The Rotterdam Study, a prospective, population-based cohort study, has been approved by the Medical Ethics Committee of the Erasmus MC (registration number MEC 02.1015) and by the Dutch Ministry of Health, Welfare and Sport (Population Screening Act WBO, license number 1,071,272-159,521-PG). The Rotterdam Study Personal Registration Data collection is filed with the Erasmus MC Data Protection Officer under registration number EMC1712001. The Rotterdam Study has been entered into the Dutch Trial Register (NTR; https://onderzoekmetmensen.nl ) and into the WHO International Clinical Trials Registry Platform (ICTRP https://www.who.int/clinical-trials-registry-platform , search portal https://trialsearch.who.int/ ) under shared catalogue number NL6645/NTR6831. The Rotterdam Study project persistent identifier is https://ror.org/02ac58f22 . All participants provided written informed consent to participate in the study and to have their information obtained from treating physicians. The study was conducted in accordance with the tenets of the Declaration of Helsinki.

This study utilizes data from the Rotterdam Study, a prospective population-based cohort study, designed to assess determinants of age-related diseases. The Rotterdam Study consists of four cohorts (RS-I, RS-II, RS-III, and RS-IV), of which data of the first 3 were included in this study. Data collection started in 1991 with repeated examinations every 3-4 years. Details of the Rotterdam Study have been published previously.

All participants underwent visual field testing. Visual field loss had to be reproducible on 2 screening visual field tests (HFA II 740 or FDT 710, Carl Zeiss, Oberkochen, Germany) and a subsequent HFA 24–2 full threshold or SITA standard test. If no other cause could be identified during the ophthalmic assessment, and no homonymous defects or recognizable patterns like rim artefacts were observed, the defect was considered glaucomatous visual field loss. Further details were published previously. We defined OAG as glaucomatous visual field loss in at least one eye, independent of IOP.

IOP was measured using Goldmann applanation tonometry (Haag-Streit AG, Bern, Switzerland). Three measurements were taken from each eye, the median value of which was recorded. For iOAG cases, we used IOP measurements of the affected eye. If both eyes were affected or unaffected, a random eye was selected. Untreated IOP levels were not recorded for participants receiving IOP-lowering medication. Therefore, if a participant used IOP-lowering medication, we divided the measured IOP by 0.7 to estimate the untreated IOP, as an average reduction of 30% has been reported in meta-analyses. ^, In participants with a history of IOP lowering surgery, pre-treatment IOP was assumed to be at least 30 mm Hg.

Additionally, optical coherence tomography (OCT) images were acquired using different Topcon (Tokyo, Japan) devices over time: 3D OCT-1000 (2007–2008), 3D-1000 Mark II (2008–2011), 3D-2000 FA Plus (2011–2022), and DRI OCT Triton Plus (2022–2025). RNFL and GCL thicknesses were determined using an in-house artificial intelligence deep learning model trained on OCT scans (Supplemental Methods A). Axial length was measured with the Lenstar LS-900 (Haag-Streit AG, Bern, Switzerland). Genetic information was available for all participants (Supplemental Methods B) and was used to calculate a genetic risk score (GRS) for OAG; individuals with higher GRS values were considered at increased (i.e., ‘high’) risk of OAG.

Daily dietary intake was assessed at baseline using validated food frequency questionnaires (FFQs). ^,,,, The total daily intake of a vitamin (mg/d) was determined by summing the vitamin content of all consumed foods and drinks. This was calculated by multiplying the vitamin concentration (mg/g) in each food by the corresponding daily intake in grams (g). The vitamin concentration was derived from a Dutch government database. Alcohol intake (g/d) was obtained by multiplying the average amount of ethanol in one drink of an alcoholic beverage by the number of drinks. We calculated the overall diet quality as a score ranging from 0 to 14 according to the Dutch nutritional guidelines. Participants with extreme caloric intakes (<500 kcal/d or >5000 kcal/d) were excluded. The residual method was used to account for the fact that nutrient intake is correlated with total energy intake, differences in metabolism, body size, and more. Energy-adjusted intake was expressed on the original unit scale by adding the population mean intake to the residuals. These values should be interpreted as energy-adjusted intake values rather than observed crude intake. Because the residual method reflects intake relative to that expected for total energy intake, negative values may occur and indicate intake lower than expected based on total energy intake, rather than an observed negative intake. From this point onward, ‘vitamin intake’ will refer to energy-adjusted vitamin intake, unless otherwise specified. Crude and energy-adjusted intakes are depicted in Supplemental Table S1. Data was available for thiamine (vitamin B1), riboflavin (vitamin B2), niacin (vitamin B3), pyridoxine (vitamin B6), and cobalamin (vitamin B12). Total fasting homocysteine levels were measured using isotope-dilution liquid chromatography–tandem mass spectrometry (LC–MS/MS; Waters Acquity UPLC Quattro Premier XE).

