Risk factors for endothelial cell damage in diabetics after phacoemulsification.

BACKGROUND: To assess the corneal endothelium, central corneal thickness and the factors associated with endothelial cell damage after phacoemulsification in diabetics in comparison with non-diabetics.
METHODS: It was a case control study with 80 eyes each in the diabetic group and the control group. Intraoperative mydriasis, effective phaco time (EPT) and postoperative inflammation were measured. Preoperative, 1st week, 6th week and 3rd month postoperative endothelial cell density (ECD), coefficient of variation (CV), hexagonality and central corneal thickness (CCT) were also measured using Konan noncon robo specular microscope (Model - NSP 9900).
RESULTS: In the control group, patients in the age group of 60-69 years were 3.8 times more at risk of ECL compared to patients in the age group of 50-59 years. Patients in whom EPT was ≥0.50 min, were 8.8 times more at risk of ECL when compared to patients in whom EPT <0.25 min. In the diabetic group, patients who had an inflammatory score of 1+ in the first postoperative week; also had 5.7 times more risk of ECL when compared to patients in whom the inflammatory score was 0.5+ in the first postoperative week. There was a significant increase in CV (p-0.03) and CCT (p-0.03), significant decrease in the hexagonality (p-0.01) and no statistically significant difference in the endothelial cell loss (ECL) (p-0.34) in diabetics after phacoemulsification when compared to controls.
CONCLUSION: The present study reveals postoperative inflammation as a risk factor for ECL in diabetics and not intraoperative mydriasis and EPT.

Introduction

The endothelial cell morphology is abnormal in diabetics showing polymegathism and pleomorphism. Diabetic hyperglycemia inhibits the function of Na+-K+ ATPase and thereby the corneal endothelial function.[1] Chronic hyperglycemia causing an increase in the level of reactive oxygen species is implicated in the pathogenesis of diabetic cataract.[2] The cornea is also found to be thicker in diabetics due to the slower recovery from corneal edema.[34]

Diabetic patients undergoing cataract surgery are likely to exhibit more intraocular inflammation. Increased incidence of corneal edema in diabetic patients may relate to poor endothelial pump function, the extent of intraocular inflammation, and duration of surgery. Many parameters such as the technique of phacoemulsification,[5678] phaco tip positions, vacuum, power, flow rate, ophthalmic viscosurgical devices used, nucleus density,[9] and axial length[10] also influence the endothelial cell count after phacoemulsification. Diabetes itself causing damage to endothelial cell damage was studied by Yamazoe et al.[11] Hence, the present study is planned to evaluate the endothelial morphology in diabetics after phacoemulsification.

Subjects and Methods

Ethics committee approval was obtained and this study adhered to the tenets of the Declaration of Helsinki. Clinical trial registration has been done for this study (NCT02548273). It was a case–control study with two groups – diabetics and nondiabetics who underwent phacoemulsification. The purpose and details of the study were explained to each participant and after obtaining informed consent, patients were recruited into the study.

All the diabetic patients were noninsulin dependent type 2 diabetics with good control of blood sugars, which was ensured by assessing the fasting and postprandial sugar values. The sample size of the study was 80 patients in each group (80 eyes). Patients >40 years with nuclear sclerosis cataract were included in the study. Patients with high myopia (>−6D), corneal opacities, pseudoexfoliation, and uveitis were excluded from the study. Drugs used in the study were tropicamide, phenylephrine, flurbiprofen, and ofloxacin-prednisolone combination. Patients were not on any other topical medication before the study period.

Preoperatively, detailed ocular examination, corneal endothelial count, morphology assessment – (endothelial cell density [ECD], coefficient of variation of cell size [CV], hexagonality), and central corneal thickness (CCT) were measured using Konan noncon robo specular microscope (Model-NSP 9900) by the same observer.

