J Clin Periodontol 2000; 27: 53–60 Copyright C Munksgaard 2000

Printed in Denmark . All rights reserved

ISSN 0303-6979

D. Sakellari1, J. M. Goodson2,
Concentration of 3 tetracyclines in S. S. Socransky2, A. Kolokotronis1
and A. Konstantinidis1

plasma, gingival crevice fluid and

1
Department of Periotontology, Dental
School, Aristotelian University of
Thessaloniki, Greece; 2Department of
Periodontology, Forsyth Dental Center,

saliva Boston MA, USA

Sakellari D, Goodson JM, Socransky SS, Kolokotronis A, Konstantinidis A: Con-

centration of 3 tetracyclines in plasma, gingival crevice fluid and saliva. J Clin
Periodontol 2000; 27: 53–60. C Munksgaard, 2000.

Abstract
Background: Systemically-administered tetracyclines have been used widely for
treatment of periodontal diseases with little understanding of their delivery charac-
teristics to periodontal tissues. This study was designed to measure concen-
trations of 3 tetracyclines in gingival crevice fluid (GCF), plasma and saliva of
following systemic administration.
Method: The concentration of tetracycline (TC), minocycline (MN) and doxycy-
cline (DX) was measured in gingival crevice fluid (GCF), plasma and saliva of 20
subjects following single sequential standard oral systemic doses. Gingival crevice
fluid concentration was measured at 4 sites (2 shallow and 2 deep) before ad-
ministration, and at 1 h and 2 h following administration. Plasma and saliva
concentrations were measured from in samples at the same time points. No
antibacterial activity was detected before administration. The highest concen-
trations were measured 2 h after administration.
Results: The average concentrations at 2 h were highest in plasma (TCΩ1.02,
MNΩ2.18, DXΩ2.35 mg/ml). Intermediate concentrations were measured in
GCF (TCΩ0.61, MNΩ1.49, DXΩ1.65 mg/ml). Saliva concentrations (TCΩ0.09
MNΩ0.31, DXΩ0.47 mg/ml) were the lowest of the 3 fluids monitored. Data are
presented indicating that the average GCF concentration of systemically adminis-
tered tetracyclines is less than the that of plasma concentration. The concentration
of tetracyclines in GCF was strongly associated with plasma concentration, indi-
cating a primary rôle of drug absorption in the delivery of these systemically ad-
ministered antibiotics to the site of action in periodontal therapy. The average
GCF concentration in individuals varied widely (between 0 and 8 mg/ml) with ap- Key words: tetracycline; minocycline;
proximately 50% of samples not achieving levels of 1 mg/ml. doxycycline; tetracyclines/analysis; blood;
antibiotics/therapeutic use; gingival crevicular
Conclusions: These observations suggest that poor absorption of orally-adminis- fluid/analysis
tered tetracyclines in many individuals may account for much of the variability in
clinical response to antibiotics observed in practice. Accepted for publication 22 March 1999

Systemically administered tetracyclines,
most often tetracycline (TC), minocyc-
line (MN) or doxycycline (DX) are
often frequently used in the adjunctive
treatment of periodontal diseases (Van
Winkelhoff et al. 1996). Concentrations
of these tetracyclines appearing in gin-
gival crevice fluid (GCF) following sys-
temic administration have been re-
ported in studies of different patient
groups (Goodson 1994).
Tetracycline differs from minocycline
and doxycycline in many pharmaco-
kinetic characteristics (Benet et al.
1996). Although all tetracyclines are in-
completely absorbed, the oral availabil-
ity of tetracycline is considerably lower
than that of MN or DX. In addition,
MN and DX are not affected by con-
comitant food ingestion. TC differs from
MN and DX in having a higher percen-
tage of urinary excretion, a lower % of
plasma binding, clearance that is more
rapid, a larger volume distribution and a
shorter half-life.
It has been recognized for some time
that individuals differ widely in their
peak plasma levels of tetracyclines,
which has been attributed to variability
in absorption. Generally, a peak plasma
concentration of 2 to 2.5 mg/ml occurs
2 to 4 h following oral administration
of repeated doses of TC. Oral adminis-
54 Sakellari et al.

line. 4 days after the minocycline ad-

ministration, the procedure was re-
peated with doxycycline.

Antibiotic administration

On the appointed day at each of 3

visits, each subject was given their his/
her antibiotic under supervision. The
usual recommended dosage of each
antibiotic was administered. Single oral
doses of tetracycline (TerramycinA, te-
tracycline hydrochloride, 250 mgΩ231
mg baseΩ520 mmoles), minocycline
(MinocinA, minocycline hydrochloride,
100 mg baseΩ219 mmoles), and doxycy-
cline (VibramycinA, doxycycline hycl-
ate, 100 mg baseΩ225 mmoles) were
given sequentially to each patient.
72 h elapsed between the administra-
Fig. 1. Protocol of tetracycline administration study. Subjects were administered either tetracy-
tion of tetracycline and minocycline
cline, minocycline or doxycycline at 1 of the 3 visits. Samples of blood, saliva and GCF were
taken at baseline (just prior to administration) at 1 h following administration and at 2 h
providing approximately 6.7 half times
following administration for analysis of antibiotic concentration. GCF samples were taken for TC excretion. 120 h elapsed between
from two deep (>6mm) and two shallow (<4 mm) pockets. GCF samples were taken from the administration of minocycline and
the same sites following each antibiotic administration. Visits were scheduled to allow at least doxycycline providing approximately
5 half-times for drug clearance (T1/2(TC)Ω10.6 hours, T1/2(MN,DX)Ω16 h). 7.5 half times for MN excretion (Fig.
1).

