Prospective Analysis of Traits Related to 6-Year Change in Sodium-Lithium Countertransport
Abstract—Sodium-lithium countertransport (Na-Li CT) activity in red blood cells relates cross-sectionally and longitudinally to blood pressure and hypertension. Lifestyle and metabolic factors relate cross-sectionally to this sodium transporter. The aim of this study was to conduct a prospective analysis of 6-year Na-Li CT change and of traits related to Na-Li CT change. In 2183 participants in the Gubbio Population Study (972 men and 1211 women; baseline ages, 18 to 74 years), the following data collected at baseline and 6-year follow-up were analyzed: Na-Li CT; gender; age; body mass index (BMI); blood pressure; antihypertensive treatment; alcohol intake; smoking habits; urinary sodium-to-potassium ratio; and plasma cholesterol, glucose, uric acid, sodium, potassium, and triglycerides (measured only at follow-up). Six-year changes were defined as follow-up minus baseline values. Na-Li CT was higher at follow-up than at baseline in both genders (P<0.001). Baseline Na-Li CT; baseline and change values of BMI; and change values of alcohol intake, plasma potassium, and plasma glucose related to Na-Li CT change significantly and independently with control for other variables. Follow-up plasma triglyceride levels also related independently to Na-Li CT change. Coefficients were positive for BMI, alcohol intake, and plasma glucose and triglyceride levels and were negative for baseline Na-Li CT and plasma potassium levels. Baseline and change values of other variables did not relate significantly to Na-Li CT change. In conclusion, in prospective analyses, BMI, alcohol intake, plasma glucose, and lipids were directly related to Na-Li CT change; baseline Na-Li CT and plasma potassium levels were inversely related. The data support the concept that lifestyle and related metabolic factors influence Na-Li CT.
- sodium-lithium countertransport
- body mass index
- plasma glucose
- plasma lipids
- plasma potassium
- Gubbio Population Study
Sodium-lithium countertransport (Na-Li CT) is a membrane transport system involving a one-to-one exchange of sodium for lithium, usually measured as sodium-dependent lithium efflux in lithium-loaded red blood cells (RBCs). Na-Li CT is considered a marker of hypertension proneness because its activity relates to blood pressure (BP),1 2 BP change over time,3 and the prevalence and incidence of hypertension.2 4 5 Na-Li CT relates also to diabetic nephropathy in insulin-dependent diabetes mellitus, possibly through its association with high BP.6 The physiological role of Na-Li CT is not defined. There is circumstantial evidence that Na-Li CT activity relates to sodium proton exchange.7 8 Conflicting data are reported on the relation of Na-Li CT to sodium reabsorption at the renal tubular level.9 10 Na-Li CT activity is influenced by genetic,11 12 13 14 environmental, and metabolic factors.14 15 16 17 Cross-sectional population-based data suggest that Na-Li CT activity increases over time because the means of Na-Li CT in adults are progressively higher with age.16 Hunt et al18 reported that Na-Li CT increased slightly after 30-month follow-up and that Na-Li CT change related to change in body mass index (BMI), plasma triglycerides, and cholesterol. The present study investigates these matters prospectively using data collected at entry and 6-year follow-up in the Gubbio Population Study.
The Gubbio Population Study is an ongoing cross-sectional and prospective investigation in the hill town of Gubbio in north central Italy. Time, setting, methods, and laboratory procedures with technical error have been described in previous articles for baseline examination and 6-year follow-up.2 3 5 16 19 20 At baseline and follow-up, data collection included Na-Li CT activity in RBCs; gender; age; BMI (weight in kilograms/height in meters squared); BP; antihypertensive treatment status; alcohol intake; smoking habits; urinary sodium-to-potassium ratio (taken as an index of dietary salt and potassium intake); and plasma concentrations of glucose, uric acid, sodium, potassium, total cholesterol, and high-density lipoprotein (HDL) cholesterol (used to calculate non-HDL cholesterol as total minus HDL cholesterol). Plasma triglyceride levels were also measured at follow-up only.
