Abstract
Purpose
To determine if incidence of contrast sensitivity (CS) impairment differs by generation and identify factors to explain these differences.
Methods
The Beaver Dam Eye Study (BDES) and Beaver Dam Offspring Study (BOSS) are cohort studies of aging adults in Beaver Dam, Wisconsin. Baseline examinations occurred from 1993–1995 (BDES) and 2005–2008 (BOSS). Follow-up examinations occurred in five-year intervals. CS testing was conducted with Pelli-Robson letter sensitivity charts; Incident impairment was a log CS score <1.55 in either eye at follow-up. Associations of incidence with generation were investigated using estimated hazard ratios (HR) with 95% confidence intervals (CI).
Results
Participants (N=3185) had a mean age of 51.9 years at baseline (standard deviation=9.9) and 51.9% were female. Ten-year cumulative incidence of CS impairment was 40.1%, was higher among women (41.7%) than men (38.8%) and increased by age group. The risk of incident CS impairment decreased by 39% per generation. In multivariable models, the Baby Boom Generation (HR=0.42, 95%CI= 0.31, 0.58) and Generation X (HR=0.56, 95%CI=0.34, 0.91) had a significantly decreased risk of CS impairment compared to the Greatest Generation. Results were similar in sensitivity analyses excluding those with cataract, age-related macular degeneration, or visual acuity impairment.
Conclusion
The risk of incident CS impairment decreased by birth cohort, with the greatest reduction in the Baby Boom Generation. The difference in risk suggests that there are unknown modifiable risk factors that may help to further explain the etiology of CS impairment and provide potential pathways for prevention in the future.
Introduction
Differences in incidence of disease by generation of birth has been noted in a number of conditions associated with aging including cardiovascular disease, cognitive impairment, hearing loss, and age-related macular degeneration.1–5 These differences in incidence by generation or birth cohort effects (BCE), can be a result of changing exposures due to public health policy, changes in social norms and lifestyle, advances in prevention and treatment of disease, access to healthcare and care seeking behavior, or yet unknown influences. Understanding the differing disease incidence by birth cohort can provide information on the pathogenesis of the disease as well as help to design strategies and interventions to reduce the burden in the population and in future generations.
The current literature on the effect of birth cohort on vision is limited to age-related macular degeneration (AMD). In two cohort studies, the Beaver Dam Eye Study (BDES) and the Beaver Dam Offspring Study (BOSS), it was found that the incidence of AMD declined by birth cohort, with a 30% decrease in risk of early AMD per birth cohort in the BDES alone and a 60% decrease in risk of any AMD per generation in combined analysis of BDES and BOSS.5–6 This decline was independent of known risk factors for AMD. This means there is likely some unknown systematic difference between later generations’ exposures compared to earlier generations that account for differences in vision-related health. It is unknown if a similar pattern exists for other vision-related outcomes.
Contrast sensitivity (CS) is a measure of visual function that captures an aspect of vision not measured by the more commonly reported distance visual acuity (VA) and may better represent ability to identify real-world targets.7 CS describes how well a person can distinguish between light and dark, or how well they can see faint objects against a light background. This measure has been shown to affect tasks of daily living, safety, and autonomy.8–11 Incidence of CS impairment has been shown to be common among aging adults, even those with good VA, indicating it may reflect more subtle changes in visual function.12–14 CS impairment has been shown to have a number of potentially modifiable risk factors including smoking status, cardiovascular disease, and inflammation.12 It is not known if incidence of CS impairment differs by birth cohort. The current study aimed to determine if a birth cohort effect influences the risk of CS impairment incidence.
Materials and Methods
Study population
Data for the current investigation come from ten years of follow-up from two longitudinal cohort studies, the BDES and the BOSS. The BDES is a population-based study of adult residents (aged 43–84 in 1987–88) of Beaver Dam, Wisconsin. The BDES examinations began in 1988 with the first phase ending in 1990. Follow-up examinations occurred every 5 years.6 Participants of the population-based Epidemiology of Hearing Loss Study (EHLS), all of whom participated in the BDES, were asked permission to contact their adult children for enrollment in the BOSS.15 As such, recruitment into the BOSS required at least 1 parent to be a BDES participant. The BOSS began in 2005 with the first examinations ending in 2008 (baseline participant age 21–84) and follow-up examinations occurring every 5 years.16 The second examination phase of BDES and first examination phase of BOSS serve as baseline for this investigation of incident CS impairment in the 10 subsequent years of follow-up. Informed written consent was obtained from participants prior to each examination phase and all studies were approved by the University of Wisconsin Health Sciences Institutional Review Board and were conducted in accordance with the Declaration of Helsinki.
