Clinical Science (1998) 95, 649–657 (Printed in Great Britain)
R
E
V
I
E
W
Erythrocyte Na+/Li+ countertransport and
Na+/K+–2Cl− co-transport measurement in
essential hypertension: useful diagnostic
tools or failure? A meta-analysis
of 17 years of literature
T. Tepper, W. J. Sluiter*, R. M. Huisman and D. de Zeeuw†
Department of Internal Medicine, Division of Nephrology, P.O. Box 30.001, 9700 RB Groningen, The Netherlands, *Department
of Endocrinology, P.O. Box 30.001, 9700 RB Groningen, The Netherlands and †Department of Clinical Pharmacology,
Groningen Institute for Drug Studies (GIDS), State University Groningen, Groningen, The Netherlands
A
B
S
T
R
A
C
T
1. A meta-analysis of 17 years of literature on erythrocyte Na+/Li+ countertransport (NLCT) and
Na+/K+ co-transport (COT) measurements in relation to essential hypertension is presented.
The analysis aimed to answer two questions : (i) Which clinical or laboratory variables influence
NLCT and COT flux values ? (ii) How useful are NLCT and COT measurements as a diagnostic aid
in essential hypertension ?
2. Regression analysis was performed on the mean flux values and relevant clinical and laboratory
values. Studies in both normotensive and hypertensive subjects were stratified for variables
which showed a significant association with the measured flux. For hypertensive subjects the
studies were also stratified for medication. Means of strata were calculated after weighing the
mean of a study by the inverse of its own variance and were compared in normotensive as well
as hypertensive subjects using a t-test.
3. The analysis did not demonstrate systematic effects of laboratory variables for either NLCT or
COT. It was found that essential hypertension, family history of hypertension, gender and
antihypertensive medication are main determinants for the flux values of both transport
systems. After stratification for these determinants, significant differences in weighed mean flux
values between normotensive and hypertensive subjects were demonstrated. However, these
differences are much smaller than the variance in the weighed mean flux values, suggesting the
existence of other unknown variables that strongly affect the flux rates.
4. In conclusion, NLCT and COT measurements cannot be of diagnostic use in essential
hypertension.
INTRODUCTION
In 1960, Losse and co-workers [1] observed increased
sodium concentrations in erythrocytes of patients with
essential hypertension as compared with normal individuals. They hypothesized that this phenomenon was
related to an altered sodium transport over the erythrocyte membrane. In the years thereafter, assays for
sodium transport measurement in erythrocytes, incu-
Key words : erythrocyte, essential hypertension, sodium–lithium countertransport, sodium–potassium co-transport.
Abbreviations : COT, Na+\K+ co-transport ; NLCT, Na+\Li+ countertransport.
Correspondence : Dr R. M. Huisman.
#1998 The Biochemical Society and the Medical Research Society
649
650
T. Tepper and others
bated in artificial media, became available to test this
hypothesis.
Among the first promising assays were the Na+\K+2Cl− co-transport (COT) and the Na+\Li+ countertransport (NLCT) assays published by Dagher and
Garay [2] and Canessa et al. [3] respectively in 1980. They
used the assays for clinical studies in normal subjects and
patients with essential hypertension. Dagher and Garay
reported lower COT flux rates in hypertensive subjects,
whereas Canessa et al. found higher NLCT flux rates in
hypertensive subjects compared with normal individuals.
Both studies revealed almost total discrimination between normal and hypertensive subjects. These findings
pointed to the potential use of both transport assays as an
aid in the diagnosis of hypertension.
After these first promising results, inconsistent and
often conflicting findings were reported. This raised the
question of the applicability of the assays as diagnostic
tools. In the discussion section of many of these reports
the inconsistencies have been considered to be the results
of differences in methodology and patient selection.
Indeed, both the measurements of transport and the
selection of the population under study varied between
the investigations.
Many reviews tried to deal with this issue. However,
the conclusions were based on qualitative [4,5] or semiquantitative [6] rather than on quantitative analyses of
the available flux rate data. A reliable comparison of the
results of clinical studies is hampered by the large
variability in methodology of flux rate measurements and
the fact that clinical subject characteristics are not always
given in full detail. In our opinion, therefore, the validity
of the conclusions drawn thus far is still debatable.
The present report tries to deal with this problem and
describes a quantitative analysis of flux data in relation to
the variabilities mentioned above. The data are extracted
from clinical studies using NLCT and COT measurements in normotensive and essential hypertensive subjects, presented between 1980 and 1996. The time seems
to be ripe for such analysis as the initial stream of
publications has almost come to an end during the last
few years, and substantial new data are unlikely to be
forthcoming.
Usually, meta-analyses are applied in order to integrate
the results of directly comparable clinical studies. However, the variability of methods in flux rate measurements
and subject selection complicates direct comparison of
the flux data. Therefore, the approach of analysis,
although satisfying the general definition of meta-analysis [7], differs from the usual method of meta-analysis.
The present report tries to answer two main questions :
1. Which variables influence NLCT and COT flux rate
values, and to what extent ?
2. What are the consequences of these results for judging
the applicability of NLCT or COT measurements as a
diagnostic aid in essential hypertension ?
