C-Reactive Protein and Age-Related Macular Degeneration and Cataract: The Singapore Malay Eye Study

Abstract

Purpose.: To describe associations between C-reactive protein (CRP) and age-related macular degeneration (AMD) and cataract in an Asian population.

Methods.: A population-based, cross-sectional study of 3280 (78.7% response) Malay persons aged 40 to 80 years was conducted in Singapore. Wisconsin grading protocols were used to grade retinal photographs for signs of early and late AMD, and lens photographs were assessed for nuclear, cortical, and posterior subcapsular (PSC) cataract. Venous blood samples were assessed for serum CRP.

Results.: Of 3100 (94.5%) subjects with gradable fundus photographs, 177 (5.7%) had any AMD, 155 (5.0%) early AMD, and 22 (0.7%) late AMD. No association was found between CRP and AMD. After subjects were stratified by diabetes status, the data showed that higher CRP was associated with any AMD in 2385 persons without diabetes (multivariate-adjusted odds ratio [OR], 1.73; 95% confidence interval [CI], 1.03–2.91, comparing the 4th versus the 1st quartiles of CRP). Cataract was present in 1431 (48.5%) subjects, and 634 (21.5%), 881 (29.9%) and 443 (15.0%) had nuclear, cortical, and PSC cataract, respectively. After multivariate analysis, no association was found between CRP and any cataract (OR, 0.99; 95% CI, 0.73–1.34), nuclear (OR, 0.98; 95% CI, 0.68–1.40), cortical (OR, 0.94; 95% CI, 0.71–1.23), or PSC (OR, 1.12; 95% CI, 0.81–1.55) cataract.

Conclusions.: In the general population, there were no associations between CRP and AMD or cataract. In persons without diabetes, higher CRP was associated with AMD. These data suggest only a weak link between systemic inflammation and AMD and cataract in Asian people.

Age-related macular degeneration (AMD) and cataract are two major causes of vision impairment in the elderly population. ¹ Both AMD and cataract are strongly related to age. In addition, risk factors for AMD include cigarette smoking, ^1,2 lack of dietary nutrients, ³ and cardiovascular disease. ⁴ For cataract, risk factors include diabetes, ⁵ higher body mass index, ⁶ and cigarette smoking. ⁷  

Inflammation is now increasingly thought to be a key risk factor for AMD ^8–10 and, to a lesser extent, for cataract. ^11,12 The major acute-phase reactant, C-reactive protein (CRP), is present in human serum at low levels under normal conditions, but can be markedly elevated in inflammatory conditions. ¹³ The association of CRP with AMD has been reported in only a few studies, with somewhat inconsistent results. ^8–10 For example, Seddon et al. ⁸ reported that elevated CRP was an independent risk factor for AMD, but this relationship was not seen in the Cardiovascular Health Study. ⁹ Similarly, for cataract, whereas the relationship between systemic markers of inflammation and age-related cataract has been studied, ^11,12 no conclusive trend has been demonstrated. None of these studies included Asian populations. 

The purpose of our study was to examine the association of CRP with AMD and age-related cataract in an Asian Malay population in Singapore. 

Methods

Study Population

The Singapore Malay Eye Study (SiMES) is a population-based, cross-sectional, epidemiologic study of 3280 Malay adults, aged between 40 and 79 years, living in Singapore. Details of the SiMES study design, sampling plan, and methodology have been reported elsewhere. ¹⁴ The sampling frame consisted of all Malays aged 40 to 80 living in 15 residential districts across southwestern Singapore. From an initial list of 16,069 Malay names provided by the Ministry of Home Affairs, an age-stratified, random-sampling procedure was used to select 5,600 names (1,400 people from each decade of 40–49, 50–59, 60–69, and 70–80 years). Of the 5600 names initially identified, 4168 (74.4%) participants were determined to be eligible to participate based on the inclusion criteria mentioned earlier. ¹⁴  

Of these, 3280 (78.7%) were examined in the clinic, and the remaining 888 (21.3%) were classified as nonparticipants. The study adhered to the Declaration of Helsinki, and ethics approval was obtained from the Singapore Eye Research Institute Institutional Review Board. Written informed consent was obtained from all subjects before enrollment. 

Study Measurements

AMD Grading.

