486 Original article
Urinary cortisol to cortisone metabolites ratio in prednisone-
treated and spontaneously hypertensive patients
Oliviero Olivieria, Francesca Pizzoloa, Viviana Ravagnanib, Lorenzo Morettia,
Antonio Carlettob, Giovanni Faccinic, Francesco Pasinid, Simonetta Frisoa
and Roberto Corrochera
Objective and methods Prednisone and its active
metabolite prednisolone, both substrates for 11b-
hydroxysteroid dehydrogenase type 2 (11b-HSD2),
may represent a pharmacological challenge for the
enzyme. The aim of the present work was to define
the possible role of abnormal cortisol/cortisone
handling, as revealed by an urinary tetrahydrocortisol R
allotetrahydrocortisol (THFs)/tetrahydrocortisone (THE)
ratio between 1.5 and 3.0, by measuring urinary cortisol
and cortisone metabolites in: normotensive individuals
(n U 100) who received 78 mg/day of oral prednisone
and were then followed for development of hypertension;
essential hypertensive (EH) participants from primary care
(n U 103); and EH hypertensive patients referred to the
Hypertension Unit (n U 141).
Results About one-third (14 out of 47, 30%) of glucorticoid-
treated patients who developed hypertension showed a
THFs/THE ratio >1.5, which was seen in 3% (n U 3) and 14%
(n U 19) of primary and tertiary care hypertensive patients,
respectively. A THFs/THE ratio >1.5 was associated with a
3.8-fold incremental risk of hypertension after
glucocorticoid therapy, regardless of duration and intensity
of prednisone therapy.
Introduction
Normal activity of 11b-hydroxysteroid dehydrogenase
type 2 (11b-HSD2), a microsomal enzyme responsible
for the conversion of 11-hydroxy glucocorticoids to their
11-keto metabolites, is essential for preventing the acti-
vation of mineralcorticoid receptor (MR) by cortisol [1].
Since cortisol and aldosterone have the same affinity for
MR and the plasma concentration of the former is much
higher than that of the latter, MR activation by aldoster-
one depends strictly on the inactivation of cortisol into
cortisone by 11b-HSD2 [2].
Inappropriate MR stimulation by cortisol is the patho-
genetic mechanism responsible for the apparent miner-
alocorticoid excess (AME) syndrome [3], an autosomal
recessive form of salt-sensitive hypertension caused by
mutations in the HSD11B2 gene leading to substantial or
complete loss of enzymatic activity [4]. In spite of its
intrinsic interest, classic AME has for a long time been
considered a rare condition with a limited impact on
clinical practice.
0263-6352 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Conclusions A number of EH patients and glucocorticoid-
treated participants shared a similar phenotype,
characterized by both arterial hypertension and elevated
urinary THFs/THE ratio. Such a phenotype is more common
in severely, rather than in mildly, hypertensive patients.
J Hypertens 26:486493 Q 2008 Wolters Kluwer Health |
Lippincott Williams & Wilkins.
Journal of Hypertension 2008, 26:486493
Keywords: cortisol, cortisone, glucocorticoids, 11b-hydroxysteroid
dehydrogenase, hypertension, prednisone
Abbreviations: AME, Apparent Mineralocorticoid Excess; 11beta-HSD2,
11-hydroxysteroid dehydrogenase type 2; EH, Essential hypertensive; MR,
Mineralcorticoid receptor; RA, Rheumatoid arthritis; THFs,
Tetrahydrocortisol allotetrahydrocortisol; THE, Tetrahydrocortisone; UFF,
Urinary free cortisol; UFE, Urinary free cortisone
a
Unit of Internal Medicine, bUnit of Rheumatology, Department of Clinical and
Experimental Medicine, cInstitute of Clinical Chemistry and dDepartment of
Medicine and Public Health, University of Verona, Italy
Correspondence to Oliviero Olivieri, Dipartimento Medicina Clinica e
Sperimentale, Cattedra di Medicina Interna, Universita di Verona, Policlinico
Borgo Roma, 37134 Verona, Italy
Tel: +39 45 8074401; fax: +39 45 580111; e-mail: oliviero.olivieri@univr.it
Received 4 June 2007 Revised 22 September 2007
Accepted 1 October 2007
More recently, however, evidence that partially defective
11b-HSD2 activity may play a role in hypertension has
changed this view [5]. Since the 1990s, moderate impair-
ment of 11b-HSD2 has been recognized in some patients
with essential hypertension (EH) [68], and several
genetic mutations or polymorphisms with milder effects
on 11b-HSD2 activity than those described in classic
AME may be present in some EH patients [914].
Although the results of these investigations have not
been conclusive, and genetic mutations with a definite
causal role have not been identified, a milder phenotype
of AME (also called AME type II) has been documented
in some hypertensive patients and a form of hypertension
indistinguishable from EH was described in their
relatives [15]. In addition, impaired 11b-HSD2 activity
has been claimed to be associated with salt sensitivity in
healthy individuals [16], but this finding has not been
confirmed and still remains the object of controversy [17].
In clinical practice, 11b-HSD2 activity is not curren-
tly evaluated in EH patients, probably reflecting the

uncertain pathogenetic role and prevalence of mildly
impaired enzyme activity in the hypertensive population.
Traditionally, the profile of urinary steroids has served as
the main tool for detecting AME or AME-like phenotype,
typically characterized by a relative increase in the
excretion of A-ring-reduced metabolites of cortisol [tetra-
hydrocortisol (THF) allotetrahydrocortisol (aTHF)],
plus the corresponding decrease of the levels of A-ring-
reduced metabolites of cortisone [tetrahydrocortisone
(THE)] [18,19]. While normal activity of 11b-HSD2 pro-
duces similar urinary amounts of THF aTHF (THFs)
and THE, with a resulting THFs/THE ratio
1, in
patients with impaired activity this ratio increases propor-
tionally to the extent of reduction in enzyme activity.
