- Top of page
- Materials and Methods
Purpose: To assess cataract surgery visual outcomes 12 months postoperatively in patients with diabetes, with or without diabetic retinopathy (DR), compared to patients without diabetes.
Methods: We followed 1192 cataract surgical patients aged ≥65 for 12 months postoperatively. Standardised pre- and postoperative pinhole LogMAR visual acuity (VA) measurements were taken. Mean VA improvement was determined by comparing VA after 12 months to preoperative VA.
Results: Of 1192 surgical patients, 324 (27.2%) had diabetes, of whom, 136 (42.0%) had DR. After adjusting for age, gender, diabetes duration and preoperative pinhole VA, the average VA gained 12 months after surgery was 10.8 letters among 868 patients without diabetes, 10.6 letters among 188 patients with diabetes but no DR, 10.0 letters among 95 patients with DR but no past laser treatment, and no letters among 41 patients with DR plus past laser treatment (p < 0.0001, compared to the other three groups). Diabetes duration ≥20 years was associated with mean VA gain of 3 fewer letters than duration <10 years (7 versus 10 letters, p = 0.023), after adjusting for age, gender, DR and preoperative pinhole VA.
Conclusion: Cataract surgery improved VA by an average two lines for patients both with and without diabetes, or with DR but no past laser treatment. No significant VA improvement was evident for patients who had preoperative DR and laser therapy.
Materials and Methods
- Top of page
- Materials and Methods
The Cataract Surgery and Age-Related Macular Degeneration Study is a clinic-based cohort study of older patients undergoing cataract surgery at Westmead Hospital in Sydney. Westmead Hospital is a major tertiary hospital and provides eye services to the suburban population of western Sydney, Australia. The study was approved by the University of Sydney and the Sydney West Area Health Service Human Research Ethics Committees and conducted according to the provisions of the Declaration of Helsinki. Written, informed consent was obtained from all participants. We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research.
During 2004–2007, we recruited 1995 patients aged ≥65 who underwent cataract surgery at Westmead Hospital. Past medical histories were collected, and visual acuity was assessed before surgery. All cataract surgical procedures were performed using phacoemulsification with intraocular lens (IOL) implantation. Several types of IOLs were used, including the Acrysof SA60AT (72.8%; Alcon Laboratories, Fort Worth, TX, USA), Sensar AR40e (10.4%; Advanced Medical Optics, Santa Ana, CA, USA), MA50BM (4.7%, Alcon Laboratories), Akreos Adapt (4.7%, Bausch & Lomb), Akreos Adapt AO (1.8%, Bausch & Lomb) and Quatrix (0.70%, Corneal/Croma, Austria). Postoperatively, patients received a course of acetazolamide (Diamox; Sigma Pharmaceuticals, Victoria, Australia), 250 mg × 4 doses, topical chloramphenicol 0.5% (Chlorsig eye drops, Sigma Pharmaceuticals) four times a day for 4 weeks and topical dexamethasone 0.1% (Maxidex, Alcon Laboratories) four times a day for 6 weeks. If cystoid macular oedema was observed at the 1-month postoperative visit, patients received a course of topical dexamethasone 0.1% (Maxidex, Alcon Laboratories) eight times a day, and ketorolac trometamol (Acular, Allergan Australia, NSW, Australia) four times a day for 4 weeks. If macular oedema had not resolved by 4 weeks, patients would receive focal laser and/or intravitreal triamcinolone acetonide 4 mg (Bristol-Myers Squibb Pharmaceuticals, Vic., Australia) and followed to be assessed for treatment effects.
Of the 1995 patients, 1525 (76.4%) had 12-month postoperative examinations, and of these, 1349 (88.5% of patients who completed 12-month follow-up) had complete data on VA and baseline diabetes status. Of the 470 patients (23.6%) who had not completed 12-month postoperative examinations, 194 (41.3%) were either still waiting for surgery or had had surgery but had not had their 12-month assessment, 64 (13.6%) had died, 53 (11.3%) withdrew because of health reasons, 11 (2.3%) were no longer contactable and 148 (31.0%) declined to continue participation.
