Erythrocyte Sodium–Lithium Countertransport and Proliferative Diabetic Retinopathy

Abstract

purpose. To investigate whether elevated erythrocyte Na⁺/Li⁺ countertransport (Na⁺/Li⁺ CT) activity is present in patients with proliferative diabetic retinopathy (PDR).

methods. The rate of Na⁺/Li⁺ CT activity assayed in 21 patients with type 1 diabetes mellitus (DM) presenting PDR was compared with 10 patients with nonproliferative retinopathy (NPDR) and with 11 patients with normal fundi. Twelve normal volunteers with no family history of hypertension were used as a control group. The albumin excretion rate was determined by nephelometry, and the glomerular filtration rate was measured by the plasma clearance of eidetic acid labeled with chromium-51.

results. Patients with PDR showed higher diastolic blood pressure levels (mean ± SD) compared with those with NPDR or normal fundi (95 ± 13 versus 90 ± 09 and 82 ± 19 mm Hg, P = 0.02, respectively). The albumin excretion rate was higher [geometric mean (range)], and the glomerular filtration rate was lower (mean ± SD) in patients with PDR than in those with NPDR or normal fundi [333 (2 to 5140) versus 32 (5.9 to 2200) and 6 (1.5 to 306) μg/min, P = 0.01, and 63 ± 33 versus 99 ± 37 and 93 ± 43 ml/min, P = 0.02, respectively]. The mean Na⁺/Li⁺ CT in patients with PDR was significantly higher than in patients with NPDR or normal fundi and control group (0.46 ± 0.20 versus 0.32 ± 0.12, 0.32 ± 11, and 0.21 ± 0.07 mM/L red blood cells (RBC)/h, respectively, P = 0.0001). In a multiple logistic regression analysis, with PDR as the dependent variable, Na⁺/Li⁺ CT (odds ratio [OR]: 4.7, confidence interval [CI]: 1.2–17.6, P = 0.02), diastolic blood pressure (OR, 3.4; CI, 1.3 to 9.6; P = 0.018), and glomerular filtration rate (OR, 5.1; CI, 1.6–17.7; P = 0.007) were the only variables that were maintained in the equation, indicating that they were the main determinants of PDR.

conclusions. Patients with type 1 DM and proliferative retinopathy have elevated erythrocyte Na⁺/Li⁺ CT.

Sodium–lithium countertransport activity (Na⁺/Li⁺ CT) in red blood cells has been found to be abnormal in subjects with essential hypertension. ¹ The role of Na⁺/Li⁺ CT in cell physiology is still not clear, although this transport bears similarities to the physiological sodium–hydrogen exchanger. ² The Na⁺/H⁺ exchanger is a membrane transporter that regulates intracellular pH, cellular volume, and growth and bicarbonate reabsorption by the proximal tubules in the kidney. ² In patients with type 1 diabetes mellitus (DM), cross-sectional studies have demonstrated that Na⁺/Li⁺ CT is associated with micro or macroalbuminuria. ^{3 4 5} This finding was recently confirmed by a prospective study. ⁶ Because this abnormality of cell membrane transport is probably the best reproducible cellular cation transport abnormality phenotype associated with essential hypertension in white individuals, ¹ it has been suggested that a genetic predisposition to essential hypertension is an important factor in the susceptibility to diabetic renal disease. ^{3 4 5} To the best of our knowledge, no study has investigated whether elevated Na⁺/Li⁺ CT is present in patients with proliferative diabetic retinopathy (PDR). 

The aim of the present study was to compare the rate of erythrocyte Na⁺/Li⁺ CT activity in patients with type 1 DM presenting with PDR to that observed in patients with nonproliferative retinopathy (NPDR) or normal fundi, and to explore the interactions between this abnormality in cellular transport and other risk factors for PDR. 

Methods

Patients

We estimated that to detect a difference of 0.20 ± 0.18 mM/L red blood cells (RBC)/h, which has been reported in the literature for white diabetic patients with and without microvascular complications, ³ a sample size of 17 patients in the studied group will have a 90% power to detect that difference with a two-tailed significance level of 0.05. 