Considering THFs/THE ratio as a quantitative marker,
a pathogenetic hypertensive role is consistently associated
in vivo with a urinary THFs/THE ratio  3, a value
corresponding to a 50% reduction of enzyme activity
[5]; in contrast, the practical relevance of a less dramatic
functional impairment, as witnessed by an urinary THFs/
THE ratio higher than normal but < 3, remains unclear.
Based on the observation that children with AME chal-
lenged with exogenous cortisol develop further sodium
retention and marked increase in blood pressure (BP) [20],
impairment of 11b-HSD2 enzyme activity is considered as
one of the possible mechanisms that may explain the
occurrence of glucocorticoid-associated hypertension
[21,22]. However, the issue has not been object of specific
clinical investigation and the proportion of patients who
develop hypertension after chronic glucocorticoid
exposure and/or also have increased THFs/THE ratios
is not known [22]. Hypertensive patients affected by
Cushings syndrome have a high (approximately twofour
times normal) cortisol to cortisone metabolite ratio because
11b-HSD2 is insufficient to handle the elevated circulat-
ing levels of cortisol [23]. It is therefore reasonable to
suppose that chronic glucorticoid therapy may increase
BP as well as THFs/THE ratio in a similar manner. Since
prednisolone, the active metabolite of prednisone,
has been demonstrated to be a better substrate for
11b-HSD2 oxidation than cortisol [24], and it is in turn
inactivated to prednisone in the kidney by 11b-HSD2 [24],
patients receiving such pharmacological treatment may
represent a suitable model to prove the hypothesis of a
competitive inhibition mechanism of 11b-HSD2. In
addition, the similarities of structure between cortisol/
cortisone and prednisolone/prednisone represent intuitive
support to this hypothesis.
Based on all these considerations, we planned the present
study with the purpose of evaluating the prevalence of
individuals with an increased THFs/THE ratio in a group
of normotensive individuals who developed hypertension
after chronic treatment with a moderatelow dose of
prednisone. We also estimated the prevalence of EH
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THFs/THE in different forms of hypertension Olivieri et al. 487
patients with a similarly high ratio, selected from primary
and tertiary care hypertensive populations. In this way, it
was possible to identify patients sharing an identical
phenotype (arterial hypertension an elevated urinary
THFs/THE ratio) determined by different factors: an
acquired factor in the case of glucocorticoid-treated
patients, and spontaneous factor(s) in the case of essen-
tial hypertensive patients.
Methods
Patient selection: glucocorticoid-treated patients
(group A)
Patients were selected among those referred to the Rheu-
matology Unit of the University Hospital of Verona for
rheumatoid arthritis (RA). We retrospectively analysed all
the folders of individuals who received chronic glucocor-
ticoid treatment for this disorder and were still on therapy
in December 2004. To be included in the study, patients
had to show normal BP (< 140/90 mmHg) on the first two
ambulatory examinations, sequentially scheduled before
starting any (including glucocorticoid) treatment; more-
over, since patients were examined on a regular basis,
persistently elevated or normal BP had to be confirmed
during the subsequent follow-up. To assess whether there
is any evidence of a significant BP fluctuation over the
whole period of observation, the average of the recorded
BP values before glucocorticoid treatment was used as a
baseline measure and classification for persistent normo-
tension or glucocorticoid-associated hypertension was con-
firmed when an increase of systolic BP (SBP)/diastolic
BP (DBP) was < or > 11/7.5 mmHg, respectively, that is
below or above the standard deviation values of baseline
BP average, to avoid the so-called regression on the mean
phenomenon. This arbitrary cut-off was chosen because
recorded BP values of patients were often lower at late than
at early measurements during the follow-up period. Actu-
ally, all of such patients classified as normotensives also
showed BP < 140/90 mmHg at the last examination, while
all patients who were classified as hypertensives
were taking BP-lowering drugs [angiotensin-converting
enzyme (ACE) inhibitors, sartans, diuretics, calcium
blockers] at the last examination. Given general agreement
on 140/90 mm Hg as cut-off BP values in terms of risk, all
subsequent considerations imply that patients with pre-
dnisone-induced hypertension were at least 140/90 mmHg
and had a  11/7.5 mmHg increase from baseline
average BP.
According to a previously approved rheumatological pro-
tocol, all patients were treated with oral prednisone at
doses ranging from 5 to 15 mg/day. The patients meeting
the requirements described above were recalled for bio-
chemical tests and 24-h urine collection. BP-lowering
drugs and glucorticoids were not withdrawn before
sampling. When assayed, all patients were in stable
remission of disease and were taking the same dose of
oral prednisone for at least 3 months (according to the
 488 Journal of Hypertension 2008, Vol 26 No 3

Table 1 Clinical and biochemical features of glucorticoid-treated in BP values > 5 mmHg. To establish a firm diagnosis,
patients high BP also had to be confirmed 2 weeks later (the
Normotensive Hypertensive same procedure was repeated as described above); or
Variable patients (n 53) patients (n 47) P (2) current therapy for a previous diagnosis of hyperten-
Age (years) 50.6  12.6 56.7  12.3 < 0.05 sion.