Contrast sensitivity assessment
Standardized protocols were followed in both studies and CS testing was conducted in the same manner in the BDES and BOSS. Monocular contrast sensitivity was measured in each eye with the Pelli-Robson Letter Sensitivity Charts viewed at a distance of one meter.17 Participants were tested wearing trial frames containing the appropriate distance correction as determined by Humphrey 530 Automatic Refractor (Humphrey, Inc., San Leandro, CA, USA) in the BDES and by Grand-Seiko autorefractor (WR-5001K; Grand Seiko, Hiroshima, Japan) in the BOSS, and refinement by subjective refraction when VA was worse than a Snellen value of 20/40. The contrast sensitivity charts consist of 16 letter triplets where the contrast in each successive triplet decreases by a factor of 0.15 log units. Participants were encouraged to progress as far as possible, making a best guess if unsure about a specific letter. The last triplet where the participant correctly identified at least two of three letters was used to assign a log CS score. A log CS score less than 1.55 was considered impaired and only those free of impairment at baseline were included in this investigation. Incident cases of CS impairment occurred when either eye was impaired at follow-up.
Generational classification
The birth-year of each participant was used to assign a birth cohort as follows: Greatest Generation - 1901 through 1924, Silent Generation –1925 through 1945, Baby Boom Generation –1946 through 1964, and Generation X –1965 through 1984.
Covariates
Additional information on potential confounders was collected. Blood pressure, height, weight, and waist circumference were measured following standard protocols. Hypertension was defined as a measured systolic blood pressure ≥ 140 mmHg, diastolic BP ≥90 mmHg, or physician diagnosis with current blood pressure medication. Body Mass Index was calculated (kg/m2) and classified as normal (< 25), overweight (25–29), or obese (30+). Retinal photographs were taken using a fundus camera (Zeiss FF4 in BDES or Canon Dgi-45NM in BOSS), lens images were taken using a slit-lamp camera (Topcon Sl-3 microscope and 5D Topcon slit-lamp camera in BDES or TopCon SL-D7 slit-lamp and TopCon DG-1 camera back in BOSS) and a retroillumination cataract screener (Neitz CT-R in BDES and Neitz CT-S in BOSS). Presence of any AMD was determined by fundus image grading by the University of Wisconsin Ocular Epidemiology Reading Center using the Wisconsin Age-Related Maculopathy Grading System, and presence of any cataract (cortical, nuclear sclerosis, or posterior subcapsular) by slit-lamp and retro-illumination lens image grading.18–20 Best corrected VA was measured monocularly using the Early Treatment Diabetic Retinopathy Study charts and protocol.21 Impaired VA was defined as an equivalent Snellen value of 20/40 or worse. Carotid artery ultrasound scans were used to measure intima-media thickness (IMT; mean of up to 12 wall thicknesses in the carotid arteries) and count of plaque in the carotid arteries (0 to 6 sites: common carotid, carotid bulb, and internal carotid, right and left sides).22 Measures of IMT in the BDES participants were collected as part of the EHLS examinations. Whole blood glycosylated hemoglobin A1C was measured in BDES using the Glyc-Affin GHb kit (Isolab, U.S.A.) and glycated hemoglobin A1C was measured in BOSS using an automated high-performance liquid chromatography method (Tosoh A1C G7 Glycohemoglobin Analyzer, Tosoh Medics). Diabetes was defined by an A1C ≥ 6.5% or a self-reported physician diagnosis. Inflammatory markers, interleukin 6 (IL-6) and high sensitivity C-reactive protein (hsCRP), were measured in stored serum samples by the University of Minnesota Advanced Research and Diagnostic Lab. The BDES participant inflammatory markers were measured as a part of the EHLS examination protocol. For both BDES and BOSS participants, IL-6 levels were measured by a quantitative sandwich enzyme technique using the ELISA QuantiKine High Sensitivity kit from R&D Systems (Minneapolis, MN). The IL-6 reference range was 0.45 to 9.96 pg/mL and the coefficient of variation (CV) was 11.7%. In BDES participants, hsCRP was measured using a latex particle-enhanced immunoturbidmetric method (Kamiya Biomedical Company) read on a Roche Hitachi 911 analyzer. BOSS hsCRP levels were measured using a latex particle-enhanced immunoturbidimetric assay from Roche Diagnostics (Indianapolis, IN) and read on a Roche Modular P800 analyzer. For both methods the reference range was 0 to 5 mg/L and the CV was 4.5%. Age, sex, socio-economic status (household income and education), smoking status (never, past, or current), alcohol consumption (in past year), and exercise (at least once a week long enough to work up a sweat) were assessed by in-person interview or a self-administered questionnaire following standard protocols.