Sex was defined as a biological variable (male/female) as recorded at baseline. Weight and height were measured at the research center. BMI was calculated by dividing the participants weight in kilograms by their height in meters squared. Blood pressure was measured on the right brachial artery, using the mean of 2 consecutive seated measurements. Hypertension was defined as a resting blood pressure exceeding 140/90 mm Hg or usage of blood pressure-lowering medication. Diabetic status was categorized into healthy, pre-diabetic and diabetic, as described previously. ^, Total serum cholesterol was determined by an automated enzymatic procedure. Other covariates were acquired through questionnaires. Education level was categorized into primary, lower, intermediate or higher education. Smoking status was categorized into never smoked, former smoker or current smoker. Physical activity was recalculated into a z-standardized score based on metabolic equivalent of task (MET)-hours per week. ^,

Differences in baseline characteristics between participants with and without iOAG were assessed using independent samples t -tests and chi-square tests. Multivariable logistic regression analysis was used to estimate odds ratios (ORs) with corresponding 95% confidence intervals (CIs) for iOAG, using energy-adjusted vitamin intake (residual method) with additional adjustment for total energy intake. All analyses were further adjusted for age, sex, diet quality, and follow-up time (model 1).Follow-up time was calculated from baseline until the last visit with reliable ophthalmic examination or the first visit with iOAG. Model 2 additionally adjusted for lifestyle factors (i.e., educational level, smoking status, alcohol intake, and physical activity). Model 3 additionally adjusted for comorbidities (i.e., hypertensive status, diabetic status, and serum cholesterol). Model 4 included adjustment for IOP. Moreover, we stratified the analyses on the GRS of OAG. We modelled vitamin intake in quintiles with the first quintile (Q1) serving as the reference category to test for linear trends, with the median value for each category as continuous variables in separate logistic regression models. The dose–response relations between B-vitamins intake and predicted iOAG probability were examined using generalized additive modeling. Multivariable linear regression analysis was used to study the association between B-vitamins and continuous OAG-associated parameters (at the incident follow-up visit), including IOP, RNFL and GCL thickness. In the layer thickness analyses, we additionally adjusted for OCT acquisition device and axial length. Given the influence of age, ^, sex, ^, and (potentially) BMI ^,, on both the exposure and outcome, we also performed propensity score matching (Supplemental Methods C), after which the analyses were repeated. The covariates that were matched for were excluded as covariates. Lastly, to assess potential reverse causality, we analyzed the association between B-vitamins and iOAG in cumulative follow-up intervals. Statistical analyses were performed using SPSS v28.0.1.0 (SPSS Inc.) and R v3.6.1 (R Inc.), with packages DescTools, mgcv, ggplot2, dplyr and ggforestplot. A 2-sided p -value <.05 was considered statistically significant.

Each milligram higher niacin intake per day was associated with a 6% decrease in OAG risk ( Table 2 , model 1, OR [95%CI]: 0.94 [0.90 0.98]). Participants with the highest niacin intake (Q5: mean 23.27 mg/d, crude intake) had the largest risk reduction ( Figure 1 A, model 1, OR [95%CI]: 0.43 [0.21 0.86]) compared to participants with the lowest intake (Q1: mean 9.98 mg/d, crude intake)(p-trend=0.02). The dose-response relation between niacin intake and iOAG showed a linear dose-response relation after a small initial increase of iOAG risk ( Figure 2 A; crude intake, Supplemental Figure S1A; energy-adjusted intake). We observed that niacin intake was specifically associated with a lower risk of iOAG in participants with the highest GRS ( Figure 3 C, Model 1, OR [95%CI]: 0.90 [0.83 0.97] per mg/d). Although the association between niacin intake and IOP itself was borderline significant (Supplemental Table S2, model 1, Beta [95%CI]: −0.02 [−0.05–0.00] mm Hg per mg/d), there was a significant trend towards lower IOP with higher niacin intake (p-trend=0.04). We did not observe clear associations between niacin intake and RNFL (Supplemental Table S3) or GCL thickness (Supplemental Table S4).

Table 2

Multivariable-Adjusted Odds Ratios With Corresponding 95% Confidence Intervals for Incident Open-Angle Glaucoma Per Unit Increase in Vitamin Intake.