#1998 The Biochemical Society and the Medical Research Society
METHODS
Literature search
Publications in English between January 1980 and
December 1996 were collected manually (Current Contents Citation Index, Index Medicus) and by computerized searches (MEDLINE, EMBASE). The following
key words or text were used : sodium–lithium countertransport, Na+\Li+ countertransport, Na+\K+ co-transport, sodium–potassium co-transport, Na+–K+-Cl− (or
2Cl−) co-transport, essential hypertension. From the
references obtained in these searches, reports were
selected that gave transport data in both normotensive
and essential hypertensive individuals. In order to extend
the number of studies with transport values in normotensive subjects an additional computerized search was
performed. Thirty clinical journals were screened in
which studies could be expected on normotensive subjects and subjects with other diseases such as psoriasis,
obesitas etc. The same key words as mentioned above
(except essential hypertension) and the ISSN numbers of
the journals were used in this search.
Inclusion criteria for meta-analysis
From the references retrieved by the literature search
those satisfying the following criteria were selected :
(a) Because of the relatively few studies in black individuals only reports on Caucasians were included.
(b) Measurement of NLCT or COT must essentially
have been carried out according to the method of
Canessa et al. [3] or that of Dagher and Garay [2]
respectively.
‘ NLCT measurement essentially according to
Canessa et al. [3] ’ involved :
– washing of the isolated erythrocytes
– ‘ loading ’ of the cells with lithium by means of
diffusion exchange
– measurement of lithium efflux by assessment of
the difference in appearance rate of lithium in
sodium-free and sodium-rich media either in the
presence or absence of ouabain at 37 mC.
‘ COT measurement essentially according to Dagher
and Garay [2] ’ included :
– washing of isolated erythrocytes
– diffusion exchange of sodium and potassium in
high sodium and low potassium medium, resulting
in sodium ‘ loaded ’ and potassium-depleted cells
– incubation of these cells at 37 mC in sodium-free,
ouabain-containing media in the presence or absence of frusemide respectively
– following the appearance rate of sodium and
potassium in the incubation medium.
For both NLCT and COT assays, the composition
of the various media may vary qualitatively and\or
quantitatively between different studies.
(c) The results of the transport measurements had to be
Ion transport in hypertension : a meta-analysis
given as Vmax in mmol:h−":l−" cellspS.D. or S.E.M.
For analysis, S.E.M. was converted into S.D.
(d) Only studies in adult subjects, but not in subgroups
like ‘ elderly ’, ‘ adolescents ’, ‘ students ’, ‘ obese ’ or
‘ lean ’, were included.
(e) Abstracts were excluded from the meta-analysis.
It cannot be excluded a priori that in different reports
from the same authors flux data are given that were
obtained in (partly) the same populations. Therefore, we
carefully compared subject characteristics in order to
avoid, as far as possible, inclusion of these populations on
more that one occasion in the analyses.
(P 0n01, one-sided) was used to disclose any outliers,
which were removed from further analysis. The F-test
was repeated to judge homogeneity of a stratum with
regard to the remaining studies.
Means of strata were calculated after weighing the
mean of a study by the inverse of its own variance
(1\S.E.M.#). No correction was made for heterogeneity,
because of lack of relevant information.
The weighed means of the flux values of the strata were
compared in normotensive as well as hypertensive
subjects using a t-test. Furthermore, the weighed difference between normotensive and hypertensive subjects
was analysed in the same way.
Analysis of association of variables with
measured flux
Regression analysis was performed only when data were
available from at least 10 studies. If a variable was of a
continuous nature, the mean flux value of the population
was entered ; otherwise it was entered as a category.
Univariate regression analysis was performed on the
mean flux values and relevant clinical and laboratory
variables.
Multiple regression analysis was carried out with the
variables that showed significant (P 0n05) association
in the univariate analysis. Variables that independently
contributed significantly to the flux were retained in the
model. All other variables were added one by one to the
model to check their non-significant contribution.
Pooling of data for statistical analysis
Integration of published data was hampered by the fact
that the extent of information on the characteristics of
interest varied considerably between the studies. In an
attempt to ‘ quantify ’ this observation we investigated on
which potential relevant variables, and to what extent,
data are given in the reports under study. Subsequently,
those populations sharing one reported characteristic can
be combined in pools (strata). This stratification of
populations allows subsequent statistical processing of
the strata and comparisons between strata.
Before statistical analysis of the NLCT strata we
removed populations of normal pregnant women and
normal contraceptive pill users. From references [8–11]
it appeared that these populations were associated with
significantly higher mean flux values compared with
populations of non-pregnant women or non-pill users.
RESULTS
Studies available for statistical analysis
Thirty-three references of NLCT and 20 of COT studies
were retrieved in which flux values of NLCT or COT
were compared in normotensive and essential hypertensive subjects in relation to diagnostic purposes. Six
references were found reporting additional NLCT flux
values of normotensive subjects in relation to other
diseases. For COT studies this number amounted to
three. The data in those reports were used for statistical
analyses and are from references [2,3,8,9,11–64].
