Association between statin use and rates of structural and functional loss in glaucoma

Abstract

Background/Aims:

To evaluate the association between statin use and rates of standard automated perimetry (SAP) and retinal nerve fiber layer (RNFL) change in glaucoma patients and glaucoma suspects.

Methods:

This retrospective cohort study included subjects from the Duke Glaucoma Registry with primary open angle glaucoma and glaucoma suspects. Subjects were assigned to groups according to history of statin use. Rates of change in SAP mean deviation (MD) and spectral-domain optical coherence tomography (SD OCT) RNFL thickness over time were estimated using linear mixed models and compared in the statin versus control groups. The effect of duration of statin use was also assessed. Patients with glaucoma versus suspects were analyzed separately. Analyses were adjusted for potential confounding factors of age, gender, race, intraocular pressure, and follow-up time.

Results:

The study included 10,049 SAP tests and 14,198 SD OCT tests from 3,007 eyes (1,976 patients) followed for an average of 4.7±2.0 years. Of these, 775 subjects (1,179 eyes) had history of statin use. No difference in rates of change were seen between the statin versus control groups for MD (−0.07±0.16 dB/year vs. −0.07±0.15 dB/year; P = 0.873, respectively) or RNFL thickness (−0.70±0.60 μm/year vs. −0.70±0.61 μm/year; P = 0.923, respectively). Multivariable models controlling for potential confounders showed no significant association between duration of statin use and rates of MD or RNFL thickness change.

Conclusions:

We did not find a statistically significant association between statin use or duration of statin use and rates of structural and functional change in those with glaucoma or glaucoma suspects.

Introduction

Glaucoma is a neurodegenerative disease characterized by progressive retinal ganglion cell death.[1] Continued disease progression in some patients despite intraocular pressure (IOP) lowering suggests pressure-independent mechanisms of glaucomatous damage, and the need for pressure-independent interventions.[2 3] Statins, 3-hydroxy-3-methylglutaryl co-enzyme A (HMG-CoA) reductase inhibitors, have been demonstrated to have neuroprotective effects in other neurodegenerative diseases,[4] and have thus been investigated for their potential role in glaucoma.

Some studies suggest statins may decrease risk of primary open-angle glaucoma (POAG).[5–8] The proposed mechanisms for the protective nature of statins include neuroprotection,[9–11] increased blood flow via upregulation of endothelial nitric oxide synthase,[12] and inhibition of rho kinase activity with subsequent increase of aqueous outflow facility.[13–17] However, the evidence for the role of statins in glaucoma is inconsistent. Several studies have demonstrated no benefit to statin use or even increased risk of glaucoma.[18–24] In general, longitudinal studies investigating the role of statins in glaucoma have been restricted to a limited number of follow-up time points or to small sample sizes. The few large-scale longitudinal studies on statins and glaucoma determined progression by the need for additional medications and/or surgery, which may be unreliable endpoints to determine disease progression.[7 21] An assessment of the relationship between statins and glaucoma progression should include quantification of the relationship between drug use and objective metrics of functional and structural deterioration in the disease.

In the present study, we investigated the impact of statin use on rates of glaucomatous progression on a large cohort of patients followed over time. The cohort was followed with both standard automated perimetry (SAP) and optical coherence tomography (OCT), allowing a better quantification of the relationship between statin use and progressive glaucoma.

Materials and Methods

Study Population

The data for this study was collected from the Duke Glaucoma Registry (DGR),[25 26] a database of electronic medical and research records at the Vision, Imaging and Performance Laboratory of the Duke Eye Center. The Duke University Institutional Review Board approved this study and provided a waiver of informed consent due to the retrospective nature of the study. All methods adhered to the tenants of the Declaration of Helsinki for research involving human participants, and the study was conducted in accordance with regulations of the Health Insurance Portability and Accountability Act.