Assessment of AMD in this study population has been described previously. ¹⁵ A digital retinal camera (CR-DGi with a 10-D SLR backing; Canon, Tokyo, Japan) was used to obtain color photographs of Early Treatment of Diabetic Retinopathy Study (ETDRS) standard field 1 (centered on the optic disc) and ETDRS standard field 2 (centered on the fovea) in each eye. ¹⁶ The photographs were graded by using the Wisconsin age-related maculopathy grading system. ¹⁷  

Drusen type was classified based on the size and the sharpness of the edges. Drusen were classified as hard or soft: soft, if the maximum diameter was ≥250 μm. The soft drusen were then divided into categories of distinct and indistinct. Retinal pigmentary abnormalities, including hypopigmentation and hyperpigmentation, were identified. Early AMD was defined as either soft indistinct or reticular drusen, or soft, distinct drusen with retinal pigment epithelium (RPE) abnormalities, using the Blue Mountains Eye Study (BMES) definition. ¹⁸  

Neovascular AMD lesions were defined as the presence of neurosensory or RPE detachment; subretinal or sub-RPE hemorrhage; or intraretinal, subretinal, or sub-RPE fibrosis. Subretinal hemorrhage or hard exudates within the macular area were also considered signs of neovascular AMD, if other retinal vascular diseases were ruled out as alternate causes. Geographic atrophy was defined as the presence of visible choroidal vessels and a discrete atrophic area with a sharp border, with an area of ≥175 μm diameter. Late AMD was defined as the presence of either neovascular AMD or geographic atrophy. Any AMD was defined as the presence of either early or late AMD. 

Lens Opacities.

A digital slit lamp camera (model DC-1, with FD-21 flash attachment; Topcon, Tokyo, Japan) and a Scheimpflug retroillumination camera (model EAS-1000; Nidek, Gamagori, Japan) were used to photograph the lens through the dilated pupil for assessment of nuclear, cortical, and posterior subcapsular (PSC) cataract, according to the Wisconsin Cataract Grading System. ¹⁹ The slit beam was adjusted to completely fill the pupil, bisecting the lens from the 12- to 6-o'clock position at a 45° angle. Focus was placed on the sulcus of the lens, if visible, to allow for the grading of the presence and extent of a central optically clear zone. If the sulcus was not visible, focus was placed on the estimated center of the lens so that it was possible to grade the degree of nuclear opalescence in comparison with standard photographs. In addition, with a diffused light source, the slit-lamp camera was used to photograph the anterior segment with the focus on the iris margin. All photographs captured by the slit lamp camera were in color. If an eye was aphakic or pseudophakic, only the anterior segment shot was taken. Black-and-white Scheimpflug retroillumination photographs were then taken of each eye for grading of cortical and posterior subcapsular opacities. ¹⁹  

A five-point scale was used to assess the presence and severity of nuclear cataract. The scale was determined by comparing participant photographs with four standards of nuclear photographs with increasing opacity. Nuclear cataract was defined as nuclear opacity worse than standard 3 of the Wisconsin Cataract Grading System. ¹⁹ The presence and severity of cortical and PSC cataracts were graded by using a circular grid divided into eight equal wedges and a central circle. The grader estimated the percentage area in each of the nine segments involved in opacities. These percentages were then summated to give an overall percentage of the whole lens area that was involved by cataract. Cortical cataract was considered present when at least 5% of the lens area was involved, and PSC was defined if any was present. Any cataract was defined as having at least one subtype of cataract (nuclear, cortical, or PSC) in either eye or any previous cataract surgery. 

For both AMD and cataract, the worse eye was considered for analysis. 

Measurement of Other Risk Factors.

Participants underwent a standardized interview, systemic and ocular examination, and laboratory investigations. A detailed interviewer-administrated questionnaire was used to collect information about medical history (e.g., hypertension, diabetes, angina, myocardial infarction, and stroke), cigarette smoking (defined as current, past, or never), and alcohol consumption (defined as ever or never). Education was recorded as the highest number of years of schooling completed and categorized as: (1) primary or lower (≤6 years), (2) secondary (7–10 years), and (3) postsecondary (≥11 years, including university education). 