Figure 1 illustrates the time course of appearance of
these publications during the period 1980–1996. It
demonstrates the declining, and eventually almost disappearing, clinical interest in NLCT and COT measurements for diagnostic purposes.
Availability of relevant information in the
literature
From Table 1 it appears that in most of the studies
information relating to gender, age, body weight and
blood pressure is given but in different degrees of
Analysis of strata
Studies in both normotensive and hypertensive subjects
were stratified for variables which showed a significant
association with the measured flux. For hypertensive
subjects we also stratified for medication. Within-study
and between-study variances were estimated for each
stratum. Homogeneity was tested by F-test. When the
strata proved to be (P 0n05) heterogeneous, a t-test
Figure 1 Number of studies comparing NLCT or COT flux
values in normotensive and hypertensive individuals in the
period 1980–1996
#1998 The Biochemical Society and the Medical Research Society
651
652
T. Tepper and others
Table 1 Availability of information relating to relevant
clinical subject characteristics and flux measurement in
Na+/Li+ countertransport (NLCT) and Na+/K+ co-transport
(COT) studies
Abbreviations : SBP, systolic blood pressure ; DBP, diastolic blood pressure ; MAP,
mean arterial pressure.
% of studies providing
information on the
variable
Variable
Gender
Male or female
mixed populations
Age
meanpS.D.
mean or meanjrange or range
Body weight
meanpS.D.
mean or meanjrange or range
Body mass index
meanpS.D.
Body mass indexjweight
meanpS.D.
Blood pressure
mean (SBPjDBP)pS.D.
mean SBPpS.D.
mean DBPpS.D.
MAPpS.D.
‘ normotensive ’/‘ hypertensive ’
lower DBP/upper SBP
Family history of essential hypertension
Antihypertensive medication
Contraceptive medication/pregnancy
Plasma lipids
Diet
NLCT
COT
71
24
38
42
68
32
63
25
24
30
33
0
32
13
10
4
49
2
5
12
20
12
54
53
23
24
17
66
4
0
11
20
4
63
58
23
0
29
completeness. Among other variables, family history of
essential hypertension, antihypertensive medication and
plasma lipids have been discussed by many authors in
relation to flux rate values. However, in a considerable
proportion of the studies no explicit information on these
variables was provided. In only 23 % of the studies in
which women participated was information provided on
pregnancy or the use of contraceptive medication.
Details about the measurement of cation transport
were given in all the studies. For NLCT, in 23 % of the
studies the measurement was considered to have been
carried out ‘ (exactly) according to Canessa et al. [3] ’. In
the remainder of the reports the composition of the
various media was given with minor or major (qualitative
or quantitative) modifications compared with Canessa’s
media. Similarly, the COT measurement was performed
#1998 The Biochemical Society and the Medical Research Society
‘ (exactly) according to Dagher and Garay [2] ’ in 32 %
of the reports.
The cation concentration of the lithium and sodium
‘ loaded ’ erythrocytes is reported to be of crucial
importance for the Vmax value. Information on these
concentrations was given in 27 % and 67 % of the NLCT
and COT studies respectively.
For both NLCT and COT the reproducibility of the
transport assay was given in 43 % of the reports.
Statistical analysis of NLCT studies
Detailed information on the data underlying the results
of the statistical analyses can be found at URL : http : \\
cs.portlandpress.co.uk\cs\095\cs0950649add.htm
First, we considered which clinical or laboratory
variables are associated with the measured flux values in
normotensive subjects. The clinical variables comprised :
family history of hypertension, gender, body mass index,
body weight, plasma cholesterol, age, diastolic blood
pressure and systolic blood pressure. It appeared that
family history of hypertension and gender were significantly associated with the measured mean flux values.
In addition, we analysed whether the mean flux value
measured in established normotensive subjects differed
from that of subjects participating in population studies
(which include low percentages of hypertensive subjects).
No differences were found between these two groups.
For the laboratory variables we investigated whether
the fluxes were different when measured either ‘ according to Canessa ’ or with modifications. In addition,
differences in qualitative and quantitative compositions
in wash media and incubation media were compared to
reveal associations with the flux values. None of the
laboratory variables appeared to be associated with the
flux values.
After removal of outliers, the strata of hypertensive
males, hypertensive females and hypertensive subjects
with a positive family history could be considered
homogeneous. The strata of hypertensive subjects without a positive family history, hypertensive subjects
without or with medication and normal males and
females were all heterogeneous. We could not correct for
heterogeneity.
Figure 2 presents the results of the statistical analysis of
NLCT studies in normotensive and hypertensive subjects
after stratification for gender and family history, and also
for antihypertensive medication in hypertensive subjects.
It appears that in all cases significant higher weighed
mean flux values in the hypertensive groups were found
compared with the normotensive groups.
A positive family history of hypertension is associated
with higher weighed mean fluxes in both normotensive
and hypertensive subjects compared with a negative
family history. When comparing all populations (i.e.
normotensive plus hypertensive subjects) with a negative
family history versus all with a positive family history we
Ion transport in hypertension : a meta-analysis
Figure 2 Box plot of weighed means of NLCT flux values in normotensive (open bars) and hypertensive subjects (shaded bars)
comparing gender, family history of hypertension (FH +/FH −) and antihypertensive medication (Med+/Med−)
* denotes significantly different from the normotensive group ; **, significant difference ; n, number of subjects.
computed a significant difference of 0n06 mmol:h−":l−"
cells.
