This cohort study of 5775 adults aged 40 years and older included 2881 women (weighted proportion, 52.6%) and 2894 men (weighted, 47.4%), with a weighted mean (SE) age of 56.6 (0.4) years; 3105 participants (weighted, 77.1%) were of non-Hispanic white ancestry, 1190 (weighted, 9.7%) of non-Hispanic black ancestry, 890 (weighted, 5.4%) of Mexican American ancestry, and 590 (weighted, 7.8%) of other racial/ethnical ancestry. At baseline, any RMA was present in 1251 participants (weighted, 17.9%), of whom 710 (weighted, 9.8%) had retinopathy, 635 (weighted, 9.3%) had AVN, 64 (weighted, 1.0%) had FAN, and 21 (weighted, 0.3%) had HP. For 82.9% of the subjects, AVN and retinopathy were mutually exclusive (eFigure S4 in the Supplement). During 65 205 person-years of observation (median follow-up, 12.2 years [range, 0.1–15.0]), 1488 deaths occurred, including 452 associated with CVD, 341 with cancer, and 695 with other causes.
The baseline demographic characteristics, health-related behaviours, and general health comorbidities of the participants overall and by RMA status were presented in Table 1. Participants with RMA tended to be older (mean [SE] age, 61.6 [0.5] vs 55.6 [0.4] years), less educated (< high school, 451 [24.0%] vs 1256 [16.9%]), to have less family IPR (≥ 3.5, 318 [39.2%] vs 1610 [52.6%]), to have poor general health condition (430 [27.6%] vs 1058 [17.0%]), to have higher prevalence rates of hypertension (1001 [76.5%] vs 2872 [60.0%]), diabetes (481 [31.5%] vs 741 [11.9%]), hypercholesterolemia (581 [45.9%] vs 1785 [39.6%]), history of coronary heart disease (stroke, 111 [7.8%] vs 201 [3.5%]), congestive heart failure (103 [6.5%] vs 173 [2.8%]), and heart attack (136 [8.7%] vs 237 [4.2%]). Other characteristics were comparable between the groups with and without RMA. Participants with FAN and HP were more likely to be women (for FAN, 37 [65.1%] vs 2821 [52.4%]; for HP, 13 [61.1%] vs 2862 [52.6%]). Participants with any RMA or its subtypes invariably had higher prevalence rates of medical comorbidities, including hypertension, diabetes, hypercholesteremia, congestive heart failure, coronary heart disease, angina, heart attack and stroke (Table 1).
Table 1
Baseline Characteristics of Participants With or Without Retinal Microvascular Abnormalities
Characteristic a | Total | No RMA | Presence of RMA |
Any RMA b | Retinopathy | AVN | FAN | HP |
Participants, No. | 5775 | 4524 | 1251 | 710 | 635 | 64 | 21 |
Age, mean (SE), y | 56.6 (0.4) | 55.6 (0.4) | 61.6 (0.5) | 60.1 (0.8) | 63.1 (0.6) | 73.1 (1.8) | 67.7 (2.7) |
Sex | | | | | | | |
Male | 2894 (47.4) | 2237 (46.5) | 657 (51.6) | 389 (55.0) | 329 (49.8) | 27 (34.9) | 8 (38.9) |
Female | 2881 (52.6) | 2287 (53.5) | 594 (48.4) | 321 (45.0) | 306 (50.2) | 37 (65.1) | 13 (61.1) |
Race and ethnicity | | | | | | | |
Non-Hispanic White | 3105 (77.1) | 2470 (77.5) | 635 (75.0) | 298 (67.8) | 367 (81.0) | 59 (97.0) | 13 (85.1) |
Non-Hispanic Black | 1190 (9.7) | 897 (9.2) | 293 (11.9) | 211 (15.8) | 116 (9.0) | 0 (0) | 7 (13.6) |