Sample taking

tration of 200-mg doxycycline creates a
peak plasma concentration of 3 mg/ml
at 2 h. (Kapusnik-Uner et al. 1996).
Little is known of the concentrations
occurring in GCF and saliva relative to
plasma concentrations following oral
administration. In order to investigate
mechanisms by which periodontal ther-
apy may succeed or fail in an individual,
plasma, GCF and saliva concentrations
were measured following a single oral
dose of the 3 tetracyclines. All measure-
ments were taken in the same group of
subjects, sampling the same gingival
sites in each case.
Materials and Methods
Experimental design
proximately 10 h had elapsed since their
last meal) and were administered the se-
lected antibiotic. All subjects provided
informed consent to participate in a 9-
day schedule of tetracycline administra-
tion as illustrated in Fig. 1.
At the first visit, two shallow (<4
mm) and two deep (>6 mm) sites were
selected for GCF sampling. Subject and
site characteristics are summarized in
Table 1. Saliva, blood and GCF were
collected for baseline determinations
and the subject was given 1 tetracycline
dose. At 1 and 2 h, samples were again
collected for assay. The subject was re-
appointed three days later when the
procedure was repeated with minocyc-
Prior to and following antibiotic ad-
ministration, samples of gingival fluid,
blood and saliva were taken for analysis
by biological assay. All samples were
taken at 1 h and again at 2 h following
antibiotic administration. GCF samples
were collected by intrasulcular place-
ment of filter paper strips. GCF was
allowed to collect for sufficient time at
each site until visible wetness was ob-
served. Blood samples were taken by
venipuncture. Whole saliva samples
were collected by expectoration.
The approximate time of maximum
antibiotic concentration in plasma and
GCF was determined in a separate ex-
periment by administering single oral

20 subjects were selected from those

Table 1. Characteristics and numbers of subjects and sites selected to participate in the study
being treated at the Department of Pre-
ventive Dentistry and Periodontology, Subjects
Dental School, Aristotelian University number 20
of Thessaloniki. Subjects were selected gender 12M, 8F
who had at least 2 periodontal pockets mean age 51 (39–66)
>6 mm and, had not received systemic Sites
antibiotics within the last 3 months,
and were not known to have disease Baseline measures Shallow Deep
other than periodontitis. Subjects were number 40 40
not taking antacids or vitamin supple- pocket depth 3.1∫0.4* 7.2∫1.5*
ments at the time tested. All subjects ar- attachment level 4.2∫0.8* 7.9∫1.8*
rived early in the morning of their ap- mean% bleeding on probing 87.5% 100%
pointed time in a fasting condition (ap- *Mean∫standard deviation.
 Concentration of 3 tetracyclines 55

Table 2. Average concentration of tetracycline, minocycline, and doxycycline in saliva, gingival each applied sample and concurrent
crevice fluid and plasma following administration of a single dose of each antibiotic standards.
Antibiotic Time N Plasma GCF Saliva Standards were made by dissolving
TC 1 16 0.61∫0.82 0.30∫0.66 0.00∫0.01
tetracycline hydrochloride (Sigma T
2 20 1.02∫1.20 0.61∫0.67 0.09∫0.30 3383, formula weight 480.9), minocyc-
MN 1 17 [1.37∫1.22 [0.75∫1.30 0.20∫0.33* line hydrochloride (Sigma M 9511, for-
2 20 [2.18∫1.63* [1.49∫1.59† 0.31∫0.44* mula weight 493.9) and doxycycline hy-
DX 1 15 [0.96∫1.09 [0.83∫1.44 [0.16∫0.17 drochloride (Sigma D9891, formula
2 20 [2.35∫1.88* [1.65∫2.06† [0.47∫0.64* weight 480.9). Calibration curves of
Concentrations at times with brackets are significantly different by paired t-test, aΩ0.05. standards containing 16, 8, 4, 2, 1,0.5,
*Statistically different from TC by ANOVA, aΩ0.05. 0.25, and 0.125 pg of each antibiotic
†
Borderline difference from TC by ANOVA, aΩ0.08. were included with GCF samples from
Measurements were taken at 1 h and 2 h following administration of each antibiotic. Values each patient visit (20 calibrations/anti-
represent the mean∫SD from each patient. N is the number of subjects for which determi- biotic) on each plate. Each assay re-
nations were made. sponse was fitted to a power function of
the form qΩkAc where qΩthe amount
of antibiotic and AΩthe area of inhi-
doses of 100, 200 and 400 mg doxycy- of hemolysin release was measured by bition. The parameters k and c were de-
cline to 1 volunteer. 2 weeks elapsed be- the area of the oval approximating the termined by a least squares fit of the
tween each single dose administration observed inhibition zone using major log-transformed values of q and A.
to provide sufficient time for wash out. and minor diameters of the inhibition Concentration (pg/mlΩmg/ml) of the hy-
GCF from 4 periodontal pockets (5–6 zone measured to the nearest mm on drochloride salt of each antibiotic was
mm deep) and finger stick blood
samples were taken every hour for 10
h following doxycycline administration.
GCF concentration was computed as
the average of the 4 samples taken at
each time point. Doxycycline in this
part of the study was analyzed in the
same manner as were the samples from
the larger group of subjects.