In both normotensive and hypertensive subjects lower
weighed mean flux values in females compared with
males were calculated. However, this difference did not
reach the level of significance in hypertensive subjects.
A significant difference of 0n04 mmol:h−":l−" cells was
found between all (i.e. normotensive plus hypertensive
subjects) and females and males.
Antihypertensive medication in hypertensive subjects
results in a significant decrease of the weighed mean
fluxes of 0n02 mmol:h−":l−" cells.
Hence, the NLCT fluxes in normotensive and hypertensive subjects are influenced (in decreasing order)
by family history of hypertension, gender and antihypertensive medication in hypertensive subjects.
Although there are significant differences between
normotensive and hypertensive subjects of about
0n05 mmol:h−":l−" cells they are much smaller than the
S.D. of about 0n08 in normotensive subjects and
0n11 mmol:h−":l−" cells in hypertensive subjects. Because
of this large overlap, measurement of NLCT cannot be of
diagnostic use.
Figure 3 Box plot of weighed means of COT flux values in
normotensive (open bars) and hypertensive subjects (shaded
bars) comparing gender, family history of hypertension
(FH +/FH −) and antihypertensive medication (Med+/Med−)
* denotes significantly different from Med− group ; ** significant difference ; n,
number of subjects.
Statistical analysis of COT studies
For COT studies in normotensive subjects, clinical
variables including family history of hypertension, gender, age, diastolic pressure and systolic blood pressure
were analysed to reveal associations with the flux values.
We also studied whether the flux values measured in
established normotensive subjects differed from those
found in subjects participating in population studies. As
a laboratory variable we compared the flux values
measured ‘ according to Garay ’ with the values obtained
by modifications of that method. After univariate and
multiple regression analysis, none of these variables,
except gender, showed a significant association.
After removal of outliers all strata remained heterogeneous. We could not correct for heterogeneity.
Figure 3 illustrates the results of the statistical analysis
of COT studies after stratification for gender and family
history of hypertension in normotensive subjects. In the
#1998 The Biochemical Society and the Medical Research Society
653
654
T. Tepper and others
literature, family history of hypertension has been
discussed as a determinant influencing the flux values.
Therefore, the strata of populations with and without
family history of hypertension were included in the
analysis. For the hypertensive subjects only the strata of
populations with and without antihypertensive medication comprised a sufficient number of studies to allow
meaningful analysis.
In the comparison between data of normotensive and
hypertensive strata, only the stratum of hypertensive
subjects without antihypertensive medication was considered of interest for diagnostic purposes. In comparison
with hypertensive subjects, we found higher mean flux
values (j0n06 mmol:h−":l−" cells) in normotensive males
and in normotensive subjects without a family history of
hypertension (j0n07 mmol:h−":l−" cells).
For normotensive subjects we found significantly
lower (k0n06 mmol:h−":l−" cells) mean flux values in
females compared with males.
A positive family history of hypertension showed a
significantly lowering (k0n06 mmol:h−":l−" cells) influence compared with a negative family history. In
populations of patients on antihypertensive medication,
significantly higher fluxes (j0n11 mmol:h−":l−" cells)
were calculated compared with patients without medication.
Again, the differences of about 0n06 mmol:h−":l−" cells
between hypertensive and normotensive groups are much
smaller than the S.D. of the weighed means of about
0n12 mmol:h−":l−". As with in NLCT measurements,
because of this large overlap, measurement of COT
cannot be of diagnostic use.
DISCUSSION
In the practice of meta-analysis, after the acceptable
studies have been identified according to the inclusion
criteria, each study is weighed by quality scoring.
Assessment of study quality leads to subjective judgment
as to the score of the study. In our approach no quality
scoring was used. The laws of statistics were applied to
groups of stratified populations and thus decided whether
a study, and to what extent a population, contributed to
the analysis.
Our study was hampered by the lack of availability of
all relevant information on clinical and laboratory
methods in many reports. After it became clear that the
methodologies were the subject of discussion by many
authors, one might have expected more attention to this
aspect in subsequent years. However, during the period
of 1980–1996, we did not notice a substantial improvement in this area, which is a rather disturbing finding.
It would have been interesting to include some
variables such as plasma high-density lipoprotein cholesterol or triacylglycerols in the analysis, but the number
#1998 The Biochemical Society and the Medical Research Society
of studies was lower than 10. Only plasma cholesterol
was given in a sufficient number of studies.
Meta-analysis is subject to publication bias as was
discussed earlier by Jenicek [7] and L’Abbe! et al. [65].
Inadequate reporting on clinical variables like gender,
family history of hypertension and medication or on
assay methodology results in a reduced contribution to
the analysis. In order to deal with this problem, L’Abbe!
et al. [65] recommended contacting the investigators for
missing information, but the inclusion of unpublished
data has its drawbacks. Therefore, we omitted that
procedure. Abstracts do not provide the required
methodological details and therefore were not included
in our study.