Mexican American | 890 (5.4) | 684 (5.3) | 206 (6.1) | 127 (7.2) | 101 (5.6) | 4 (2.2) | 0 (0) |
Other c | 590 (7.8) | 473 (8.0) | 117 (7.1) | 74 (9.2) | 51 (4.5) | 1 (0.8) | 1 (1.3) |
Educational level | | | | | | | |
<High school | 1707 (18.2) | 1256 (16.9) | 451 (24.0) | 277 (25.8) | 214 (22.2) | 24 (33.1) | 9 (41.0) |
High school | 1437 (26.5) | 1130 (26.3) | 307 (27.3) | 178 (29.9) | 156 (26.0) | 19 (33.5) | 5 (25.8) |
Some college or more | 2629 (55.4) | 2137 (56.8) | 492 (48.7) | 255 (44.2) | 264 (51.8) | 20 (33.4) | 7 (33.2) |
Marital status | | | | | | | |
Married or living with partner | 3662 (68.8) | 2896 (69.3) | 766 (66.7) | 443 (67.6) | 382 (66.3) | 35 (61.2) | 13 (69.6) |
Widowed, divorced, separated, or never married | 2110 (31.2) | 1626 (30.7) | 484 33.3) | 267 (32.4) | 252 (33.7) | 28 (38.8) | 8 (30.4) |
Family IPR d | | | | | | | |
<1.3 | 1374 (15.5) | 1043 (14.9) | 331 (17.9) | 184 (17.3) | 178 (18.7) | 15 (18.9) | 7 (42.3) |
1.3–3.5 | 2066 (34.3) | 1564 (32.4) | 502 (42.9) | 299 (45.9) | 247 (41.4) | 30 (56.6) | 10 (52.3) |
≥3.5 | 1928 (50.3) | 1610 (52.6) | 318 (39.2) | 175 (44.2) | 159 (40.0) | 12 (24.4) | 1 (5.4) |
BMI, kg/m2 | | | | | | | |
<25.0 | 1490 (27.8) | 1225 (29.0) | 265 (22.4) | 132 (21.1) | 136 (22.0) | 22 (37.1) | 7 (24.4) |
25.0-29.9 | 2058 (35.4) | 1603 (35.1) | 455 (36.7) | 264 (36.3) | 222 (36.4) | 31 (45.6) | 5 (46.4) |
≥ 30.0 | 2181 (36.8) | 1660 (35.9) | 521 (41.0) | 310 (42.6) | 271 (41.5) | 10 (17.3) | 9 (29.3) |
Smoking status | | | | | | | |
Non-smoker | 2717 (48.2) | 2147 (49.1) | 570 (44.4) | 340 (47.6) | 267 (38.9) | 25 (41.8) | 4 (14.6) |
Past smoker | 1892 (31.3) | 1465 (30.9) | 427 (33.1) | 233 (29.2) | 225 (36.8) | 26 (39.4) | 7 (31.5) |
Current smoker | 1162 (20.5) | 909 (20.1) | 253 (22.5) | 137 (23.2) | 142 (24.3) | 12 (17.3) | 10 (53.9) |
Alcohol consumption | | | | | | | |
Never drinker | 790 (12.9) | 600 (12.2) | 190 (15.9) | 119 (20.1) | 82 (11.9) | 17 (27.0) | 2 (5.3) |
Former drinker | 609 (10.6) | 446 (9.8) | 163 (14.3) | 103 (16.9) | 73 (11.6) | 6 (10.5) | 4 (25.2) |
Current drinker | 3371 (76.6) | 2723 (78.0) | 648 (69.8) | 334 (63.0) | 358 (76.5) | 28 (62.5) | 13 (69.4) |
General health condition | | | | | | | |
Excellent to good | 4160 (81.1) | 3365 (83.0) | 795 (72.4) | 419 (67.8) | 418 (73.3) | 45 (72.5) | 10 (54.2) |
Fair or poor | 1488 (18.9) | 1058 (17.0) | 430 (27.6) | 275 (32.2) | 204 (26.7) | 17 (27.5) | 11 (45.8) |
Medical comorbidities | | | | | | | |
Hypertension | 3873 (62.9) | 2872 (60.0) | 1001 (76.5) | 564 (74.6) | 524 (80.3) | 58 (92.2) | 20 (95.7) |
Diabetes | 1222 (15.4) | 741 (11.9) | 481 (31.5) | 363 (41.8) | 171 (22.7) | 11 (15.3) | 6 (26.9) |
Hypercholesteremia | 2366 (40.7) | 1785 (39.6) | 581 (45.9) | 342 (47.6) | 294 (46.1) | 29 (44.3) | 13 (76.7) |
Congestive heart failure | 276 (3.4) | 173 (2.8) | 103 (6.5) | 73 (8.3) | 45 (5.8) | 6 (9.0) | 2 (12.2) |
Coronary heart disease | 342 (4.9) | 218 (4.0) | 124 (9.2) | 76 (9.6) | 58 (8.9) | 8 (11.0) | 3 (12.3) |