The database included patient International Classification of Diseases (ICD) diagnostic codes and Current Procedural Terminology (CPT) codes, medical history, medication history, laboratory tests, best-corrected visual acuity, slit-lamp biomicroscopy, IOP measurement using the Goldmann applanation tonometry (GAT; Haag-Streit, Konig, Switzerland), gonioscopy, ophthalmoscopy examination, photography stereoscopic optic disc photographs, and the results of all SAP acquired with the Humphrey Field Analyzer (HFA, versions II and III; Carl Zeiss Meditec, Inc., Dublin, CA) and peripapillary RNFL thickness tests acquired with the Spectralis spectral-domain OCT (SD OCT; Heidelberg Engineering, Germany) during the study period.

Patients included in the study had been evaluated at the Duke Eye Center for POAG or suspicion of glaucoma according to ICD codes. They were required to have at least 2 SAP tests and 2 peripapillary RNFL SD OCT tests over a minimum study period of 6 months. Since SAP was already widely used in clinical care when SD OCT was introduced, all SAP tests acquired more than 6 months before the first SD OCT test or more than 6 months after the last SD OCT test were excluded from the study. This ensured that eyes were evaluated in a corresponding time period for both tests.

Eligible patients were then analyzed according to the history of statin use at any point during the study period. Statin use was defined by the record on electronic health records (EHR) of use of any HMG-CoA reductase inhibitor (i.e., atorvastatin, cerivastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, or simvastatin). Total time of statin use was defined as the time between the first reported use of the statin in the patient’s medications list up to the last reported use of the drug in the EHR or the end of follow-up (i.e., date of the last available SAP and SD OCT tests).

The baseline characteristics and demographics were drawn from the date when the first reliable SD OCT for each eye was performed. Tests performed after any diagnosis of retinal detachment, retinal or malignant choroidal tumors, non-glaucomatous disorders of the optical nerve and visual pathways, uveitis, age-related macular degeneration (exudative, atrophic, or late stage), amblyopia, and venous or arterial retinal occlusion according to ICD codes were excluded. In addition, tests performed after treatment with panretinal photocoagulation, or incisional glaucoma surgery (i.e. trabeculectomy or tube shunt implantation), according to CPT codes, were also excluded. The list of ICD and CPT codes used for the study have been detailed previously.[25]

Visual fields tests included 24–2 and 30–2 Swedish Interactive Threshold Algorithm (SITA) tests with size III white stimulus. SAP tests were excluded if they had more than 33% fixation losses or more than 15% false-positive errors. SAP mean deviation (MD) was the parameter used to assess rates of change in visual fields over time.

The Spectralis RNFL circle scan was used to determine the average circumpapillary RNFL thickness corresponding to the 360° measure automatically calculated by the SD OCT software. RNFL thickness measurements were obtained from a 12-degree (for single circle scans) or a 3.45mm-diameter peripapillary circle scan (for scans from the Glaucoma Mode Premium Edition) using the Spectralis SD OCT, as described in detail previously.[27] Tests were acquired using the latest available software version at the time of the scan and exported using the latest available version at the time of the analysis (software version 6.8). For each scan, the global average RNFL thickness was calculated as the average of thicknesses of all points from the 360 degrees around the optic nerve head (ONH). This parameter was used to assess rates of change in RNFL thickness over time. All SD OCT scans that had a quality score lower than 15 were excluded from this analysis. Furthermore, scans that had average global RNFL thickness measurements with implausible values (i.e., lower than 20 and greater than 150 μm) were excluded. Those cutoffs represent measurements above the higher range of reported RNFL thickness for normal controls and below the lower range for glaucoma subjects[28–30] and may indicate the presence of acquisition or segmentation errors in the presence of otherwise good quality scores.[31] When more than one good-quality test was available for the same date, the mean global RNFL thickness of all tests from that date was used in the analysis.