Sample analysis

All samples were analyzed by bioassay

using methods previously described
(Gordon et al. 1980). Samples were ap-
plied to standardized filter paper strips Fig. 2. Average plasma concentration of antibiotics 2 h following oral administration. Subjects
(Periotron) and placed on agar culture were sorted by their plasma concentration. The white dashed line indicates a concentration
media containing whole sheep blood of 1 mg/ml.
and Clostridium perfringens ATCC
13124. After incubation at 37æC for 4 h
under anaerobic conditions, the zone of
inhibition of hemolysis was measured
as the outcome variable. Blood samples
were centrifuged and the plasma as-
sayed by measurement of the zone of
hemolysis inhibition resulting from 1 ml
of plasma applied to a filter paper strip.
Saliva samples were centrifuged and as-
sayed by measurement of the zone of
hemolysis inhibition of 2 ml of the
supernatant applied to a Periotron filter
paper strip. The volume of gingival
crevice fluid absorbed on filter paper
strips was measured by conductance
(Periotron HAR-6000) and assayed di-
Fig. 3. Correlation between saliva or GCF and plasma. The plasma concentration of all sub-
rectly. Prior to assay, the colored tab
jects following administration of tetracycline, minocycline and doxycycline are correlated with
was removed from filter paper strips of salivary and average GCF concentrations at both 1 and 2 h sampling times. The resulting
all samples. linear relationships suggest that saliva concentrations are approximately 19% and GCF con-
Assay response was determined using centrations are approximately 68% of plasma concentrations. The squared correlation coef-
the zone of hemolysis inhibition. The ficients of 0.47 for saliva and 0.50 for plasma suggest that approximately 50% of the variability
area of the zone representing inhibition of each of these concentrations is explained by the plasma concentration.
56 Sakellari et al.

0.0058¿periotron units. By least

squares analysis, an average response of
173 periotron units/ml with a squared
correlation coefficient of 0.96 (r2) was
computed. A mean sample volume of
0.74∫0.22 ml (mean∫SD) and a range
of 0.5 to 1.25 ml were measured in the
426 GCF samples collected. 82% of
GCF samples measured within the cali-
bration range of 0.5 to 1 ml.
The mean GCF volume collected at
shallow sites (0.56∫0.14) was on aver-
Fig. 4. Average GCF concentration of tetracycline antibiotics two hours following oral admin- age 0.36 ml less than that collected at
istration. Subjects were sorted for the highest average concentration. Black bars represent the deep sites (0.92∫0.12) (values in ml∫SD
mean of 4 site measurements and the whisker represents the standard deviation. White circles
for 214 sites). Despite this observation,
indicate the concentration of individual samples (4/subject). The white dashed line indicates
an antibacterial concentration of 1 mg/ml.
the GCF volume collected was not sig-
nificantly related to the concentration
(r2Ω0.007).