The cohort for this report is 2183 individuals (972 men and 1211 women), baseline ages 18 to 74 years, with valid Na-Li CT determination2 16 at baseline and follow-up. Baseline values for plasma glucose and uric acid are missing for 174 men and 206 women of the cohort, as these determinations began in the sixth month of the baseline examination.19 For all variables, change values were defined as follow-up minus baseline value. Statistical procedures included the Wilcoxon test; product-moment, rank, and partial correlation analyses; ANOVA; and multivariate linear regression.
Previous articles reported response rates at baseline and follow-up examinations, characteristics of participants at baseline and follow-up (with and without valid Na-Li CT determinations), descriptive statistics for hypertension (prevalence and incidence), and use of antihypertensive drugs.2 3 5 19 20 Table 1⇓ shows baseline, follow-up, and 6-year change for Na-Li CT and other variables. Follow-up plasma triglyceride levels were 166±125 and 127±68 mg/100 mL (mean±SD) for men and women, respectively. In men and women, respectively, distribution of Na-Li CT was positively skewed for baseline values (+1.12 and +1.53) and follow-up values (+0.83 and +1.08), but not for change values (–0.050 and –0.212). Therefore, baseline and follow-up Na-Li CT values were log-transformed in regression analyses. Log-transformed Na-Li CT values (mean±SD) were (men and women, respectively) 2.48±0.19 and 2.39±0.19 μmol/L RBC · h−1 at baseline and 2.51±0.18 and 2.46±0.17 μmol/L RBC · h−1 at follow-up. Because of skewed distributions, alcohol intake and cigarettes per day were also log-transformed. Log-transformed alcohol intake for nondrinkers and log-transformed cigarettes per day for nonsmokers were defined as zero, as previously reported.3 5
Change and Tracking Over Time, Na-Li CT, and Other Variables
Na-Li CT was higher at follow-up than baseline in men and women (P<0.001 by the Wilcoxon test), and average 6-year Na-Li CT change showed a significant increase in both genders. With an average duration of follow-up of about 6 years in both genders (shown in Table 1⇑ as age change), Na-Li CT change per year of follow-up averaged +4.7±26.7 and +6.9±23.7 μmol/L RBC · h−1 in men and women, respectively. Baseline and follow-up Na-Li CT (log-transformed values) were directly correlated in men and women (r=0.415 and 0.408, P<0.001); findings were similar for nontransformed Na-Li CT in rank correlation analyses (r=0.425 and 0.425, P<0.001). Baseline Na-Li CT (log-transformed values) significantly and inversely related to 6-year Na-Li CT change in men and women (r-0.527 and –0.541, P<0.001); findings were similar with the use of nontransformed baseline Na-Li CT (not shown).
Table 2⇓ shows baseline, follow-up, and change for Na-Li CT by quintiles of baseline Na-Li CT. For quintiles 1 through 3 in men and quintiles 1 through 4 in women, follow-up Na-Li CT was higher than baseline; for quintiles 4 and 5 in men and quintile 5 in women, follow-up Na-Li CT was lower than baseline. In both genders, these Na-Li CT changes reflected regression toward the mean. Na-Li CT change was linearly lower over the quintiles: with small change for quintiles 2 through 4, large increases for quintile 1, and large decreases for quintile 5. Difference in mean Na-Li CT between extreme quintiles (quintile 5 minus quintile 1) decreased about 50% from baseline to follow-up in men (from 387 to 173 μmol) and women (from 333 to 139 μmol).
Mean values were higher at follow-up than baseline for BMI; prevalence of antihypertensive drug treatment; urinary Na-K ratio; and plasma cholesterol, glucose, sodium, and potassium; mean values were lower for BP and plasma uric acid. Also for other variables, baseline values related directly to follow-up values and inversely to 6-year change (not shown; P<0.001).