Height was measured in centimeters with a wall-mounted measuring tape, and weight was measured in kilograms with a digital scale (SECA, model 782 2321009; Vogel & Halke, Hamburg, Germany). Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Based on the World Health Organization standard, overweight status was defined as a BMI between 25 and 29.9 kg/m² and obesity as a BMI over 30 kg/m². ²⁰  

Diabetes mellitus was defined as (1) nonfasting glucose ≥11.1 mM, (2) use of diabetic medication, or (3) self-reported history of diabetes. ²¹ Among the persons defined as having diabetes, the number of persons satisfying the criteria was 326 (42.7%), 448 (58.6), and 658 (86.1%), respectively. Blood pressure was taken with the participant seated, after 5 minutes of rest. Systolic and diastolic blood pressure and pulse rate were measured with a digital automatic blood pressure monitor (Dinamap model Pro Series DP110X-RW, 100V2; GE Medical Systems Information Technologies, Inc., Milwaukee, WI) by the methods used in the Multi-ethnic Study of Atherosclerosis (MESA). ²² Blood pressure was measured on two occasions 5 minutes apart. If the blood pressure differed by more than 10 mm Hg systolic and 5 mm Hg diastolic, a third measurement was made. The blood pressure recorded for each individual was the mean between the two closest readings. Hypertension was defined as systolic blood pressure, ≥140 mm Hg; diastolic blood pressure, ≥90 mm Hg; or physician diagnosis. 

A 40-mL sample of venous blood (nonfasting) was collected to determine serum lipid levels (total cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides), blood glucose, glycosylated hemoglobin (HbA1c), serum creatinine, and serum high-sensitivity CRP levels. All serum biochemistry tests were sent to the National University Hospital Reference Laboratory for measurement on the same day. 

Statistical Methods

AMD and cataract were taken as binary outcomes, and CRP data were analyzed by (1) dividing into quartiles and (2) transforming to the log scale to correct for the skewed distribution of original data (all analyses: SPSS, ver. 15.0; SPSS, Chicago, IL). 

Logistic regression was performed for AMD and cataract, using the lowest CRP quartile as the referent. Initial models were adjusted for age and sex, then for other possible confounders identified in previous papers for AMD and cataract. Multivariate models of AMD were adjusted further for BMI, total cholesterol, systolic blood pressure, use of antihypertensive treatment, and current smoking, whereas multivariate models for cataract were adjusted further for BMI, systolic blood pressure, current smoking, alcohol intake, education, history of myocardial infarction, and HbA1c level. Test for linear trend was performed by treating categorical CRP quartiles as a continuous variable in logistic regression, and P ≤ 0.05 was considered statistically significant. At a prespecified power of 0.8 and α = 0.05, the minimum detectable OR for an association comparing the 4th quartile to the 1st quartile of CRP is 1.73 for any AMD and 1.35 for any cataract. Finally, we performed subgroup analyses stratified by diabetes, smoking, and BMI, as these factors have been shown to influence CRP levels. ^23,24  

Results

Of the 3280 recruited subjects, 3100 (94.5%) had retinal photographs of sufficient quality to grade AMD signs, and 2948 (89.9%) had lens photographs gradable for at least one cataract subtype or had previous cataract surgery done. There were 177 (5.7%) subjects with AMD, 155 (5.0%) had early AMD, and 22 (0.7%) had late AMD. Of 2948 subjects, 1431 (48.5%) had cataract, and 634 (21.5%), 881 (29.9%), and 443 (15.0%) had nuclear, cortical, and posterior subcapsular (PSC) cataract, respectively. The prevalence of diabetes was 21.9%. 

Table 1 shows the characteristics of the study population. The persons with any AMD were more likely to be older, male, and have hypertension and higher systolic blood pressure and triglyceride levels than did persons without AMD. The persons with any cataract were more likely to be older and have hypertension and higher systolic blood pressure. They were also more likely to have diabetes, higher blood glucose and HbA1c levels, and higher total cholesterol and LDL cholesterol levels. However, the persons with cataract were less likely to be alcohol consumers and current smokers, than were the persons without cataract. 

Table 1.

 

