Volume 71 Number 9
`
`Long-Term Use of Subantimicrobial Dose
`Doxycycline Does Not Lead to Changes in
`Antimicrobial Susceptibility
`John Thomas,* Clay Walker,t and Mark Bradshaw
`
`Background: Adjunctive subantimicrobial dose doxycycline
`(SDD) with scaling and root planing leads to improved clinical
`parameters of adult periodontitis, but has raised questions about
`potential changes in antibiotic susceptibility of the host
`microflora. Our four studies assessed whether long-term SDD
`changes antibiotic susceptibility of the oral microflora in adults
`with periodontitis.
`Methods: In studies 1 and 2, adult patients with periodonti-
`tis were randomized to receive SDD 10 mg qd, 20 mg qd, 20
`mg bid, or placebo. In study 3, patients were randomized to
`receive SDD 20 mg bid or placebo. No medication was admin-
`istered in study 4, a follow-up to study 3. Subgingival plaque
`samples were collected at baseline (all studies) and at 12, 15
`to 18, and 24 months (study 1); 12, 18, and 27 months (study
`2); 3, 6, and 9 months (study 3); and 3 months post-study 3
`(study 4). Antimicrobial susceptibility of isolated bacteria was
`assessed by: 1) minimum inhibitory concentration (MIC) levels
`(studies 1 and 2); 2) cross-resistance to non-tetracycline antibi-
`otics (studies 2 and 3); and 3) the proportion of doxycycline-
`resistant isolates (studies 3 and 4).
`Results: Organism MIC levels remained constant among all
`treatment groups at 18 and 24 months compared with baseline
`(study 1). Observed changes in susceptibility at 12 and 18
`months for the 20 mg groups were attributed to the limited num-
`ber of isolates tested (study 1). There were no statistically sig-
`nificant differences in the proportion of doxycycline- resistant
`isolates among treatment groups (studies 3 and 4), and no evi-
`dence of multi-antibiotic resistance (studies 3 and 4) or cross-
`resistance (studies 2 and 3) at any timepoint.
`Conclusion: Long-term SDD does not contribute to changes
`in antibiotic susceptibility. J Periodontol 2000;71:1472-1483.
`K
`Doxycycline/therapeutic use; drug resistance, microbial;
`antibiotics/therapeutic use; periodontitis/drug therapy;
`dose-response relationship, drug; comparison studies.
`
`* Department of Pathology/Periodontics, West Virginia University, Morgantown, WV.
`t Periodontal Disease Research Clinics, University of Florida, Gainesville, FL.
`Covance, Inc., Princeton, NJ.
`
`1472
`
`D oxycycline has been shown to effec-
`
`tively inhibit collagenase activity in
`gingival tissues and crevicular fluid,
`thereby reducing the destruction of collagen
`in adult periodontitis. 14 This doxycycline-
`induced decrease in collagenase activity is
`accompanied by a significant improvement
`in other periodontal disease parameters,
`such as improvement in periodontal
`attachment levels and decreasing probing
`depth .2-5 A long-term trial evaluating the
`efficacy of a 9-month regimen of suban-
`timicrobial dose doxycycline (SDD) found
`that a combined regimen of SDD and scal-
`ing and root planing (SRP) resulted in sta-
`tistically significant improvements in peri-
`odontal attachment level and probing depth,
`when compared to SRP plus placebo. 6,7
`The anticollagenase activity of doxycy-
`dine is independent of its antimicrobial
`activity, as first reported by Golub et al. in
`1983 8 and confirmed through subsequent
`research . 24 ’9 Effective anticollagenase
`activity is obtained in man at administered
`doses well below those routinely used for
`effective antimicrobial treatment. For exam-
`ple, the usual antimicrobial dose of doxy-
`cycline is a 200 mg initial dose followed by
`100 mg qd, which produces blood levels of
`3 .Lg/ml to 4 Ig/ml) 0 When used to sup-
`press collagenase activity, however, the
`effective dose of doxycycline is 20 mg bid,
`which produces maximum serum concen-
`trations of 0.79 .rg/ml during chronic
`administration (unpublished data).