Gender (M/F) 11/42 7/40 NS
SBP (mmHg) 122.3  10.9 138.5  15.3 < 0.001
DBP (mmHg) 78.23  7.5 87.20  9.0 < 0.001
Since the study was aimed at defining aldosterone to
S-Na (mmol/l) 141.8  2.5 141.5  2.6 NS renin ratio [25], drugs were withdrawn for the 4 weeks
S-K (mmol/l) 3.9  0.3 3.77  0,4 NS following inclusion in the study (wash-out period); if
U-Na (mmol/24 h) 159  58 151  60 NS
U-K (mmol/24 h) 56  17 56  20 NS necessary, BP reduction was obtained using agents such
S-Creatinine (mmol/l) 67  14 68  15 NS as verapamil and alpha-blockers [25]. Samples for
S-Cortisol (mg/dl) 12.62  8.7 9.5  6.9 NS
P- ACTH (pg/ml) 20.1  14.6 15.7  13.0 < 0.05
aldosterone and renin were obtained after at least 2 h
P- Aldosterone (pg/ml) 182  130 229  168 NS in the upright posture and a subsequent period of 10 min
P- Renin (pg/ml) 14.8  15 36.6  56 < 0.01 in the seated position. No patient had current or past
ARR (pg/ml:pg/ml) 21.4  18.0 21.2  20.5 NS
THF (mg/24 h) 0.85  0.6 0.93  0.8 NS
chronic glucocorticoid treatment.
a-THF (mg/24 h) 0.41  0.4 0.53  0.6 NS
THE (mg/24 h) 1.70  1.3 1.47  1.3 NS
THFs/THE ratio 0.92  0.70 1.44  1.1 < 0.001
Patient selection: tertiary care (group C)
THFs/THE ratio 1.5 9% (5/53) 30% (14/47) 0.01 All patients referred to the Hypertension Unit of the
Duration of therapy (months) 27.8  57 31.0  58 NS University Hospital of Verona for resistant hypertension
Prednisone dose (mg/day) 7.45  2.9 7.78  3.8 NS
or for possible secondary causes of hypertension over the
Data are means  SD. ACTH, adrenocorticotrophic hormone; ARR, aldosterone to previous 2 years were examined by biochemical testing
renin ratio; DBP, diastolic blood pressure; P, plasma; S, serum; SBP, systolic and 24-h urine cortisol/cortisone metabolite assay. Only
blood pressure; THE, tetrahydrocortisone; THF, tetrahydrocortisol; THFs, tetra-
hydrocortisol allotetrahydrocortisol; U, urinary. Statistical significance (P < 0.05) patients for whom a diagnosis of EH was made at the end
was defined by Students t-test, KruskalWallis or x2 test. of the work-up were included in the study; patients with
secondary forms of hypertension (in particular patients
scheduled visit interval for RA patients). Doses are with nephro-parenchymal disease in whom an abnormal
reported in the Results section (see Table 1), whereas THFs/THE ratio might occur) were therefore excluded.
the duration of therapy indicates the overall months of Similarly, patients currently treated with glucocorticoids,
exposure to the glucocorticoid treatment before THFs/ or with a clinical history of previous glucocorticoid
THE ratio assay. One hundred patients were finally treatment, were excluded. No patient showed liquorice-
included in the study. induced hypertension or was taking liquorice-containing
sweets.
Patient selection: primary care (group B)
A detailed description of the selection of patients is pro- A total of 141 EH patients were finally considered.
vided in reference [25]. Briefly, a sample of 1462 adults According to our protocol, all patients were on a
aged 3574 years, randomly selected from the population sodium-controlled diet (NaCl 110120 mmol/day) for
register of the Bussolengo Health District (northern Italy) 3 days before blood and urine sampling, and took no
and representative of the total population of the district, hypotensive drugs other than verapamil and/or alpha-
was examined by the general practitioners who partici- blockers over the previous 4 weeks [26]. Samples for
pated in the study. After a detailed explanation and after aldosterone and renin were obtained after at least 2 h in
obtaining written informed consent, patients affected by the upright posture and a subsequent period of 10 min in
hypertension (defined according to the criteria specified the seated position.
below) were included in the study and examined by
biochemical testing. Of 412 identified hypertensive Biochemical assays
patients, 103 agreed to give blood and to collect 24-h Patient blood samples for hormonal and routine
urine. These patients were similar for demographic and parameters were collected after overnight fasting between
socio-economic features to the hypertensive individuals 0800 and 0900 h. Plasma aldosterone and direct active
who refused examination (data not shown), so they should renin were measured by commercially available methods
be representative of the entire group. The inclusion (Nichols Diagnostics, San Clemente, California, USA)
criteria were: [27]. Cortisol and plasma adrenocorticotrophic hormone
(ACTH) were assayed by routine immunometric methods
(1) SBP > 140 mmHg or DBP > 90 mmHg, measured (Diagnostic Products Corporation, Los Angeles, Califor-
twice in both arms, with the patient in the sitting nia, USA); intra- and inter-assay coefficients of variation
position for 5 min, providing that he/she had not drunk were 6.27.3% for cortisol and 6.78.2% for ACTH,
any coffee or smoked during the previous 30 min; a respectively. For THFs/THE assay, urine samples were
third measurement (or further measurements, if analysed by gas chromatographymass spectrometry
necessary) was obtained in the case of a difference (Hewlett-Packard 6890-5973Hp mass spectrometer;

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Hewlett-Packard, Milan, Italy) as described previously
[27]. Intra-assay and inter-assay coefficients of variation
were 7.7 and 9.4%, respectively.
Statistical analysis
Data analysis was performed using the SPSS 13.0 for
Windows (SPSS, Chicago, Illinois, USA). Quantitative
values were expressed as means  standard deviation.
Students t-test for unpaired observations or the Krus-
kalWallis test was used for normally distributed or
skewed variables, respectively. Comparison of proportions
was carried out by cross-tabulation, Pearsons chi-squared
and Fishers exact test.
To assess the extent to which covariates were associated
with the risk of having glucocorticoid-associated hyper-
tension, odds ratios (OR) with 95% confidence intervals
(CI) were estimated by logistic-regression analysis.
Results
Glucocorticoid-treated patients (group A)
In agreement with the prevalence of rheumatoid arthritis,
most patients (82%) included in the study were women.