Angina/angina pectoris | 240 (3.4) | 157 (2.8) | 83 (5.9) | 52 (6.1) | 38 (6.0) | 2 (3.1) | 3 (10.4) |
Heart attack | 373 (5.0) | 237 (4.2) | 136 (8.7) | 82 (8.5) | 68 (9.0) | 9 (11.9) | 4 (16.0) |
Stroke | 312 (4.2) | 201 (3.5) | 111 (7.8) | 81 (9.9) | 51 (7.5) | 9 (12.0) | 3 (6.5) |
Abbreviations: RMA, retinal microvascular abnormality; AVN, arteriovenous nicking; FAN, focal arteriolar narrowing; HP, Hollenhorst plaque; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); IPR, income to poverty ratio; NHANES, National Health and Nutrition Examination Survey. |
a The number of participants was unweighted. All means and SEs for continuous variables and percentages for categorical variables were weighted considering the complex sampling design of the NHANES. |
b Any RMA was defined as the presence of retinopathy, arteriovenous nicking, focal arteriolar narrowing, or Hollenhorst plaque in the worse eye. |
c Categorized based on self-reports from the NHANES interview. |
d IPR is categorized according to the eligibility of the Supplemental Nutrition Assistance Program (formerly the Food Stamp Program). |
Association between RMA and Mortality
The associations of RMA with all-cause and cause-specific mortality are shown in Table 2. Participants with any RMA were at a higher risk of death during the follow-up period, with death from all-cause in 38.5% (unweighted, 482 of 1251 vs. 1006 of 4524 [22.2%]) and from CVD in 13.7% (unweighted, 172 of 1251 vs. 280 of 4524 [6.2%]) of all individuals. Kaplan-Meier curves showed that participants with any RMA or retinopathy had significantly greater all-cause and CVD mortality as compared to counterparts without retinal lesions (eFigure S5 and S6 in the Supplement) after multivariate adjustment. The multivariate Cox regression model showed that any RMA present at baseline was associated with a higher risk of all-cause mortality (HR, 1.26; 95%CI, 1.07–1.47), CVD mortality (HR, 1.36; 95%CI, 1.06–1.73) and other-cause mortality (HR, 1.33; 95%CI, 1.06–1.67). Retinopathy was also significantly associated with a higher risk of all-cause mortality (HR, 1.36; 95%CI, 1.09–1.71), CVD mortality (HR, 1.53; 95%CI, 1.04–2.26), and other-cause mortality (HR, 1.55; 95%CI, 1.20–2.01). In addition, FAN was significantly associated with an increased risk of other-cause mortality (HR, 2.06; 95%CI, 1.16–3.65), but it did not increase the risk of all-cause or CVD mortality. Neither any RMA nor its subtype was associated with cancer mortality. No significant associations were found between AVN or HP and all-cause or cause-specific mortality after multiple adjustments. We investigated the dose‒response relationship between eye laterality and the risk of mortality. There was a significant dose‒response relationship between the number of eyes affected by RMA and the risk of all-cause, CVD and other mortality (all Pfor trend <0.05) but not for cancer mortality (Pfor trend =0.946; Fig. 1).