Antibiotic assay

computed from the amount of anti- The amount of antibiotic on a sample

Results
biotic determined from the assay (pg) strip was determined by an approximat-
Volume calibration
divided by the volume determined from ing function of the form QΩAk (QΩthe
the gingival fluid meter response (ml). The gingival fluid meter was calibrated amount of antibiotic in pg, AΩthe area
All values reported are in mg/ml of the before each visit by application of 0.5, of the zone of hemolysis inhibition in
hydrochloride salt of each antibiotic. To 0.75 and 1 ml samples of distilled water. mm2, k is the fitted parameter). The av-
compute concentrations in mg/ml (base) The volume in ml computed as erage calibration response of each anti-
multiply by the following factors: TCΩ
0.924, MNΩ0.926 and DXΩ0.924. To
compute concentration in mmoles/l, Table 3. Analysis of saliva, gingival crevice fluid and plasma antibiotic concentration by AN-
multiply by the following factors: TCΩ COVA
2.079, MNΩ2.025 and DXΩ2.079. Source Plasma GCF Saliva
model ∞0.0001 ∞0.0001 ∞0.0001
antibiotic 0.0006 0.44 0.26
Data analysis subject ∞0.0001
While a complete data set (20 subjects, time ∞0.0001 0.79
subject *antibiotic 0.0004
80 sites) was collected 2 hours post anti-
time *antibiotic 0.1604
biotic administration, data on 12 sub- plasma concentration ∞0.0001 ∞0.0001
jects (TCΩ4, MNΩ3, DXΩ5) was ab- GCF volume collected 0.33
sent for the 1 h measurement. The mean
Values listed are p-values for the overall model and significant association with the various
of duplicate plasma concentrations was
factors tested. Plasma levels were significantly related to the type of antibiotic, the subject,
used for all comparisons. Significance the time and interaction between subjects and individual antibiotics. GCF levels and saliva
of mean differences between antibiotic levels were significantly related only to the plasma concentration.
concentrations at 1 h and 2 h were
tested by paired t-test, assigning sig-
nificance for two-tail p∞0.05. Differ- Table 4. Regression analysis of GCF and saliva antibiotic concentration compared to plasma
ences in mean concentration between concentration
antibiotics at each time was tested first
Antibiotic Time GCF/Plasma Saliva/Plasma
by analysis of variance (ANOVA) fol-
lowed by a post-hoc comparison using tetracycline 1 0.57* 0.001
paired t-tests. Analysis of plasma, GCF 2 0.24* 0.15*
and saliva concentrations was per- minocycline 1 0.51* 0.15*
formed by analysis of covariance (AN- 2 0.75* 0.16*
COVA). These results were further doxycycline 1 0.85* 0.08*
tested by repeated measures analysis. 2 0.82* 0.26*
The relationship between GCF and sal-
*Significant correlation between plasma and GCF or saliva concentration (p∞0.05) by
iva to plasma concentrations was ana- ANOVA.
lyzed for descriptive purposes by simple Values are the slope of the linear regression. The GCF/ plasma slope indicates that at 2 h, the
linear regression of site data (for GCF) GCF concentration of the antibiotics tested was 24–85% of the plasma concentration. The
and subject data (for saliva) to obtain saliva/plasma slope indicates that at 2 h, the saliva concentration of the antibiotics tested was
slope and correlation coefficients. 0.1–26% of the plasma concentration.
 Concentration of 3 tetracyclines 57

Table 5. Values reported for GCF and plasma concentrations of antibiotics in the order of highest GCF concentration for each antibiotic
GCF Plasma Dose Frequency Duration Sampling
Antibiotic N conc. conc. (mg) (h) (days) time (h) Subjects Reference
TC 6 5–12 1–2.6* 250 or 500 3.5–7 H (Gordon et al. 1981a)
3 4–10 2.2–3.4* 250 6 5 48 H (Pascale et al. 1986)
2 4–8 2–2.5* 250 6 5 48 H (Gordon et al. 1981b)
2 2–4 0.3–1.5* 250 12 5 48 H (Gordon et al. 1981b)
20 0.61 1.02 250 2 PD This paper
MN 13 8.03 2.58 200 24 7 168 PD (Ciancio et al. 1982)
15 2.87–5.97 – 100 12 8 24 and 192 PD (Freeman et al. 1992)
10 5.26 1.02 200 24 PD (Ciancio et al. 1980)
15 3.64–4.77 – 100 24 8 24 and 192 PD (Freeman et al. 1992)
10 3.98 1.5 150 24 PD (Ciancio et al. 1980)
20 1.49 2.18 100 2 PD This paper
DX 4 3–10 2.1–2.9* 200 48 H (Pascale et al. 1986)
20 1.65 2.35 100 2 PD This paper
5 ?.07–1.07 0.91–1.48 100 2,4,6,8,12 PD (Stoller et al. 1998)
*Concentration in blood.
NΩnumber of subjects, GCF conc.Ωantibiotic concentration (mg/ml) in GCF, plasma conc.Ωantibiotic concentration (mg/ml) in plasma, doseΩ
amount of antibiotic given at each administration (mg), frequencyΩtime between administrations (hours) and durationΩtime over which
administration was continued (days). Note: no frequency or duration entries appear for a single dose studies. Sampling timeΩtime (h) after
initial therapy that the GCF sample was taken, subjectsΩdisease status (healthy (H) or periodontally diseased(PD)), referenceΩinformation
source.