`Prior stt"11 (cid:9)
`filcJ t.- ,t*,-’t nn
`hir (cid:9)
`antimicrol Amneal 1064
`gingival m Amneal v. Supernus
`ing SDD i. 1PR2013-00371
`
`

`

`Volume 71 (cid:149) Number 9
`
`Long-Term Use of Subantimicrobial Dose
`Doxycycline Does Not Lead to Changes in
`Antimicrobial Susceptibility
`John Thomas,* Clay Walker,t and Mark Bradshaw
`
`Background: Adjunctive subantirnicrobial dose doxycycline
`(SDD) with scaling and root planing leads to improved clinical
`parameters of adult periodontitis, but has raised questions about
`potential changes in antibiotic susceptibility of the host
`microflora. Our four studies assessed whether long-term SDD
`changes antibiotic susceptibility of the oral microflora in adults
`with periodontitis.
`Methods: In studies 1 and 2, adult patients with periodonti-
`tis were randomized to receive SLDD 10 mg qd, 20 mg qd, 20
`mg bid, or placebo. In study 3, patients were randomized to
`receive SDD 20 mg bid or placebo. No medication was admin-
`istered in study 4, a follow-up to study 3. Subgingival plaque
`samples were collected at baseline (all studies) and at 12, 15
`to 18, and 24 months (study 1); 12, 18, and 27 months (study
`2); 3, 6, and 9 months (study 3); and 3 months post-study 3
`(study 4). Antimicrobial susceptibility of isolated bacteria was
`assessed by: 1) minimum inhibitory concentration (MIC) levels
`(studies I and 2); 2) cross-resistance to non-tetracycline antibi-
`otics (studies 2 and 3); and 3) the proportion of doxycycline-
`resistant isolates (studies 3 and 4).
`Results: Organism MIC levels remained constant among all
`treatment groups at 18 and 24 months compared with baseline
`(study 1). Observed changes in susceptibility at 12 and 18
`months for the 20 mg groups were attributed to the limited num-
`ber of isolates tested (study 1). There were no statistically sig-
`nificant differences in the proportion of doxycycline- resistant
`isolates among treatment groups (studies 3 and 4), and no evi-
`dence of multi-antibiotic resistance (studies 3 and 4) or cross-
`resistance (studies 2 and 3) at any timepoint.
`Conclusion: Long-term SDD does not contribute to changes
`in antibiotic susceptibility. J Periodontol 2000;71:1472-1483.
`
`Doxycycline/therapeutic use; drug resistance, microbial;
`antibiotics/therapeutic use; periodontitis/drug therapy;
`dose-response relationship, drug; comparison studies.
`
`* Department of Pathology/Periodontics, West Virginia University, Morgantown, WV
`t Periodontal Disease Research Clinics, University of Florida, Gainesville, FL.
`Inc., Princeton, NJ.
`
`1472
`
`D oxycycline has been shown to effec-
`
`tively inhibit collagenase activity in
`gingival tissues and crevicular fluid,
`thereby reducing the destruction of collagen
`in adult periodontitis*4 This doxycycline-
`induced decrease in collagenase activity is
`accompanied by a significant improvement
`in other periodontal disease parameters,
`such as improvement in periodontal
`attachment levels and decreasing probing
`depth .2-5 A long-term trial evaluating the
`efficacy of a 9-month regimen of suban-
`timicrobial dose doxycycline (SDD) found
`that a combined regimen of SDD and scal-
`ing and root planing (SRP) resulted in sta-
`tistically significant improvements in peri-
`odontal attachment level and probing depth,
`when compared to SRP plus placebo. 6 ’ 7
`The anticollagenase activity of doxycy-
`dine is independent of its antimicrobial
`activity, as first reported by Golub et al. in
`19838 and confirmed through subsequent
`research .2-4 Effective anticollagenase
`activity is obtained in man at administered
`doses well below those routinely used for
`effective antimicrobial treatment. For exam-
`ple, the usual antimicrobial dose of doxy-
`cycline is a 200 mg initial dose followed by
`100 mg qd, which produces blood levels of
`3 ig/ml to 4 ig/ml. 10 When used to sup-
`press collagenase activity, however, the
`effective dose of doxycycline is 20 mg bid,
`which produces maximum serum concen-
`trations of 0.79 .Lg/ml during chronic
`administration (unpublished data).
