FP1330 : Corneal Endothelial Changes in CRF Patients : Relevance To Cataract Surgeons

Dr. (Lt Col) Alok Sati, S09925, Dr.
(Brig) Parthasarathi Moulick

Abstract

Purpose: To evaluate corneal endothelial changes in chronic renal failure (CRF) patients.

Methods: Endothelial parameters (EP), measured by specular microscope, of the three groups (dialyzed group: 32 corneas, non-dialyzed group: 34 corneas, control group: 64 corneas) were compared. Influence on corneal endothelium of blood urea, serum creatinine, calcium, and phosphorus were analyzed.

Results: Using ANOVA and post hoc test, a significant difference was found in central corneal thickness (CCT) and endothelial cell density (CD), between control  and dialyzed groups [p<0.001 (CCT); p<0.001 (CD)], control and non-dialyzed groups  [p=0.023 (CCT); p=0.016 (CD)] and dialyzed and non-dialyzed groups [p=0.002 (CCT); p=0.007 (CD)]. Using linear generalized model, a significant correlation was found between blood urea and EP.

Conclusion: Corneal endothelial alteration is present in CRF patients, more marked with increasing blood urea level. Hence, cataract surgeons need to careful in such patients.

Introduction

Chronic kidney disease (CKD) refers to a spectrum of patho-physiologic processes associated with a gradual decline in glomerular filtration rate (GFR) in patients with abnormal kidney function.¹ Based on estimated GFR, CKD comprises of five stages.² The term CRF typically refers to CKD stages 3-5.¹ Stages 3 and 4, also referred to as non-dialyzed stages, are characterized by anemia, reduced appetite, abnormal calcium and phosphorus metabolism, electrolyte disturbance and acid-base imbalance. These stages are primarily treated by evaluating and treating complications and preparing for renal replacement therapy respectively.¹ Stage 5, also referred to as dialyzed stage, is characterized by accumulation of toxins in addition to the features of stages 3 and 4 and is managed by renal replacement therapy in form of hemodialysis.³

In both the non-dialyzed and the dialyzed stages, various ophthalmic complications have been reported including calcium deposits in conjunctiva, ^{4, 5} band-shaped keratopathy, ⁵ cataract, ⁶ retinal detachment and retinopathy.⁷ It has been postulated that the above complications might be attributed to the breakdown of homeostasis of body fluids including aqueous humor. As a result, various undesirable substances including urea get accumulated in aqueous humor. Since the aqueous humor is mainly responsible for nourishing the corneal endothelium, it is expected that the patients with CRF may have corneal endothelial abnormalities and the current study is primarily aimed to prove it.

Material & Methods

A total of 128 corneas of 128 subjects have been included in the study and three groups were formed. The first, the dialyzed group, includes 32 corneas of 32 patients with stage 5 of CKD, the second, the non-dialyzed group, includes 34 corneas of 34 patients with stages 3-4 of CKD and third, the age matched control group, includes 64 corneas of 64 healthy subjects. These three groups were examined by noncontact specular microscope (Topcon, Paramus, NJ, USA) and their endothelial parameters [central corneal thickness (CCT) in micron (µm), endothelial cell density (CD) in cells/mm², coefficient of variation (CV) in percentage (%) and hexagonality (Hex) in percentage (%)] were analyzed. The dialyzed group with an enhanced level of toxins in blood was further analysed to assess the influence of toxins like blood urea, serum creatinine, serum calcium and serum phosphorus on the corneal endothelium. The influence of duration of dialysis (≤ 3 years & >3 years) was also assessed in this group.

The above sample size i.e. 32 corneas of 32 patients in dialyzed group and 34 corneas of 34 patients in non-dialyzed group, has been calculated after considering an effect size of clinical interest (d) i.e. difference in means of endothelial cell density (CD) as 100 cells/mm², standard deviation (sd) as 135 (after pilot survey), Z alpha (Zα) as 1.96 (corresponding to type I error of 5% i.e. 0.05), Z beta (Zβ) as 0.84 (corresponding to power of 80%).