After starting glucocorticoid treatment, a total of 47 (47%)
patients developed hypertension while the remaining
individuals (n 53, 53%) did not substantially modify their
BP during the period of follow-up. As reported in Table 1,
these two subgroups of patients matched for most clinical
and biochemical features; in particular, duration (27.8 
56.6 versus 30.9  58 months) and dosage (7.45  2.9 versus
7.78  3.8 mg/day) of prednisone therapy were not statisti-
cally different between subgroups of individuals. However,
hypertensive patients were older and had higher renin
values than normotensive individuals, probably reflecting
concurrent BP-lowering treatment (Table 1).
Hypertensive patients had lower ACTH values and higher
urinary cortisol to cortisone metabolites ratio than normo-
tensive individuals (Table 1). Moreover, by evaluating
how many individuals showed THFs/THE >1.5, an
altered cortisol/cortisone metabolite ratio was observed
in about one-third (30%) of hypertensive patients but in
only 9.4% of normotensive participants (Table 1).
To assess the extent to which age, ACTH and raised
THFs/THE ratio (i.e. the main covariates differently
distributed between the subgroups) were associated with
the development of hypertension, OR with 95% CI was
estimated by logistic-regression analysis (Table 2). With
THFs/THE ratio as a continuous or categorical variable,
it proved to be a predictor of development of hyperten-
sion; in particular, a THFs/THE ratio > 1.5 was associ-
ated with a 3.8-fold increased risk of hypertension after
glucocorticoid therapy. In contrast, low ACTH levels
were not significantly related to this risk. Adjustment
for sex, duration and dose of glucocorticoid therapy did
not change the results (data not shown).
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THFs/THE in different forms of hypertension Olivieri et al. 489
Table 2 Risk of development of hypertension after glucocorticoid
therapy (by logistic regression analysis)
Odds ratio (95% CI)
Covariates n 100 B Wald P value
Model 1
Age 1.05 (1.011.09) 0.048 5.68 0.017
ACTH 0.98 (0.951.01) 0.022 1.64 0.20
THFs/THE ratio 2.47 (1.195.1) 0.90 5.94 0.015
Model 2
Age 1.05 (1.011.09) 0.045 5.47 0.019
ACTH 0.98 (0.951.01) 0.022 1.58 0.209
THFs/THE ratio >1.5a 3.8 (1.1412.64) 1.33 4.70 0.030
ACTH, adrenocorticotrophic hormone; CI, confidence interval; THE, tetrahydro-
cortisone; THFs, tetrahydrocortisol allotetrahydrocortisol. a Considered as
categorical covariate, that is having or not a raised THFs/THE ratio.
The entire group of patients taking glucorticoids was then
divided into two subgroups according to THFs/THE ratio
values: patients with an elevated ratio (n 19, 19% of total
group) more often had hypertension (74 versus 41%,
P 0.01), higher urinary excretion of potassium (68  16
versus 54  18 mmol/24 h, P < 0.01), lower levels of ACTH
(11.1  11.5 versus 19.6  14.1 pg/ml, P 0.001), lower
cortisol (5.4  5.4 versus 12.6  8.0 mg/dl, P 0.001), and
a shorter mean time of steroid therapy (15.2  16.1 versus
32.6  63 months, P < 0.05) than patients with a normal
THFs/THE ratio. In contrast, mean prednisone dosage
taken by the patients of the two subgroups was similar
(7.4  3.2 versus 8.3  3.7 mg/day).
Since five patients showed a raised THFs/THE ratio but
normal BP, we analysed possible distinguishing features
of these individuals who were resistant to the develop-
ment of hypertension in comparison with the patients
(n 14) showing similarly increased THFs/THE ratio
but high BP. Although the small number of cases
represents a statistical limitation, the subgroups were
matched for all the features listed in Table 1. In particular
there were no differences in serum and urinary electro-
lytes, serum cortisol, ACTH, THFs/THE ratio, duration
and dose of glucorticoid therapy (data not shown). Of
note, four out of five normotensive individuals were
fertile females with a regular menstrual cycle.
Hypertensive patients from primary and tertiary care
The clinical and biochemical features of hypertensive
patients selected from primary (group B) or tertiary care
(group C) are reported for comparison in Table 3.
As expected, patients referred to the Hypertension
Unit were younger and showed more severe hyper-
tension (46% had BP values consistent with stage II
of JNC 7 classification [28], despite treatment with
verapamil and/or a-adrenergic blockers); they also had
increased levels of serum creatinine and plasma
aldosterone, and a more frequent positive family history
for hypertension than primary-care patients (Table 3).
Obesity, diabetes, plasma lipids and smoking were
similarly distributed between the two groups (data
not shown).
 490 Journal of Hypertension 2008, Vol 26 No 3

Table 3 Clinical and biochemical features of hypertensive patients observed for most clinical and biochemical features (data
selected from primary or tertiary care not shown) including age (53.8  12 versus 52.5  15 years)
Patients from Patients from and gender (male 53.7% versus 40.3%); however, remark-
primary care tertiary care
Variable (n 103) (n 141) P
ably, patients with an elevated ratio were characterized by
abnormal potassium handling, with lower serum levels
Age (years) 56.4  9.8 51.8  14.6 < 0.01 (3.8  0.47 versus 4.1  0.5 mmol/l, P < 0.05) and higher
Gender (M/F) 51/52 75/66 NS
SBP (mmHg) 145.8  11.9 161.1  21.5 < 0.001
urinary concentrations (56.1  24.9 versus 47.3 
DBP (mmHg) 89.1  9 99.8  12.8 < 0.001 18.7 mmol/l, P < 0.05) than those of the other patients.