Relation of Other Variables to 6-Year Na-Li CT Change: Analyses Controlled for Baseline Na-Li CT
Partial correlation analyses with control for baseline Na-Li CT (log-transformed) were used to study the correlation of baseline and change values of other variables with 6-year Na-Li CT change. In these analyses, positive coefficients indicate that high (positive) values of the variable related to Na-Li CT increase; negative coefficients indicate that high (positive) values of the variable related to Na-Li CT decrease. For baseline values, coefficients were significant (P<0.05) and positive in both genders for BMI (men and women, r=0.066 and 0.085) and alcohol intake (r=0.108 and 0.093), and in women only, for plasma uric acid (r=0.079) and total (r=0.088) and non-HDL (r=0.099) cholesterol. For change values, coefficients were significant and positive in both genders for BMI (r=0.130 and 0.048) and plasma glucose (r=0.151 and 0.053), and in men only, for diastolic BP (r=0.057). Coefficients were significant and negative in both genders for cigarettes per day (r=−0.061 and –0.063) and plasma potassium (r=−0.082 and –0.052), and in women only, for age (r=−0.080) and plasma HDL cholesterol (r=−0.070). For follow-up plasma triglycerides, coefficients were positive and significant in men and women (r=0.145 and 0.153, respectively; P<0.001). For other variables coefficients were not significant with use of baseline and change values.
With control for baseline Na-Li CT, 6-year Na-Li CT change was not significantly different between individuals with and without hypertension (systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg and/or antihypertensive drug treatment) at baseline or incident at follow-up (in analyses with or without hypertensives on drug treatment), as well as between hypertensive individuals with and without antihypertensive drug treatment at baseline or incident at follow-up (not shown, P>0.3). Separate analyses were done also for individuals treated with angiotensin-converting enzyme inhibitors (as a single drug or in addition to other drugs) because of the reported effects of angiotensin-converting enzyme inhibitors on Na-Li CT.21 Incidence of treatment with angiotensin-converting enzyme inhibitors at follow-up was reported by 105 hypertensive individuals (47 men and 58 women); this group was compared with the group of 695 hypertensive individuals (303 men and 392 women) untreated or treated with other drugs. With control for baseline Na-Li CT, mean 6-year Na-Li CT change was not significantly different between the two groups of men (–9.2 and +22.0 μmol/L RBC · h−1, P=0.139) and women (+14.2 and +29.2 μmol/L RBC · h−1, P=0.373). Findings were similar in analyses with inclusion of nonhypertensive individuals (not shown).
Relation of Other Variables to 6-Year Na-Li CT Change: Multiple Regression Analyses
Tables 3⇓, 4⇓, and 5⇓ show regression coefficients from different multivariate linear models with 6-year Na-Li CT change (dependent variable) regressed on other variables (independent variables). Analyses were done for men and women separately and for men and women combined with control for gender. For other variables, models in Table 3⇓ include the baseline values; models in Table 4⇓, the change values; and models in Table 5⇓, baseline and change values. Log-transformed baseline Na-Li CT and baseline age were included among the independent variables in all models. For individuals with missing values for baseline plasma glucose and uric acid (174 men and 206 women), gender-specific means were used in analyses.
For analyses with baseline values (Table 3⇑), BMI and alcohol intake related significantly to 6-year Na-Li CT change in both genders; plasma non-HDL cholesterol related significantly in women but not men.
For analysis with change values (Table 4⇑), BMI, plasma glucose, and plasma potassium related significantly to 6-year Na-Li CT change in both genders; cigarettes per day and plasma HDL cholesterol related significantly in women but not men.
Baseline and Change Values
For analyses with baseline and change values together (Table 5⇑), baseline and change in BMI, baseline and change in alcohol intake, baseline and change in plasma potassium, and change (not baseline) of plasma glucose related significantly or with borderline significance to 6-year Na-Li CT change in men and women. In women only, baseline plasma non-HDL cholesterol, change in cigarettes per day, and change in plasma HDL cholesterol related to 6-year Na-Li CT change.