Table 2
Associations of Retinal Microvascular Abnormality and its Subtypes With All-Cause and Cause-Specific Mortality
| Hazard Ratio (95% CI) |
| Any RMA | Retinopathy | AVN | FAN | HP |
Model | Absent | Present | Absent | Present | Absent | Present | Absent | Present | Absent | Present |
All-cause mortality |
Deaths, No./total No. | 1006/4524 | 482/1251 | 1174/4993 | 275/710 | 1233/5125 | 247/635 | 1423/5674 | 44/64 | 1464/5742 | 17/21 |
Model 1 a | 1 [Ref] | 1.42 (1.28–1.58) | 1 [Ref] | 1.71 (1.44–2.02) | 1 [Ref] | 1.19 (1.04–1.35) | 1 [Ref] | 1.25 (0.93–1.68) | 1 [Ref] | 2.78(1.66–4.68) |
Model 2 b | 1 [Ref] | 1.42 (1.22–1.65) | 1 [Ref] | 1.69 (1.39–2.05) | 1 [Ref] | 1.18 (0.95–1.46) | 1 [Ref] | 1.30 (0.77–2.22) | 1 [Ref] | 1.88 (1.21–2.92) |
Model 3 c | 1 [Ref] | 1.26 (1.07–1.47) | 1 [Ref] | 1.36 (1.09–1.71) | 1 [Ref] | 1.13 (0.90–1.41) | 1 [Ref] | 1.37 (0.90–2.10) | 1 [Ref] | 1.57 (0.73–3.38) |
CVD mortality |
Deaths, No./total No. | 280/4524 | 172/1251 | 338/4993 | 102/710 | 359/5125 | 91/635 | 432/5674 | 12/64 | 444/5742 | 6/21 |
Model 1 a | 1 [Ref] | 1.68 (1.39–2.05) | 1 [Ref] | 2.19 (1.71–2.82) | 1 [Ref] | 1.32 (0.98–1.77) | 1 [Ref] | 0.88 (0.52–1.48) | 1 [Ref] | 2.97 (1.35–6.53) |
Model 2 b | 1 [Ref] | 1.58 (1.27–1.98) | 1 [Ref] | 1.91 (1.37–2.66) | 1 [Ref] | 1.32 (0.97–1.79) | 1 [Ref] | 1.07 (0.61–1.87) | 1 [Ref] | 2.17 (0.99–4.73) |
Model 3 c | 1 [Ref] | 1.36 (1.06–1.73) | 1 [Ref] | 1.53 (1.04–2.26) | 1 [Ref] | 1.20 (0.82–1.74) | 1 [Ref] | 1.05 (0.62–1.78) | 1 [Ref] | 1.74 (0.51–5.93) |
Cancer mortality |
Deaths, No./total No. | 249/4524 | 92/1251 | 285/4993 | 44/710 | 290/5125 | 51/635 | 334/5674 | 5/64 | 334/5742 | 4/21 |
Model 1 a | 1 [Ref] | 1.07 (0.82–1.38) | 1 [Ref] | 0.97 (0.65–1.50) | 1 [Ref] | 1.05 (0.77–1.43) | 1 [Ref] | 0.65 (0.31–1.38) | 1 [Ref] | 4.25 (1.04–17.3) |
Model 2 b | 1 [Ref] | 1.05 (0.75–1.47) | 1 [Ref] | 0.85 (0.49–1.49) | 1 [Ref] | 1.24 (0.85–1.81) | 1 [Ref] | 0.43 (0.10–1.79) | 1 [Ref] | 1.76 (0.48–6.40) |
Model 3 c | 1 [Ref] | 1.01 (0.67–1.52) | 1 [Ref] | 0.79 (0.42–1.49) | 1 [Ref] | 1.21 (0.79–1.87) | 1 [Ref] | 0.44 (0.11–1.72) | 1 [Ref] | 1.64 (0.37–7.23) |
Other mortality |
Deaths, No./total No. | 477/4524 | 218/1251 | 551/4993 | 129/710 | 584/5125 | 105/635 | 657/5674 | 27/64 | 686/5742 | 7/21 |
Model 1 a | 1 [Ref] | 1.44 (1.22–1.70) | 1 [Ref] | 1.78 (1.46–2.17) | 1 [Ref] | 1.16 (0.91–1.50) | 1 [Ref] | 1.71 (1.15–2.54) | 1 [Ref] | 2.03 (0.94–4.36) |
Model 2 b | 1 [Ref] | 1.49(1.18–1.89) | 1 [Ref] | 1.96 (1.56–2.48) | 1 [Ref] | 1.06 (0.76–1.50) | 1 [Ref] | 1.82 (0.91–3.64) | 1 [Ref] | 1.77 (0.77–4.09) |
Model 3 c | 1 [Ref] | 1.33 (1.06–1.67) | 1 [Ref] | 1.55(1.20–2.01) | 1 [Ref] | 1.05 (0.79–1.39) | 1 [Ref] | 2.06 (1.16–3.65) | 1 [Ref] | 1.57 (0.60–4.08) |
Abbreviations: RMA, retinal microvascular abnormality; AVN, arteriovenous nicking; FAN, focal arteriolar narrowing; HP, Hollenhorst plaque; CI, confidence interval; Ref, reference. |
a Adjusted for age (continuous), sex (male or female), and race (non-Hispanic White, non-Hispanic Black, Mexican American, or other). |