Patients with CRF and with history of ocular illnesses affecting the corneal endothelium, prior ophthalmic interventions, severe ocular trauma, contact lens wearers and systemic ailment like diabetes mellitus have been excluded from the study.

A complete ophthalmic evaluation was done in all patients including assessment of visual acuity by Snellen chart, anterior segment evaluation by slit lamp examination to rule ocular illnesses that might affect corneal endothelium, posterior segment evaluation by indirect ophthalmoscopy and +90D biomicroscopy to rule out posterior segment disorders and glaucoma,  intraocular pressure measurement by noncontact tonometer,  corneal thickness and endothelial count measurement by noncontact specular microscope.

Three images were taken for each cornea by the same operator at the same time of the day. An average was taken for each endothelial parameter from the three images and this average value was subsequently subjected to analysis.

Results

The mean age of patients in dialyzed, non-dialyzed and control groups was 36±09 years (range, 25 to 69 years), 36 ±10 years (range, 23 to 69 years) and 36 ±07 years (range, 23 to 63 years) respectively. On comparison using ANOVA, an insignificant difference (p=0.931) was observed amongst the three groups.

Comparison of endothelial parameters amongst dialyzed, non-dialyzed and age matched control groups

In Table 1 ANOVA test was used to compare the three groups. On comparison, it was found that the three groups differ significantly in terms of CCT (P<0.001) and CD (P<0.001) whereas insignificant difference was seen in terms of CV (a measure of pleomorphism) and % Hex (a measure of polymegathism). Though, in terms of CV, a trend of increasing variation in size was observed from control to non-dialyzed to dialyzed groups.

Table 2 has shown a pair wise comparison involving the three groups by using post hoc analysis (Tukey’s test). When pair wise comparison was done in terms of CCT, a statistically significant difference was observed between control and dialyzed groups (p<0.001), between control and non-dialyzed groups (p=0.023) and in between dialyzed and non-dialyzed groups (p=0.002). A similar observation was seen in terms of CD i.e. between control and dialyzed groups (p<0.001), between control and non-dialyzed groups (p=0.016) and in between dialyzed and non-dialyzed groups (p=0.007).

Influence of blood toxins and duration of dialysis on endothelial parameters in dialysed group

 Table 3 highlights the correlation between the blood toxins like blood urea, serum creatinine, serum calcium and serum phosphorus on the endothelial parameters. By using correlation test, it was found that there was a significant (p=0.006) but poor positive correlation between CCT and serum urea. This poor positive correlation has been depicted in Figure 1. Similarly, by using correlation test, it was found that there was a significant [CV (p=0.002), CD (p=0.022) and % Hex (p=0.026)] but poor negative correlation between CV, CD and % Hex with serum urea. This poor negative correlation has been depicted in Figures 2, 3 and 4. However, by using the same correlation test, it was found that there was no significant correlation between endothelial parameters with serum creatinine, serum calcium and serum phosphorus.

Table 4 demonstrates the influence of duration of dialysis on corneal endothelium and by using ANOVA test; it was found that there was no significant difference amongst the patients who have undergone ≤ 1 year, 1-3 years and > 3 years of hemodialysis.

Discussion

The present study has highlighted the endothelial alteration in patients with CRF. Amongst the various endothelial parameters analyzed including CCT, we have found that the patients undergoing hemodialysis exhibit reduced CD and increased CCT as compared to patients not undergoing dialysis. We further tried to elicit the impact of varying levels of altered blood components including toxins, on corneal endothelium and found that the patients with an increasing serum urea level exhibit more pronounced endothelial changes including CCT in the dialyzed group. Evaluation of CCT is important in a study reflecting endothelial parameters. If CCT is normal and the endothelial parameters like CD, CV or % Hex are abnormal, it reflects that inspite of corneal endothelium being affected, its function has remained normal.