Characteristics of the Participating Malay Adults in Singapore

Characteristic	Any AMD			Any Cataract
	Yes (n = 177)	No (n = 2923)	P *	Yes (n = 1431)	No (n = 1517)	P *
Mean (SD)
Age, y	66.5 (9.77)	58.1 (10.92)	<0.001	66.6 (8.34)	51.4 (7.78)	<0.001
BMI, kg/m2	26.1 (5.48)	26.4 (5.08)	0.417	26.3 (5.21)	26.5 (5.02)	0.205
Systolic BP, mm Hg	154.6 (23.05)	146.5 (23.75)	<0.001	154.8 (23.57)	139.9 (21.74)	<0.001
Diastolic BP, mm Hg	80.4 (11.54)	79.7 (11.15)	0.406	79.6 (11.23)	79.9 (11.08)	0.443
Total cholesterol, mmol/L	5.5 (1.22)	5.6 (1.16)	0.257	5.7 (1.22)	5.6 (1.10)	0.010
HDL-cholesterol, mmol/L	1.4 (0.34)	1.4 (0.33)	0.458	1.4 (0.34)	1.3 (0.33)	0.166
LDL-cholesterol, mmol/L	3.5 (1.06)	3.5 (0.99)	0.720	3.6 (1.03)	3.5 (0.95)	0.081
Triglycerides, mmol/L	1.4 (1.26)	1.6 (1.32)	0.044	1.6 (1.35)	1.6 (1.29)	0.185
Blood glucose, mmol/L	6.7 (3.8)	6.8 (3.66)	0.824	7.2 (3.93)	6.3 (3.26)	<0.001
HbA1c, %	6.4 (1.57)	6.4 (1.54)	0.966	6.7 (1.67)	6.2 (1.34)	<0.001
Serum creatinine, mmol/L	96.5 (31.55)	93.4 (57.51)	0.477	100.9 (70.49)	86.3 (29.6)	<0.001
Number (%)
Sex, male	106 (59.9)	1368 (46.9)	0.001	669 (46.8)	740 (48.8)	0.270
Overweight†	96 (54.2)	1685 (57.7)	0.362	807 (56.4)	895 (59.0)	0.153
Diabetes, yes vs. no	42 (23.7)	672 (23.0)	0.828	452 (31.6)	225 (14.8)	<0.001
Hypertension, yes vs. no	149 (84.2)	1974 (67.6)	<0.001	1195 (83.6)	833 (54.9)	<0.001
Alcohol intake, yes vs. no	2 (1.1)	49 (1.7)	0.578	11 (0.8)	36 (2.4)	0.001
Current smoking, yes vs. no	28 (16.0)	588 (20.2)	0.178	219 (15.4)	376 (24.8)	<0.001

*  Difference in mean characteristics or proportions based on t-test or χ², respectively, adjusted for age and sex.

†  Overweight defined as BMI ≥25 kg/m².

Table 2 shows the characteristics of the study population, by CRP quartiles. There was no difference in age across the CRP quartiles, but CRP levels were higher in the women than in the men. The persons with higher CRP levels had higher BMI and were more likely to be overweight. Those with higher CRP levels were also more likely to have diabetes or hypertension, higher total and LDL-cholesterol and triglyceride levels, and lower HDL-cholesterol levels. CRP levels were higher in the persons with higher blood glucose and HbA1c levels. There was no difference in serum creatinine level or alcohol consumption across the CRP quartiles. CRP levels were also higher in the persons who were not current smokers. 

Table 2.

 

Characteristics of the Participating Malay Adults in Singapore, by CRP Quartiles

Characteristics	1st Quartile (n = 829)	2nd Quartile (n = 774)	3rd Quartile (n = 741)	4th Quartile (n = 766)	P *
Mean (SD)
Age, y	58.5 (11.4)	58.9 (11.1)	57.9 (10.5)	59.2 (11.1)	0.624
BMI, kg/m2	24.0 (4.2)	26.0 (4.3)	27.5 (4.7)	28.4 (6.0)	<0.001
Systolic blood pressure, mm Hg	143.6 (24.6)	146.2 (22.8)	147.8 (23.0)	150.8 (24.2)	<0.001
Diastolic blood pressure, mm Hg	78.5 (11.2)	79.7 (10.8)	80.6 (10.9)	80.3 (11.7)	<0.001
Total cholesterol, mmol/L	5.5 (1.1)	5.6 (1.1)	5.7 (1.2)	5.7 (1.2)	<0.001
HDL-cholesterol, mmol/L	1.4 (0.4)	1.4 (0.3)	1.3 (0.3)	1.3 (0.3)	<0.001
LDL-cholesterol, mmol/L	3.4 (1.0)	3.5 (0.9)	3.6 (1.0)	3.7 (1.1)	<0.001
Triglycerides, mmol/L	1.4 (1.2)	1.6 (1.3)	1.7 (1.4)	1.7 (1.4)	<0.001
Blood Glucose, mmol/L	6.2 (3.2)	6.7 (3.4)	6.9 (3.7)	7.4 (4.2)	<0.001
HbA1c, %	6.2 (1.4)	6.4 (1.5)	6.5 (1.5)	6.7 (1.7)	<0.001
Serum creatinine, mmol/L	93.6 (47.0)	92.8 (53.8)	93.9 (58.5)	94.2 (65.4)	0.734
Number (%)
Sex, male	422 (50.9)	398 (51.4)	356 (48.0)	304 (39.7)	<0.001
Overweight†	315 (38.0)	444 (57.4)	509 (68.7)	521 (68.0)	<0.001
Diabetes, yes vs. no	167 (20.1)	180 (23.3)	167 (22.5)	204 (26.6)	0.005
Hypertension, yes vs. no	492 (59.3)	524 (67.7)	530 (71.5)	587 (76.7)	<0.001
Alcohol intake, yes vs. no	19 (2.3)	12 (1.6)	13 (1.8)	7 (0.9)	0.051
Current smoking, yes vs. no	189 (22.8)	152 (19.7)	162 (21.9)	115 (15.1)	0.001