Statistical analyses
Age-, sex-, and generation-specific risk of ten-year cumulative incidence and 95% confidence intervals were calculated using Kaplan-Meier survival analysis.23 A discrete-time hazard model with a complementary log-log transformation link and a binomial distribution was used to investigate the association between generation and cumulative incidence of CS impairment. This model is analogous to the continuous time hazard model in Cox proportional hazards regression.24–25 First, the effect of generation, adjusted for age and sex, was modelled with generation as an ordinal variable and then as an indicator variable with the Greatest Generation as a reference. A plot of estimated 10-year incidence of CS impairment by age and generation was generated from the age- and sex-adjusted indicator variable model. Potential confounders of the relationship between generation and contrast sensitivity impairment were explored using the same modelling technique. Confounders with results suggestive of an association with CS impairment (p<0.20) were used to build a final multivariable model to investigate BCEs independent of these confounders. As participants from the same family are likely to be more similar due to shared genetics and environmental exposures, additional analyses were conducted adjusting for familial clusters. Two sensitivity analyses were conducted removing those with known ocular conditions, AMD, cataract, and VA impairment, at baseline or at any examination. While there is sufficient age overlap between each generation and the successive one, age overlap across all generations is limited, with the youngest age group represented only in Generation X. Although multivariable modelling adjusts for factors with differing distributions among risk groups, we conducted an additional analysis limited to those aged 40 to 69 years at baseline in the Silent and Baby Boom Generations, where there was greatest sample overlap, to see if findings were consistent. All analyses were conducted using SAS version 9.4 (SAS Institute, Inc. Cary, NC).
Results
Retention of the group at-risk for CS impairment was high with 3185 (85%) of 3759 participants returning for follow-up visits. The distribution of age, sex and generation was similar between those included and those lost to follow-up (Table 1). Losses to follow-up for this investigation (N=572) occurred due to completing only a questionnaire at follow-up (N=185), death prior to examinations (N=134), and dropout (N=253). Among the 3185 participants at risk for incident CS impairment with follow-up, the mean age at baseline was 51.9 years (SD=9.9), and 1653 (51.9%) were women. The overall 10-year cumulative incidence of CS impairment was 40.1%, was higher among women than men, 41.7% and 38.8% respectively, and increased by age group overall and within generations (Table 2). Incidence was highest among the Greatest Generation, 96.0%, and lowest among Generation X, 13.2%.
Table 1:
Baseline characteristics for those at risk of contrast sensitivity impairment by inclusion status.
| Included | Lost to Follow-up | |
|---|---|---|
| Total N | 3185 | 572 |
| Mean baseline age, years (SD) | 51.9 (9.9) | 52.1 (11.5) |
| Sex, n (%) | ||
| Female | 1653 (51.9) | 301 (52.6) |
| Male | 1532 (48.1) | 271 (47.4) |
| Generation, n (%) | ||
| Greatest (1901–1924) | 141 (4.4) | 54 (9.4) |
| Silent (1925–1945) | 1260 (39.6) | 187 (32.7) |
| Baby Boom (1946–1964) | 1323 (41.5) | 224 (39.2) |