From January 1997 to July 1998, consecutive patients with type 1 DM attending the diabetic outpatient clinic at the University Hospital of the State University of Campinas were asked to participate in the study. The criteria for inclusion in the study were patients with type 1 DM, defined as age at diagnosis less than 35 years, a history of sudden onset of severe hyperglycemic symptoms, marked weight loss, spontaneous sustained ketosis or ketonuria, an age ranging from 18 to 45 years old, white, with DM for at least 10 years (to allow enough time for developing of diabetic retinal disease) and free of any endocrine, hepatic, metabolic, or cardiac disease, and nondiabetic renal disease. The exclusion criteria were pregnancy and use of contraceptives or estrogen. Such criteria were deemed appropriate to exclude patients with factors that may influence the determination of Na⁺/Li⁺ CT. ⁷ Patients with high myopia, chorioretinitis scars, posterior uveitis, and glaucoma were also excluded, as were those who had undergone a previous ocular surgery, because such events may influence the development of diabetic retinopathy. ⁸ Of the potential 80 patients for the study attended at the outpatient clinic, 38 (48%) did not meet the inclusion and exclusion criteria. The remaining 42 patients gave their informed consent before participating in the study. Twelve normal volunteers with no family history of hypertension were used as a control group. The study was carried out in conformity with the tenants of the Declaration of Helsinki, which was approved by the Ethics Committee of the University Hospital. All patients were treated with insulin, and 17 (41%) who were taking antihypertensive drugs were asked to discontinue them for 48 hours before the protocol. 

Clinical and Ophthalmologic Measurements

The patients attended the outpatient clinic after an overnight fast. Blood pressure (BP) was measured twice, in the supine position, by a single observer (JMLF), to the nearest 2 mm Hg, using a standard cuff, 5 minutes apart after the subject had been resting for at least 5 minutes. Mean systolic and diastolic (fifth Korotkoff’s sound) blood pressures were obtained by averaging the two measurements. A blood venous sample was taken for the measurements of RBC Na⁺/Li⁺ CT activity and biochemical analysis. A complete ophthalmologic examination was then performed, including indirect fundus ophthalmoscopy with stereoscopic color fundus photographs of seven standard fields of both eyes. The stereoscopic color photographs were examined by a single observer (JMLF) unaware of the patient’s Na⁺/Li⁺ CT status. The level of DR was classified according to the Early Treatment Diabetic Retinopathy Study retinopathy severity classification. ⁹ The eye with the most severe level of DR in each subject was considered for analysis. ¹⁰ Eyes presenting level 1, levels 2 to 5, and levels 6 to 8 were grouped into normal fundi, NPDR, and PDR, respectively. 

Methodology

RBC Na⁺/Li⁺ CT activity was measured according to the original method described by Canessa et al., ¹ as previously reported by our laboratory. ¹¹ In our laboratory, the normal range for normotensive subjects with negative familial history of hypertension (n = 12) was 0.21 ± 0.07 mM/L RBC/h, ranging from 0.12 to 0.38 mM RBC/h. ¹¹ The intraindividual and interassay variations were approximately 9.2%, a value similar to that reported previously. 

Other Measurements.

Three consecutive timed overnight urine samples per patient were collected for albumin measurement. Albumin concentrations were determined by nephelometry. The median value of the three specimens was used for classifying the patients into three categories: normoalbuminuric (albumin excretion rate [AER] ≤ 20 μg/min), microalbuminuric (20 < AER ≤ 200 μg/min), and macroalbuminuric (AER > 200 μg/min). Glomerular filtration rate (GFR) was measured by the plasma clearance of eidetic acid labeled with chromium-51 (⁵¹CR-EDTA). For comparison between patients, GFR was corrected for body surface area (1.73 m²). The percentage of the A_1c component of glycosylated hemoglobin was assessed by high-performance liquid chromatography. Normal values in our laboratory were 5.6% ± 1.3%. Serum creatinine, total cholesterol, and triglycerides were measured by an automated method. In our laboratory, the normal values for serum cholesterol and triglycerides were up to 5.20 and 2.30 mM, respectively. 

Statistical Analysis.