biotic and its approximating function tude of average antibiotic concen- concentrations at 2 h. The standard de-
produced an acceptable fit within the trations was plasma±GCF±saliva. Av- viation of paired differences in dupli-
range of 0 to 8 ng. 98.4% of GCF erage saliva concentrations were cate determinations was 0.46 mg/ml
samples assayed contained less than 8 approximately 13% of plasma concen- from which the standard deviation of
ng of the antibiotic. No blood or saliva trations. Average gingival crevice fluid plasma determinations could be esti-
samples exceeded 8 ng. The amount of concentrations were approximately 62% mated to be 0.33 mg/ml.
tetracycline hydrochloride (ng) was of plasma concentrations. Subjects varied in their plasma anti-
computed as 0.0016¿area of inhibition biotic concentrations (Fig. 2). The
in mm2 raised to the 1.78 power. The plasma concentration in one subject
Plasma concentration
amount of minocycline hydrochloride (.17) did not reach 1 mg/ml with any of
was computed as 0.00056¿area of inhi- Average plasma levels were higher at 2 h the systemically administered anti-
bition in mm2 raised to the 1.94 power. than at 1 hour for all antibiotics tested biotics (5% of subjects). The plasma
The amount of doxycycline hydro- (Table 2). There was no significant dif- antibiotic concentration in 25% (nΩ5)
chloride was computed as ference between plasma levels of any of subjects (.3, 4, 8, 10 and 16) achieved
0.00052¿area of inhibition in mm2 the antibiotics tested at 1 h. Both mino- at least 1 mg/ml with all antibiotics. Te-
raised to the 1.93 power. cycline and doxycycline achieved sig- tracycline concentrations reached 1 mg/
Antibiotic detection limits were esti- nificantly higher average plasma levels ml in 8 subjects (40%), minocycline and
mated as the theoretical amount of at two hours than did tetracycline. doxycycline reached plasma levels of 1
antibiotic in a 2 mm diameter circle of There was no significant difference be- mg/ml in 16 subjects (80%). In three sub-
inhibition (3.14 mm2). These values tween minocycline and doxycycline jects (15%), only one of the 3 tested
were 0.02 ng for the average volume of
GCF samples taken (0.74 ml), 0.01 ng
for plasma samples (1 ml), and 0.007 ng
for saliva samples (2 ml). No significant Table 6. Concentration of doxycycline in GCF from four sites collected from a healthy subject
2 hours following systemic administration of 100 mg
difference in detection limits between
the tetracyclines was observed. Collection Volume Collected Concentration
time (min) (ml) (mg/ml)
Site short long short long short long
Average concentration
146 0.5 3.3 0.36 0.88 2.40 3.80
Antibiotic activity was not detected in 266 0.5 3.0 0.18 0.87 4.69 5.40
any of the baseline blood, saliva or GCF 271 0.5 3.1 0.23 1.19 2.90 12.79
samples taken. The average concen- 361 0.5 3.4 0.18 0.92 4.79 4.01
trations of tetracycline, minocycline and average 0.5 3.2 0.2 1.0 3.7 6.5
doxycycline in saliva, gingival crevice Each site was sampled twice in quick succession, first for 0.5 minutes (short) and then for
fluid and plasma at 1 and 2 h following approximately 3 minutes (long). By the short sampling protocol 0.2 ml was collected with an
administration of a single dose are illus- average concentration of 3.7 mg/ml. The long sampling protocol resulted in 1.0 ml samples
trated in Table 2. The order of magni- with an average of concentration of 6.5 mg/ml.
58 Sakellari et al.
Fig. 5. Concentration of doxycycline in plasma and GCF following single doses of 100, 200
and 400 mg. Average GCF concentrations were computed from measurements at 4 sites with
periodontal pockets 5–6 mm deep. Both plasma and GCF samples were taken every hour for
10 h. 2 weeks elapsed between times of the single dose administrations.
antibiotics achieved average plasma
values of 1 mg/ml (.5 – tetracycline
only; .19 – minocycline only, .20 –
doxycycline only).
Factors that would likely affect drug
absorption from an oral route of ad-
ministration were tested in relation to
plasma concentration; specifically of
the antibiotic used, subject differences