| Generation X (1965–1984) | 461 (14.5) | 107 (18.7) |
Table 2:
10-year cumulative incidence of contrast sensitivity impairment by generation, sex, and age group (years)a
| Overall | Greatest (1901–1924) | Silent (1925–1945) | Baby Boom (1946–1964) | Generation X (1965–1984) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| At Risk | Incidence (95% CI) | At Risk | Incidence (95% CI) | At Risk | Incidence (95% CI) | At Risk | Incidence (95% CI) | At Risk | Incidence (95% CI) | |
| Overall | 3185 | 40.1 (38.4, 41.9) | 141 | 96.0 (90.9, 98.7) | 1260 | 60.6 (57.8, 63.4) | 1323 | 22.8 (20.6, 25.3) | 461 | 13.2 (10.4, 16.8) |
| Sex | ||||||||||
| Men | 1532 | 38.8 (36.3, 41.4) | 63 | 93.1 (83.3, 98.2) | 626 | 57.7 (53.6, 61.7) | 623 | 22.9 (19.7, 26.5) | 220 | 10.8 (7.3, 15.8) |
| Women | 1653 | 41.7 (39.3, 44.2) | 78 | 98.2 (91.8, 99.9) | 634 | 63.4 (59.5, 67.3) | 700 | 22.8 (19.7, 26.2) | 241 | 15.4 (11.4, 20.8) |
| Age group | ||||||||||
| 22–39 | 357 | 13.0 (9.8, 17.0) | 357 | 13.0 (9.8, 17.0) | ||||||
| 40–49 | 848 | 19.0 (16.4, 22.0) | 68 | 41.3 (30.4, 54.3) | 676 | 17.3 (14.6, 20.5) | 104 | 14.1 (8.6, 22.8) | ||
| 50–59 | 1319 | 42.1 (39.3, 44.9) | 699 | 53.6 (49.8, 57.5) | 620 | 28.1 (24.6, 31.9) | ||||
| 60–69 | 522 | 72.4 (68.4, 76.4) | 27 | 88.9 (70.8, 98.0) | 468 | 73.5 (69.3, 77.6) | 27 | 38.4 (22.7, 59.9) | ||
| 70–89 | 139 | 94.0 (88.4, 97.5) | 114 | 97.5 (92.3, 99.5) | 25 | 79.4 (60.8, 93.1) | ||||
Displayed as incidence per 100 population with 95% confidence intervals (CI) estimated using Kaplan-Meier survival analysis
In age- and sex-adjusted models, the risk of incident CS impairment decreased by 39% per generation when analyzed as an ordered factor (HR=0.61, 95% CI=0.53, 0.70). In models comparing each generation to the Greatest Generation, the Baby Boom Generation had a 58% reduction in risk while Generation X had a 44% reduction. The Silent Generation had a 12% reduction in risk, though this result was not statistically significant (Table 3). Figure 1 displays the 10-year incidence of CS impairment by age and generation adjusted for sex.
Table 3:
Age- and sex-adjusted risk of 10-year incidence of CS impairment by generation; Hazard Ratio (HR) and 95% Confidence Interval (CI)
| Generation | HR (95% CI) |
|---|---|
| Greatest | Reference |
| Silent | 0.88 (0.68, 1.13) |
| Baby Boom | 0.42 (0.31, 0.58) |
| Generation X | 0.56 (0.34, 0.91) |
Figure 1:
Estimated 10-year incidence of impaired contrast sensitivity by age and generation, adjusted for sex
Several factors previously associated with incident CS impairment were significantly associated with incidence in this study. In models adjusted for age and sex, current smoking (HR=1.37, 95%CI=1.16, 1.61), larger waist circumference (per 5 cm: HR=1.02, 95%CI=1.00, 1.04), hypertension (HR=1.14, 95%CI=1.01, 1.29), higher arterial plaque count (per site; HR=1.16, 95%CI=1.11, 1.22), diabetes (HR=1.66, 95%CI=1.34, 2.06), and higher levels of inflammation (IL6 T3 versus T1: HR=1.27, 95%CI=1.09, 1.47; CRP >3 mg/L versus <1 mg/L: HR=1.44, 95%CI=1.23, 1.69), were associated with increased risk, while higher income ($60K+ versus <$30K: HR=0.65, 95%CI=0.55, 0.76), college education or greater (HR=0.85, 95%CI=0.74, 0.97), and exercising at least once per week (HR=0.85, 95%CI=0.76, 0.96) were associated with decreased risk. (Table 4) Additionally, other eye conditions, cataract (HR=1.44, 95%CI=1.21, 1.70) and AMD (HR=1.23, 95%CI=1.02, 1.47) were associated with increased risk of incidence.