A standardised questionnaire was administered at the preoperative survey by trained interviewers. The mean duration between the study recruitment and surgery was 108 days (95% CI 102–114 days). History of diabetes was assessed using the following questions:
‘Have you ever been told by a doctor that you have diabetes?’
‘If yes, when was it first diagnosed?’
‘What type of treatment have you had for diabetes? Choices of answer included: no treatment, diet alone, oral hypoglycaemic, and/or insulin.’
‘Have you ever been told by a doctor that you have an eye disease or eye damage related to your diabetes (diabetic retinopathy)?’
‘Have you ever had laser treatment for your diabetic eye disease?’
The most recent preoperative blood glucose level (mmol/l) was recorded, either self-reported by the patient or reported by the patient’s doctor for those who did not perform glucose testing at home. The most recent glycosylated Hb (HbA1c) levels and the presence of microalbuminuria were reported by the patient’s doctor.
Comprehensive eye examinations were conducted, including the measurement of presenting VA (with habitual correction) and pinhole-corrected VA, using a back-illuminated LogMAR (Vectorvision CSV 1000, Vectorvision Inc, Dayton, OH, USA) chart. VA was the number of letters read correctly at 2.44 m (8 ft). If no letters could be read at 2.44 m (8 ft), VA was assessed as counting fingers at 0.61m (2 ft), hand movements, perception of light or no perception of light. The maximum number of letters read correctly was 70, while −15 letters represented either no letters read correctly or a VA of Counting Fingers, Hand Movements, Perception or No Perception of Light. Mydriatic retinal photographs were also taken of both eyes, using either a Topcon TRC 50 IA retinal camera (Topcon Optical, Tokyo, Japan) with Kodachrome 64 35 mm slide film or a Canon CF-60DSi mydriatic digital camera. A Canon CR-45 NM non-mydriatic camera was only used when patients refused pupil dilation (both were from Canon, Tokyo, Japan). Retinal photographs were taken at preoperative and 1-month postoperative visits, with both visits considered to provide information about the baseline retinal status (Cugati et al. 2007). The photographs included Early Treatment Diabetic Retinopathy Study (ETDRS) standard fields 1 (optic disc) and 2 (macula) for patients without diabetes, and additional fields 4 (supero-temporal), 5 (infero-temporal) and a nasal field for all patients with diabetes. A grader initially assessed the photographs for the presence of DR and a second grader (TH) assessed DR according to ETDRS criteria (Davis et al. 1998), with adjudication provided by a retinal specialist (PM). DR was determined by the presence of any microaneurysms, haemorrhages, hard or soft exudates, venous beading, intraretinal microvascular abnormalities or new vessels in eyes of patients with diabetes. Clinically significant macular oedema (CSME) was defined by one or more of the following ETDRS criteria: (i) apparent thickening of the retina <500 μm from centre of the macula, (ii) hard exudates with apparent thickening of adjacent retina <500 μm from centre of the macula or (iii) a zone of apparent retinal thickening 1 disc area or larger <1 disc diameter from centre of the macula (Early Treatment Diabetic Retinopathy Study Research Group. 1991), and adjudication by PM. Proliferative DR (PDR) was defined if retinal or optic disc neovascularisation was present or if pan-retinal laser scars were present. Only one eye per patient was included that did not have surgery before study recruitment but was operated soon after recruitment (the ‘study eye’). Past laser treatment was self-reported and confirmed using retinal photographic grading.
Analysis of covariance (ancova, F-test) was used to compare the mean improvement in pinhole VA between the preoperative level with the level at the 12-month follow-up visits among four groups of patients: patients without diabetes, patients with diabetes but no DR, patients with DR but no past laser treatment and patients with DR who had signs of previous laser treatment. We also compared the mean improvement in pinhole VA by the duration of diabetes. Comparisons by diabetes status were made after adjusting for age, gender, duration of diabetes and preoperative pinhole VA. Comparisons by diabetes duration were made after adjusting for age, gender, presence of DR and preoperative pinhole VA. Odds ratios (OR) and 95% confidence intervals (CI) were estimated for the dichotomous outcome of postoperative VA ≥ 6/12, after adjusting for age, gender, duration of diabetes and preoperative pinhole VA, using multivariate logistic regression models. sas software (V9.1; SAS, Cary NC, USA) was used for statistical analysis.