Because of lack of clinical data, the heterogeneity of the
strata could not be corrected. For instance, the strata of
hypertensive subjects with and without medication could
not be corrected for differences in gender and family
history. As even the strata of normal males and females
show heterogeneity, there must be other factors that lead
to heterogeneity.
Strictly speaking, our findings cannot be compared
with earlier ‘ narrative ’ overviews because of our fundamentally different approach in using meta-analysis.
Nevertheless, we checked whether our results and conclusions confirm or contradict opinions given in the
‘ narrative ’ review of Lijnen et al. [66], which is the most
comprehensive review on the subject covering the period
of time under study. Lijnen et al. [66] noticed discrepant
data on the activity of NLCT in hypertensive subjects :
some authors reported higher flux values compared with
normotensive subjects, whereas others failed to confirm
that finding. Our results obviate this uncertainty ; we
found definitely higher flux values in hypertensive
subjects. The results of COT measurement in hypertensive subjects as summarized in the review were also
controversial, whereas we were able to demonstrate
significant lower flux values in hypertensive subjects. In
accordance with our findings, Lijnen et al. [66] mentioned
lower COT flux rates in normotensive women compared
with men. No conclusive standpoint was reached by
Lijnen et al. as to the effect of a positive family history on
COT fluxes, whereas we calculated significantly lower
values in normotensive subjects with a positive history.
The influence of gender and familial background on
NLCT in normotensive subjects was not considered by
the reviewer.
Comprehensive reviews on methodology of NLCT
and COT measurement in relation to the flux values were
not available. In contrast to the alleged possibility, as
discussed in the literature, our analysis does not demonstrate systematic effects of laboratory variables (i.e.
composition of media used in transport measurement) on
both NLCT and COT. In a recent report, Hardman and
Land [67] point to the higher NLCT fluxes when
measured in magnesium-containing efflux media com-
Ion transport in hypertension : a meta-analysis
pared with media based on choline. Our meta-analysis
did not confirm a significant association of these variables
with the flux values. The same authors [67] state that
kinetic studies have shown independent variation of Vmax
and the affinity constant km. Measurement of these
parameters of NLCT transport separately may be of
diagnostic value.
A huge amount of effort and cost has been invested in
the attempts to use NLCT and COT measurement as
diagnostic tools in hypertension. After analysis of the
available published data we have to accept, however, that
(unlike the early expectations) these investments did not
lead to the desired result. However, the value of NLCT
measurement in predicting the development of essential
hypertension in non-hypertensive adults has been shown
by Laurenzi et al. [68]. We did not consider other disease
states, although for example NLCT measurement may
have importance in the prediction of nephropathy in
diabetes [30]. Also, the subgroup of essential hypertensive subjects that is characterized by a high NLCT not
only has a family history of hypertension, but also
vascular complications [69].
In conclusion, for the first time, herewith a quantitative
integration of studies on NLCT and COT measurement
in normotensive and hypertensive Caucasian populations
is presented. For both NLCT and COT, we could not
demonstrate a significant effect of the varying composition of the incubation media on the outcome of flux
rate measurement. We found that essential hypertension,
family history of hypertension, gender and antihypertensive medication are main determinants for the
flux values of both transport systems. However, the large
S.D.s of the flux values after taking into account gender,
family history of hypertension and antihypertensive
medication, suggest the existence of other, unknown,
variables affecting the flux rates and contributing to
heterogeneity in the strata. Significant differences in flux
values between subgroups of normotensive and hypertensive subjects could be demonstrated. However, these
differences are much smaller than the variance in the flux
values in normotensive and hypertensive subjects. Hence,
NLCT or COT measurement cannot be of diagnostic use
in essential hypertension.
ACKNOWLEDGMENT
GIDS is part of the Groningen Utrecht Institute for
Drug Exploration (GUIDE).