 Absolute CRP levels (mg/dL) were as follows: 1st quartile 0–0.9, 2nd quartile 0.91–2.00, 3rd quartile: 2.01–4.30, 4th quartile 4.31– 119.70.

*  Linear trend in mean characteristics across quartiles based on χ² or ANOVA.

†  Overweight defined as BMI ≥ 25 kg/m².

Table 3 shows the relationship of CRP and AMD in the study population, stratified by diabetes status. In the overall population, no association was found between CRP and AMD (age–sex-adjusted odds ratio [OR], 1.40; 95% confidence interval [CI], 0.93–2.11, comparing 4th versus 1st quartile of CRP level). Multivariate analysis (which was further adjusted for BMI, total cholesterol, systolic blood pressure, use of antihypertensive treatment, and current smoking) also showed no association between CRP and any AMD (OR, 1.45; 95% CI, 0.94–2.24), early AMD (OR, 1.34; 95% CI, 0.85–2.11), or late AMD (OR, 2.21; 95% CI, 0.62–7.93). After the subjects were stratified by diabetes status, higher CRP level was associated with any AMD in the persons without diabetes (OR, 1.72; 95% CI, 1.05–2.80) in the initial model that adjusted for age and sex. This association persisted after further adjustment for BMI, total cholesterol, systolic blood pressure, use of antihypertensive treatment, and current smoking (OR, 1.73; 95% CI, 1.03–2.91). 

Table 3.

 

Relationship of CRP and Any, Early, or Late AMD, in All, Diabetic, and Nondiabetic Persons

CRP	N	Any AMD			Any Early AMD		Any Late AMD
		n (%)	Age-Sex OR (95% CI)	Multivariate OR* (95% CI)	n (%)	Multivariate OR* (95% CI)	n (%)	Multivariate OR* (95% CI)
All persons
1st Quartile	827	46 (5.6)	1.00	1.00	42 (5.1)	1.00	4 (0.5)	1.00
2nd Quartile	771	39 (5.1)	0.87 (0.56–1.36)	0.90 (0.57–1.42)	33 (4.3)	0.83 (0.51–1.35)	6 (0.8)	1.57 (0.43–5.77)
3rd Quartile	739	34 (4.6)	0.87 (0.55–1.39)	0.88 (0.54–1.43)	30 (4.1)	0.84 (0.51–1.40)	4 (0.5)	1.29 (0.30–5.48)
4th Quartile	763	58 (7.6)	1.40 (0.93–2.11)	1.45 (0.94–2.24)	50 (6.6)	1.34 (0.85–2.11)	8 (1.0)	2.21 (0.62–7.93)
			P trend = 0.111	P trend = 0.087		P trend = 0.185		P trend = 0.263
LogCRP			1.09 (0.96–1.24)	1.10 (0.96–1.26)		1.09 (0.94–1.26)		1.12 (0.79–1.59)
With diabetes
1st Quartile	166	15 (9.0)	1.00	1.00	12 (7.2)	1.00	3 (1.8)	1.00
2nd Quartile	180	8 (4.4)	0.44 (0.18–1.07)	0.46 (0.19–1.13)	6 (3.3)	0.44 (0.16–1.21)	2 (1.1)	0.50 (0.08–3.25)
3rd Quartile	167	4 (2.4)	0.26 (0.08–0.79)	0.28 (0.09–0.87)	3 (1.8)	0.26 (0.07–0.95)	1 (0.6)	0.39 (0.04–4.12)
4th Quartile	202	15 (7.4)	0.82 (0.39–1.75)	0.98 (0.44–2.16)	14 (6.9)	1.13 (0.49–2.63)	1 (0.5)	0.26 (0.02–3.04)
			P trend = 0.573	P trend = 0.891		P trend = 0.779		P trend = 0.254
LogCRP			0.86 (0.65–1.14)	0.91 (0.68–1.21)		0.98 (0.72–1.34)		0.58 (0.28–1.20)
No diabetes
1st Quartile	661	31 (4.7)	1.00	1.00	30 (4.5)	1.00	1 (0.2)	1.00
2nd Quartile	591	31 (5.2)	1.13 (0.67–1.90)	1.15 (0.67–1.97)	27 (4.6)	1.02 (0.58–1.78)	4 (0.7)	4.70 (0.51–43.80)
3rd Quartile	572	30 (5.2)	1.23 (0.72–2.08)	1.23 (0.70–2.14)	27 (4.7)	1.12 (0.63–1.99)	3 (0.5)	3.90 (0.38–40.05)
4th Quartile	561	43 (7.7)	1.72 (1.05–2.80)	1.73 (1.03–2.91)	36 (6.4)	1.44 (0.84–2.48)	7 (1.2)	8.80 (1.00–76.82)
			P trend = 0.029	P trend = 0.038		P trend = 0.168		P trend = 0.040
LogCRP			1.16 (1.01–1.34)	1.16 (0.99–1.35)		1.11 (0.95–1.31)		1.42 (0.95–2.12)