and time of sampling (Table 3). Analy-
sis of covariance (ANCOVA) revealed
that plasma concentration was signifi-
cantly related to all of the factors tested
with a significant interaction between
the subject and the type of antibiotic
taken. By contrast, GCF and saliva
concentrations were significantly re-
lated only to the plasma concentration.
Simple linear correlation between sal-
iva and plasma concentration of all
antibiotics irrespective of sample time
or antibiotic type suggests that saliva
levels were 19% of plasma levels. Corre-
lation between GCF and plasma con-
centration suggests that GCF levels
were 68% of plasma levels (Fig. 3). The
squared correlation coefficient values of
0.47 and 0.50 respectively suggest that
approximately 50% of the variability of
both the saliva and GCF concentration
was accounted for by the plasma con-
centration.
GCF concentration
The average GCF concentrations of
MN and DX were greater than TC but
this was of borderline statistical signifi-
cance (at 2 h TCΩ0.61, MNΩ1.49,
DXΩ1.65 mg/ml, Table 2). Mean GCF
concentrations were higher at 2 h than
at 1 h for all antibiotics tested. This dif-
ference was statistically significant for
minocycline and doxycycline.
Considerable variation in the average
GCF concentration of antibiotics was
observed among subjects (Fig. 4). The
average GCF antibiotic concentration
in 20% (nΩ4) of subjects (1, 14, 17, and
20) did not reach 1 mg/ml with any of
the systemically administered anti-
biotics. The average GCF antibiotic
concentration in 15% (nΩ3) of subjects
(4, 8, and 9) achieved at least 1 mg/ml
with all antibiotics. Mean TC concen-
trations reached 1 mg/ml in 5 subjects
(25%), MN in twelve subjects (60%) and
DX in 9 subjects (45%). In 9 subjects
(45%), only 1 of the 3 tested antibiotics
achieved average GCF values of 1 mg/
ml (.5 – TC only; .7, 11, 12, 16, and
19 – MN only, .3, 15, and 18 – DX
only).
Of 80 sites monitored in 20 subjects
at 2 h, 16.3% (13) achieved >1 mg/ml
following administration of tetracy-
cline, 47.5% (38) following minocycline
and 42.5% (34) following doxycycline.
Within subjects, numerous examples of
sites with extremely low antibiotic con-
centrations were detected while other
sites in the same mouth had relatively
high concentrations. The average range
(the highest concentration minus the
lowest) of GCF values in sites within
subjects measured at 2 h was 1.9 mg/ml.
The highest range of GCF measured
was 9.6 mg/ml (doxycycline, subject 3).
This variability is indicated in Fig. 4 by
the height of the standard deviation
whisker.
Several factors that would likely af-
fect gingival fluid concentration were
tested; specifically the plasma concen-
tration, the antibiotic used, the volume
of GCF collected and the pocket depth
of the site being sampled. As indicated
in Table 3, only the plasma concen-
tration was found significantly associ-
ated with GCF concentration. The type
of antibiotic, the volume of GCF col-
lected for analysis and the depth of the
pocket being sampled were not signifi-
cantly related to the GCF concen-
tration.
All gingival fluid levels exhibited a
significant correlation with plasma
levels by regression analysis (Table 4).
The GCF levels at 2 h were 24% (TC),
75% (MN) and 82% (DX) of plasma
levels. Although the concentration of
tetracycline at 1 h was less than at 2 h
(Fig. 4), both the correlation with
plasma levels and the slope were con-
siderably greater at 1 h than at 2 h.
Sampling time
The GCF concentration profile ob-
tained from periodontal pockets of a
single DX-treated subject (Fig. 5) indi-
cates that maximum concentration oc-
curred over a broad peak extending
from 2 to 4 h following administration.
These data suggest that the 2-h sam-
pling protocol was a reasonable time to
determine peak concentration of the te-
tracyclines although it is possible that 3
h samples may have had slightly higher
concentrations.
Saliva concentration
Salivary concentrations of all anti-
biotics measured were generally lower
(Table 2) than that considered antibac-
terial (ca. 1 mg/ml). Mean salivary levels
were higher at 2 h than at 1 h for all