REFERENCES
1 Losse, H., Weymeijer, H. and Wessels, F. (1960) Der
Wasser- und Elektrolyt-Gehalt von Erythrocyten bei
arterieller Hypertonie. Klin. Wochenschr. 38, 393–395
2 Dagher, G. and Garay, R. P. (1980) A Na+,K+ cotransport assay for essential hypertension. Can. J.
Biochem. 58, 1069–1074
3 Canessa, M., Adragna, N., Solomon, H. S., Conolly, T. M.
and Tosteson, D. C. (1980) Increased sodium–lithium
countertransport in red cells of patients with essential
hypertension. N. Engl. J. Med. 302, 772–776
4 Parker, J. C. and Berkowitz, L. R. (1983) Physiologically
instructive genetic variants involving the human red cell
membrane. Phys. Rev. 82, 341–352
5 Blaustein, M. P. (1984) Sodium transport and
hypertension. Where are we going ? Hypertension 6,
445–453
6 Williams, R. R., Hunt, S. C., Wu, L. L., Hasstedt, S. J.,
Hopkins, P. N. and Owen Ash, K. (1988) Genetic and
epidemiological studies on electrolyte transport systems in
hypertension. Clin. Physiol. Biochem. 6, 136–149
7 Jenicek, M. (1989) Meta-analysis in medicine. Where we
are and where we want to go. J. Clin. Epidemiol. 42, 35–44
8 Worley, R. J., Hentschel, W. M., Cormier, C. et al. (1982)
Increased sodium–lithium countertransport in erythrocytes
of pregnant women. N. Engl. J. Med. 307, 412–416
9 Logan, A. G., Soldin, S. J. and Ryan, D. (1985) Alterations
in red cell cation transport in hypertensive disorders of
pregnancy. Klin. Wochenschr. 63 (Suppl. III), 16–19
10 M’Buyamba-Kabangu, J. R., Lijnen, P., Fagard, R. et al.
(1985) Erythrocyte concentrations and transmembrane
fluxes of sodium and potassium and biochemical
measurements during the menstrual cycle in normal
women. Am. J. Obstet. Gynecol. 151, 687–693
11 Gless, K.-H., Su$ tterlin, U., Schaz, K., Schu$ tz, V. and
Hunstein, W. (1986) Intracellular sodium concentration
and transport in red cells in essential hypertension,
hyperthyroidism, pregnancy and hypokalaemia. Clin.
Physiol. Biochem. 4, 199–209
12 Canali, M., Borghi, L., Sani, E. et al. (1981) Increased
erythrocyte lithium–sodium countertransport in essential
hypertension : its relationship to family history of
hypertension. Clin. Sci. 61, 13S–15S
13 Cusi, D., Barlassina, M., Ferrandi, M., Palazzi, P., Celega,
E. and Bianchi, G. (1981) Relationship between altered
Na+–K+ cotransport and Na+–Li+ countertransport in the
erythrocytes of ‘ essential ’ hypertensive patients. Clin. Sci.
61, 33S–36S
14 Canessa, M., Bize, I., Solomon, H. et al. (1981) Na
countertransport and cotransport in human red cells :
function, dysfunction, and genes in essential hypertension.
Clin. Exp. Hypertens. 3, 783–795
15 Cusi, D., Barlassina, C., Ferrandi, M., Brevi, R., Vazzola,
A. and Bianchi, G. (1982) Erythrocyte membrane
transport systems as possible markers for essential
hypertension. Clin. Sci. 63, 57S–59S
16 Adragna, N. C., Canessa, M. L., Solomon, H., Slater, E.
and Tosteson, D. C. (1982) Red cell lithium–sodium
countertransport and sodium–potassium cotransport in
patients with essential hypertension. Hypertension 4,
795–804
17 Weder, A. B., Torretti, B. A. and Julius, S. (1984) Racial
differences in erythrocyte cation transport. Hypertension
6, 115–123
18 Smith, J. B., Owen Ash, K., Hunt, S. C. et al. (1984)
Hypertension 6, 159–166
19 Dagher, G. and Canessa, M. (1984) The Li–Na exchange in
red cells of essential hypertensive patients with low Na–K
cotransport. J. Hypertens. 2, 195–201
20 Wiley, J. S., Clarke, D. A., Bonacquisto, L. A., Scarlett,
J. D., Harrap, S. B. and Doyle, A. E. (1984) Erythrocyte
cation cotransport and cotransport in essential
hypertension. Hypertension 6, 360–368
21 Beuckelmann, D. and Erdmann, E. (1984) Exogenous
factors influencing the human erythrocyte sodium–lithium
countertransport system. Eur. J. Clin. Invest. 14, 392–397
22 Weder, A. B. (1986) Red-cell lithium–sodium
countertransport and renal lithium clearance in
hypertension. N. Engl. J. Med. 314, 198–201
23 Weder, A. B., Fitzpatrick, M. A., Torretti, B. A.,
Hinderliter, A. L., Egan, B. M. and Julius, S. (1987) Red
blood cell Li+–Na+ countertransport, Na+–K+ cotransport,
#1998 The Biochemical Society and the Medical Research Society
655
656
T. Tepper and others
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
and the hemodynamics of hypertension. Hypertension 9,
459–466
Millunpalo, S., Saarinen, R., Marniemi, J. and Lehtonen, A.
(1989) Erythrocyte cation transport in obesity,
hypertension, and during antihypertensive drug therapy.
Clin. Physiol. Biochem. 7, 176–183
Carr, S. J., Thomas, T. H., Laker, M. F. and Wilkinson, R.
(1990) Elevated sodium–lithium counter-transport : a
familiar marker of hyperlipidaemia and hypertension ? J.
Hypertens. 8, 139–146
Rutherford, R. A., Thomas, T. H., Laker, M. F. and
Wilkinson, R. (1990) Increased erythrocyte
sodium–lithium countertransport activity in essential
hypertension is due to an increased affinity for
extracellular sodium. Clin. Sci. 79, 365–369
Cusi, D., Fossali, E., Piazza, A., Tripodi, G., Barlassina, L.
and Bianchi, G. (1991) Heritability estimate of erythrocyte
Na–K–Cl cotransport in normotensive and hypertensive
families. Am. J. Hypertens. 4, 725–734
Canessa, M., Morgan, K., Goldszer, R., Moore, T. J. and
Spalvins, A. (1991) Kinetic abnormalities of the red blood
cell sodium–proton exchange in hypertensive patients.