				Quintile vitamin intake
Vitamin	Model	OR (95% CI) a	P- value	Q1	Q2	Q3	Q4	Q5	p – trend b
Thiamine (mg/d)	1	0.99 (0.50-1.80)	.97	Ref.	0.92 (0.57-1.50)	0.76 (0.45-1.29)	0.82 (0.47-1.42)	1.01 (0.55-1.88)	.97
	2	0.98 (0.53-1.82)	.95	Ref.	0.94 (0.57-1.55)	0.75 (0.44-1.29)	0.82 (0.46-1.44)	1.00 (0.53-1.87)	.93
	3	1.01 (0.55-1.84)	.98	Ref.	0.93 (0.57-1.51)	0.74 (0.44-1.26)	0.82 (0.47-1.42)	1.03 (0.55-1.91)	>.99
	4	1.07 (0.52-2.22)	.85	Ref.	1.07 (0.62-1.85)	0.88 (0.48-1.60)	1.01 (0.53-1.90)	1.08 (0.52-2.24)	.92
Riboflavin (mg/d)	1	0.79 (0.55-1.14)	.21	Ref.	0.54 (0.32-0.91) *	0.85 (0.53-1.36)	0.76 (0.47-1.25)	0.54 (0.32-0.93) *	.10
	2	0.80 (0.55-1.16)	.23	Ref.	0.49 (0.29-0.85) *	0.83 (0.51-1.34)	0.77 (0.47-1.26)	0.53 (0.30-0.91) *	.10
	3	0.80 (0.55-1.15)	.22	Ref.	0.53 (0.31-0.91) *	0.85 (0.53-1.36)	0.74 (0.45-1.21)	0.55 (0.32-0.94) *	.10
	4	0.78 (0.50-1.23)	.28	Ref.	0.42 (0.23-0.79) *	0.79 (0.46-1.35)	0.76 (0.44-1.33)	0.52 (0.27-0.98) *	.20
Niacin (mg/d)	1	0.94 (0.90-0.98)	.005 *	Ref.	0.85 (0.55-1.32)	0.66 (0.37-1.09)	0.74 (0.43-1.27)	0.43 (0.21-0.86) *	.02 *
	2	0.95 (0.91-0.99)	.02 *	Ref.	0.85 (0.54-1.34)	0.70 (0.42-1.17)	0.85 (0.49-1.48)	0.49 (0.24-0.99) *	.07
	3	0.94 (0.90-0.98)	.007 *	Ref.	0.85 (0.55-1.32)	0.64 (0.38-1.06)	0.75 (0.43-1.29)	0.44 (0.21-0.88) *	.02 *
	4	0.95 (0.90-1.00)	.04 *	Ref.	0.76 (0.45-1.28)	0.70 (0.39-1.23)	0.82 (0.43-1.56)	0.45 (0.20-1.00)	.08
Pyridoxine (mg/d)	1	0.60 (0.35-1.03)	.06	Ref.	0.89 (0.54-1.45)	0.98 (0.60-1.60)	0.88 (0.52-1.47)	0.68 (0.38-1.21)	.17
	2	0.63 (0.36-1.09)	.10	Ref.	0.95 (0.58-1.57)	0.95 (0.57-1.58)	0.91 (0.54-1.55)	0.73 (0.41-1.32)	.31
	3	0.62 (0.36-1.06)	.08	Ref.	0.87 (0.53-1.42)	0.99 (0.61-1.62)	0.87 (0.53-1.49)	0.70 (0.39-1.25)	.21
	4	0.67 (0.36-1.24)	.21	Ref.	0.89 (0.51-1.56)	1.03 (0.58-1.84)	0.94 (0.52-1.71)	0.88 (0.46-1.71)	.67
Cobalamin (µg/d)	1	0.90 (0.83-0.97)	.005 *	Ref.	0.70 (0.44-1.12)	0.37 (0.20-0.67) *	0.21 (0.10-0.44) *	0.25 (0.12-0.52) *	<.001 *
	2	0.91 (0.85-0.98)	.02 *	Ref.	0.72 (0.45-1.17)	0.38 (0.21-0.71) *	0.24 (0.12-0.52) *	0.30 (0.14-0.63) *	<.001 *
	3	0.90 (0.83-0.97)	.005 *	Ref.	0.70 (0.43-1.11)	0.37 (0.20-0.67) *	0.21 (0.10-0.44) *	0.25 (0.12-0.52) *	<.001 *
	4	0.81 (0.72-0.91)	<.001 *	Ref.	0.74 (0.43-1.28)	0.40 (0.20-0.78) *	0.22 (0.09-0.55) *	0.26 (0.11-0.63) *	<.001 *

Only gold members can continue reading. Log In or Register to continue

Share this:

• X
• Facebook

Related

Related posts:

1. Unsuccessful Clinical Trials in Retina: Lessons Learned
2. Glucagon-Like Peptide-1 Receptor Agonists and Risk of Systemic and Ocular Vascular Complications in Patients With Type 2 Diabetes and Diabetic Retinopathy
3. Altitude-Specific IOP Risks After Vitrectomy With Gas Tamponade: Modeling and Travel Guidance Based on Calculated Elevation Profiles
4. One-Year Tumor Volume Regression Predicts Survival After I-125 Plaque Brachytherapy for Posterior Uveal Melanoma

Stay updated, free articles. Join our Telegram channel

Join