All patients underwent phacoemulsification using “stop and chop” technique. Surgeries were performed by two surgeons. Viscodispersives were used during surgery. Ringer lactate was used as the irrigating solution. Hydrophobic acrylic nonheparin coated foldable intraocular lenses were used for all patients. Subconjunctival dexamethasone and gentamycin was injected after the surgery. Intraoperative mydriasis, phacoemulsification time, and power were noted. Postoperatively, visual acuity, ocular inflammation scores, corneal thickness, CD, CV, and hexagonality were measured on the 1st week, 6 weeks and 3 months. Ocular inflammation score was calculated based on the standardization of uveitis nomenclature[12] working group grading scheme for anterior chamber cells. Patients were on a tapering course of ofloxacin-prednisolone eye drops over 6 weeks' time (starting with 6 times a day) along with a course of oral ciprofloxacin postoperatively.

Analysis

Test for differences between groups in terms of demographic and clinical characteristics was done using the Chi-square test for categorical variables. Independent Student's t-test or Mann–Whitney's U-test was used for comparing continuous variables. To identify the independent factors associated with the outcome, multiple logistic regression analysis was used. Preoperative versus postoperative modifications within the groups were verified using two-way repeated measures analysis of variance. All statistical tests were carried out at 5% level of significance and P < 0.05 was considered statistically significant. Data analysis was performed using SPSS (version 20.0, SPSS Inc. IBM, Armonk, NY, USA).

Results

The baseline parameters are shown in Table 1. There was no statistically significant difference in the mean mydriasis (P = 0.23), mean effective phaco time (EPT) (P = 0.19) and the median inflammatory score (P = 0.23) between both the groups. The difference in the baseline mean ECD, CV, hexagonality, and CCT were not statistically significant between both the groups (P = 0.20, 0.08, 0.11, and 0.66 for mean ECD, CV, hexagonality, and CCT, respectively).

Table 1

Demographic characteristics and baseline parameters in controls and diabetics

Parameters	Mean (SD)		P
	Controls (n=80)	Diabetics (n=80)
Age (year)	61.0 (6.3)	58.7 (5.5)	0.06*
Sex (males %)†	62.5	65	0.74‡
Intraoperative mydriasis (mm)	6.7 (0.78)	6.5 (0.66)	0.23§
EPT (min)	0.46 (0.34)	0.38 (0.31)	0.19§
Inflammation score||	1¶	1¶	0.23§
ECD (cell/mm2)	2337.4 (363.6)	2408.1 (345.3)	0.20*
CV	49.9 (8.5)	47.8 (6.6)	0.08*
Hexagonality (%)	37.9 (8.9)	39.1 (5.9)	0.11§
CCT (µm)	572.0 (48.3)	575.0 (40.2)	0.66*

Data expressed as mean (SD) except †Expressed as percentage and ||Expressed as score. *P value measured using independent t-test, ‡P value measured using Chi-square test, §P value measured using Mann-Whitney test, ¶Median value, in the first postoperative visit. EPT: Effective phaco time, ECD: Endothelial cell density, CV: Coefficient of variation, CCT: Central corneal thickness, SD: Standard deviation

In the control group, patients in the age group of 60–69 years were 3.8 times more at risk of endothelial cell loss compared to patients in the age group of 50–59 years. There was a statistically significant endothelial cell loss in the age group 60–69 years (P = 0.03). Furthermore, patients in whom EPT was ≥0.50 min, were 8.8 times more at risk of endothelial cell loss when compared to patients in whom EPT <0.25 min and it was statistically significant (P = 0.002). Rest of the parameters such as gender, mydriasis, and inflammation did not cause significant endothelial cell loss in the control group as shown in Table 2. In the diabetic group, patients who had an inflammatory score of 1+ in the 1st postoperative week also had 5.7 times more risk of endothelial cell loss compared to patients with an inflammatory score of 0.5+ and it was statistically significant (P = 0.006). Rest of the parameters such as age, gender, mydriasis, and EPT did not cause significant endothelial cell loss in the diabetic group as shown in Table 2.