*  Adjusted for age, sex, BMI, total cholesterol, systolic blood pressure, use of antihypertensive treatment and current smoking.

Table 4 shows the relationship of CRP and cataract. After multivariate analysis (for age, sex, BMI, systolic blood pressure, smoking, alcohol consumption, education, history of myocardial infarction, and HbA1c), no significant association was found between CRP and any cataract (OR, 0.99; 95% CI, 0.73–1.34, comparing 4th versus 1st quartile of CRP level). No significant association was found between CRP and the various cataract subtypes: nuclear (OR, 0.98; 95% CI, 0.68–1.40), cortical (OR, 0.94; 95% CI, 0.71–1.23), or PSC (OR, 1.12; 95% CI, 0.81–1.55). No significant associations were found after stratifying for diabetes status. 

Table 4.

 

Relationship of CRP with Any Cataract and the Three Types of Cataract, in All, Diabetic, and Nondiabetic Persons

	Any Cataract			Any Nuclear Cataract		Any Cortical Cataract		Any PSC Cataract
	n (%)	Age-Sex OR (95% CI)	Multivariate OR* (95% CI)	n (%)	Multivariate OR* (95% CI)	n (%)	Multivariate OR* (95% CI)	n (%)	Multivariate OR* (95% CI)
All persons
1st Quartile	384 (48.5)	1.00	1.00	188 (25.2)	1.00	239 (31.7)	1.00	117 (15.6)	1.00
2nd Quartile	353 (48.4)	0.94 (0.71–1.25)	0.90 (0.67–1.20)	161 (23.2)	0.87 (0.61–1.23)	213 (30.5)	0.88 (0.68–1.15)	105 (15.1)	0.96 (0.70–1.32)
3rd Quartile	316 (45.3)	0.96 (0.72–1.28)	0.89 (0.66–1.20)	123 (18.8)	0.86 (0.59–1.25)	198 (29.9)	0.94 (0.72–1.24)	100 (15.1)	1.06 (0.76–1.48)
4th Quartile	378 (51.8)	1.13 (0.85–1.50)	0.99 (0.73–1.34)	162 (24.3)	0.98 (0.68–1.40)	231 (33.6)	0.94 (0.71–1.23)	121 (17.7)	1.12 (0.81–1.55)
		P trend = 0.421	P trend = 0.921		P trend = 0.898		P trend = 0.767		P trend = 0.400
LogCRP		1.05 (0.96–1.14)	1.00 (0.91–1.10)		0.98 (0.88–1.10)		0.98 (0.90–1.07)		1.03 (0.93–1.14)
With diabetes
1st Quartile	103 (65.6)	1.00	1.00	40 (29.2)	1.00	67 (46.5)	1.00	29 (20.3)	1.00
2nd Quartile	121 (72.0)	1.36 (0.76–2.41)	1.34 (0.75–2.41)	46 (31.9)	1.03 (0.53–1.97)	78 (51.7)	1.30 (0.79–2.14)	32 (22.2)	1.04 (0.56–1.94)
3rd Quartile	103 (64.4)	1.08 (0.62–1.90)	1.04 (0.58–1.85)	34 (25.8)	1.18 (0.60–2.34)	63 (44.1)	1.03 (0.62–1.72)	32 (22.4)	1.29 (0.69–2.44)
4th Quartile	125 (65.1)	0.87 (0.50–1.50)	0.79 (0.45–1.40)	38 (24.2)	0.67 (0.34–1.30)	80 (47.6)	1.04 (0.63–1.71)	41 (24.6)	1.25 (0.68–2.32)
		P trend = 0.430	P trend = 0.274		P trend = 0.294		P trend = 0.866		P trend = 0.384
LogCRP		0.93 (0.79–1.11)	0.89 (0.74–1.07)		0.88 (0.71–1.08)		0.98 (0.83–1.14)		1.06 (0.87–1.29)
No diabetes
1st Quartile	281 (44.3)	1.00	1.00	148 (24.3)	1.00	172 (28.2)	1.00	88 (14.5)	1.00
2nd Quartile	232 (41.4)	0.82 (0.59–1.15)	0.78 (0.55–1.10)	115 (20.9)	0.81 (0.53–1.23)	135 (24.6)	0.75 (0.55–1.03)	73 (13.3)	0.93 (0.64–1.36)
3rd Quartile	213 (39.6)	0.91 (0.65–1.28)	0.83 (0.59–1.19)	89 (17.1)	0.75 (0.48–1.17)	135 (26.0)	0.90 (0.65–1.24)	68 (13.1)	0.99 (0.67–1.47)
4th Quartile	253 (47.0)	1.19 (0.85–1.68)	1.06 (0.74–1.51)	124 (24.4)	1.17 (0.75–1.81)	151 (29.0)	0.87 (0.63–1.21)	80 (15.5)	1.08 (0.74–1.58)
		P trend = 0.281	P trend = 0.701		P trend = 0.595		P trend = 0.638		P trend = 0.644
LogCRP		1.07 (0.96–1.19)	1.03 (0.93–1.16)		1.03 (0.90–1.18)		0.97 (0.88–1.07)		1.02 (0.91–1.14)