Table 4:
Potential baseline confounders of relationship between generation and CS impairment incidence: Age- and sex- adjusted Hazard Ratios (HR) and 95% Confidence Intervals (CI)
| Risk factor | HR (95% CI) | p-value |
|---|---|---|
| Income | ||
| $0-$29K | Reference | |
| $30-$59K | 0.81 (0.70, 0.94) | 0.006 |
| $60K+ | 0.65 (0.55, 0.76) | <0.001 |
| Education | ||
| <16 years | Reference | |
| >=16 years | 0.85 (0.74, 0.97) | 0.02 |
| Smoking status | ||
| Never | Reference | |
| Past | 0.98 (0.86, 1.12) | 0.80 |
| Current | 1.37 (1.16, 1.61) | 0.0002 |
| Alcohol consumption (g/week) | ||
| 0 | 1.15 (0.96, 1.37) | 0.13 |
| >0–14 | Reference | |
| 15–74 | 0.93 (0.79, 1.09) | 0.36 |
| 75–140 | 1.04 (0.86, 1.26) | 0.67 |
| ≥141 | 1.03 (0.84, 1.25) | 0.78 |
| History of heavy drinking | 1.08 (0.92, 1.26) | 0.35 |
| Exercise at least 1/week | 0.85 (0.76, 0.96) | 0.006 |
| Waist | ||
| Per 5 cm | 1.02 (1.00, 1.04) | 0.02 |
| Hypertension | 1.14 (1.01, 1.29) | 0.03 |
| Plaque Count | ||
| Per site | 1.16 (1.11, 1.22) | <0.0001 |
| Diabetes | 1.66 (1.34, 2.06) | <0.0001 |
| Interleukin 6 (pg/mL) | ||
| Tertile 1 (<1.2) | Reference | |
| Tertile 2 (1.2-<2.4) | 1.02 (0.88, 1.19) | 0.78 |
| Tertile 3 (≥2.4) | 1.27 (1.09, 1.47) | 0.002 |
| C-reactive Protein (mg/L) | ||
| <1 | Reference | |
| 1–3 | 1.33 (1.15, 1.55) | 0.0002 |
| >3 | 1.44 (1.23, 1.69) | <0.0001 |
| Cataract | 1.44 (1.21, 1.70) | <0.0001 |
| Age-related macular degeneration | 1.23 (1.02, 1.47) | 0.03 |
| Visual acuity impairment | 1.28 (0.85, 1.93) | 0.23 |
In a model adjusting for multiple confounders, the reduced risk observed in the Baby Boom Generation and Generation X in the age- and sex-adjusted models remained, 57% (HR=0.43, 95%CI=0.28, 0.66) and 47% (HR=0.53, 95%CI=0.29, 0.97) respectively. (Table 5) No significant reduction of risk was found in the Silent Generation (HR=0.91, 95%CI=0.65, 1.27). Adjustment for family clusters was also conducted to test whether heritability of vision could have affected results. This additional adjustment had no impact on these results. Two additional sensitivity analyses were conducted, removing anyone with cataract, AMD, or VA impairment at baseline (N=531) and removing anyone with any of these conditions at any examination phase (N=1202). The results of the analysis removing those with the conditions at baseline showed strengthened BCEs, with a 66% (HR=0.34, 95%CI=0.18, 0.62) and 61% (HR=0.39, 95%CI=0.18, 0.86) reduction in risk for the Baby Boom Generation and Generation X, respectively (Table 5). Removing anyone with these conditions at any examination phase had similar results, though the effect for Generation X was no longer statistically significant (HR=0.42, 95%CI=0.13, 1.28).
Table 5:
Multivariable models of risk of 10-year incidence of CS impairment by generation; Hazard Ratios (HR) and 95% Confidence Intervals (CI)a
| Generation | All Participants | Excluding participants with cataract, AMD, or VA impairment at: | |
|---|---|---|---|
| Baseline | Any examination | ||
| HR (95%CI) | HR (95%CI) | HR (95%CI) | |
| Greatest | Reference | Reference | Reference |
| Silent | 0.91 (0.65, 1.27) | 0.76 (0.44, 1.31) | 0.90 (0.37, 2.17) |
| Baby Boom | 0.43 (0.28, 0.66) | 0.34 (0.18, 0.62) | 0.37 (0.14, 0.97) |
| Generation X | 0.53 (0.29, 0.97) | 0.39 (0.18, 0.86) | 0.42 (0.13, 1.28) |
Adjusted for age, sex, income, smoking, exercise, hypertension, diabetes, waist circumference, plaque count, interleukin 6, C-reactive protein, cataract, age-related macular degeneration (AMD), visual acuity (VA) impairment
An additional analysis was conducted among only those between 40 to 69 years of age in the Silent and Baby Boom generations to ensure that the extremes were not driving the observed relationship. As data from all ages in each generation were not available, this approach investigates the trend only in ages with the most generational overlap. In this analysis the later generation, specifically the Baby Boom Generation, had a significantly lower risk of incident CS impairment compared to the Silent Generation (HR=0.45, 95%CI=0.38, 0.54).