Hypertension stage IIa 10.7% 46.3% < 0.001 This aspect was consistent with the supposed mineralo-
S-Na (mmol/l) 141.7  2 141.3  2.7 NS
S-K (mmol/l) 4.59  0.49 3.84  0.4 < 0.001 corticoid effect induced by an impaired renal 11b-HSD2
S-Creatinine (mmol/l) 68.44  17.13 80.8  18.5 < 0.001 activity. Renin (15.5  12.3 vs. 15.9  22.3 pg/ml) and
S-Cortisol (mg/dl) Not available 25  64
P- ACTH (pg/ml) Not available 30.7  23.6
aldosterone (185  155 vs. 199 119 pg/ml) were not
P-Aldosterone (pg/ml) 168.9  93 218.2  121.4 0.001 different in the two groups of patients. The significantly
P-Renin (pg/ml) 14.3  19.7 15.7  18.1 NS lower cortisol (12.1  8.5 versus 20.7  54 mg/dl) and
ARR (pg/ml:pg/ml) 30  25 27.3  24.9 NS
THF (mg/24 h) 1.92  0.95 1.99  1.1 NS higher ACTH (18.7  18.5 versus 26.1  20.9 pg/ml)
a-THF (mg/24 h) 0.98  0.60 1.44  1.2 < 0.001 values, although within the normal range, are consistent
THE (mg/24 h) 3.25  1.46 3.32  1.7 NS with an inhibition of the adrenalpituitary axis in patients
THFs/THE ratio 0.91  0.26 1.06  0.5 < 0.01
THFs/THE ratio 1.5 3 (3%) 19 (14%) < 0.01 with an elevated ratio.
Family history of hypertension 27% 62% < 0.001

Data are means  SD. ACTH, adrenocorticotrophic hormone; ARR, aldosterone to

renin ratio; DBP, diastolic blood pressure; P, plasma; S, serum; SBP, systolic
blood pressure; THE, tetrahydrocortisone; THF, tetrahydrocortisol; THFs, tetra-
hydrocortisol allotetrahydrocortisol. Statistical significance (P < 0.05) was
defined by Students t-test, KruskalWallis or x2 test. a Classification of BP
values according to JNC 7 (SBP >160 mmHg or DBP >100 mmHg) [28] in
patients treated with verapamil and/or a-adrenergic blockers.
On average, a-THF and, in turn, the THFs/THE ratio
were higher in the urine of patients from tertiary care than
in individuals selected from primary care (Table 3).
Accordingly, a ratio > 1.5 was observed in 19 of 141
(14%) patients examined in the Hypertension Unit but
only in three out of 103 (3%) hypertensive individuals
selected by the general practitioners.
The total population of hypertensive patients was also
analysed in order to find out possible clinical hallmarks
characterizing the patients showing an elevated THFs/
THE ratio. In the hypertensive population considered as
a whole, 10 patients had previously documented brain
ischaemic damage, four of them showed an urinary
THFs/THE ratio > 1.5, that is 18.1% (4/22) of this
subgroup. At echo-Doppler ultrasonography, a significant
(> 50% lumen reduction) carotid artery stenosis was
demonstrated in 11 patients, five of them showed an
urinary THFs/THE ratio > 1.5, that is 22.7% (5/22) of
this subgroup. Differences in distribution of both these
clinical features between high and normal ratio patients
were statistically significant (P < 0.05). In contrast, dia-
betes was no more frequent, and body mass index and
blood glucose concentrations were similar in normal
patients and those with a high THFs/THE ratio.
Comparison between all the patients with elevated
versus normal THFs/THE ratio
The entire study population (groups A B C) was then
divided in two subgroups according to THFs/THE ratio
values (< or  1.5, respectively). No differences were
Discussion
The main purpose of the present work was to establish
the possible role of abnormal cortisol/cortisone handling,
as shown by urinary THFs/THE ratios ranging from 1.5
to 3.0, in human hypertension. We obtained evidence
that a significant number of individuals who received
glucocorticoid treatment share a phenotype similar to
that previously described in some EH patients [68],
characterized by both arterial hypertension and a mildly
elevated (> 1.5) urinary THFs/THE ratio. Moreover, for
the first time, we quantified the prevalence of such EH
patients in different clinical settings, showing that a
definite proportion may be found in both primary and
tertiary care. Overall, these results support the view an
acquired factor (in the case of glucocorticoid-treated
patients) or constitutive factor(s) (in the case of EH
patients) similarly effect a common pathway involved in
BP regulation.
It has been supposed that chronic glucocorticoid therapy
may challenge the functional capability of 11b- hydroxy-
steroid dehydrogenases and facilitate the development of
hypertension [20,21]. Prednisone and its active metab-
olite prednisolone are substrates for both 11b-HSD type
1 and 2 [23]; compared with cortisol, prednisone and
especially prednisolone are preferentially oxidized by
11b-HSD2 in an in vitro model, providing the pharma-
cokinetic explanation for their reduced mineralcorticoid
activity [24]. Chemical structures of cortisol/cortisone
and prednisolone/prednisone are indeed very similar,
differing only for the D1-dehydroconfiguration that
characterizes prednisolone/prednisone but not the adre-
nal hormones. Possible inhibitory competition for renal
11b-HSD2 is therefore plausible, although not yet
demonstrated in vivo or in a clinical context.
In our population, an unbalanced increase in tetrahydro-
metabolites of cortisol versus tetrahydro-metabolites of

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

cortisone occurred in about 30% of individuals chronically
treated with lowmoderate doses (on average 78 mg/day)
of prednisone. Most of these individuals with initially
normal BP became hypertensive during treatment,
whereas only 5 patients with a THFs/THE ratio >1.5
remained normotensive, so that this condition was associ-
ated with a 3.8-fold higher risk of developing hypertension
after glucocorticoid therapy. Although dose and duration of
such therapy did not affect the overall risk of developing
hypertension (Table 1), patients with an elevated THFs/
THE ratio had significantly lower ACTH and cortisol
values than individuals with a normal ratio, thus suggesting
an increased sensitivity to the glucocorticoids and a more
marked inhibition of the adrenalpituitary axis despite
similar prednisone therapy. Long-term competition of
prednisolone with cortisol for b-oxidation by 11b-HSD2
may allow for a prolonged half-life of the residual cortisol
not transformed into cortisone, and a more elevated
(though normal) concentration of active cortisol available
at the pituitary cellular level, thus possibly suggesting that
some individuals are more exposed to such a competitive
mechanism than others.