antibiotics tested. The average maxi-
mum saliva concentration of DX at 2 h
was significantly greater than TC but
not significantly different from MN
(TCΩ0.087, MNΩ0.310, DXΩ0.470
mg/ml, Table 2). The only factor that
was significantly associated with saliva
concentration was the plasma concen-
tration (Table 3). The salivary levels at
2 h were 15% (TC), 16% (MN) and 26%
(DX) of plasma levels (Table 4).
Discussion
One of the most important findings of
this study was the frequency with which
tetracyclines failed to achieve antibac-
terial levels in GCF. The mean GCF
concentration of approximately half of
the subjects given single oral doses of 3
tetracyclines did not reach levels that
are normally considered antibacterial (1
mg/ml). The most important single fac-
tor controlling the ability of the tetracy-
clines to establish an antibacterial con-
centration within the periodontal en-
vironment was the plasma
concentration. In other words, if the
drug was not well absorbed, it did not
reach antibacterial levels in the peri-
odontal pocket. In this study, subjects
were administered the antibiotic under
supervision in a fasting state preventing
interference from concomitant divalent
cation consumption. The relatively
poor delivery to the periodontal en-
vironment focuses our attention on the
intrinsic ability of individuals to absorb
tetracyclines. To knowledge, no factors
have been identified that explain indi-
vidual differences in tetracycline ab-
sorption.
Of the tetracyclines, MN and DX
consistently achieved higher concen-
trations than TC. Nevertheless, 20% of
subjects failed to establish antibacterial
levels following oral administration of
these tetracyclines. These observations
may explain why literature reports of
the clinical efficacy of orally adminis-
tered tetracyclines have been so variable
(Slots & Rams 1996).
The TC and MN concentration in
GCF measured in this study was lower
than that reported by several investi-
gators (Table 5). In some studies, higher
concentrations might be attributed to
higher dosage administration, while in
others, accumulation with repeated
dosage may account for higher meas-
ured concentrations. It is also possible
that higher concentrations could have
occurred between 2 and 4 h after ad-
ministration. In some cases, however,
the differences are were so great that
other explanations must be sought.
A fundamental issue is how GCF
concentrations could exceed plasma
concentrations. Considering the passive
mechanism by which GCF is thought to
be produced (Pashley 1976), it is diffi-
cult to imagine how GCF concen-
tration would ever exceed that of
plasma. In the current study of 20 sub-
jects, the average GCF concentration
was found to be 20–50% lower than the
plasma concentration.
A One possible explanation for the
higher GCF concentration in the cited
studies could lie in the nature of the
subjects selected. Several of the reports
of elevated tetracycline levels relative to
plasma came from studies using healthy
subjects. Since healthy subjects have
much lower GCF flow rates than sub-
jects with periodontal disease, long
sampling times may have been used to
collect reasonable sample volumes. In
this situation, it is possible that samples
could have concentrated on filter paper
strips by evaporation. To test this hypo-
thesis, 4 sites of a healthy subject given
100 mg DX were sampled for short and
long time periods (Table 6). The average
concentration measured for the long
sampling protocol was approximately
twice that of the short sampling proto-
col. These data support the hypothesis
that elevated TC concentrations relative
to plasma may be the result of sample
concentration due to evaporation.
All data from this study support the
hypothesis that following systemic ad-
ministration of tetracyclines, GCF con-
centration is established by passive dif-
fusion from plasma. As such, it is con-
ceivable that failure to appreciate
factors, such as evaporation, that can
affect GCF samples has led to reports
of levels exceeding those of plasma.
Evaporation as a source of error was
minimized in this study by selecting
subjects with high GCF flow rates as-
sociated with moderate to advanced
periodontal disease. Experimental error
was minimized by considering a suffi-
ciently large number of subjects to
allow meaningful statistical analysis.
Under these conditions, the GCF con-
centration of systemically administered
tetracyclines was seen to approach
plasma concentration as a maximum
upper limit.
The administration of higher doses
of DX produced proportionally higher
plasma and GCF concentrations (Fig.
Concentration of 3 tetracyclines 59
5). Considering the low and erratic
levels produced by the usual 100 mg
dose, it is possible that higher dosage
schedules may be justified for antibac-
terial therapy.
Acknowledgements
The authors gratefully acknowledge the
editorial assistance of Amy Roberts
and the partial support of research
grants DE-11814 and DE-04881 from
the National Institute of Dental Re-
search.
Zusammenfassung
Konzentrationen von 3 Tetracyclinen in Plas-
ma, Sulkusflüssigkeit und Speichel
Die Konzentrationen von Tetracyclin (TC),
Minocyclin (MN) und Doxicyclin (DX) wur-
den in der Sulkusflüssigkeit (SF), dem Plas-
ma und dem Speichel von 20 Probanden
nach sequentieller einmaliger Gabe der ora-
len Standarddosierung untersucht. Die SF-
Konzentrationen wurden an 4 Stellen (2 fla-
che, 2 tiefe) vor Verabreichung sowie 1 und 2
Stunden danach bestimmt. Die Plasma- und
Speichelkonzentrationen in den Proben wur-
den zu den gleichen Zeitpunkten bestimmt
Vor Verabreichung war keine antibakterielle
Aktivität feststellbar. Die höchsten Konzen-
trationen wurden nach 2 Stunden gemessen.
Die mittleren Konzentrationen nach 2 Stun-
den waren am höchsten im Plasma (TCΩ
1.02; MNΩ2.18; DXΩ2.35 mg/ml). Mittlere
Konzentrationen wurden in der SF festge-
stellt (TCΩ0.61; MNΩ1.49; DXΩ1.65 mg/
ml). Die Speichelkonzentrationen (TCΩ0.09;