Discussion
Incidence of CS impairment was common in this study, 40.1% overall, though risk was greatly reduced in later generations, specifically 57% lower for the Baby Boom Generation and 47% lower for Generation X, compared to the Greatest Generation. To our knowledge this is the first study of the birth cohort effect on the incidence of CS impairment. Trends for reduction in prevalence and incidence by generation have been observed for other age-related conditions, such as cardiovascular disease, dementia, age-related macular degeneration, and hearing impairment.1–5 Understanding these trends is important not only for understanding the etiology and progression of disease, but also to aid in estimation of future burden of disease and to develop strategies for prevention. Impaired contrast sensitivity has been shown to impact tasks of daily living, safety, and autonomy and reduction of incidence in future generations would have an important public health impact.8–11
Rapidly declining risk of incidence is evidence that a disorder is at least partially preventable as genetic changes are slow. Adjusting for familial cluster in the present study also had no effect suggesting that non-genetic factors are responsible for the observed declines in incidence. Several known risk factors were included in this study as confounders. The fact the BCE remained significant suggests there are unidentified modifiable risk factors for CS impairment which may help to explain the observed decline in incidence. One potential factor could be nutrition or food scarcity. The Greatest Generation lived through the Great Depression and as a result may have experienced prolonged periods of malnutrition which could have had a negative impact on long-term ocular health. Two recent trials found enrichment of diet with carotenoids improved CS function.26–27 Long-term effects of nutrition on CS are unknown but could explain part of the BCE observed in the present study. Exposure to infectious diseases in childhood has also changed by generation. With increasing rates of use of antibiotics and vaccines, rates of childhood infectious diseases have decreased greatly over time.28 The neural mechanisms which regulate CS mature during childhood and although the long-term effects of childhood infections on vision are unknown, reduction in risk due to vaccinations could potentially explain part of the BCE in the present study.29 Additionally, possible non-specific effects of vaccinations have been postulated, including long-term upregulation of innate immune cells, which could contribute to better overall health and potentially to longer retention of CS function.30–31 Another potential factor for the observed BCE could be exposure to neurotoxins. A recent study in the BOSS population found exposure to cadmium (Cd) but not lead (Pb) was associated with increased incidence of CS impairment, though levels of Pb in the study were generally quite low.12 The older generations represented by the participants of the BDES may have greater levels of Cd and Pb exposure and bioaccumulation which may contribute to our findings, though these measures were not available in the current study. Rates of these and other unmeasured environmental factors may have changed over time and differences in exposure by generation could potentially have influenced our results. Future studies should investigate if differences in nutrition, childhood diseases, vaccination, and exposure to environmental toxins by generation help to explain part of the observed birth cohort effect on incident CS impairment.
Another potential explanation for the decreased risk of CS impairment by generation could be a decrease in comorbid eye conditions. The noted decreases in incidence of AMD, increases in access to care for cataracts, including earlier extractions, and low prevalence of un- or under-corrected visual impairments could potentially help explain the observed differences.5–6, 32–34 However, in this study the effect of generation remained when excluding anyone with cataract, AMD, or VA impairment at baseline, and was similar when excluding anyone with these conditions at any time point, though the association of reduced risk for Generation X was no longer statistically significant. This change in significance may be a function of power as sample size was greatly reduced due to these exclusions and the point estimate remained relatively unchanged. The BCE remaining when removing those with other eye conditions implies that while incidence of CS impairment is decreasing, it may not be solely due to decreases in concomitant eye disease. However, if subclinical manifestations of AMD or other ocular pathologies affect performance on CS tests and these are decreasing by generation then the observed effect could be due to the decreases in incidence of other ocular pathologies. The present study cannot determine with certainty whether the reduction of CS impairment by generation is completely independent of the reduction seen in AMD incidence or other ocular pathologies.
Although both the Baby Boom Generation and Generation X had a reduced risk of incidence of CS impairment, the protective effect was greater in the Baby Boom Generation (Figure 1). While the difference between these two groups was not statistically significant, the reason for the observed difference could be important. If the generational trend is slowing, it is important to determine the factors that explain this trend. Future studies should monitor generational trends to investigate a potential plateau in the benefit of later generational status in risk of CS impairment.
The present study has several strengths including long term follow-up of a well-defined population-based cohort, with standardized measurement of CS and numerous potential confounders. The large sample size and wide age range provides the power necessary to detect differences in risk between the generations. The familial relationships between the BDES and BOSS, specifically the genetic homogeneity of the multigenerational design, reduce the likelihood the observed generational differences in risk are due to heritable sources and are more likely due to modifiable risk factors. A limitation of the present study is while we were able to control for AMD, cataract, and VA impairment status, other unmeasured ocular conditions could have had an impact on results. The racial and ethnic homogeneity of the BDES and BOSS is another limitation as nearly all participants were of non-Hispanic white background. As a result, these findings may not be generalizable to other racial and ethnic groups and generational differences in risk of CS impairment should be investigated in diverse populations to see if this trend exists among other backgrounds. In order to be included, participants needed to be free of CS impairment at baseline. As a result, those in the older generations may have been systematically different from their peers who did not make it to the baseline examinations free of impairment. However, they should be representative of the at-risk population in this age range. Finally, the birth cohorts observed have a limited amount of age-specific overlap, however the analytic approach used adjusts for age and allows for differing age distributions in the comparison groups. In order to verify these results, however, we conducted an analysis among the Silent and Baby Boom Generations with the greatest age overlap (40–69 years). The results of this analysis were consistent as the risk remained lower for the Baby Boom generation. While previous studies have used similar methods in investigating birth-cohort effects,4–6 care should be taken with interpretation of results as comparison of these generations assumes that the trends continue across the age continuum.