`Prior studies have failed to detect an
`antimicrobial effect of doxycycline on sub-
`gingival microflora. 10 ’ 11 In a study evaluat-
`ing SDD and placebo treatment with and
`
`

`

`Periodontol September 2000 (cid:9)
`
`Thomas, Walker, Bradshaw
`
`without accompanying SRP, Walker et al. found no sta-
`tistically significant or microbiological differences
`between or among treatment groups for motile rods,
`coccoid forms, non-motile rods, fusiforms, and fila-
`mentous rods. Only levels of small and large spiro-
`chetes were found to be significantly lower in the SDD
`group than the placebo group. 10 This decrease in spiro-
`chetes was not attributed to an antimicrobial effect.
`Rather, it was explained by an overall improvement in
`periodontal health due to the anti-inflammatory and
`anticollagenolytic properties of doxycycline, which alter
`the microenvironment of the periodontal pocket, to
`which spirochetes are known to be particularly sensi-
`tive. 12
`However, any long-term use of antibiotics raises ques-
`tions of changes in antibiotic susceptibility. 9’’1316 In
`addition, there are concerns that the suppression of a
`normal, susceptible microflora could lead to overgrowth
`of more resistant and potentially pathogenic microor-
`ganisms in reservoirs such as the oral cavity. 11
`To confirm the results of previous research, four
`studies were performed to assess whether the long-
`term use of SDD changes antibiotic susceptibility of the
`subgingival microflora in adults with periodontitis. Indi-
`cations of altered susceptibility were monitored by
`measuring changes in minimum inhibitory concentra-
`tions (MICs) of specific species; examining suscepti-
`bility patterns to specific antibiotics (doxycyclirie,
`minocycline, tetracycline, amoxicillin, erythromycin,
`
`penicillin, ampicillin, cefoxitin, metronidazole, and din-
`damycin) according to the National Committee for
`Clinical Laboratory Standards (NCCLS) categories; 17
`and by determining alterations in the distribution of
`susceptibility.
`
`MATERALS AND METHODS
`The methods and treatments used in each of the four
`studies are summarized in Table 1. In studies 1 and 2,
`patients were eligible for study participation if they
`were between 35 and 75 years of age; in studies 3 and
`4, patients were included if they were between 30 and
`75 years of age. All patients had a clinical diagnosis
`of periodontitis. Periodontitis was defined as both clin-
`ical attachment levels >5 mm and <9 mm and prob-
`ing depths ~!5 mm and <9 mm in at least two subgin-
`gival tooth sites within the full mouth (studies I and
`2) or in 2 tooth sites in each of 2 quadrants (i.e., 4
`sites) (studies 3 and 4).
`Patients were excluded if they had received any
`antibiotics within 6 weeks of the baseline visit or if
`they required chronic antibiotic treatment (i.e., more
`than 2 weeks) or prophylactic antibiotics for routine
`dental therapy. Women who were pregnant or lactat-
`ing were also excluded, as were patients diagnosed
`with diabetes mellitus, a serious medical illness (e.g.,
`kidney or liver disease), or a systemic infection. Addi-
`tionally, patients with known hypersensitivity to tetra-
`cyclines or who were taking significant concomitant
`
`Table I.