- Top of page
- Materials and Methods
We excluded 153 patients with signs of glaucoma or late age-related macular degeneration and 4 who received extracapsular cataract extraction, leaving 1192 patients for this report. Included were 324 (27.2%) with diabetes (3 with type 1 and 321 with type 2) and 868 (72.8%) without diabetes. Of those with diabetes, 136 (42.0%) had DR, including 41 (12.7%) with the signs of past laser treatment. Table 1 shows selected baseline characteristics of the groups by diabetes status. Patients with DR were more likely to report a history of random blood glucose ≥ 8.0 mmol/l and HbA1c ≥ 7.0%, than those with diabetes but no DR. They were also more likely to be on insulin treatment (Table 1).
Table 1. Selected characteristics in diabetic patients (with or without diabetic retinopathy) and in non-diabetic patients.
|Characteristics*||With diabetes||Without diabetes|
|With DR (n = 136)||Without DR (n = 188)||p-value†||(N = 868)||p-value‡|
|Age (years), mean (SD)||73.5 ± 5.5||74.1 ± 5.9||0.35||74.9 ± 5.7||0.0060|
|Ever smoked cigarettes (n = 1186) (%)||45.1||49.7||0.41||49.7||0.56|
|BMI (kg/m2) (n = 1039) mean (SD)||28.1 ± 6.0||28.7 ± 6.1||0.48||26.7 ± 5.2||<0.0001|
| Angina (n = 1189) (%)||18.4||22.8||0.34||11.3||<0.0001|
| Myocardial infarction (n = 1190) (%)||21.3||16.9||0.32||11.9||0.0022|
| Stroke or TIA (n = 1189) (%)||14.0||9.5||0.21||10.0||0.49|
| Hypertension (n = 1190) (%)||69.1||75.7||0.19||54.7||<0.0001|
| High cholesterol (n = 1189) (%)||46.3||56.6||0.067||43.4||0.0059|
|History of random plasma glucose ≥ 8 mmol/l (n = 287) (%)||83.2||70.1||0.010|| || |
|History of HbA1c ≥ 7% (n = 194) (%)||64.1||37.6||0.0002|| || |
|Presence of microalbuminuria (n = 171) (%)||44.3||35.7||0.25|| || |
|Using oral hypoglycaemic (%)||82.4||81.0||0.75|| || |
|Using insulin (%)||35.3||10.1||<0.0001|| || |
Table 2 shows that the mean postoperative LogMAR VA 12 months after surgery improved from 37.6 (95% CI 36.7–38.4) to 48.2 (47.6–48.9) letters in patients without diabetes, from 38.5 (36.7–40.2) to 48.3 (47.0–49.6) letters in patients with diabetes but no DR, from 37.2 (34.4–39.9) to 47.4 (45.3–49.5) letters in patients with DR and no signs of past laser treatment, but only from 30.1 (25.9–34.4) to 35.7 (32.4–38.9) letters in patients with DR who had signs of past laser treatment, after adjusting for age, gender and diabetes duration. After additional adjustment for preoperative pinhole VA, there was no improvement in mean VA (95% CI −3.1 to 3.1 letters) in eyes with laser-treated DR compared to an improvement of approximately 10.3 letters (95% CI 7.5–13.1) in the other three groups (all p-values <0.0001).