Findings in Tables 3 through 5⇑⇑⇑ were similar with the exclusion of 174 men and 206 women who had missing values for baseline plasma glucose and uric acid, or with use of diastolic instead of systolic BP, or with exclusion of individuals on antihypertensive drug treatment. In additional models similar to those shown in Table 5⇑ with inclusion of follow-up plasma triglycerides among the independent variables, coefficients for plasma triglycerides were significant (P<0.002) for men (+0.118), women (+0.208), and men and women combined (+0.136). For other variables, coefficients were similar to those shown in Table 5⇑ except for those of plasma HDL and non-HDL cholesterol, which were not significant.
Relation of BMI and Alcohol Intake to 6-Year Na-Li CT Change
With the use of coefficients shown in Table 5⇑ for men and women combined, 6-year Na-Li CT change was estimated for individuals in the population with BMI and alcohol intake, both baseline and change values, higher by about 1 SD. Compared with individuals with BMI of 23 kg/m2 at baseline and follow-up (BMI change=0), nonalcohol drinker at baseline and follow-up (baseline alcohol intake and change in alcohol intake=0), and with control for other variables, 6-year Na-Li CT change was estimated to be 59.0 μmol higher for individuals with BMI of 27 kg/m2 at baseline and +2 kg/m2 change, and with alcohol intake of 6 g/d at baseline and +6 g/d change (0.8 log-transformed g/d).
To analyze 6-year Na-Li CT change in people with similar traits and without adjustment for other variables, two subgroups, defined as high and low, were selected with differences in BMI and alcohol intake, baseline and change values, similar to those indicated above. In the high subgroup (22 men and 39 women), baseline BMI was 26.8±0.5 kg/m2 (mean±SEM); BMI change, +2.00±0.15 kg/m2; baseline alcohol intake, 7.5±0.2 g/d; and alcohol intake change, +6.5±0.1 g/d. In the low subgroup (24 men and 94 women), baseline BMI was 23.1±0.2 kg/m2; BMI change, –0.12±0.11 kg/m2; and baseline and change in alcohol intake, 0±0 g/d (ie, nonalcohol drinker at baseline and follow-up). In analysis for men and women combined, 6-year Na-Li CT change differed by 43.7±20.9 μmol between the high and low subgroups (+64.9±16.0 and +21.2±13.4 μmol, respectively; P=0.038) in the presence of not significantly different baseline Na-Li CT (282±15 and 257±13 μmol, P=0.227).
This study shows that Na-Li CT change over time is related not only to BMI and plasma triglycerides18 but also to baseline values and change over time in alcohol intake, plasma glucose, and plasma potassium. These relationships were independent of baseline Na-Li CT, which was strongly related to 6-year Na-Li CT change. Moreover, the study confirms in a large series of adults that Na-Li CT significantly changes over time, with an increase in the majority of individuals. A complete kinetic analysis of Na-Li CT activity was not done in the Gubbio Population Study. Therefore, Na-Li CT change could be due either to changes in the number of sodium transporters expressed per cell (ie, Na-Li CT units per stem or erythroid cell) or to the modulation of Na-Li CT activity in mature, circulating RBCs. Evaluation of technical errors was used to exclude data from days with laboratory inadequacies. Nevertheless, the lack of reference values that are stable over time for Na-Li CT makes it likely that Na-Li CT change in a long-term population-based study may reflect low precision in some determinations.