*  Adjusted for age, sex, BMI, systolic blood pressure, current smoking, alcohol intake, education, history of myocardial infarction and, in all participants by HbA1c.

Associations of CRP levels with both AMD and cataract were largely similar in subgroup analyses stratified by smoking status (current versus past/never smokers) and BMI (≥25 kg/m² vs. <25 kg/m²; data not shown). 

Discussion

In this population-based study of Asian Malays aged 40 to 80 years, we did not find a consistent pattern of association of CRP with either AMD or cataract. In subgroup analysis among persons without diabetes, higher CRP levels were associated with AMD, particularly late AMD, but this association was not seen in people with diabetes. Our data suggest at most a weak relationship between systemic inflammation and AMD in nondiabetic Asian people. 

Inflammation is now considered a key pathophysiological process in the development of AMD. ^25,26 CRP is a highly conserved molecule that is secreted by the liver in response to inflammatory cytokines and may contribute to tissue damage through its role in the induction of complement activation. ²⁷ Complement factors and CRP have been identified in drusen as well as adjacent to Bruch's membrane, ²⁸ and it has been proposed that repeated attack on retinal pigment epithelial (RPE) cells by complement leads to AMD. ^29,30 More recently, the association of complement factor H (CFH) and risk of AMD has been described, where common CFH polymorphisms are strongly associated with AMD. ^31–33 Some CRP genetic haplotypes predispose to higher CRP levels and, when found in combination with a particular CFH polymorphism, result in an enhanced effect of the CFH gene on AMD. ³⁴ However, different genetic CRP variants per se were not found to be directly related to neovascular AMD. ^34–36 CRP levels have also been associated with atherosclerosis, stroke, diabetes mellitus, and Alzheimer disease, ^37–40 but there is a major debate on whether the associations with these vascular diseases are causal. For example, CRP levels confers a moderate risk of coronary disease, ⁴⁰ but no causal relationship has been shown in genetic studies of CRP variants and coronary disease. ^38,39 It has therefore been suggested that previously demonstrated associations between CRP and heart disease may be due to confounding by other inflammatory markers, other cardiovascular risk factors associated with CRP, or reverse causality bias. ³⁸ Similarly, although CRP has long been associated with type 2 diabetes, Lee et al., ³⁷ in a recent meta-analysis, noted that CRP's previously reported associations may be confounded by factors such as weight–height ratio, markers of liver dysfunction, and other serum biomarkers. 