The first SAP and SD OCT in the series for each eye was used to classify eyes at baseline into glaucoma versus suspect groups based on an objective reference standard previously described.[32] In brief, the objective criteria for glaucomatous optic neuropathy accounts for the presence and correspondence of structural and functional defects, defined from parameters from SD OCT and SAP. To be considered glaucoma, an eye had to meet the criteria for global loss (i.e., global RNFL thickness from SD OCT outside normal limits and abnormal SAP, as defined by Glaucoma Hemifield Test [GHT] outside normal limits or pattern standard deviation [PSD] with P < 5%); or localized loss (at least one sector in the RNFL thickness outside normal limits with a corresponding abnormality on the opposite SAP hemifield, defined as a hemifield MD with P < 5%). Eyes with normal SAP and SD OCT parameters were considered normal and were not included in the analysis. SD OCT-SAP pairs that did not meet criteria for glaucoma or normal were considered suspects (i.e., only structural, or only functional damage). This classification was preferred to avoid relying on ICD codes for the classification into glaucoma suspects vs. POAG, and to instead allow the assessment of the effect of statin in eyes with and without established structural and functional defect.

Data Analysis

Normality assumption was assessed using histograms and Shapiro-Wilk test. Wilcoxon rank-sum test was used to compare demographical and clinical characteristics between groups according to use of statins, since continuous variables were non-normally distributed. Categorical variables were analyzed using Fisher’s exact test.

Longitudinal linear mixed models were used to estimate rates of change in SAP MD and SD OCT RNFL thickness over time. In brief, mixed models take into account the natural correlation of such data over time, as well as the fact that each patient may contribute two eyes in the analysis.[33] Individual best linear unbiased predictions of slopes of MD and RNFL thickness measurements were obtained for each eye included in the study, taking into account the individual measurements and the parameter estimates of the linear mixed-effects model applied to each group (i.e., records of use of statin or not). Subsequently, the relationship was studied using a multivariable analysis to adjust for potentially confounding factors, including age at baseline, sex, race, IOP, and time of follow-up. For each eye, a mean IOP value was calculated as the average of all measurements with GAT during the corresponding follow-up and peak IOP was determined as the highest IOP value during follow-up. Additional analyses were also performed to evaluate the effect of statins separately on POAG versus suspect patients.

All statistical analyses were performed with Stata, software version 16 (StataCorp LP, College Station, TX) within the Protected Analytics Computing Environment (PACE), a highly protected virtual network space developed by Duke University for analysis of identifiable protected health information.

Results

This retrospective longitudinal study included 10,049 SAP tests and 14,198 SD OCT tests of 3,007 eyes of 1,976 patients with POAG or suspected of having the disease followed for an average of 4.7 ± 2.0 years. Of the 1,976 patients, 775 (39%) patients had history of statin use. The mean duration of statin use was 5.0 ± 2.3 years. The statin group had a significantly lower percentage of females (50% vs. 57%; P < 0.001, respectively) and a higher percentage of African Americans (36% vs. 29%; P < 0.001, respectively) compared to the control group (Table 1). Patients in the statin group were also significantly older at baseline (66.9 ± 9.5 vs. 62.6 ± 12.8 years; P < 0.001, respectively) and had longer mean follow-up time per eye (5.0 ± 2.0 vs. 4.6 ± 2.0 years; P < 0.001). The statin group had a higher percentage of glaucoma suspects compared to the control group (62% vs 56%; P = 0.004) (Table 1).

Table 1.

Demographics and Baseline Clinical Characteristics

Characteristic	Group Control 1,828 eyes of 1,203 patients	Group Statin 1,179 eyes of 775 patients	P Value
Sex, female	686 (57%)	388 (50%)	<0.001 a
Race, AA	344 (29%)	276 (36%)	<0.001 a
Age at baseline, years	62.6 ± 12.8	66.9 ± 9.5	<0.001 b
Mean IOP, mmHg	15.6 ± 3.2	15.6 ± 3.1	0.602b
Peak IOP, mmHg	19.1 ± 5.4	18.9 ± 4.9	0.508b
Diagnosis per eye
Suspect	1,033 (56%)	728 (62%)	0.004 a
Glaucoma	795 (44%)	451 (38%)
Follow-up time, year
Mean ± SD	4.6 ± 2.0	5.0 ± 2.0	<0.001 b
Median [IQR]	4.5 [2.9; 6.2]	5.1 [3.5; 6.7]
Statins, years of use
Mean ± SD	---	5.0 ± 2.3	---
Median [IQR]		5.4 [3.2; 6.7]
Number of SAP tests per eye	3.3 ± 1.5	3.4 ± 1.6	<0.010 b
Baseline MD, dB
Mean ± SD	−3,66 ± 4.69	−3.76 ± 4.74	0.407b
Median [IQR]	−2.26 [−4,63; −0.83]	−2,26[−4,79; −0.93]
Number of SD OCT tests per eye	4.6 ± 1.9	4.8 ± 1.9	0.006 b
Baseline RNFL thickness, μm
Mean ± SD	76.1 ± 15.6	76.9 ± 14.9	0.171b
Median [IQR]	76.0 [66.0; 87.0]	77.0 [67.0; 87.0]