MNΩ0.31; DXΩ0.47 mg/ml) waren die ge-
ringsten von den untersuchten Körperflüs-
sigkeiten. Diese Ergebnisse zeigen, daß die
mittleren SF-Konzentrationen systemisch
verabreichter Tetracycline geringer sind als
die Plasmakonzentrationen. Die SF-Konzen-
trationen der Tetracycline waren stark mit
der Plasmakonzentration assoziiert, was auf
eine vorrangige Rolle der Wirkstoffresorp-
tion für die Freigabe dieser systemisch verab-
reichten Antibiotika an ihrem Wirkungsort
in der Parodontaltherapie hinweist. Die mitt-
lere SF-Konzentration variierte stark zwi-
schen einzelnen Probanden (zwischen 0 und
8 mg/ml) mit etwa 50% der Proben, die eine
Konzentration von 1 mg/ml nicht erreichten.
Diese Beobachtungen legen den Schluß nahe,
daß schlechte Absorption der oral verab-
reichten Tetracycline bei vielen Individuen
für die starke Variabilität der klinischen Er-
gebnisse nach Anwendung von Antibiotika
in der Praxis verantwortlich ist.
Résumé
Concentration de 3 tétracyclines dans le plas-
ma, le fluide créviculaire gingival et la salive
La concentration de tétracycline (TC), mino-
60 Sakellari et al.
cycline (MN) et doxyclycline (DX) a été me-
surée dans le fluide créviculaire gingival
(GCF), le plasma et la salive de vingt sujets
suivant une séquence standard de prise de
doses systémiques. La concentration GCF
était mesurée au niveau de 4 sites (2 profonds
et 2 peu profonds) avant l’administration, et
une et deux heures après. Les concentrations
plasmatiques et salivaires ont été mesurées
dans des échantillons prélevés au même mo-
ment. Aucune activité antibactérienne n’a été
détectée avant la prise du médicament. Les
concentrations les plus hautes ont été mesu-
rées 2 h après. Les concentrations moyennes
à 2 heures étaient les plus importantes dans
le plasma (TCΩ1.02, MNΩ2.18, DXΩ2.35
mg/ml). Les concentrations intermédiaires
ont été mesurées dans le GCF (TCΩ0.61,
MNΩ1.49, DXΩ1.65 mg/ml). Les concentra-
tions salivaires (TCΩ0.09, MNΩ0.31, DXΩ
0.47 mg/ml) étaient les plus faibles des 3 flui-
des analysés. La concentration moyenne dans
le GCF des tétracyclines prises par voie sys-
témique est inférieure à celle dans le plasma.
La concentration de tétracycline dans le
GCF était fortement associée avec celle du
plasma, indiquant un rôle essentiel de l’ab-
sorption de la drogue dans la libération de
ces antibiotiques administrés par voie systé-
mique au site d’action du traitement paro-
dontal. La concentration moyenne dans le
GCF chez les individus variait fortement de
0 à 8 mg/ml avec environ 50% des échantillons
n’arrivant pas à des taux de 1 mg/ml. Ces ob-
servations suggèrent qu’une mauvaise ab-
sorption des tétracyclines administrées par
voie buccale chez beaucoup d’individus peut
entraı̂ner cette forte variation dans la répon-
se clinique aux antibiotiques observée en pra-
tique clinique.
References
Benet, L. Z., Oie, S. & Schwartz, J. B. (1996)
Design and optimization of dosage regi-
mens; pharmacokinetic data. In: J. G.
Hardman, L. E. Limbird, P. B. Molinoff,
R. W. Ruddon, & A. G. Gilman (eds.).
The pharmacological basis of therapeutics,
9th edition (pp. 1707–1792). New York:
McGraw-Hill.
Ciancio, S. G., Mather, M. L. & McMullen,
J. A. (1980) An evaluation of minocycline
in patients with periodontal disease.
Journal of Periodontology 51, 530–534.
Ciancio, S. G., Slots, J., Reynolds, H. S.,
Zambon, J. J. & McKenna, J. D. (1982)
The effect of short-term administration of
minocycline HCL on gingival inflam-
mation and subgingival microflora.
Journal of Periodontology 53, 557–561
(abstr).
Freeman, E., Ellen, R. P., Thompson, G.,
Weinberg, S. E., Song, M. & Lazarus, R.
H. (1992) Gingival crevicular fluid concen-
tration and side effects of minocycline: a
comparison of two dose regimens. Journal
of Periodontology 63, 13–18.
Goodson, J. M. (1994) Antimicrobial strat-
egies in periodontal treatment. In: Micro-
biology and immunology of periodontal dis-
eases. Periodontology 2000 5, 142–
157.
Gordon, J. M., Walker, C. B., Goodson, J.
M. & Socransky, S. S. (1980) Sensitive as-
say for measuring tetracycline levels in
gingival crevice fluid. Antimicrobial Agents
and Chemotherapy 17, 193–198.
Gordon, J. M., Walker, C. B., Murphy, J. C.,
Goodson, J. M. & Socransky, S. S. (1981a)
Concentration of tetracycline in human
gingival fluid after single doses. Journal of
Clinical Periodontology 8, 117–121.
Gordon, J. M., Walker, C. B., Murphy, J. C.,
Goodson, J. M. & Socransky, S. S. (1981b)
Tetracycline: levels achievable in gingival
crevice fluid and in vitro effect on subgin-
gival organisms. Part I. Concentrations in
crevicular fluid after repeated doses.
Journal of Periodontology 52, 609–612.
Kapusnik-Uner, J. E., Sande, M. A. &
Chambers, H. F. (1996) Tetracyclines,
chloramphenicol, erythromycin and mis-
cellaneous antibacterial agents. In: J. G.
Hardman, L. E. Limbird, P. B. Molinoff,
R. W. Ruddon, & A. G. Gilman (eds.):
The pharmacological basis of therapeutics
(pp. 1123–1153).
Pascale, D., Gordon, J., Lamster, I., Mann,
P., Seiger, M. & Arndt, W. (1986) Concen-
tration of doxycycline in human gingival
fluid. Journal of Clinical Periodontology
13, 841–844.
Pashley, D. H. (1976) A mechanistic analysis
of gingival fluid production. Journal of
Periodontal Research 11, 121–134.
Slots, J. & Rams, T. E. (1996) Antibiotics in
periodontal therapy: advantages and dis-
advantages. Journal of Clinical Periodonto-
logy 17, 479–493.
Stoller, N. H., Johnson, L. R., Trapnell S.,
Harrold, C. Q. & Garrett, S. (1998) The
pharmokinetic profile of a biodegradable
controlled-release delivery system contain-
ing doxycycline compared to systemically
delivered doxycycline in gingival crevicular
fluid, saliva and serum. Journal of Period-
ontology 69, 1085–1091.
Van Winkelhoff, A. J., Rams, T. E. & Slots,
J. (1996) Systemic antibiotic therapy in
periodontics. Periodontology 2000 10, 45–
78.
Address:
J.Max Goodson
Forsyth Dental Center
140 The Fenway
Boston, MA 02115
USA
Fax: π1 617 262 4021
e-mail: mgoodson/forsyth.org