Table 2. Adjusted mean pinhole visual acuity (letters read correctly) in operated eyes of 868 non-diabetic patients and 324 diabetic patients [188 without diabetic retinopathy (DR), 95 with DR without past laser treatment and 41 with past laser treatment] in the Cataract Surgery and Age-Related Macular Degeneration Study before and 12 months after cataract surgery.
|Diabetes status||Preoperative VA*||VA at 12-month follow-up*||Difference in VA*||Difference in VA†||p-value‡|
|No diabetes (n = 868)||37.6 (36.7–38.4)||48.2 (47.6–48.9)||10.6 (9.7–1.5)||10.8 (10.1–11.4)||<0.0001|
|Diabetes without DR (N = 188)||38.5 (36.7–40.2)||48.3 (47.0–49.6)||9.8 (8.0–11.7)||10.6 (9.4–11.9)||<0.0001|
|With DR but no past laser treatment (N = 95)||37.2 (34.4–39.9)||47.4 (45.3–49.5)||10.2 (7.4–13.1)||10.0 (8.1–12.0)||<0.0001|
|DR with past laser treatment (N = 41)||30.1 (25.9–34.4)||35.7 (32.4–38.9)||5.5 (1.1–10.0)||0.0025 (−3.1 to 3.1)||Reference|
Of the 41 patients who had signs of past laser treatment, 9 (22.0%) had pan-retinal laser scars, 16 (39.0%) had macular laser scars, 12 (29.3%) had evidence of both types of treatment and 4 (9.8%) had only localised peripheral laser scars. Table 3 shows the mean VA change 12 months after surgery in patients with signs of past laser treatment. There was only a 3-letter improvement (range −3 to 9) in eyes with macular laser scars but this VA change was not significant compared to the other groups who had signs of past laser treatment (0.25 ≤ p ≤ 0.6).
Table 3. Adjusted* mean preoperative and 12-month postoperative pinhole visual acuity (letters read correctly) of operated eyes of 41 patients who had signs of past laser treatment, stratified by laser type, in the Cataract Surgery and Age-Related Macular Degeneration Study.
|Type of laser scar||Preoperative VA||VA at 12-month follow-up||Difference in VA||p-value†|
|Pan-retinal laser scars (N = 9)||30.6 (22.7–38.4)||36.3 (30.3–42.2)||5.7 (−2.4 to 13.9)||0.60|
|Macular laser scars (N = 16)||32.5 (26.8–38.3)||35.7 (31.3–40.1)||3.2 (−2.8 to 9.2)||Reference|
|Both pan-retinal and macular laser scars (N = 12)||28.3 (21.4–35.1)||36.5 (31.3–41.7)||8.2 (1.1–15.4)||0.25|
|Localised peripheral laser scars (N = 4)||24.7 (13.3–36.1)||33.3 (24.6–41.9)||8.6 (−3.3 to 20.4)||0.41|
Only three patients in our study cohort had cystoid macular oedema, including two without diabetes and one with macular laser scars. Of the three patients, two were treated successfully with topical eye drops, while one had further treatment with intravitreal triamcinolone and focal laser. On average, their mean postoperative VA 12 months after surgery increased by three letters (from 28 to 31).
After adjusting for age, gender, presence of DR at baseline and preoperative pinhole VA, diabetes duration ≥20 years was associated with a mean VA gain of 6.8 letters (95% CI 4.9–8.8) 12 months after cataract surgery (Table 4). This was similar to patients with diabetes duration 15–19 years (mean 7.4 letters gained; 95% CI 4.2–10.6), but less than the gain by patients with diabetes duration 10–14 years (mean 10.0 letters gained; 95% CI 7.9–12.0) or those with diabetes duration <10 years (mean 9.9 letters gained; 95% CI 8.4–11.5) (Table 4). There was a significant difference in VA gain between patients with diabetes duration ≥15 years and those with diabetes duration <15 years (p = 0.0074), after controlling for age, gender, diabetic retinopathy and preoperative pinhole visual acuity.
Table 4. Adjusted mean preoperative and 12-month postoperative pinhole visual acuity (letters read correctly) of operated eyes of 324 patients with diabetes, stratified by diabetic duration, in the Cataract Surgery and Age-Related Macular Degeneration Study.