b Further adjusted for educational level (< high school, high school, or ≥ some college), marital status (married/living with partner or widowed/divorced/separated/never married), family income level (IPR < 1.3, 1.3–3.5, or ≥ 3.5), body mass index (calculated as weight in kilograms divided by height in meters squared: <25.0, 25.0-29.9, or ≥ 30.0), smoking status (never, past, or current), and alcohol consumption (never, former, or current). |
c Further adjusted for general health condition (excellent to good, or fair/poor), diabetes (yes or no), hypertension (yes or no), hypercholesteremia (yes or no), history of congestive heart failure (yes or no), coronary heart disease (yes or no), angina/angina pectoris (yes or no), heart attack (yes or no), and stroke (yes or no). |
Stratified Analyses for Association of RMA with All-Cause Mortality
In terms of any RMA, we found significant interactions between obesity, history of CVD and any RMA with the risk of all-cause mortality (P = 0.026 and P = 0.014 for interactions, respectively) after multiple adjustment (Table 3). For the subgroup of obese individuals (BMI ≥ 30 kg/m2), the HR of all-cause mortality was 1.68 (95% CI, 1.19–2.39) for those with any RMA compared with that of the subgroup without obesity (BMI < 30 kg/m2; HR, 1.14; 95% CI, 0.96–1.35). In the subgroup without a history of CVD, the HR of all-cause mortality was 1.47 (95% CI, 1.25–1.73) for individuals with any RMA, while the risk was not increased for those with a history of CVD (HR, 1.06; 95% CI, 0.78–1.45). Similar patterns of interactions between obesity, history of CVD and AVN were found for the risk of all-cause mortality (Table 3). For retinopathy, we only found interaction between the history of CVD and retinopathy with the risk of all-cause mortality. The HRs were 1.72 (95% CI, 1.38–2.15) and 1.06 (95% CI, 0.78–1.45) for the subgroups without and with a history of CVD, respectively. No significant interactions were found between any RMA, retinopathy, AVN or any other strata variable and the risk of all-cause mortality (Table 3).
Table 3
Stratified Analyses for the Association between Retinal Microvascular Abnormalities and All-Cause Mortality a
Variable | Any RMA, HR (95% CI) | Retinopathy, HR (95% CI) | AVN, HR (95% CI) |
Absent | Present | P Value for Interaction | Absent | Present | P Value for Interaction | Absent | Present | P Value for Interaction |
Age | | | 0.184 | | | 0.347 | | | 0.216 |
<60 y | 1 [Ref] | 1.39 (0.90–2.13) | | 1 [Ref] | 1.26 (0.82–1.93) | | 1 [Ref] | 1.35 (0.69–2.65) | |
≥ 60 y | 1 [Ref] | 1.25 (1.06–1.47) | | 1 [Ref] | 1.41 (1.11–1.79) | | 1 [Ref] | 1.09 (0.86–1.38) | |
Sex | | | 0.284 | | | 0.854 | | | 0.856 |
Male | 1 [Ref] | 1.37 (1.09–1.71) | | 1 [Ref] | 1.36 (0.98–1.90) | | 1 [Ref] | 1.16 (0.91–1.49) | |
Female | 1 [Ref] | 1.13 (0.91–1.41) | | 1 [Ref] | 1.35 (1.07–1.72) | | 1 [Ref] | 1.12 (0.79–1.57) | |
Race/ethnicity | | | 0.680 | | | 0.779 | | | 0.725 |
Non-Hispanic White | 1 [Ref] | 1.26 (1.04–1.52) | | 1 [Ref] | 1.41 (1.07–1.86) | | 1 [Ref] | 1.11 (0.86–1.43) | |
Non-Whiteb | 1 [Ref] | 1.19 (0.89–1.59) | | 1 [Ref] | 1.29 (0.92–1.82) | | 1 [Ref] | 1.12 (0.79–1.59) | |