^a= Fishers’s exact

^b= Wilcoxon ranksum

Values are mean ± SD unless otherwise noted.

AA = African American; IQR = Interquartile Range; MD = Mean Deviation; RNFL = Retinal Nerve Fiber Layer; SAP = Standard Automated Perimetry; SD = Standard Deviation; SD OCT = Spectral Domain Optical Coherence Tomography

The distribution of the individual slopes calculated from linear mixed models for each group is shown in Figure 1. No statistically significant difference was seen in the rates of MD change between eyes of patients who used statin versus those who did not (−0.07 ± 0.16 dB/year vs. −0.07 ± 0.15 dB/year, respectively; P = 0.873). The rate of change in RNFL thickness also did not have a statistically significant difference between the two groups (−0.70 ± 0.60 μm/year vs. −0.70 ± 0.61 μm/year, respectively; P = 0.923). Both univariable and multivariable models showed no statistically significant effect of statin use on rates of change in SAP MD or RNFL thickness (Table 2).

Figure 1.

Histograms Showing the Distribution of the Individual Slopes of Change for Standard Automated Perimetry (SAP) Mean Deviation (MD) and Spectral Domain Optical Coherence Tomography (SD OCT) Retinal Nerve Fiber Layer (RNFL) Thickness for Groups Control and Statin.

MD = Mean Deviation; RNFL = Retinal Nerve Fiber Layer; SAP = Standard Automated Perimetry; SD OCT = Spectral Domain Optical Coherence Tomography

Table 2.

Univariable and Multivariable Linear Mixed Models of the Effect of Each Clinical Characteristic on the Rate of Change of Standard Automated Perimetry (SAP) Mean Deviation (MD) and Spectral-Domain Optical Coherence Tomography (SD OCT) Retinal Nerve Fiber Layer (RNFL) Thickness Over Time.

	Univariable Models				Multivariable Models
	SAP MD (dB/year)		SD OCT RNFL thickness (μm/year)		SAP MD (dB/year)		SD OCT RNFL thickness (μm/year)
Characteristic	Coefficient	P value	Coefficient	P value	Coefficient	P Value	Coefficient	P Value
Use of statin, yes	−0.004	0.873	0.004	0.923	0.025	0.342	−0.049	0.419
Sex, female	0.017	0,504	0.024	0.569	0.021	0.413	0.118	0.040
Race, AA	0.084	0.003	0.060	0.184	0.057	0.046	−0.021	0.738
Peak IOP, per 1 mmHg higher	−0.005	0.043	−0.055	<0.001	−0.007	0.007	−0.058	<0.001
Mean IOP, per 1 mmHg higher	−0.002	0.609	−0.039	<0.001	---	---	---	---
Age at baseline, per 10 years older	−0.071	<0.001	−0.018	0.343	−0.074	<0.001	−0.043	0.097
Follow-up time, per 1 year longer	−0.013	0.083	0.019	0.111	−0.017	0.022	0.044	0.012
Diagnosis at baseline, glaucoma	−0.143	<0.001	0.049	0.254	---	---	---	---

Boldface indicates statistical significance (P<0.05).

AA = African American; IOP = intraocular pressure

We further evaluated if the duration of statin use had a significant impact on the rates of change. Figure 2 demonstrates a scatterplot of rates of change in SAP MD and global RNFL thickness versus duration of statin use. Longer duration of statin use did not have a statistically significant effect on rates of change for SAP MD nor RNFL thickness changes over time (P = 0.446 and 0.138, respectively; Table 3).

Figure 2.