`Summary of Studies
`
`Study
`
`I
`
`2
`
`3
`
`4
`
`N
`
`40
`
`40
`
`171
`
`Method
`
`Target Organisms
`
`Randomized, double-blind (12 months),
`placebo-controlled, parallel study
`
`Actinomyces spp
`Fusobacterium spp
`
`Randomized, 12-month blinded, 6
`months open-label, 9 months no
`treatment
`
`Actinomyces spp
`Fusobocterium spp
`
`SDD
`Treatmentst
`
`10 mg qd doxy
`20 mg qd doxy
`20 mg bid doxy
`Placebo
`
`10 mg qd doxy
`20 mg qd doxy
`20 mg bid doxy
`Placebo
`
`Assessment Periods
`
`Baseline, 12 months,
`15-18 months (study exit),
`6 months post-treatment
`
`Baseline, 12 months,
`18 months, 27 months
`
`Randomized, placebo-controlled,
`double-blind, parallel group. In each
`patient, two quadrants treated with
`SRP; two without SRR
`
`Predominant oral flora
`
`20 mg bid doxy
`Placebo
`
`Baseline, 3 months,
`6 months, 9 months
`
`146+
`
`Follow-up of study 3
`
`Predominant oral flora
`
`Totals
`
`251
`
`* Total number of patients.
`t Doxy = doxycycline.
`+ One hundred forty-six (146) patients in study 4 were not unique patients but a subpopulation of study 3 and were not counted in the total
`
`1473
`
`No medication
`administered
`
`3 months post-study 3
`
`45 months (active
`treatment)
`
`

`

`Subantimicrobial Doxycycline and Microbial Susceptibility (cid:9)
`
`Volume 71 Number 9
`
`therapy were also excluded from the studies. Clinical
`and microbial data were collected at the University of
`Florida, Gainesville, and West Virginia University, Mor-
`gantown.
`Study 1 Methods
`Study I was a double-blind, placebo-controlled, par-
`allel study with an open-label extension. Patients were
`administered SDD (20 mg qd, 20 mg bid, or 10 mg
`qd) or placebo for 12 months, after which they could
`opt to enter an open-label phase lasting 3 to 6 months.
`Patients continued their dosing regimen during the
`open-label phase. Patients received a supragingival
`prophylaxis at baseline, at 6 and 12 months, and at
`study exit (15 to 18 months). Additionally, patients
`were asked to return for further analysis 6 months after
`treatment ended (21 to 24 months post-baseline).
`Subgingival plaque samples were collected with ster-
`ile endodontic paper points at baseline and at 12
`months, 15 to 18 months (on cessation of treatment),
`and 6 months post-treatment. Each sample was
`assessed for total cultivable anaerobic bacteria, total
`Actinomyces spp isolates, and total Fusobacterium spp
`isolates to determine the effect of the SDD or placebo
`regimen on microbial flora. Actinomyces spp and
`Fusobacterium spp were selected as representative
`Gram-positive and Gram-negative isolates known to be
`encountered in a patient population with moderate to
`severe periodontal disease. Previous research has
`demonstrated that these are appropriate marker organ-
`isms for evaluating the development of resistance when
`using SDD) 8
`Representative Actinomyces and F’usobacterium
`isolates were obtained from each sample to test sus-
`ceptibility to doxycycline (i.e., minimum inhibitory con-
`centration [MIC] values). Trypticase-soy agar supple-
`mented with 5% whole defibrinated sheep blood,
`0.005% hemin, and 0.0005% menadione was used as
`a non-selective medium for total anaerobic counts.
`Susceptibility testing was performed by an agar dilu-
`tion method. 1 7,19,20 To assess clinical efficacy, clinical
`attachment levels and probing depths were measured
`at 6 sites around each tooth in the whole mouth by
`manual probing at baseline, 6 months, and 12 months.