In conclusion, the risk of CS impairment was reduced in later generations compared to the Greatest Generation. Part of this reduction in risk is likely due to changes in modifiable risk factors, although these factors could not be specifically identified in the present study. Future study should further investigate modifiable factors which may explain the reduction in risk observed, to monitor trends to see if the protective effect of generation is in decline, and to investigate if there are similar protection and trends in other populations.
Funding details
This work was supported by the National Institute on Aging and the National Eye Institute under Grant R01AG021917 (Dr. Cruickshanks); the National Institute on Aging under Grant R37AG011099 (Dr. Cruickshanks); the National Eye Institute under Grant U10EY06594 (Drs. B.E.K. Klein and R. Klein); Research to Prevent Blindness under an unrestricted grant to the Department of Ophthalmology and Visual Sciences at the University of Wisconsin School of Medicine and Public Health.
Footnotes
Disclosure of Interest
The authors report no conflict of interest. The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institute on Aging, the National Eye Institute, the National Institutes of Health, or Research to Prevent Blindness.
The authors warrant that this article is original, is not simultaneously under consideration by another journal, and has not been previously published. This manuscript was previously submitted to JAMA Ophthalmology, and despite a favorable review, was not accepted. The authors used the feedback from this review to clarify the sample selection in the current version of the manuscript. We conducted additional analyses to address concerns of overlap of age ranges among the birth cohorts.
References
- 1.Ford ES, Roger VL, Dunlay SM, Go AS, Rosamond WD. Challenges of Ascertaining National Trends in the Incidence of Coronary Heart Disease in the United States. J Am Heart Assoc. 2014;3:e001097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Aparicio HJ, Himali JJ, Satizabal CL et al. Temporal Trends in Ischemic Stroke Incidence in Younger Adults in the Framingham Study. Stroke. 2019; May 14:STROKEAHA119025171 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Satizabal CL, Beiser AS, Chouraki V, Chêne G, Dufouil C, Seshadri S. Incidence of Dementia over Three Decades in the Framingham Heart Study. N Engl J Med. 2016;374:523–532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zhan W, Cruickshanks KJ, Klein BEK, et al. Generational Differences in the Prevalence of Hearing Impairment in Older Adults. Am J Epidemiol. 2010;171:260–266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cruickshanks KJ, Nondahl DM, Johnson LJ, et al. Generational Differences in the 5-Year Incidence of Age-Related Macular Degeneration. JAMA Ophthalmol. 2017;135(12):1417–1423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Klein R, Knudtson MD, Lee KE, Gangnon RE, Klein BEK. Age-Period-Cohort Effect on the Incidence of Age-Related Macular Degeneration: The Beaver Dam Eye Study. Ophthalmology. 2008;115(9):1460–1467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Owsley C, Sloane ME. Contrast sensitivity, acuity, and the perception of ‘real-world’ targets. Br J Ophthalmol. 1987;71(10):791–796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rubin GS, Bandeen-Roche K, Huang GH, et al. The association of multiple visual impairments with self-reported visual disability: SEE project. Invest Ophthalmol Vis Sci. 2001;42(1)64–72. [PubMed] [Google Scholar]
- 9.West SK, Rubin GS, Broman AT, Muñoz B, Bandeen-Roche K, Turano K. How does visual impairment affect performance on tasks of everyday life? The SEE project. Salisbury Eye Evaluation. Arch Ophthalmol. 2002;120(6):774–780. [DOI] [PubMed] [Google Scholar]
- 10.Klein BEK, Klein R, Lee KE, Cruickshanks KJ. Performance-based and self-assessed measures of visual function as related to history of falls, hip fractures, and measured gait time. The Beaver Dam Eye Study. Ophthalmology. 1998;105(1):160–164. [DOI] [PubMed] [Google Scholar]
- 11.de Boer MR, Pluijm SMF, Lips P, et al. Different aspects of visual impairment as risk factors for falls and fractures in older men and women. J Bone Miner Res. 2004;19(9):1539–1547. [DOI] [PubMed] [Google Scholar]