Extensive MEDLINE search revealed scant literature related to endothelial alteration in patients with CRF. Ohguro et al, ⁸ analyzed the corneal endothelium by wide field microscopy in 20 corneas of 20 patients with CRF undergoing hemodialysis and compared with age matched control. They found that the patients of CRF undergoing hemodialysis exhibit polymegathism and pleomorphism whereas the CD remains normal. The above findings are in discordance with our results where a significant difference was seen in CD whereas no significant change was seen in terms of polymegathism and pleomorphism. However, we have observed an increased tendency towards pleomorphism in patients undergoing hemodialysis. This difference in results could be attributed to the difference in the number of patients analysed in the two studies. On the contrary, Couchoud et al.⁹ compared the dialyzed and non-dialyzed groups. In line with our results, they have found a statistically significant difference in CD between the two groups. However, no significant difference was seen in terms of polymegathism and pleomorphism. However, unlike our study, they have found that this reduction in CD was associated with the length of dialysis.

The above results including that of our study clearly indicate that there is an insult to the corneal endothelium. We hypothesize that this insult is a result of accumulation of multiple toxins in blood which in turn are raised in aqueous humor and cause endothelial alterations. The second factor could be an increased oxidative stress in patients with CRF. It is reflected by an enhanced plasma level of oxidized glutathione in addition to the reduction in level of vitamin E. As evident from our study, there lies no influence of serum calcium on endothelial parameters.  It is in contrast to the study by Jansen, ¹⁰ in which it has been stated that the needle-like crystals of calcium hydroxyapatite precursors were found intracellular in the corneal endothelium when viewed on electron microscope in patients with primary hyperthyroidism, a hypercacemic state similar to CRF. However, this contrast is secondary to the well controlled hypercalcemic state as seen in our patients due to administration of activated Vitamin D.

As highlighted above, a number of factors have been implicated by the different studies. We believe that it is not the single factor rather a conglomeration of multiple factors that is responsible for endothelial alteration in CRF patients.

On assessing the influence of various blood components on corneal endothelium, we found a significant increase in CCT and a significant decrease in CD with the increasing concentration of blood urea level. This change is being expected by us as we have hypothesized above, urea being an important and measurable toxin in CRF patients. However, we did not find the similar changes with increasing levels of creatinine, calcium and phosphorus including the duration of dialysis. .

The uniqueness of the current study lies in the fact that this is the only study existing in the literature that compares the three groups i.e. patients undergoing  hemodialysis , patients not undergoing dialysis and the age matched healthy subjects. Secondly, this study makes us aware that we should be careful in dealing with the eyes of the patients with CRF especially the intraocular procedure like phacoemulsification. Like any other study, this study is not without limitations. Firstly, due to logistic constraints, we failed to evaluate the influence of serum parathyroid hormone, one of the blood component in dialyzed patient, on the corneal endothelium. Secondly, the different stages of CKD may affect the results and conclusion.  Thirdly, the influencing presumption of toxins especially urea, is solely based on previous studies, to make the study more authentic, it is better to estimate the aqueous urea level and to evaluate its influence on corneal endothelium. This arm can be performed in the future studies.

References

1. Bargman JM. Chronic kidney disease. In: Fauci AS, ed. Harrison’s principles of internal medicine. New York: McGraw Hill Proessions, 2008:1811.
2. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, classification and stratification. Am J Kidney Dis 2000;39:2000.
3. Bargman JM. Chronic kidney disease. In: Fauci AS, ed. Harrison’s principles of internal medicine. New York: McGraw Hill Proessions, 2008:1822.
4. Easterbrook M, Mortimer CB. Ocular signs in chronic renal failure. Br J Ophthalmol 1970; 54:724.
5. Porter R, Crombie AL. Corneal and conjunctival calcification in chronic renal failure. Br J Ophthalmol 1973; 57:339-343.
6. Berlyne GM, Danovitch GM, Ari JB, et al. Cataracts of chronic renal failure. The Lancet 1972; 299:509-511.
7. Mullaem G, Rosner MH. Ocular problems in the patients with end stage renal disease. Semin Dial 2012; 25:403-407.
8. Ohguro N, Matsuda M, Fukuda M. Corneal endothelial changes in patients with chronic renal failure. Am J Ophthalmol 1999; 128:234-236.
9. Diaz CP, Bordas FD, Garcia JRF, et al. Corneal Disease in Patients with chronic renal insufficiency undergoing hemodialysis. Cornea 2001; 20: 695-702.
10. Jensen OA. Ocular calcifications in primary hyperparathyroidism. Histochemical and ultrastructural study of a case. Comparison with ocular calcifications in idiopathic hypercalcaemia of infancy and in renal failure. Acta Ophthalmol 1975; 53:173-186