Hypertension 17, 340–348
Rutherford, P. A., Thomas, T. H., Carr, S. J., Taylor, R.
and Wilkinson, R. (1992) Kinetics of sodium–lithium
countertransport activity in patients with uncomplicated
type I diabetes. Clin. Sci. 82, 291–299
Rutherford, P. A., Thomas, T. H., Carr, S. J., Taylor, R.
and Wilkinson, R. (1992) Changes in erythrocyte
sodium–lithium countertransport in diabetic nephropathy.
Clin. Sci. 82, 301–307
Andronico, G., Mangano, M. T., Nardi, E., Mule' , G.,
Piazza, G. and Cerasola, G. (1993) Insulin-like growth
factor 1 and sodium–lithium countertransport in essential
hypertension and in hypertensive left ventricular
hypertrophy. J. Hypertens. 11, 1097–1101
Dowd, A., Thomas, T. H., Taylor, R. and Wilkinson, R.
(1994) Erythrocyte sodium–lithium counter-transport
activity is related to fluidity in IDDM patients.
Diabetologia 37, 394–400
Brugnara, C., Corrocher, R., Foroni, L., Steinmayr, M.,
Bonfanti, F. and De Sandre, G. (1983) Lithium–sodium
countertransport in erythrocytes of normal and
hypertensive subjects. Relationship with age and plasma
renin activity. Hypertension 5, 529–534
Trevisan, M., Ostrow, W., Cooper, R. et al. (1983)
Abnormal red blood cell ion transport and hypertension.
The people’s gas company study. Hypertension 5, 363–367
Trevisan, M., Ostrow, D., Cooper, R. S., Sempas, C. and
Stamler, J. (1984) Sex and race differences in
sodium–lithium countertransport and red cell sodium
concentration. Am. J. Epidemiol. 120, 537–541
Turner, S. T., Johnson, M., Boerwinkle, E., Richelson, E.,
Taswell, H. F. and Sing, C. F. (1985) Sodium–lithium
countertransport and blood pressure in healthy blood
donors. Hypertension 7, 955–962
Adragna, N. C., Chang, J. L., Morey, M. C. and Williams,
R. S. (1985) Effects of exercise on cation transport in
human red cells. Hypertension 7, 132–139
Trevisan, M., De Santo, N., Laurenzi, M. et al. (1986)
Intracellular ion metabolism in erythrocytes and uraemia :
the effect of different dialysis treatments. Clin. Sci. 71,
545–552
Smith, J. B., Wade, M. B., Fineberg, N. S. and Weinberger,
M. H. (1988) Sodium transport parameters in erythrocytes
of patients with primary aldosteronism. Hypertension 11,
141–146
Smith, J. B., Wade, M. B., Fineberg, N. S. and Weinberger,
M. H. (1988) Influence of race, sex, and blood pressure on
erythrocyte sodium transport in humans. Hypertension 12,
251–258
Trevisan, M., Vaccaro, O., Laurenzi, M. et al. (1988)
Hypertension, non-insulin-dependent diabetes, and
intracellular sodium metabolism. Hypertension 1, 264–268
Weder, A. B. and Egan, B. (1988) Erythrocyte water,
Na+–K+-cotransport, and forearm vascular function in
humans. Hypertension 12, 199–205
#1998 The Biochemical Society and the Medical Research Society
43 Semplicini, A., Canessa, M., Mozzato, M. G. et al. (1989)
Red blood cell Na+\H+ and Li+\Na+ exchange in patients
with essential hypertension. Am. J. Hypertens. 2, 903–908
44 Strazzullo, P., Cappuccio, F. P., Trevisan, M. et al. (1989)
Erythrocyte sodium\lithium countertransport and renal
lithium clearance in a random sample of untreated middle
aged men. Clin. Sci. 77, 337–342
45 Persky, V., Ostrow, D., Langenberg, P., Ruby, E.,
Bresolin, L. and Stamler, J. (1990) Hypertension and
sodium transport in 390 healthy adults in Chicago. J.
Hypertens. 8, 121–128
46 Turner, S. and Michels, V. V. (1991) Sodium–lithium
countertransport and hypertension in Rochester,
Minnesota. Hypertension 18, 183–190
47 Trevisan, M. and Laurenzi, M. (1991) Correlates of
sodium–lithium countertransport. Findings from the
Gubbio epidemiological study. Circulation 84, 2011–2019
48 Strazzulo, P., Cappucio, F. P., Trevisan, M. et al. (1993)
The relationship of erythrocyte sodium–lithium
countertransport to blood pressure and metabolic
abnormalities in a sample of untreated middle-aged male
workers. J. Hypertens. 11, 815–822
49 Turner, S. T., Boerwinkle, E., Johnson, M., Richelson, E.
and Sing, C. F. (1987) Sodium–lithium countertransport in
ambulatory hypertensive and normotensive patients.