`ANOVA was used to determine whether statistically
`significant differences in total organisms isolated
`existed among sample periods. Fisher’s PLSD analy-
`sis was used to determine whether differences in total
`organisms isolated existed between two treatments at
`each sample period.
`Study 2 Methods
`Study 2 was similar in design to study 1. Patients were
`administered SDD or placebo for 12 months, after
`which they entered an open-label phase (i.e., treat-
`ment was not blinded) lasting 3 to 6 months. Patients
`continued their dosing regimen during the open-label
`
`phase and received a supragingival prophylaxis at
`baseline, at 6 and 12 months, and at the cessation of
`treatment. Patients returned up to 9 months after ces-
`sation of treatment for further analysis (i.e., up to 27
`months post-baseline).
`Subgingival plaque samples were collected by curet
`at baseline, 12 months (at the end of blinded treat-
`ment), 18 months (after the 6-month open-label
`phase), and up to 27 months (after 9 months without
`treatment). As in study 1, Actinomyces spp and
`Fusobacterium spp were isolated using a non-selective
`complex medium21 ’22 as representative taxa common
`in the periodontal microflora. 18 The isolates were tested
`for resistance (i.e., MIC values) to tetracycline, eryth-
`romycin, penicillin, ampicilliri, cefoxitin, and metro-
`nidazole using antibiotic-impregnated strips (used by
`West Virginia University) or agar dilution (used by the
`University of Florida). 19,20,23 To assess efficacy, clini-
`cal attachment levels and probing depths were mea-
`sured at 6 sites around each tooth by manual probing
`at baseline, 6 months, and 12 months.
`Study 3 Methods
`Study 3 was a multi-center, randomized, parallel-group,
`placebo-controlled study, and study 4 was a 3-month
`follow-up of study 3. In study 3, all patients were
`treated in two qualifying quadrants with SRP at base-
`line. To be considered a qualifying quadrant, two tooth
`sites in each quadrant had to have both clinical attach-
`ment level and probing depth ~5 mm and :!~9 mm, as
`measured by manual probing. SRP was performed by
`the same therapist at each study center and lasted for
`up to 1 hour per quadrant. Both ultrasonic and uni-
`versal or area-specific curets were permitted, as was
`local anesthesia. After SRP was performed, patients
`in study 3 were randomly assigned to receive either
`SDD 20 mg bid or placebo for 9 months.
`In each patient, two sites with probing depths >5
`mm from the SRP quadrants and two sites with prob-
`ing depths >5 mm from non-SRP quadrants were
`selected for microbiological sampling. Subgingival
`plaque samples were taken from these sites using ster-
`ile endodontic paper points at baseline and after 3, 6,
`and 9 months of treatment. The same sites were sam-
`pled throughout the study. Samples from the SRP sites
`were pooled by subject and then processed; the same
`was done for non-SRP sites. Samples were examined
`by direct microscopy, culture on selective and non-
`selective media, and predominant cultivable technique.
`Samples were also tested for susceptibility to doxycy-
`dine, minocycline, tetracycline, amoxicillin, ery-
`thromycin, and clindamycin by either agar dilution
`method 19’20’23 or antibiotic-impregnated strips.§ To
`assess efficacy as an adjunct to SRP, clinical attach-
`ment levels and probing depths were measured by
`
`§ Etest, AB Biodisk, Some, Sweden.
`
`1474
`
`

`

`Periodontol September 2000 (cid:9)
`
`Thomas, Walker, Bradshaw
`
`manual probing at each of 6 tooth sites around each
`tooth in the qualifying quadrants at baseline and at 3,
`6, and 9 months. To check for rapid attachment loss
`requiring additional intervention, similar measurements
`were also made around the teeth in the non-qualify-
`ing quadrants.