Table 1. Comparison of endothelial parameters amongst the three groups

Parameter	Control (n=62)			Dialysed (n=32)			Non Dialysed (n=34)			p-value
	Mean	SD	Range	Mean	SD	Range	Mean	SD	Range
CCT (µm)	506	29	430 – 537	549	30	500 – 598	524	27	470 – 565	< 0.001*
CV (%)	35.94	3.77	28 – 45	38.00	5.28	28 – 47	36.07	3.60	29 – 42	0.064
CD (cells/mm2)	2760	304	2332 – 3506	2337	324	1802 – 2856	2574	260	2066 – 3037	< 0.001*
Hex (%)	51.56	6.43	42 – 66	48.50	9.10	32 – 68	52.23	9.96	36 – 68	0.136

CCT: Central corneal thickness, CV: Coefficient of variation, CD: Cell density, Hex:  Percentage of Hexagonality3

*Significant (p-value < 0.05)

Table 2. Pair wise comparison of endothelial parameters in three groups

		Control	Dialysed	Non Dialysed
CCT (µm)	Control	–	< 0.001*	0.023*
	Dialysed	–	–	0.002*
	Non Dialysed	–	–	–
CV (%)	Control	–	0.062	0.989
	Dialysed	–	–	0.166
	Non Dialysed	–	–	–
CD (cells/mm2)	Control	–	< 0.001*	0.016*
	Dialysed	–	–	0.007*
	Non Dialysed	–	–	–
Hex (%)	Control	–	0.192	0.926
	Dialysed	–	–	0.169
	Non Dialysed	–	–	–

CCT: Central corneal thickness, CV: Coefficient of variation, CD: Cell density,  Hex:  Percentage of Hexagonality, *Significant (p-value < 0.05)

Table 3. Linear correlation between endothelial parameters with blood parameters for dialysed group.

		Blood Urea	Serum creatinine	Serum calcium	Serum phosphorus
CCT (µm)	Pearson Correlation coefficient (r)	0.474	0.300	0.349	0.078
	p-value	0.006*	0.095	0.050	0.672
	N	32	32	32	32
CV (%)	Pearson Correlation coefficient (r)	-0.519	-0.291	0.104	0.007
	p-value	0.002*	0.106	0.573	0.968
	N	32	32	32	32
CD (cells/mm2)	Pearson Correlation coefficient (r)	-0.404	0.016	0.156	-0.170
	p-value	0.022*	0.933	0.392	0.353
	N	32	32	32	32
Hex (%)	Pearson Correlation coefficient (r)	-0.394	-0.204	-0.231	0.194
	p-value	0.026*	0.263	0.204	0.288
	N	32	32	32	32

CCT: Central corneal thickness, CV: Coefficient of variation, CD: Cell density, Hex: Percentage of Hexagonality, *Significant (p-value < 0.05)

Table 4. Influence of duration of dialysis on endothelial parameters

	≤ 1 year (n=24)		1 – 3 year (n=5)		> 3 years (n=3)		p-value
	Mean	SD	Mean	SD	Mean	SD
CCT (µm)	548	32	546	27	564	25	0.715
CV (%)	39.08	5.42	37.80	2.79	38.00	2.65	0.648
CD (cells/mm2)	2321	338	2397	278	2371	383	0.884
Hex (%)	47.79	8.11	50.80	15.21	50.33	6.35	0.758