Scatterplot of rates of change in Standard Automated Perimetry (SAP) Mean Deviation (MD) and Spectral Domain Optical Coherence Tomography (SD OCT) Retinal Nerve Fiber Layer (RNFL) Thickness for different duration of statin use in years

MD = Mean Deviation; RNFL = Retinal Nerve Fiber Layer; SAP = Standard Automated Perimetry; SD OCT = Spectral Domain Optical Coherence Tomography

Table 3.

Multivariable Linear Mixed Models of the Effect of Each Clinical Characteristic on the Rate of Change of Standard Automated Perimetry (SAP) Mean Deviation (MD) and Spectral-Domain Optical Coherence Tomography (SD OCT) Retinal Nerve Fiber Layer (RNFL) Thickness Over Time for the Group with Statin Exposure Only.

	Multivariable Models Group Statin n = 1,179 eyes of 775 subjects
	SAP MD (dB/year)		SD OCT RNFL thickness (μm/year)
Characteristic	Coefficient	P value	Coefficient	P Value
Exposure to statin, per 5 years	0.036	0.446	0.110	0.138
Sex, female	0.034	0.371	0.076	0.215
Race, AA	0.011	0.795	0.067	0.306
Peak IOP, per 1 mmHg higher	−0.007	0.098	−0.039	<0.001
Age at baseline, per 10 years older	−0.063	0.003	−0.064	0.062
Follow-up time, per 1 year longer	−0.019	0.109	0.006	0.765

Boldface indicates statistical significance (P<0.05).

AA = African American; IOP = intraocular pressure

Analyses were also performed separately for glaucoma and suspect eyes. When looking at these groups separately, there was still no statistically significant effect of statin use (Supplemental Table 1). Duration of statin use also had no statistically significant effect on the rates of change in MD and RNFL in either group (data not shown). Finally, to evaluate whether our conclusions held true for eyes with extended follow-up and number of tests, we performed a sub-analysis of patients with 5 or more SAP and OCT tests (n = 476 eyes of 338 subjects). In this sample, 197 eyes (41%) had history of statin use with an average of 5.8 ± 2.1 years of exposure. Multivariable analyses in this subgroup also showed similar rates of change in both groups with no significant association between exposure to statins and rate of change in MD or RNFL thickness (Supplemental Table 2).

Discussion

In this large retrospective study, we were not able to find statistically significant differences in rates of structural and functional loss in patients with history of statin use versus those without. We also were not able to find a statistically significant effect of the length of statin use on rates of change. Our findings suggest that statin use may have a limited role, if any, in preventing progression of glaucoma.

Distributions of rates of progression in SD OCT global RNFL thickness were almost identical in groups with and without a history of statin use (−0.70 ± 0.60 and −0.70 ± 0.61 μm/year, respectively). A study by de Castro et al. previously measured change in optic nerve parameters in control and statin groups.[34] In their study, confocal scanning laser ophthalmoscopy (CSLO) with the Heidelberg Retinal Tomograph (HRT) II was used to image 76 POAG suspect patients for an average follow-up time of approximately 2 years. Patients had an average number of only 2.8 HRT tests. Although the authors reported differences in some HRT parameters between the groups, the results tended to be inconsistent and indicative of large variability. Control eyes were noted to have a change of −13.7%/year for cup volume compared to +26.7%/year for the group using statins. Similar findings were seen for global RNFL, with corresponding values of −10.3%/year and +26.6%/year, respectively. The large decline in rim volume or loss of RNFL thickness per year would be very surprising for suspects. More importantly, the positive increases in rim volume or RNFL thickness over time for the statin group are unrealistic and most likely point to large variability or perhaps artifacts related to reference planes used in the HRT imaging technology. Given the differences in methods, it is difficult to draw any comparisons between the studies. However, the much larger sample size of our study, along with lengthier follow-up and use of SD OCT provide stronger evidence for a non-clinically significant effect of statins on structural progression in glaucoma.

