records. Nevertheless, there are limitations to consider. Firstly, health practices and concurrent diseases were documented only once, implying that the status of these factors for participants might have altered during the course of the follow-up period. Secondly, while efforts were made to account for multiple confounding factors, including anti- vascular endothelial growth factor treatment, the possibility of unmeasured residual confounders still exists. Nonetheless, the consistency of the HR values above 1 across all subgroup analyses strengthens the validity of our findings.

Conclusions

The investigation revealed that separate categories of soft drusen, including those categorized as distinct or indistinct soft drusen and those measuring 500 μm or more, were individually linked to all-cause mortality, CVD mortality, and mortality resulting from non-cardiovascular and non-cancerous conditions. These outcomes propose a potential shared etiology between soft drusen and CVD. Furthermore, soft drusen might serve as an indicator of systemic pathological comorbidities, symbolizing frailty and aging processes. Elucidation of these connections and the underlying mechanisms warrant further investigation.