The inverse relation between baseline Na-Li CT and Na-Li CT change can be reasonably explained by the combination of two factors: First, the statistical phenomenon of regression toward the mean associated with repeated Na-Li CT measurement; and second, the biological unlikelihood of registering increases or decreases in Na-Li CT in individuals with baseline Na-Li CT in the high or low tail of the distribution, respectively. The direct relationships of BMI and alcohol intake to Na-Li CT change are consistent with and extend cross-sectional findings.1 4 16 17 22 The cumulative effects of BMI and alcohol intake were significant not only statistically; the combination of moderate differences in BMI and alcohol intake related to sizable differences in Na-Li CT change. The data support the concept that lifestyle influences Na-Li CT. Corresponding with results of cross-sectional analyses for baseline data,16 Na-Li CT change related directly to metabolic indexes such as plasma glucose change from baseline to follow-up and plasma lipids. In multivariate analyses, regression coefficients were significant for triglycerides but not for cholesterol, suggesting that triglycerides may play a predominant role. In keeping with this interpretation, clinical studies reported that Na-Li CT is reduced by lowering plasma triglycerides but not by lowering plasma cholesterol.23 24 For plasma lipids, as well as plasma glucose, it is uncertain whether the relation to Na-Li CT is explained by nutritional, genetic, or other factors (such as insulin sensitivity).25 In keeping with the observations of an inverse association between plasma potassium and Na-Li CT,16 26 changes in plasma potassium related inversely to changes in Na-Li CT. Various mechanisms could explain this relationship. It seems unlikely that Na-Li CT has a role in the control of plasma potassium. Alternative possibilities are that potassium homeostasis influences Na-Li CT, that Na-Li CT relates to renal potassium handling, or that a third factor causes opposite changes in Na-Li CT and plasma potassium. For circulating factors such as alcohol and plasma lipids, glucose, and potassium, influences on Na-Li CT could be directly induced on mature RBCs and/or on premature erythroid cells at the marrow level. For alcohol, a stimulation of Na-Li CT was in fact reported also with in vitro experimentation.27 In the case of BMI, the influences on Na-Li CT must be mediated by other (eg, humoral) factors, perhaps related to insulin sensitivity, as previously hypothesized.25
The practical implications of these findings are speculative at present. The fact that Na-Li CT can change over time suggests the need for caution when this sodium transporter is used as an intermediate phenotype in studies on hypertension or diabetic nephropathy. The physiological role of Na-Li CT remains unclear despite the consistent association of the activity of this sodium transporter with hypertension. A recent study showed that Na-Li CT activity is detectable also in human skin fibroblasts where it might reflect an amiloride-insensitive isoform of sodium proton exchange.28 If the association of high Na-Li CT with hypertension reflects a role of Na-Li CT in the control of BP, nongenetic and genetic factors modulating the activity of the sodium transporter could influence BP and the risk of hypertension also via effects on Na-Li CT. Alternatively, if high Na-Li CT is a marker of hypertension proneness without a pathogenetic role, the association of the sodium transporter with hypertension could be strengthened by lifestyle and metabolic factors that increase Na-Li CT and BP.
As for other variables related to Na-Li CT change in some analyses only—age, diastolic pressure, cigarettes per day, and plasma uric acid—findings were not consistent in both genders and in multivariate analyses. This could indicate lack of independent relation, confounding effects, overadjustment, low statistical power, or a combination of these factors.
In summary, the present study shows that Na-Li CT activity in adults changes over time, with significant tracking; on average, the tendency is toward a slight increase. BMI, alcohol intake, plasma glucose, plasma lipids, and plasma potassium are independently related to Na-Li CT change. Data support the concept that nongenetic factors play an important role in Na-Li CT regulation. If Na-Li CT has a role in BP control, it is reasonable to hypothesize that factors related to Na-Li CT and Na-Li CT change may affect BP also via effects on this sodium transporter.