Another possible interpretation of these findings could be
that of a greater prednisone-related induction of liver
11b-HSD1 isoform in these patients, with an increased
conversion of cortisone into cortisol. For some authors
[29,30] THFs/THE ratio represents a global indicator of
both 11b-HSD isoenzyme activities rather than that of
the type 2 isoform alone, while urinary free cortisol
(UFF) and cortisone (UFE) might be more specifically
informative about 11b-HSD2 activity [18,29,30].
We did not measure UFF/UFE ratio and this certainly
may represent a limitation of the present study. Such a
limitation, however, does not substantially lessen the
value of the results because it does not change the fact
that an increased THFs/THE ratio (ranging from 1.5 to
3.0) was consistently and statistically associated with
higher levels of clinically relevant hypertension. Another
possible limitation of our conclusions may originate from
the hypothetical confounding role exerted by inflamma-
tory cytokines on 11b-HSD2 activity. A recent report has
indeed identified the interleukins IL1b and IL6 and
tumour necrosis factor-a (TNFa) as in-vitro enzymatic
inhibitors of placental 11b-HSD2 [31]; however, in our
patients affected by arthritis, markers of inflammation
were similar in patients with or without impaired 11b-
HSD2 activity (data not shown), and rheumatic disease
was in remission phase at the time of the study.
Both the type (prednisone) and doses of the glucorticoid
therapy in our patients are currently used in rheumato-
logy and in internal medicine. If one considers that
patients who receive glucocorticoids in equivalent doses
(prednisolone equivalent  7.5 mg/day) are more than
2.5 times as likely as patients who do not use glucocorti-
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THFs/THE in different forms of hypertension Olivieri et al. 491
coids to experience a cardiovascular event [32], the
practical consequences are evident. Consistent with this
view, greater BP reduction after a thiazide diuretic was
recently observed in hypertensive patients with an elev-
ated THFs/THE ratio [12].
As expected on the basis of previously reported evidence
[21,22], our data also show that inhibition of 11b-HSD2
activity is not the only mechanism leading to glucocorti-
coid-associated hypertension. Correspondingly, impaired
11b-HSD2 activity may not necessarily result in hyper-
tension but other factors may interfere with this mechan-
ism. As previously reported, in our population, a normal
THFs/THE ratio was found in more than two-thirds of
prednisone-treated hypertensive patients, whereas five
women remained normotensive despite an elevated
THFs/THE ratio. Of note, four out of these women were
rather young and had a regular menstrual cycle, suggesting
that female steroid hormones may to some extent offset or
block the mechanism leading to hypertension.
The second relevant result of the work is that an increase
of urinary THFs/THE ratio ranging between values
> 1.5 and 3 characterizes some spontaneously hyperten-
sive patients, with a substantially higher prevalence in
more severely affected individuals (14%) than in patients
with mild hypertension (3%, see Table 3). In 60 healthy
normotensive controls, nobody showed an urinary THFs/
THE ratio > 1.5.
To our knowledge, this is the first work exploring the
ratio of cortisol to cortisone metabolites in hypertensive
patients from primary care and, in turn, comparing this
prevalence with that observed in patients referred to a
Hypertension Unit. A clear gradient between these
different settings was observed, suggesting that an abnor-
mal urinary ratio of cortisol to cortisone metabolites is
preferentially associated with more severe hypertension.
This conclusion is consistent with a recent report showing
that a mild impairment of 11b-HSD2 activity is associ-
ated with left ventricular mass in essential hypertension
[33]. Interestingly, an increased THFs/THE ratio (above
the mean normal value 2SD 1.54) and left ventricular
hypertrophy at echocardiography were found in the
same percentage of hypertensive patients [33]. In our
population of EH patients, though quantitatively limited,
the proportion of patients with an elevated THFs/THE
ratio who also presented previous cerebrovascular
damage (4/22, 18%) or carotid artery stenosis (5/22,
23%) was striking. Overall, these findings may suggest
that the phenotype of hypertension plus a high ratio may
have a poorer cardiovascular prognosis than EH patients
with a normal ratio. This is also consistent with the
epidemiological evidence that glucocorticoid exposure
(i.e. a prednisolone equivalent  7.5 mg/day, a dose
quite similar to that taken by our arthritic patients)
is associated with an elevated risk of heart failure,
 492 Journal of Hypertension 2008, Vol 26 No 3
myocardial infarction and stroke [32]. Thus, if caution is
usually suggested in glucocorticoid treatment of patients
with hypertension, alternative therapeutic options should
probably be considered in the case of an associated
increase of THFs/THE ratio.
Hypertensive patients with an increased THFs/THE ratio
were clinically indistinguishable from the other hyperten-
sive patients. Lower (but still normal) potassium levels and
male gender were the only variables characterizing hyper-
tensive individuals with elevated THFs/THE ratio, but
obviously both these variables are not sufficiently specific
to suspect the problem in practice. Thus, subtly altered
cortisol metabolism does not necessarily affect traditional
markers of mineralocorticoid hypertension such as renin or
aldosterone, although quantitatively more obviously
expressed defects such as classic AME do. In other reports,
individuals heterozygous for mutations inactivating the
11BHSD2 gene, who were studied as first-degree relatives
of AME patients, and having modestly increased THFs/
THE ratio, were clinically indistinguishable from EH
patients, and their renin activity was normal or not
suppressed [15,26]. In our RA patients, active renin was
not different between individuals with an increased or a
normal THFs/THE ratio, but BP-lowering treatment
(mainly based on ACE inhibitors and diuretics) could have
influenced the result. This is a clear limitation of our study.