Hypertension 9, 24–34
50 Beuckelmann, D., Meyer, H. D. and Erdmann, E. (1985)
Na+–Li+ countertransport in human erythrocytes. Effects
of hypokalaemia, oral contraceptives and antihypertensive
medication. Klin. Wochenschr. 63 (Suppl. III), 143–146
51 Weinberger, M. H., Smith, J. B., Fineberg, N. S. and Luft,
F. C. (1989) Red-cell sodium–lithium countertransport and
fractional excretion of lithium in normal and hypertensive
humans. Hypertension 13, 206–212
52 Hunt, S. C., Williams, R. R., Smith, J. B. and Owen Ash,
K. (1986) Associations of three erythrocyte cation
transport systems with plasma lipids in Utah subjects.
Hypertension 8, 30–36
53 Garay, R. P., Dagher, G., Pernollet, M.-G., Devynck, M.A. and Meyer, P. (1980) Inherited defect in a Na+, K+-cotransport system in erythrocytes from essential
hypertensive patients. Nature (London) 284, 282–283
54 Garay, R. P., Hannaert, G., Dagher, G., Nazaret, C.,
Maridonneau, I. and Meyer, P. (1981) Abnormal
erythrocyte Na+\K+ cotransport system. A proposed
genetic marker of essential hypertension. Clin. Exp.
Hypertens. 3, 851–859
55 Ghione, S., Buzzigoli, G., Bartolini, V. and Balzan Donato,
L. (1981) Comparison of outward and inward Na+\K+
cotransport-medicated fluxes in erythrocytes of essential
hypertensive patients. Preliminary results. Clin. Exp.
Hypertens. 3, 809–814
56 Davidson, J. S., Opie, L. H. and Keding, B. (1982)
Sodium–potassium cotransport activity as genetic marker
in essential hypertension. Br. Med. J. 284, 539–541
57 Gless, K.-H., Roelcke, U., Fiehn, W. et al. (1983)
Kationenfluxe am Erythrozyten bei Patienten mit
essentieller Hypertonie. Klin. Wochenschr. 61, 517–521
58 De Zeeuw, D., Jilderda, J. F. and Tepper, T. (1983)
Erythrocyte sodium–potassium cotransport in
hypertension. Proc. EDTA 20, 507–513
59 Lijnen, P., M’Buyamba-Kabangu, J.-R., Fagard, R. H.,
Groeseneken, D. R., Staessen, J. A. and Amery, A. K.
(1984) Intracellular concentration and transmembrane
fluxes of sodium and potassium in erythrocytes of white
normal male subjects with and without a family history of
hypertension. J. Hypertens. 2, 25–30
60 Delva, P., De Gasperi, M., Degan, M., Covi, G. and Lechi,
A. (1985) Kinetics of bumetanide-sensitive Na+–K+ cotransport in erythrocytes of essential hypertensive patients.
Clin. Sci. 69, 607–612
61 Krzesinski, J. M. and Rorive, G. L. (1985) Erythrocyte
sodium–potassium cotransport in hypertensive patients
advantages and limitations. Clin. Exp. Hypertens. A7,
553–572
62 M’Buyamba-Kabangu, J.-R., Lijnen, P., Fagard, R.,
Groeseneken, D., Staessen, J. and Amery, A. (1985)
Ion transport in hypertension : a meta-analysis
Intracellular concentration and transmembrane fluxes of
sodium and potassium in erythrocytes of normal men and
women. Arch. Gynecol. 236, 219–224
63 Lijnen, P., M’Buyamba-Kabangu, J.-R., Fagard, R.,
Fiocchi, R., Staessen, J. and Amery, A. (1985) Effects of
ketanserin on the intracellular concentration and
transmembrane fluxes of sodium and potassium in
erythrocytes of normal male subjects. J. Cardiovasc.
Pharmacol. 7 (Suppl. 7), S38–S40
64 Tuck, M. L., Corry, D. B., Maxwell, M. and Stern, N.
(1987) Kinetic analyses of erythrocyte Na+–K+ pump and
cotransport in essential hypertension. Hypertension 10,
204–211
65 L’Abbe! , K. A., Detsky, A. S. and O’Rourke, K. (1987)
Meta-analysis in clinical research. Ann. Intern. Med. 107,
224–233
66 Lijnen, P., M’Buyamba-Kabangu, J.-R., Fagard, R.,
Staessen, J. and Amery, A. (1990) Erythrocyte
concentrations and transmembrane fluxes of sodium and
potassium in essential hypertension : role of intrinsic and
environmental factors. Cardiovasc. Drugs Ther. 4, 321–333
67 Hardman, T. C. and Land, A. F. (1996) Controversies
surrounding erythrocyte sodium–lithium countertransport.
Hypertension 14, 695–703
68 Laurenzi, M., Cirillo, M., Panarelli, W. et al. (1997)
Baseline sodium–lithium countertransport and 6-year
incidence of hypertension. The Gubbio population study.
Circulation 95, 581–587
69 Engelman, B., Duhm, J. and Schonthier, U. M. (1993)
Relations of sodium–lithium countertransport kinetics to
plasma and red cell membrane phospholipids in
hyperlipidemia. Atherosclerosis 99, 151–163
#1998 The Biochemical Society and the Medical Research Society
657