`
`Statistical Methods
`The null hypothesis (i.e., the distribution of the strains
`identified as resistant to 4 ltg of doxycycline per mil-
`liliter across 3 resistance categories did not differ
`between placebo and SDD treatment groups) was
`tested separately for plaque samples from SRP and
`non-SRP quadrants. The results of the 2 analyses were
`similar. Separate analyses were carried out for each
`visit of the study. Data from the 2 centers were pooled,
`and the center was treated as a factor in the analyses.
`Of particular interest was the alternative hypothesis
`that at some study visits, strains isolated from plaque
`samples from SDD-treated patients may have exhib-
`ited greater resistance to doxycycline than those taken
`from patients treated with placebo. This tendency, if
`present, would result in a greater proportion of resis-
`tant strains falling in the higher resistance categories
`for SDD-treated patients. Therefore, analyses focused
`on the degree of resistance observed among resistant
`strains and differences in the degree of resistance over
`time and between treatment groups, rather than the
`raw number or proportion of resistant strains isolated.
`Two types of analyses were performed: analyses
`based on frequency distributions within and between
`patients, and non-parametric analyses based on scores
`derived from within-patient frequency distributions.
`The number of resistant strains isolated from a single
`patient’s plaque sample could vary from 0 to 3; how-
`ever, plaque samples with no resistant isolates were
`dropped from the analysis. As mentioned above, it is
`important to note that these analyses focus on shifts
`in the degree of resistance among resistant strains
`rather than changes in the number or proportion of
`resistant strains.
`Frequency distribution analysis. Because multiple
`strains could be isolated at a given visit within a patient,
`but the basic unit of analysis was considered to be the
`patient rather than the microorganism, standard chi-
`square or Cochran -Mantel -Haenszel analyses could not
`be used without an adjustment to the degrees of free-
`dom. Frequency distribution analyses were therefore
`performed by first normalizing the data. Organism
`counts for each patient visit were transformed into pro-
`portions of strains within each resistance category for
`that patient visit. Normalized individual patient/visit
`cell counts therefore always summed to unity across
`the 3 resistance categories, yielding correct overall
`degrees of freedom. Cochran -Mantel-Haenszel tests
`were performed on the normalized data, treating cen-
`
`ters as strata and treating the resistance category as
`an ordinal parameter with equal spacing between the
`categories. Scores of 2, 3, and 4 were used for each
`resistance category based on log 2 of the value of the
`category in micrograms per milliliter. However, the
`results of the analysis would not differ using any arbi-
`trary, equally spaced set of numbers for category
`values.
`Non-parametric (ranked) analysis of variance. A
`single resistance score for each patient visit was cre-
`ated by multiplying the above category scores (2, 3,
`or 4) by the proportion of resistant strains in each cat-
`egory, then adding the results together. The resulting
`scores were transformed into ranks, and the ranks were
`analyzed using an analysis of variance model with fac-
`tors for treatment group, center, and the interaction of
`treatment-by-center. This non-parametric procedure
`produces results that are invariant with respect to the
`choice of category scores, given that scores are equally
`spaced. As with the frequency of distribution analysis,
`the individual patient was taken as the unit of analy-
`sis. All analyses reflect the assumption that increased
`doxycycline resistance would be reflected in a ten-
`dency for more strains to fall within the higher resis-
`tance categories within a patient/visit. P values !~0.05
`were considered statistically significant. Post-hoc power
`analyses suggested that actual differences greater than
`15% would typically have been detected with a prob-
`ability of 0.80. All statistical tests were 2-tailed. All
`statistical analyses and data manipulations were per-
`formed using statistical software.
`
`Study 4 Methods
`No medication was administered during study 4. Sub-
`gingival plaque samples were collected from the same
`sites as those used in study 3 (two sites from the SRP
`quadrants and two from the non-SRP quadrants) at
`baseline and at 3 months for analysis of oral flora and
`antibiotic susceptibility. The two samples from the two
`SRP quadrants were pooled by subject and then
`processed; the same was done for non-SRP sites.