|Duration of diabetes (years)||Preoperative VA*||VA at 12-month follow-up*||Difference in VA*||Difference in VA†||p-value‡|
|<10 (N = 142)||37.3 (35.2–39.4)||47.1 (45.4–48.8)||9.8 (7.6–12.0)||9.9 (8.4–11.5)||0.023|
|10–14 (N = 72)||36.2 (33.5–38.9)||46.8 (44.7–49.0)||10.6 (7.8–13.5)||10.0 (7.9–12.0)||0.033|
|15–19 (N = 29)||37.4 (33.2–41.6)||44.6 (41.2–48.0)||7.2 (2.7–11.6)||7.4 (4.2–10.6)||0.78|
|20+ (N = 81)||36.5 (33.9–39.1)||43.8 (41.7–45.9)||7.3 (4.6–10.1)||6.8 (4.9–8.8)||Reference|
Patients with DR and signs of past laser treatment were slightly younger (mean age 73.7 years versus 75.3 years, p = 0.048) but had a longer mean duration of diabetes (24.6 years versus 16.8 years, p = 0.0004) than those with DR but no past laser treatment. Eyes that had past laser treatment for DR before surgery were 6 times more likely to have postoperative VA less than 6/12 after 12 months follow-up, compared to eyes with DR and no previous laser treatment (adjusted OR 5.84, 95% CI 2.27–15.03). Further, only 34% (n = 14) of the operated eyes that had past laser treatment gained ≥2 LogMAR lines after cataract surgery, compared to 42% (n = 40) of operated eyes with DR but no past laser treatment (adjusted OR 0.24, 95% CI 0.07–0.74 for eyes with past laser treatment to gain VA ≥ 2 LogMAR lines, compared to eyes with no past laser treatment) (Table 5).
Table 5. Diabetic retinopathy (DR) and postcataract surgery visual acuity levels, comparing eyes with DR and previous laser treatment to eyes with DR but no previous laser treatment.
|Change in visual status||DR with no previous laser treatment (n = 95) N (%)||DR treated by laser previously (n = 41) N (%)||OR adjusted for age, sex and diabetes duration (95% CI)||Multivariate adjusted OR (95% CI)*|
|Loss of ≥1 LogMAR line|| 7 (7)||10 (24)||3.45 (1.11–10.69)||5.90 (1.66–20.93)|
|Gain of ≥1 LogMAR line||60 (63)||20 (49)||0.56 (0.25–1.25)||0.26 (0.10–0.69)|
|Gain of ≥2 LogMAR lines||40 (42)||14 (34)||0.80 (0.35–1.82)||0.24 (0.07–0.74)|
|Gain of ≥3 LogMAR lines||24 (25)||8 (20)||0.90 (0.34–2.38)||0.10 (0.02–0.54)|
|Visual acuity <6/12 at 12-month follow-up||13 (14)||23 (56)||7.38 (2.98–18.32)||5.84 (2.27–15.03)|
- Top of page
- Materials and Methods
We report that cataract surgery improved mean VA 12 months after surgery equally for eyes of patients with and without diabetes or with DR but no past laser treatment, by an average 2 LogMAR lines. However, in eyes with DR treated by laser before surgery, there was, on average, no gain in postoperative VA after 12 months.
Previous reports indicate that poor cataract surgery visual outcomes are related to the preoperative retinal status, particularly when PDR and CSME are present (Antcliff et al. 1996; Zaczek et al. 1999; Dowler et al. 2000; Mittra et al. 2000). Our study findings are consistent with these previous reports, showing around 2-line lower VA improvement in eyes that had undergone preoperative laser treatment (probably because of PDR or CSME), than other eyes that had not had prior laser treatment. Despite the differences in VA (lines of letters) gained after cataract surgery, all groups, on average, had VA improvement after the surgery.
Previous studies were equivocal about the influence of diabetes duration on the visual outcome after cataract surgery (Mittra et al. 2000; Squirrell et al. 2002). In our study sample, we found that, after adjusting for the presence of DR, a diabetes duration ≥15 years was associated with a significantly lower mean VA gain than a diabetes duration shorter than 15 years. This suggests that the effect of diabetes duration on cataract surgery visual outcomes could be independent of the severity of DR.