Weight status | | | 0.026 | | | 0.148 | | | 0.002 |
BMI < 30 | 1 [Ref] | 1.14 (0.96–1.35) | | 1 [Ref] | 1.25 (0.96–1.64) | | 1 [Ref] | 0.96 (0.79–1.18) | |
BMI ≥ 30 | 1 [Ref] | 1.68 (1.19–2.39) | | 1 [Ref] | 1.70 (1.14–2.54) | | 1 [Ref] | 1.68 (1.12–2.52) | |
Smoking status | | | 0.323 | | | 0.229 | | | 0.339 |
Never | 1 [Ref] | 1.10 (0.79–1.53) | | 1 [Ref] | 1.18 (0.79–1.75) | | 1 [Ref] | 0.95 (0.67–1.35) | |
Past or current | 1 [Ref] | 1.38 (1.17–1.62) | | 1 [Ref] | 1.51 (1.13–2.03) | | 1 [Ref] | 1.27 (0.99–1.64) | |
Alcohol consumption | | | 0.695 | | | 0.547 | | | 0.338 |
Never | 1 [Ref] | 1.42 (0.98–2.08) | | 1 [Ref] | 1.64 (1.10–2.43) | | 1 [Ref] | 1.11 (0.80–1.54) | |
Past or current | 1 [Ref] | 1.26 (1.06–1.49) | | 1 [Ref] | 1.41 (1.09–1.83) | | 1 [Ref] | 1.17 (0.92–1.49) | |
Hypertension | | | 0.709 | | | 0.528 | | | 0.852 |
No | 1 [Ref] | 1.09 (0.66–1.80) | | 1 [Ref] | 1.46 (0.76–2.83) | | 1 [Ref] | 1.03 (0.57–1.87) | |
Yes | 1 [Ref] | 1.31 (1.11–1.54) | | 1 [Ref] | 1.37 (1.09–1.71) | | 1 [Ref] | 1.15 (0.91–1.45) | |
Diabetes | | | 0.806 | | | 0.804 | | | 0.507 |
No | 1 [Ref] | 1.22 (1.03–1.45) | | 1 [Ref] | 1.32 (1.01–1.73) | | 1 [Ref] | 1.08 (0.86–1.35) | |
Yes | 1 [Ref] | 1.38 (1.03–1.84) | | 1 [Ref] | 1.55 (1.10–2.18) | | 1 [Ref] | 1.24 (0.83–1.87) | |
Hypercholesterolemia | | | 0.641 | | | 0.790 | | | 0.045 |
No | 1 [Ref] | 1.16 (0.95–1.42) | | 1 [Ref] | 1.37 (1.02–1.85) | | 1 [Ref] | 1.01 (0.77–1.34) | |
Yes | 1 [Ref] | 1.39 (1.11–1.74) | | 1 [Ref] | 1.35 (1.01–1.81) | | 1 [Ref] | 1.42 (1.04–1.94) | |
History of CVD c | | | .014 | | | .013 | | | .014 |
No | 1 [Ref] | 1.47 (1.25–1.73) | | 1 [Ref] | 1.72 (1.38–2.15) | | 1 [Ref] | 1.25 (1.00-1.55) | |
Yes | 1 [Ref] | 1.06 (0.78–1.45) | | 1 [Ref] | 1.06 (0.78–1.45) | | 1 [Ref] | 1.06 (0.78–1.45) | |
Abbreviations: RMA, retinal microvascular abnormality; AVN, arteriovenous nicking; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CVD, cardiovascular disease; HR, hazard ratio; Ref, reference. |
a Analyses were adjusted for age (continuous), sex (male or female), race (non-Hispanic White, non-Hispanic Black, Mexican American, or other), educational level (< high school, high school, or ≥ some college), marital status (married/living with partner, or widowed/divorced/separated/never married), family income level (IPR < 1.3, 1.3–3.5, or ≥ 3.5), body mass index (calculated as weight in kilograms divided by height in meters squared: <25.0, 25.0-29.9, or ≥ 30.0), smoking status (never, past, or current), alcohol consumption (never, former, or current), general health condition (excellent to good, or fair/poor), diabetes (yes or no), hypertension (yes or no), hypercholesteremia (yes or no), histories of congestive heart failure (yes or no), coronary heart disease (yes or no), angina (yes or no), heart attack (yes or no), and stroke (yes or no). Stratification was not conducted for focal arteriolar narrowing or Hollenhorst plaque due to insufficient samples for subgroup analyses. |
b Non-white includes non-Hispanic Black, Mexican American, and other races and ethnicities. |