The Gubbio Population Study, made possible thanks to the people of Gubbio, was supported, planned, and carried out by the Center for Epidemiologic Research, Merck Sharp & Dohme–Italy (MSD-I), in cooperation with the Center for Preventive Medicine in Gubbio (CPM); the Institute of Internal Medicine and Metabolic Diseases, University of Naples, Italy (IIMMDUN); the Istituto Superiore di Sanità, Rome, Italy (ISS); and the Department of Preventive Medicine, Northwestern University Medical School, Chicago, Ill (DPMNUMS). Determination of Na-Li CT was done at IIMMDUN at baseline (supervision of M.T. and M.C.) and in the Gubbio Civil Hospital (GCH) at follow-up (supervision of M.C.); determination of plasma lipids was done at ISS at baseline and in the laboratory of GCH at follow-up. The research was supervised by a Scientific Policy Board, whose members were Prof Piero Angeletti (chair, MSD-I, deceased), Dr Umberto Mortari (MSD-I), Dr Luigi Carratelli (MSD-I), Prof Mario Mancini (IIMMDUN), Prof Alessandro Menotti (ISS), Prof Rose Stamler (DPMNUMS, deceased), Prof Jeremiah Stamler (DPMNUMS), and Prof Alberto Zanchetti, Institute of Internal Medicine, University of Milan, Italy. Thanks are expressed for their fine cooperation to the staff of CPM and GCH, particularly Dr Mario Angeletti and Dr Ondina Cardoni (deceased). The Gubbio Population Study has been funded also by grant No. R01HL-40397-02 from the National Heart, Lung, and Blood Institute.
- Received July 30, 1998.
- Revision received August 24, 1998.
- Accepted October 28, 1998.
Laurenzi M, Trevisan M. Sodium-lithium countertransport and blood pressure: the Gubbio Population Study. Hypertension. 1989;13:408–415.
Cirillo M, Laurenzi M, Panarelli W, Trevisan M, Dyer AR, Stamler R, Stamler J. Sodium-lithium countertransport and blood pressure change over time: the Gubbio Study. Hypertension.. 1996;27:1305–1311.
Turner ST, Micherls VV. Sodium-lithium countertransport and hypertension in Rochester, Minnesota. Hypertension. 1991;18:183–190.
Laurenzi M, Cirillo M, Panarelli W, Trevisan M, Stamler R, Dyer AR, Stamler J. Baseline sodium-lithium countertransport and 6-year incidence of hypertension: the Gubbio Study. Circulation. 1997;95:581–587.
Kahn AM, Allen JC, Cragoe EG Jr, Shelat H. Sodium-lithium exchange and sodium-proton exchange are mediated by the same transport system in sarcolemmal vesicles from bovine superior mesenteric artery. Circ Res. 1989;65:818–828.
Weinberger MH, Smith JB, Fineberg NS, Luft FC. Red-cell sodium-lithium countertransport and fractionary excretion of lithium in normal and hypertensive humans. Hypertension. 1989;13:206–212.
Lewitter FI, Canessa M. Red cell sodium transport studies in adult twins. Am J Hum Genet. 1984;36:172s.
Trevisan M, Laurenzi M. Correlates of sodium-lithium countertransport: findings from the Gubbio Epidemiological Study. Circulation. 1991;84:2011–2019.
Hunt SC, Williams RR, Owen KA. Changes in sodium-lithium countertransport correlate with changes in triglycerides levels and body mass index over 21/2 years of follow-up in Utah. Cardiovasc Drugs Ther. 1990;4:357–362.
Laurenzi M, Cirillo M, Angeletti M, Buongiorno A, Morisi G, Panarelli W, Panfili M, Stamler J, Terradura O, Trevisan M, Stamler R. Gubbio Population Study: baseline findings. Nutr Metab Cardiovasc Dis. 1991;1:S1–S18.
Sanchez RA, Gimenez MI, Migliorini M, Giannone C, Ramirez A, Weder AB. Erythrocyte sodium-lithium countertransport in non-modulating offspring and essential hypertensive individuals: response to enalapril. Hypertension. 1997;30(part 1):99–105.
Weder AB, Serr C, Torretti BA, Bassett DR, Zweifler AJ. Effect of lovastatin treatment on red blood cell and platelet cation transport. Hypertension. 1991;17:203–209.
Weder AB. Is there a metabolic link between increased RBC lithium-sodium countertransport and hypertension? Nutr Metab Cardiovasc Dis.. 1993;3:38–45.
Zerbini G, Mangili R, Gabellini D, Pozza G. Modes of operation of an electroneutral Na+/Li+ countertransport in human skin fibroblasts. Am J Physiol. 1997;272:C1373–C1379.