Our findings for AMD and cataract should be compared with those in the few other studies in white populations. Seddon et al. ⁸ reported that elevated CRP is associated with intermediate and advanced stages, as well as with progression of AMD. ⁴¹ Schaumberg et al. ¹⁰ also reported that increased high-sensitivity CRP levels precede the development of AMD in women. However, several other studies have found no relationship between CRP and AMD, ^9,42,43 with at least one limited by inadequate power ⁴³ (n = 396). Our finding of an association of higher CRP and any AMD, but only among persons without diabetes, appears at odds with the concept that diabetes is a proinflammatory condition. The small number of subjects with diabetes may explain this equivocal finding. We can also offer the following explanations. First, it is well recognized that the relationships of risk factors for many vascular diseases are different in people with and without diabetes, because diabetes is a strong and dominant risk factor for cardiovascular disease. ⁴⁴ For example, the relationship of retinal arteriolar caliber and blood pressure is weaker in people with diabetes, because it has been shown that retinal vessels lose autoregulatory ability in diabetics. ⁴⁵ In the same manner, we suggest that the relationship between CRP and AMD may be masked by systemic and ocular processes in persons with diabetes. Second, studies have now found that additional cardiovascular risk conferred by CRP levels may not become apparent unless the inflammatory component related to the disease is very prominent. ⁴⁶ Our study suggests that it is possible that in diabetic patients with AMD, other risk factors and mechanisms (for example, atherosclerosis) are more important than inflammation in determining the presence of AMD. Thus, systemic inflammatory markers may only reflect ocular inflammatory processes in persons without diabetes. Such a hypothesis should be confirmed in further studies. 

The relationship between other markers of inflammation and AMD is not well defined. Seddon et al. ⁸ reported that interleukin (IL)-6 is related to AMD progression, while other inflammatory markers, including tumor necrosis factor (TNF)-α2 receptor, intracellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM)-1, lipoprotein (a), and apolipoprotein (b) are not significantly related. ⁴¹ Klein et al. ⁴² reported that IL-6, TNF-α, and ICAM-1 were not associated with prevalent AMD. 

Among the many risk factors for cataract, ocular and systemic inflammatory conditions have been reported to be associated with an increase in risk. ^47–49 However, no one specific inflammatory marker has been identified and various inflammatory markers other than CRP have previously been associated with cataract, including IL-6, ICAM-1, ¹² and plasma fibrinogen. ⁵⁰ In our study, we found no significant relationship between CRP and age-related cataract. Few previous studies have reported on this relationship, ^11,12 and no conclusive trends have been proposed. The Physician's Health Study reported a threefold increased risk of CRP and cataract in healthy male physicians, ¹¹ but the Beaver Dam Eye Study reported no association of CRP with any cataract subtype. ¹² There was an increased risk of cataract with increasing quartile of CRP, but the trend was not significant. Further studies are warranted. 

Our strengths include a large sample size, a high response rate, and the use of standardized protocols to measure both AMD and cataract, based on the Blue Mountains Eye Study protocol. We had adequate power to detect a minimum OR of 1.73 for any AMD and 1.35 for any cataract. Limitations should be noted. First, because our analysis was cross-sectional, it is difficult to establish the temporal sequence of events. For example, the CRP levels could have been raised after the development of AMD in nondiabetics which was not evident at the time of our data collection. Second, we did not have adequate power to detect CRP associations for late AMD and in some of the subgroup analysis. In particular, the relationships seen for late AMD and in people without diabetes require further confirmation. Finally, we did not measure other inflammatory markers such as plasma fibrinogen, IL-6, and albumin. 

In conclusion, our study showed no consistent overall relationship of CRP with either AMD or age-related cataract in this Asian Malay population. In persons without diabetes, a higher CRP level was associated with any AMD, but the levels in late AMD and in the subgroup analyses were generally low. Future studies are therefore needed to clarify whether there are consistent associations between CRP and AMD in people without diabetes. 

Footnotes

 Supported by National Medical Research Council Grants 0796/2003, 0863/2004, and CSI/0002/2005 and Biomedical Research Council Grant 501/1/25-5. The sponsors had no role in the study design, acquisition of data, statistical analysis, and interpretation, and the final presentation and publication of the study.

Footnotes

 Disclosure: P.Y. Boey, None; W.T. Tay, None; E. Lamoureux, None; E.S. Tai, None; P. Mitchell, None; J.J. Wang, None; S.M. Saw, None; T.Y. Wong, None