Table 2

Pre- and post-operative corneal parameters

Parameters	Controls - mean (SD) (n=80)				Diabetics - mean (SD) (n=80)				P
	Baseline	Postop 1	Postop 2	Postop 3	Baseline	Postop 1	Postop 2	Postop 3
ECD (cells/mm2)	2667.4 (279.5)	2337.3 (363.6)	2223.2 (398.5)	2182.4 (413.6)	2687.9 (232.7)	2408.1 (345.3)	2307.3 (366.1)	2241.8 (400.1)	0.34
CV	42.7 (6.4)	49.9 (8.5)	51.3 (8.8)	50.8 (8.5)	42.5 (6.7)	47.8 (6.6)	49.7 (6)	50.7 (7.2)	0.03
Hexagonality (%)	45.2 (8.3)	37.9 (8.9)	37.0 (8.4)	37.0 (8.2)	44.5 (6.5)	39.1 (5.9)	36.8 (5.3)	35.1 (5.7)	0.01
CCT (µm)	545.0 (43.2)	572.0 (48.3)	558.7 (45.1)	552.8 (44.2)	551.5 (39.3)	575.0 (40.2)	571.5 (40.9)	567.0 (38.5)	0.03

Data expressed as mean (SD). The data were analysed using two way repeated measures ANOVA. Postop 1 was measured at 7 days, postop 2 at 6 weeks and postop 3 at 3 months. Postop: Postoperative day, ECD: Endothelial cell density, CV: Coefficient of variation, CCT: Central corneal thickness, ANOVA: Analysis of variance, SD: Standard deviation

There was a significant increase in CV and CCT in diabetics after phacoemulsification when compared to controls (P = 0.03 for both CV and CCT). There was a significant decrease in the hexagonality in diabetics after phacoemulsification when compared to controls (P = 0.01). There was no statistically significant difference in the endothelial cell loss after phacoemulsification between the diabetic group and the control group though; there was a decrease in ECD in both the groups as shown in Table 3.

Table 3

Risk factors for endothelial cell damage

Parameters	Controls		Diabetics
	OR (CI)	P	OR (CI)	P
Age (years)
50-59
60-69	3.8 (1.111-13.024)	0.03*	1.1 (0.353-3.545)	0.84
≥70	1.2 (0.141-11.698)	0.82	5.7 (0.288-114.749)	0.25
Sex
Female
Male	3.3 (0.943-12.098)	0.06	0.8 (0.255-2.896)	0.80
Mydriasis (mm)
8
5	2.1 (0.103-43.899)	0.62	1.4 (0.137-15.727)	0.75
6	2.7 (0.443-16.705)	0.28	1.6 (0.168-15.500)	0.67
7	0.42 (0.105-2.592)	0.42	4.5 (0.094-222.124)	0.44
Inflammationa
0.5+
1+	1.8 (0.538-6.688)	0.31	5.7 (1.646-19.970)	0.006$
2+b	-	-	-	-
EPT (min)
<0.25
0.25-<0.50	1.7 (0.362-8.818)	0.47	1.9 (0.602-6.363)	0.26
≥0.50	8.8 (2.261-34.849)	0.002#	4.4 (0.937-21.293)	0.06

aBased on the SUN working group grading scheme for anterior chamber cells, bSample could not be analysed because of very less number in this group, *P<0.05 when compared to 50-59 years group, #P<0.05 when compared to 0.25 min group, $P<0.05 when compared to 0.5+. The data were analysed using multiple logistic regression. EPT: Effective phaco time, OR: Odds ratio, CI: Confidence interval, SUN: Standardization of uveitis nomenclature