Clinical data and laboratory determinations were compared for significance using ANOVA, Fisher and Scheffé F tests from a commercial software package (SPSS for Power Macintosh). Because of the interrelationship of several variables, the univariate logistic analysis was followed by multiple logistic regression analysis, with PDR as the dependent variable. The results were expressed as the mean ± SD, unless otherwise stated. The null hypothesis was rejected below the conventional (two-tailed) 0.05 level. 

Results

Among the studied group, 21 patients had PDR, 10 patients presented NPDR, and 11 patients had normal fundi. The demographic and laboratory characteristics of the patients are summarized in Table 1 . The three groups of patients had similar distributions for sex, age, body mass index, duration of DM, and systolic BP. However, patients with PDR had diastolic BPs significantly higher than those with NPDR and normal fundi (95 ± 13 versus 90 ± 09 and 82 ± 19 mm Hg, P = 0.02, respectively) and were more likely to be on antihypertensive therapy (P = 0.001; Table 1 ). The metabolic control, evaluated by HbA_1c levels, was similar in all groups of studied patients (Table 1) . AER (geometric mean [range]) was higher, and GFR (mean ± SD) was lower in patients with PDR than in those with NPDR and normal fundi (333[ 2–5140] versus 32 [5.9–2200] and 6 [1.5 to 306] μg/min, P = 0.01; and 63 ± 33 versus 99 ± 37 and 93 ± 43 ml/min, P = 0.02, respectively). The frequency of patients with micro or macroalbuminuria was higher in patients with PDR (80%) than in those with NPDR (40%) and normal fundi (18%) PDR, P = 0.001. In addition, the mean level of the total cholesterol was significantly higher in patients with PDR than in the other groups (6.53 ± 1.80 versus 5.10 ± 0.80 and 4.8 ± 1.2 mM, P = 0.008). 

Figure 1 shows the individual values of the erythrocyte Na⁺/Li⁺ CT activity. The mean value for Na⁺/Li⁺ CT activity in patients presenting PDR was significantly higher than that observed in patients with NPDR, normal fundi, or control group (0.46 ± 0.20 versus 0.32 ± 0.12, 0.32 ± 11, and 0.21 ± 0.07 mM/L RBC/h, respectively, P ± 0.0001). This difference in Na⁺/Li⁺ CT activity remained significant even when the micro and macroalbuminuric patients were excluded from the calculation (0.55 ± 0.29 versus 0.32 ± 0.09 and 0.34 ± 0.12 mM RBC/h, P < 0.05 for patients with PDR, n = 4; NPDR, n = 6; and normal fundi, n = 9; respectively), suggesting that the difference in mean Na⁺/Li⁺ CT activity cannot be accounted for solely by this group of patients. The presence of laser therapy had no effect on Na⁺/Li⁺ CT because the activity of this cation transport was similar in patients with PDR that had laser therapy (n = 8) and those without laser therapy (n = 13; 0.48 ± 0.21 versus 0.42 ± 0.20 mM/L RBC/h, P = 0.60). The Na⁺/Li⁺ CT activity was above the upper limit of normal (>0.38 mM RBC/h) in 12 of 21 patients with PDR (58%), in 2 of 10 patients with NPDR (20%), and in 2 of 11 patients with normal fundi (18%; P = 0.03). 

To determine the relationship between PDR and other variables (serum creatinine, AER, GFR, cholesterol, BP, and Na⁺/Li⁺ CT activity) in the univariate analyses, binary logistic regression models were used while controlling for multiple potential cofounders. Subsequently, a multiple logistic regression analysis was carried out, with PDR as the dependent variable. In this model, Na⁺/Li⁺ CT (odds ratio[ OR]: 4.7, confidence interval [CI]: 1.2–17.6; P = 0.02), diastolic BP (OR, 3.4; CI, 1.3–9.6; P = 0.018) and glomerular filtration rate (OR, 5.1; CI, 1.6–17.7; P = 0.007) were the variables that were maintained in the equation, indicating that they were the main determinants of the presence of PDR (Table 2) . 