Moreover, it is also possible that the use of less precise,
direct active renin rather than the plasma reninangioten-
sin (PRA) assay have influenced this aspect. Some previous
literature findings reported a weak association between
THFs/THE ratio and recumbent but not upright renin
activity in hypertensive patients [12,29]. Since we studied
upright active renin, it is possible that the relationship
between these parameters was missed.
Several other studies suggest that an altered vasoactive
response, derived by an increased vasoconstrictor sensi-
tivity to catecholamines and/or endothelin-1, rather than
an altered volume, are the mechanisms leading to hyper-
tension during 11b-HSD2 inhibition [34,35].
If the present results undoubtedly show a link between
hypertension and slightly impaired 11b-HSD2, so con-
tributing to understanding the role that even subtle
alterations of glucocorticoid metabolism plays on BP
regulation, they do not provide unambiguous evidence
for the consequent mechanism of activation of MR. This
finding, and the selective susceptibility to the competi-
tive inhibition by prednisone/prednisolone on 11b-HSD2
in a proportion of individuals, remain questions to be
ascertained by future investigation.
Acknowledgements
We are grateful to the following general practitioners for
their important contribution in the data collection and
recruitment of patients in the primary-care setting: Paolo
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Adami, Carlo Ballarini, Anna Barbera, Gaetano Benati,
Bruno Benedetti, Claudio Bertaiola, Tiziano Bonaldi,
Annarosa Bovo, Umberto Cacchi, Marco Cherubini,
Damiano Chiesa, Massimo Chincarini, Alberto Ciacciar-
elli, Giorgio Ciampalini, Roberto Ciresa, Giuseppe Coccia,
Chiara Cressoni, Giuliana Dal Pozzo, Roberto Esposti,
Francesco Faraci, Marina Ferrarini, Guido Filippini, Anna
Fioretta, Adelheid Fischer, Carlo Andrea Franchini,
Guglielmo Frapporti, Federica Galvani, Beatrice Gargiulo,
Angelo Garofalo, Walter Idolazzi, Giorgio Laciniati,
Moreno Leoncini, Livio Libardi, Serena Losi, Italo
Lovato, Silvio Mantovani, Raffaella Marrocchella,
Innocenzo Maurelli, Marco Pietro Mazzi, Silvia
Menegazzi, Ernesto Menini, Alessio Micchi, Mauro
Montana, Osvaldo Morbioli, Franco Motta, Antonio
Panzino, Adriano Ramanzini, Giandomenico Righetti,
Arianna Rizzi, Rudi Santini, Gustavo Sartori, Paolo
Scarpolini, Francesco Schiera, Maurizio Sciortino,
Alessandro Severi, Lionello Signorati, Cesare Testi,
Mariella Varaschin and Matteo Venturi. We are also
grateful to Gianstefano Blengio and Denise Signorelli
for their rigorous work in selecting individuals from the
Demographic Register of Bussolengo, and to Giorgio
Parise for his valuable help in collecting blood and urine
samples.
This work has been supported by grants from the Min-
istry of the University and Scientific and Technological
Research (O.O.), the Veneto Region Department of
Health (O.O.) and Cariverona Foundation (S.F.).
The authors report no conflict of interest.
References
1 Quinkler M, Stewart PM. Hypertension and the cortisol-cortisone shuttle.
J Clin Endocrinol Metab 2003; 88:23842392.
2 Funder JW, Pearce PT, Smith R, Smith AT. Mineralocorticoid action: target
tissue specificity is enzyme, not receptor mediated. Science 1987;
242:583585.
3 New MI, Levine LS, Biglieri EG, Pareira J, Ulick S. Evidence for an
unidentified steroid in a child with apparent mineralcorticoid hypertension.
J Clin Endocrinol Metab 1977; 49:924933.
4 Wilson RC, Krozowski ZS, Li K, Obeyesekere VR, Razzaghy-Azar M,
Harbison MD, et al. A mutation in the HSD11B2 gene in a family with
apparent mineralcorticoid excess. J Clin Endocrinol Metab 1995;
80:22632266.
5 Wilson RC, Dave-Sharma S, Wei JQ, Obeyesekere VR, Li K, Ferrari P, et al.
A genetic defect resulting in mild low-renin hypertension. Proc Natl Acad
Sci U S A 1998; 95:1020010205.
6 Walker BR, Stewart PM, Shackleton CH, Padfield PL, Edwards CR.
Deficient inactivation of cortisol by 11 beta-hydroxysteroid dehydrogenase
in essential hypertension. Clin Endocrinol (Oxf) 1993; 39:221227.
7 Walker BR, Stewart PM, Padfield PL, Edwards CRW. Increased
vascular sensitivity to glucocorticoids in essential hypertension:
11b-hydroxysteroid dehydrogenase deficiency revisited. J Hypertens
1991; 9:10821083.
8 Soro A, Ingram MC, Tomolo G, Glorioso N, Fraser R. Evidence of coexisting
changes in 11b-hydroxysteroid dehydrogenase and 5b-reductase activity
in subjects with untreated essential hypertension. Hypertension 1995;
25:6770.
9 Agarwal AK, Giacchetti G, Lavery G, Nikkila H, Palermo M, Ricjetts M, et al.
CA-repeat polymorphism in intron 1 of HSD11B2 effects on gene
expression and salt-sensitivity. Hypertension 2000; 36:187194.
10 Ferrari P, Krozosky Z. Role of the 11beta-hydroxysteroid dehydrogenase
type 2 in blood pressure regulation. Kidney Int 2000; 57:13741381.
 THFs/THE in different forms of hypertension Olivieri et al. 493

11 White PC, Agarval AK, Li A, Nikkila H, Pratt JH, Caulfield M, et al. Possible 33 Glorioso N, Filigheddu F, Parpaglia PP, Soro A, Troffa C, Argiolas G,
association but no linkage of the HSD11B2 gene encoding the kidney Mulatero P. 11b-hydroxysteroid dehydrogenase type 2 activity is
isoenzyme of 11beta-hydroxysteroid dehydrogenase to hypertension in associated with left ventricular mass in essential hypertension. Eur Heart J
black people. Clin Endocrinol 2001; 55:249252. 2005; 26:498504.