VA levels ≥6/12 after a postsurgical period of 12 months were attained by 44% of eyes with DR that had past laser treatment, compared to 86% of those with DR but no past laser treatment (Table 5). Few studies have reported the postsurgical visual outcome of eyes with diabetic retinopathy that had been treated with laser prior to phacoemulsification cataract surgery. An ETDRS report found that 46% of eyes assigned to photocoagulation before cataract surgery achieved best-corrected VA better than 6/12 (Chew et al. 1999). However, the technique of cataract surgery was not specified in that study. We included in this report only patients who received phacoemulsification cataract surgery. In addition, the prospective nature of our study and its adjustment for age, gender, diabetes duration and preoperative pinhole VA provide a more realistic estimate for postoperative VA in eyes with diabetic damage.
Previous studies have shown that VA improved by ≥2 lines in between 74% and 86% of patients with diabetes before surgery (Antcliff et al. 1996; Krepler et al. 2002). Our finding that 42% of patients, with DR but not previously treated by laser before cataract surgery, gained 2-LogMAR lines is less than these previous reports. These previous studies had relatively small numbers of patients (n < 64). The study by Antcliff et al. (1996) was retrospective, a study design dependent on the accuracy of medical records. Krepler et al. (2002) excluded patients with CSME and ocular diseases other than cataract and DR, identified before cataract surgery, which could explain a higher proportion achieving VA improvement ≥ 2 lines after surgery.
A previous study showed that prophylactic topical non-steroidal, anti-inflammatory medication was more efficacious than topical steroids at suppressing increased retinal thickness in patients with non-proliferative diabetic retinopathy at four postoperative weeks and at 6 weeks after cataract surgery, but both eye drops were equally efficacious at suppressing increased retinal thickness in patients without DR (Endo et al. 2009). In our study, only three patients (including 1 with DR) had cystoid macular oedema after cataract surgery, and two were successfully treated with a combination of topical steroidal and non-steroidal, anti-inflammatory medications.
Strengths of our study include its longitudinal nature, the presence of non-diabetic controls for the comparison of baseline characteristics and records of 12-month postoperative VA improvement levels. Our questionnaire was administered by trained interviewers who ensured consistency in the way the information was collected. In addition, self-reported history of laser treatment was confirmed at the retinal photographic grading. However, we could have underestimated or overestimated the visual outcome of cataract surgical patients with underlying CSME, because of the very small number of cases with CSME in this sample. Second, the number of patients with laser-treated DR before surgery in our sample was relatively small (n = 41), which is likely to have decreased the precision of our estimates. We recently reported that DR was more likely to progress by one or more steps of the modified ETDRS classification system in eyes 12 months after cataract surgery, compared to non-operated fellow eyes of the same patients (Hong et al. 2009). Data from this current report suggest that such progression did not outweigh the gain in vision 12 months after surgery. Third, subjective refraction was not performed in our study. A previous report showed that best-corrected VA after subjective refraction improved pinhole visual acuity by 3–4 letters (Attebo et al. 1996). Hence, pinhole VA, as a surrogate for best-corrected VA, probably underestimates subjective VA by less than one LogMAR line.
In summary, our study showed that in patients with diabetes, those with a history or signs of preoperative laser treatment, likely indicating prior CSME or PDR, had none or less VA improvement compared to patients with or without diabetes or with DR but no previous laser treatment, who gained an average 2 LogMAR lines 1 year after phacoemulsification cataract surgery. Longer duration of diabetes had a significant adverse influence on the surgical visual outcome, after considering DR status. These findings provide realistic expectations for diabetic patients needing cataract surgery.
- Top of page
- Materials and Methods
An abstract of this manuscript has been presented at the Association for Research in Vision and Ophthalmology (ARVO) 2009 Annual Meeting, May 3–7, 2009 (abstract published as: Fong CS, Mitchell P, Rochtchina E, de Loryn T, Wang JJ. Diabetes and visual outcome after cataract surgery. Investigative Ophthalmology and Visual Science 50: e-abstract 4426, 2009). The study was supported by the Australian National Health & Medical Research Council, Canberra, Australia (Grant No 302010, 2004–2006), and Retina Australia (2005). The authors declare no competing interests. The authors wish to thank Ms Kirsten B Jakobsen, Ms Mireille Moffitt and Ms Ava G Tan for their valuable contributions.