Discussion

We have demonstrated, for the first time, that patients with type 1 DM and proliferative retinopathy have elevated erythrocyte Na⁺/Li⁺ CT activity. The more severe retinopathy in these patients with increased Na⁺/Li⁺ CT activity cannot be attributed either to the duration of diabetes or to the level of metabolic control as indexed by glycohemoglobin values because these parameters did not differ between the three group of patients. This observation is important because it may provide further insight into the pathogenesis and susceptibility to this complication of DM.  

At present, the nature of the association between PDR and elevated Na⁺/Li⁺ CT is unclear. We have observed that patients with type 1 DM, PDR, and elevated Na⁺/Li⁺ CT also have a higher frequency of diabetic nephropathy (micro/macroalbuminuria). Although this finding represents a confounding factor, this is not surprising because in these patients a close association between PDR and micro and macroalbuminuria, abnormalities associated with elevated Na⁺/Li⁺ CT, ^{3 4 5} has been established by large epidemiologic studies. ¹² However, we believe that the increased Na⁺/Li⁺ CT observed in our patients cannot be accounted for solely by the presence of nephropathy. In agreement with this hypothesis, the mean Na⁺/Li⁺ CT rates remained higher in patients with PDR even when patients with micro or macroalbuminuria were eliminated from the analysis, although the number of patients in the group with PDR was quite small. 

In summary, patients with type 1 DM presenting PDR have a higher mean erythrocyte Na⁺/Li⁺ CT activity than patients without PDR. Further studies are needed with a larger number of patients to determine the validity of this preliminary observation. 

 Supported by a grant from the FAEP/UNICAMP, Brazil. JMLF is the recipient of a scholarship from CAPES, Brazil.

 Submitted for publication April 26, 1999; revised August 9 and December 1, 1999; accepted December 30, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Clinical and Laboratory Characteristics of the Study Patients

Variables	With PDR	With NPDR	Normal Fundi	P
Sex, M:F	11:10	7:3	6:5	0.75
Age, y	33 ± 9	37 ± 7	37 ± 8	0.28
BMI, kg/m2	23.9 ± 4.2	24.7 ± 3.5	25.2 ± 3.7	0.63
Duration of DM, y	18 ± 6	18 ± 5	15 ± 4	0.29
Antihypertensive therapy, P:A	11:10	5:5	1:10	0.001*
Systolic BP, mm Hg	141 ± 34	141 ± 18	127 ± 19	0.33
Diastolic BP, mm Hg	95 ± 13	90 ± 09	82 ± 19	0.02*
Glycated hemoglobin A1c, %	9.3 ± 2	8.1 ± 2.5	8.4 ± 2.7	0.32
Serum creatinine, μmol/L	124 ± 79	74 ± 12	70 ± 10	0.02, †
Cholesterolemia, mmol/L	6.53 ± 1.80	5.10 ± 0.80	4.8 ± 1.20	0.008, †
Triglyceridemia, mmol/L	4.8 ± 4.8	2.5 ± 1.3	5.4 ± 6.5	0.36
AER, μg/min, ‡	333 (2–5140)	32 (5.9–2200)	6 (1.5–306)	0.01, †
Normo/micro/macroalbuminuria, n	4/1/16	6/2/2	9/1/1	0.001
GFR, mL/min	63 ± 33	99 ± 37	93 ± 43	0.02, †

 M, male; F, female; P, presence; A, absence for antihypertensive therapy; BMI, body mass index = [weight (kg)/height² (m²)].

*  Comparison between PDR and normal fundus.

†  Comparison between PDR and normal fundus and PDR and NPDR.

‡  Geometric mean (range).

Figure 1.

 

Rates of RBC Na⁺/Li⁺ CT in 21 patients with type 1 DM with PDR, 10 patients with NPDR, in 11 patients with normal fundi, and in 12 normal volunteers without family history of essential hypertension used as a control group.

Table 2.

 

Variables that Were Main Determinants of PDR in the Multiple Logistic Regression Model

Variables	OR	95% CI
Na+/Li+ CT	4.7	1.2–17.6
Diastolic BP	3.4	1.3–9.6
GFR	5.1	1.6–17.7

The authors thank the Clinical Pathology Laboratory for the biochemical analyses, the Department of Nuclear Medicine for measurement of glomerular filtration rates with use of ⁵¹Cr-labeled EDTA, and Stephen Hyslop for editing the manuscript.