Along with rates of structural change, we also investigated rates of visual field loss in our study and did not find a significant association with statin use. In a previous study, Leung et al.[35] found that patients with normal-tension glaucoma who had visual field progression had a lower prevalence of simvastatin use compared to patients whose visual fields remained stable in a cohort study with average follow-up of 36 months. In their study, 8 of 121 subjects who had visual field progression (6.6%) were using simvastatin, versus 23 of 135 subjects (17.0%) who did not have progression (P = 0.011). In a multivariable logistic regression model, use of simvastatin remained a statistically significant predictor when adjusted for age, history of disc hemorrhage, history of cerebrovascular accident and corneal thickness. No attempt was made to adjust for baseline disease severity, though. Also, the study did not provide rates of change in eyes of patients that had used simvastatin versus those that did not. Whigham et al. found a higher percentage of patients with visual field progression in non-statin users in a VA population, but did not include specific information regarding rates of SAP MD.[36] Inconsistencies among these studies may be due to different criteria to define glaucoma progression, as well as difficulties in controlling for all possible confounding variables such as glaucoma treatment, systemic comorbidities, use of other medications, and variable prescribing patterns of statins.[19 22] It has also been suggested that statins may only play a role early in the disease process, and therefore it is possible the effect would only be seen in patients who are glaucoma suspects as opposed to those already diagnosed with POAG.[7] However, sub-analyses of the glaucoma suspect eyes in our study also did not show statistically significant difference in rates of change between statin and control groups.

This study has limitations. The study was retrospective in nature and relied on data recorded on EHR. Data may be subject to errors in miscoding of glaucoma and statin use. Since statin information was collected from the EHR, it was not confirmed whether the patient was actually taking the medication for the duration it was listed in the record. Statins are often prescribed to patients who have metabolic syndrome, a cluster of conditions including diabetes, hypertension, hyperlipidemia, and obesity. As these conditions and systemic medications to treat these conditions may all have effects on glaucoma that have not been clearly elucidated, the effect of any single component can be difficult to assess.[37] Therefore, differentiating the effect of statin use as a medication from the effect of the disease requiring statin treatment is not easily studied and may have affected the results of our study.

It is also possible that a longer follow up time is needed in order to see significant differences between the statin and control groups, although the mean follow-up time in our study was close to 5 years and generally longer than in previous investigations. Additionally, the statin group had statistically significant longer follow up with more OCT and SAP tests. While our multivariable analyses adjusted for follow-up time, it is possible that residual confounding may not have been eliminated. Of note, while our primary analysis included a minimum of 2 SAP and SD OCT tests, sub-analyses of participants with 5 or more SAP and SD OCT data produced very similar results. We were also not able to consider potential differences in medical or laser treatment for glaucoma between the groups. This may be a confounder if treatment differed between groups. However, the effects of therapy were likely reflected into changes in IOP during follow-up, for which the regression models were adjusted. Another potential limitation of the study is that the statin group was older at baseline compared to the control group. It is known that older age is a risk factor for faster progression on visual fields.[38 39] However, age differences were relatively small and we also included age in our multivariable analysis, making it unlikely that this played a significant confounding role. Future randomized prospective studies may be beneficial in assessing the potential effect of statins on glaucoma progression.

In conclusion, we were not able to detect any statistically significant protective effect of statin use on rates of structural and functional progression in glaucoma. While the study was retrospective in nature, the large sample size and analyses of both structural and functional progression strengthen our findings. Should statins exert some metabolic effect that could protect eyes from faster glaucoma progression, these changes do not seem to be manifest in a clinically significant manner.

KEY MESSAGES.

What is already known on this topic:

The existing evidence for the role of statins in glaucoma is inconsistent and lacks quantification of the relationship between drug use and objective metrics of both functional and structural deterioration in glaucoma.

What this study adds:

This large retrospective study did not find statistically significant differences in rates of structural and functional loss in patients with history of statin use versus those without. Should statins exert some metabolic effect that could protect eyes from faster glaucoma progression, these changes did not manifest in a clinically significant manner.

How this study might affect research, practice or policy

Future randomized prospective studies may be beneficial in assessing the potential effect of statins on glaucoma progression.

Synopsis:

This large retrospective longitudinal study did not find a relationship between the use of statins and rates of structural and functional disease progression in glaucoma.