c A history of CVD was defined as a self-reported history of having a previous physician diagnosis of congestive heart failure, coronary heart disease, angina, heart attack, or stroke. |
Stratified Analyses for the Association of RMA with CVD Mortality
For CVD mortality, stratified analyses showed significant interaction between any RMA and smoking status with the risk of CVD mortality (P = 0.015 for interaction, Table 4). For the subgroup of patients who were past or current smokers, the HR of CVD mortality was 1.82 (95% CI, 1.26–2.62), and for the subgroup of non-smokers, the HR of CVD mortality was 0.79 (95% CI, 0.43–1.46). Similar interaction was found between retinopathy and smoking status with the risk of CVD mortality after multiple adjustment (P = 0.011 for interaction, Table 4). The interaction between AVN and obesity with the risk of CVD mortality was significant (P = 0.035 for interaction), showing HR = 1.96 (95% CI, 1.23–3.13) in the obesity subgroup and HR = 0.99 (95% CI, 0.64–1.55) in the non-obesity subgroup (Table 4). No other significant interactions were found between RMA and strata factors in our analyses.
Table 4
Stratified Analyses for the Association between Retinal Microvascular Abnormalities and CVD Mortality a
Variable | Any RMA, HR (95% CI) | Retinopathy, HR (95% CI) | AVN, HR (95% CI) |
Absent | Present | P Value for Interaction | Absent | Present | P Value for Interaction | Absent | Present | P Value for Interaction |
Age | | | 0.824 | | | 0.478 | | | 0.904 |
<60 y | 1 [Ref] | 1.07 (0.46–2.50) | | 1 [Ref] | 1.42 (0.46–4.39) | | 1 [Ref] | 0.78 (0.30–2.06) | |
≥ 60 y | 1 [Ref] | 1.38 (1.04–1.83) | | 1 [Ref] | 1.50 (0.96–2.36) | | 1 [Ref] | 1.25 (0.86–1.82) | |
Sex | | | 0.545 | | | 0.538 | | | 0.338 |
Male | 1 [Ref] | 1.61 (1.04–2.47) | | 1 [Ref] | 1.87 (1.20–2.92) | | 1 [Ref] | 1.06 (0.64–1.75) | |
Female | 1 [Ref] | 1.20 (0.74–1.94) | | 1 [Ref] | 1.25 (0.65–2.41) | | 1 [Ref] | 1.57 (0.86–2.86) | |
Race/ethnicity | | | 0.697 | | | 0.567 | | | 0.752 |
Non-Hispanic White | 1 [Ref] | 1.40 (1.01–1.94) | | 1 [Ref] | 1.69 (1.00-2.86) | | 1 [Ref] | 1.16 (0.75–1.78) | |
Non-Whiteb | 1 [Ref] | 1.09 (0.63–1.88) | | 1 [Ref] | 1.10 (0.63–1.93) | | 1 [Ref] | 1.25 (0.77–2.03) | |
Weight status | | | 0.329 | | | 0.826 | | | 0.035 |
BMI < 30 | 1 [Ref] | 1.27 (0.93–1.73) | | 1 [Ref] | 1.55 (0.97–2.46) | | 1 [Ref] | 0.99 (0.64–1.55) | |
BMI ≥ 30 | 1 [Ref] | 1.82 (1.26–2.62) | | 1 [Ref] | 1.73 (1.07–2.81) | | 1 [Ref] | 1.96 (1.23–3.13) | |
Smoking status | | | 0.015 | | | .011 | | | 0.157 |
Never | 1 [Ref] | 0.79 (0.43–1.46) | | 1 [Ref] | 0.84 (0.36–1.93) | | 1 [Ref] | 0.89 (0.52–1.52) | |
Past or current | 1 [Ref] | 1.90 (1.43–2.51) | | 1 [Ref] | 2.24 (1.47–3.41) | | 1 [Ref] | 1.46 (0.93–2.28) | |
Alcohol consumption | | | 0.481 | | | 0.279 | | | 0.462 |
Never | 1 [Ref] | 1.53 (0.86–2.71) | | 1 [Ref] | 1.79 (1.00-3.18) | | 1 [Ref] | 1.30 (0.60–2.81) | |
Past or current | 1 [Ref] | 1.40 (1.07–1.83) | | 1 [Ref] | 1.66 (1.09–2.53) | | 1 [Ref] | 1.25 (0.88–1.78) | |
Hypertension | | | 0.688 | | | 0.401 | | | 0.265 |