12 Williams TA, Mulatero P, Filigheddu F, Troffa C, Milan A, Argiolas G, et al. 34 Zhang Z, Kang Y, Yu Y, Wei S, Schmidt TJ, Johnson AK, Felder RB.
Role of HSD11B2 polymorphism in essential hypertension and the diuretic 11b-Hydroxysteroid dehydrogenase type 2 activity in hypothalamic
response to thiazides. Kidney Int 2005; 67:631637. paraventricular nucleus modulates sympathetic excitation. Hypertension
13 Mariniello B, Ronconi V, Sardu C, Pagliericcio A, Galletti F, Strazzullo P, 2006; 48:127133.
et al. Analysis of the 11b-hydroxysteroid dehydrogenase type 2 gene 35 Frey FJ, Odermatt A, Frey BM. Glucocorticoid-mediated mineralocorticoid
(HSD11B2) in human essential hypertension. Am J Hypertens 2005; receptor activation and hypertension. Curr Opin Nephrol Hypertens 2004;
18:10911098. 13:451458.
14 Carvajal CA, Romero DG, Mosso LM, Gonzalez AA, Campino C, Montero J,
Fardella CE. Biochemical and genetic characterization of 11b-
hydroxysteroid dehydrogenase type 2 in low-renin essential hypertensives.
J Hypertens 2005; 23:7177.
15 Li A, Tedde R, Krozoswi ZS, Pala A, Li KXZ, Shackelton CHL, et al.
Molecular basis for hypertension in the type II variant of apparent
mineralcorticoid excess. Am J Hum Genet 1998; 63:370379.
16 Lovati E, Ferrari P, Dick B, Jostarndt K, Frey FJ, Schorr U, Sharma AM.
Molecular basis of human salt-sensitivity: the role of the 11 beta-
hydroxysteroid dehydrogenase type 2. J Clin Endocrinol Metab 1999;
84:37453749.
17 Melander O, Frandsen E, Groop L, Hulthen UL. No evidence of a relation
between 11b-hydroxysteroid dehydrogenase type 2 activity and salt
sensitivity. Am J Hypertens 2003; 16:729733.
18 White PC, Mune T, Agarwal AK. 11b-Hydroxysteroid dehydrogenase and
the syndrome of apparent mineralocorticoid excess. Endocrine Rev 1997;
18:135156.
19 Ferrari P, Sansonnens A, Dick B, Frey FJ. In vivo 11beta-HSD2 activity:
variability, sensitivity and effect of licorice. Hypertension 2001; 38:1330
1336.
20 Oberfield SE, Levine LS, Carey RM, Greig F, Ulick S, New MI.
Metabolic and blood pressure responses to hydrocortisone in the
syndrome of apparent mineralcorticoid excess. J Clin Endocrinol Metab
1983; 56:332339.
21 Whitworth JA, Mangos GJ, Kelly JJ. Cushing, cortisol, and cardiovascular
disease. Hypertension 2000; 36:912916.
22 Brem AS. Insights into glucocorticoid-associated hypertension. Am J
Kidney Dis 2001; 37:110.
23 Stewart PM, Walker BR, Holder G, OHalloran D, Shackleton CHL.
11 beta-hydroxysteroid dehydrogenase activity in Cushings syndrome:
explaining the mineralocorticoid excess state of the ectopic
adrenocorticotropin syndrome. J Clin Endocrinol Metab 1995; 80:3617
3620.
24 Diederich S, Eigendorff E, Burkhardt P, Quinkler M, Bumke-Vogt C,
Roche M, et al. 11b-Hydroxysteroid dehydrogenase types 1 and 2: an
important pharmacokinetic determinant for the activity of synthetic
mineralo- and glucocorticoids. J Clin Endocrinol Metab 2002; 87:5695
5701.
25 Olivieri O, Ciacciarelli A, Signorelli D, Pizzolo F, Guarini P, Pavan C, et al.
Aldosterone to renin ratio in a primary care setting: the Bussolengo study.
J Clin Endocrinol Metab 2004; 89:42214226.
26 Lavery GG, Ronconi V, Draper N, Rabbit EH, Lyons V, Chapman KE, et al.
Late-onset apparent mineralcorticoid excess caused by novel compound
heterozygous mutations in the HSD11B2 gene. Hypertension 2003;
42:123129.
27 Shackleton CH. Mass spectrometry in the diagnosis of steroid-related
disorders and in hypertension research. Steroid Biochem Mol Biol 1993;
45:127140.
28 Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr,
et al., The National High Blood Pressure Education Programme
Coordinating Committee. Seventh Report of the Joint National Committee
on Prevention, Detection, Evaluation, and Treatment of High Blood
Pressure. Hypertension 2003; 42:12061252.
29 Palermo M, Shackelton CH, Mantero F, Stewart PM. Urinary free cortisone
and the assessment of 11b-hydroxysteroid dehydrogenase activity in man.
Clin Endocrinol 1996; 45:605611.
30 Quinkler M, Oelkers W, Diederich S. In vivo measurement of renal
11b-hydroxysteroid dehydrogenase type 2 activity. J Clin Endocrinol
Metab 2000; 85:49214922.
31 Kossinteva I, Wong S, Johnstone E, Guilbert L, Olson M, Mitchell BF.
Pro-inflammatory cytokines inhibit human placental 11beta-hydroxysteroid
dehydrogenase type 2 activity through calcium and cAMP pathways.
Am J Physiol Endocrinol Metab 2006; 290:E282E288.
32 Wei L, MacDonald TM, Walker BR. Taking glucocorticoids by prescription
is associated with subsequent cardiovascular disease. Ann Intern Med
2004; 141:764770.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.