- 12.Paulsen AJ, Schubert CR, Johnson LJ, et al. Association of Cadmium and Lead Exposure With the Incidence of Contrast Sensitivity Impairment Among Middle-aged Adults. JAMA Ophthalmol. 2018;136(12):1342–1350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pelli DG, Bex P. Measuring contrast sensitivity. Vision Res. 2013;90:10–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rubin GS, West SK, Muñoz B. et al. A comprehensive assessment of visual impairment in a population of older Americans. The SEE Study. Salisbury Eye Evaluation Project. Invest Ophthalmol Vis Sci. 1997;38(3):557–568. [PubMed] [Google Scholar]
- 15.Zhan W, Cruickshanks KJ, Klein BEK, et al. Modifiable determinants of hearing impairment in adults. Prev Med. 2011;53(4–5):338–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nash SD, Cruickshanks KJ, Klein R, et al. The prevalence of hearing impairment and associated risk factors: the Beaver Dam Offspring Study. Arch Otolaryngol Head Neck Surg. 2011;137(5):432–439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clin Vis Sci. 1988;2(3):187–199. [Google Scholar]
- 18.Klein R, Meuer SM, Moss SE, Klein BE, Neider MW, Reinke J. Detection of age-related macular degeneration using a nonmydriatic digital camera and a standard film fundus camera. Arch Ophthalmol. 2004;122(11):1642–1646. [DOI] [PubMed] [Google Scholar]
- 19.Klein R, Cruickshanks KJ, Nash SD, et al. The Prevalence of Age-Related Macular Degeneration and Associated Risk Factors: The Beaver Dam Offspring Study. Arch Ophthalmol. 2010;128(6):750–758. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Klein BEK, Klein R, Linton KL, Magli YL, Neider MW. Assessment of cataracts from photographs in the Beaver Dam Eye Study. Ophthalmology. 1990;97(11):1428–1433. [DOI] [PubMed] [Google Scholar]
- 21.Ferris FL, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol. 1982;94(1):91–96. [PubMed] [Google Scholar]
- 22.Zhong W, Cruickshanks KJ, Huang GH, et al. Carotid atherosclerosis and cognitive function in midlife: the Beaver Dam Offspring Study. Atherosclerosis. 2011;219(1):330–333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kaplan EL, Meier P. Nonparametric Estimation from Incomplete Observations. Journal of the American Statistical Association. 1958;53(282):457–481. [Google Scholar]
- 24.Singer JD, Willet JB. Applied longitudinal data analysis: modeling change and event occurrence. Oxford, New York: Oxford University Press; 2003. [Google Scholar]
- 25.Alisson PD. Survival analysis using the SAS system: a practical guide. Cary, NC: SAS Institute; 1995. [Google Scholar]
- 26.Nolan JM, Power R, Stringham J, et al. Enrichment of Macular Pigment Enhances Contrast Sensitivity in Subjects Free of Retinal Disease: Central Retinal Enrichment Supplementation Trials – Report 1. Invest Ophthalmol Vis Sci. 2016;57(7):3429–3439. [DOI] [PubMed] [Google Scholar]
- 27.Stringham JM, O’Brien KJ, Stringham NT. Contrast Sensitivity and Lateral Inhibition are Enhanced with Macular Carotenoid Supplementation. Invest Ophthalmol Vis Sci. 2017;58(4):2291–2295. [DOI] [PubMed] [Google Scholar]
- 28.Hamborsky J, Kroger A, Wolfe S, eds. Epidemiology and Prevention of Vaccine-Preventable Diseases. 13th edition. Centers for Disease Control and Prevention. Washington D.C. Public Health Foundation; 2015. [Google Scholar]
- 29.Pei F, Baldassi S, Tsai JJ, Gerhard HE, Norcia AM. Development of contrast normalization mechanisms during childhood and adolescence. Vision Res. 2017;133:12–20. [DOI] [PubMed] [Google Scholar]
- 30.Benn CS, Netea MG, Selin LK, Aaby P. A small jab – a big effect: nonspecific immunomodulation by vaccines. Trends Immunol. 2013;34(9):431–439. [DOI] [PubMed] [Google Scholar]
- 31.Mina MJ. Measles, immune suppression and vaccination: direct and indirect nonspecific vaccine benefits. J Infect. 2017;74(Suppl 1):S10–S17. [DOI] [PubMed] [Google Scholar]
- 32.Gollogly HE, Hodge DO, St. Sauver JL, Erie JC. Increasing incidence of cataract surgery: Population-based study. J Cataract Refract Surg. 2013;39(9):1383–1389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Klein BEK, Howard KP, Lee KE, Klein R. Changing Incidence of Lens Extraction Over Twenty Years: the Beaver Dam Eye Study. Ophthalmology. 2014;121(1):5–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Qui M, Wang SY, Singh K, Lin SC. Racial Disparities in Uncorrected and Undercorrected Refractive Error in the United States. Invest Ophthalmol Vis Sci. 2014;55:6996–7005. [DOI] [PMC free article] [PubMed] [Google Scholar]