No | 1 [Ref] | 1.66 (0.74–3.73) | | 1 [Ref] | 0.86 (0.39–1.94) | | 1 [Ref] | 2.26 (0.94–5.43) | |
Yes | 1 [Ref] | 1.38 (1.05–1.81) | | 1 [Ref] | 1.62 (1.06–2.46) | | 1 [Ref] | 1.14 (0.75–1.72) | |
Diabetes | | | 0.753 | | | 0.694 | | | 0.721 |
No | 1 [Ref] | 1.24 (0.89–1.72) | | 1 [Ref] | 1.36 (0.88–2.09) | | 1 [Ref] | 1.16 (0.76–1.76) | |
Yes | 1 [Ref] | 1.52 (0.88–2.64) | | 1 [Ref] | 1.80 (0.90–3.59) | | 1 [Ref] | 1.43 (0.72–2.86) | |
Hypercholesterolemia | | | 0.297 | | | 0.990 | | | 0.080 |
No | 1 [Ref] | 1.28 (0.90–1.83) | | 1 [Ref] | 1.61 (0.99–2.61) | | 1 [Ref] | 0.99 (0.59–1.66) | |
Yes | 1 [Ref] | 1.47 (1.04–2.07) | | 1 [Ref] | 1.37 (0.74–2.52) | | 1 [Ref] | 1.69 (1.08–2.64) | |
History of CVD c | | | 0.296 | | | 0.296 | | | 0.296 |
No | 1 [Ref] | 1.53 (1.20–1.95) | | 1 [Ref] | 1.88 (1.23–2.86) | | 1 [Ref] | 1.38 (0.95–2.01) | |
Yes | 1 [Ref] | 1.14 (0.70–1.86) | | 1 [Ref] | 1.14 (0.70–1.86) | | 1 [Ref] | 1.14 (0.70–1.86) | |
Abbreviations: RMA, retinal microvascular abnormality; AVN, arteriovenous nicking; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CVD, cardiovascular disease; HR, hazard ratio; CI, confidence interval; Ref, reference. |
a Analyses were adjusted for age (continuous), sex (male or female), and race (non-Hispanic White, non-Hispanic Black, Mexican American, or other), educational level (< high school, high school, or ≥ some college), marital status (married/living with partner, or widowed/divorced/separated/never married), family income level (IPR < 1.3, 1.3–3.5, or ≥ 3.5), body mass index (calculated as weight in kilograms divided by height in meters squared: <25.0, 25.0-29.9, or ≥ 30.0), smoking status (never, past, or current), alcohol consumption (never, former, or current), general health condition (excellent to good, or fair/poor), diabetes (yes or no), hypertension (yes or no), hypercholesteremia (yes or no), histories of congestive heart failure (yes or no), coronary heart disease (yes or no), angina (yes or no), heart attack (yes or no), and stroke (yes or no). Stratification was not conducted for focal arteriolar narrowing or Hollenhorst plaque due to insufficient samples for their subgroup analyses. |
b Non-white includes non-Hispanic Black, Mexican American, and other races and ethnicities. |
c A history of CVD was defined as a self-reported history of having a previous physician diagnosis of congestive heart failure, coronary heart disease, angina, heart attack, or stroke. |
Sensitivity Analyses
Sensitivity analyses demonstrated robustness of the results of our main analyses. The associations did not significantly differ when excluding the participants who died within 2 years of follow-up (eTable 1 in the Supplement), excluding the participants who had a history of CVD at baseline (eTable 2 in the Supplement), excluding the participants with malignancy at baseline (eTable 3 in the Supplement), performing the main analyses by matching subjects through a propensity score matching strategy (eTable 4 in the Supplement), or by changing the definition of any RMA through excluding Hollenhorst plaque (eTable 5 in the Supplement).