Discussion

The novelty in this study of assessing the risk factors for endothelial cell loss proves that postoperative inflammation is the only factor responsible for endothelial cell loss in people with diabetes, necessitating the importance of reducing postoperative inflammation. As low as 1+ cells was a potential risk for endothelial cell loss in diabetics which proves that even minimal inflammation can cause endothelial cell loss. Contrary to the fact of poor mydriasis in diabetics, our study has also proved that good dilatation can be achieved in diabetics too and hence mydriasis is not a concern for diabetics and in turn for endothelial cell loss. Although phacoemulsification time is proportional to the endothelial loss, our study has proved that diabetics are not at increased risk of endothelial cell loss due to this factor. The type of phacoemulsification also influences the endothelial cell loss as studied by Wang et al.,[5] Reuschel et al.,[6] Park et al.,[7] and Park et al.[8] Lee et al.[13] compared the outcomes of phacoemulsification time on damage to the endothelial cells and found that it does not induce a difference in endothelial cell loss in controls and diabetics. However in our study, we compared the outcome of EPT on damage to endothelial cells in controls and diabetes. To the best of our knowledge, there are no studies comparing age, sex, EPT, and inflammation score with endothelial cell loss in controls and diabetics after phacoemulsification as done in the study conducted by us.

Sudhir et al.[14] studied the corneal morphology and ECD in diabetic patients and found that the mean corneal ECD (cells/mm2) was lower in diabetics than in nondiabetics. Lee et al.,[13] Hugod et al.,[15] Yang et al.,[16] and Sahu et al.[17] found a decrease in ECD in people with diabetes after phacoemulsification when compared to controls which were in contrast to the results obtained in our study. The follow-up period in the study done by Lee et al.[13] was 6 months and it was 3 months in the study done by Hugod et al.[15] A study done by Misra et al.[18] concluded that diabetes is not a factor for endothelial cell loss after phacoemulsification and that the lower subbasal nerve plexus density contributes to the endothelial loss. The study showed an increase in CV in diabetics after phacoemulsification when compared to controls, which was similar to the studies done by Lee et al.,[13] Yang et al.,[16] and Itoi et al.[19] but, Hugod et al.[15] reported no difference in CV among controls and diabetics. The finding of decreased hexagonality in people with diabetes after phacoemulsification in our study was similar to the studies done by Hugod et al.,[15] Yang et al.[16] and Itoi et al.[19] Shakya et al.[20] found an increase in CCT in diabetics after phacoemulsification when compared to controls, which was similar to the results obtained in our study. However, Hugod et al.[15] found no difference in CCT among controls and diabetics after phacoemulsification.

The reason for the lack of difference in endothelial cell loss in diabetics after phacoemulsification in comparison to controls can be attributed to factors such as the intraoperative mydriasis and EPT. In a study conducted by Zaczek and Zetterström,[21] they showed that miosis during surgery caused damage to the endothelium. However, in our study, there was no variation in mydriasis between diabetics and controls. It was also found that mydriasis was not a significant factor causing damage to the endothelial cells in both the groups. EPT was a noteworthy factor causing endothelial cell loss in controls, but not in diabetics. Hence, low EPT in diabetics could be one of the reasons for no difference in endothelial cell loss in diabetics when compared to controls (though inflammation was a significant factor causing endothelial cell loss in diabetics in our study).

The increase in CV and decrease in the hexagonality in diabetics shows that diabetic corneas take a longer time for the repair process compared to controls. The fact that CV and hexagonality has not stabilized over a period of 3 months shows the study needs further follow-up. The increase in CCT in the 1st postoperative week and a gradual decline in the values by 6 weeks and 3 months postoperatively can be attributed to the inflammatory score at 1st week, 6 weeks, and 3 months. As the inflammation decreases, the corneal edema also decreases, but the CCT did not reach the baseline values even by the end of 3 months.

To make up for the limitations of the previous studies, we analyzed the factors that can influence the endothelial cell loss in controls and diabetics, namely – age, sex, intraoperative mydriasis, EPT, and inflammatory score in the 1st postoperative week. The study takes into account a large sample size (80 controls and 80 diabetics) with the power of 80%. However, factors such as nuclear density, HbA1c, duration of diabetes, time from the onset of diabetes to cataract surgery were not studied.

The study provides evidence for the endothelial morphological changes in diabetics in comparison to controls (except ECD) and also for the factors associated with endothelial cell loss. Probably, 3 months follow-up was inadequate to prove a significant decrease in the ECD in diabetics and many more factors are to be studied to narrow down the cause of endothelial cell loss in diabetics.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.