`Microbial samples were examined by either darkfield
`or phase-contrast microscopy, culture on selective
`and non-selective media, and predominant cultivable
`technique. To assess efficacy, clinical attachment lev -
`els and probing depths were measured by manual
`probing.
`To determine whether statistically significant differ-
`ences existed between the SDD and placebo treatment
`groups, the microbial parameters associated with
`microscopic evaluation and culture enumeration were
`analyzed using the unpaired t test and the Mann-
`Whitney test, a non-parametric version of the two sam-
`ple, unpaired t test.
`
`II SAS 6.12, SAS Institute, Inc., Cary, NC.
`
`1475
`
`

`

`Subantimicrobial Doxycycline and Microbial Susceptibility (cid:9)
`
`Volume 71
`
`(cid:149) Number 9
`
`Ll
`Study 1 Results
`A total of 40 subjects were included in study 1, 10 in
`each of 4 treatment groups. All subjects provided
`microbiological samples at baseline. At 12 months,
`29 subjects were sampled (8 in the SDD 20 mg qd
`treatment group; 7 in the SDD 20 mg bid treatment
`group; 7 in the SDD 10 mg qd treatment group; and
`7 in the placebo group). At 15 to 18 months (study
`exit), 33 subjects were sampled (9 in the SDD 20 mg
`qd group; 10 in the SDD 20 mg bid group; 8 in the
`SDD 10 mg qd group; and 6 in the placebo group).
`Six months post-treatment, samples were obtained
`from 24 subjects (7 in the SDD 20 mg qd group; 7 in
`the SDD 20 mg bid group; 4 in the SDD 10 mg qd
`group; and 6 in the placebo group).
`Effect on microbial flora. Treatment effect on micro-
`bial flora was assessed in study I by measuring total
`cultivable anaerobic bacteria, total Actinomyces spp
`isolates, and total Fusobacterium spp isolates. Assess-
`ments were made for each treatment at each sample
`timepoint based on the number of cultivable bacterial
`counts obtained for each target genus.
`No statistically significant differences in total anaer-
`obic bacteria, Actinomyces spp, and Fusobacterium
`spp isolates were found in study 1 among sample
`periods (Figs. IA through ic; Table 2) or between treat-
`ments at baseline, 12 months, or 15 to 18 months
`(P
`>0.12). Statistically significant differences were
`detected, however, in three instances at the 6-month
`post-treatment sample period (21 to 24 months post-
`baseline; P!~0.05). The difference in mean total anaer-
`obic bacteria was higher (P :!~0.05) in the 10 mg qd
`treatment group compared with the 20 mg qd treat-
`ment group. Differences in total Actinomyces spp iso-
`lates were significant between the 10 mg qd treat-
`ment group and the 20 mg qd treatment group, as
`well as between the 10 mg qd and placebo treatment
`groups. It was not feasible to conduct a longitudinal
`analysis of these differences because an inadequate
`number of the same patients were present in each
`treatment cell at each sample period to use a repeated
`measure ANOVA.
`Susceptibility. The MIC was evaluated, as were the
`minimum inhibitory concentrations needed to inhibit
`growth of 50% (M10 50 ) or 90% (M1C 90 ) of the target
`organisms. M10 50 or M1C 90 summaries were deter-
`mined for comparative purposes. At the 12-month
`assessment and at 15 to 18 months (study exit), no
`changes were apparent in the doxycycline MIC 50 (Fig.
`2A) or MIC 90 (data not shown) for the Actinomyces spp
`isolates in the 10 mg qd treatment group compared
`with the placebo. At 12 months post-baseline, there
`was an apparent change in susceptibility to doxycy -
`dine in the Actinomyces spp isolates in the 20 mg qd
`and 20 mg bid treatment groups compared with base-
`
`1476
`
`Total anaerobic counts recovered
`r (cid:149) (cid:9)
`
`20qd1
`----0-- Placebo
`20 bid
`10 q
`
`lO
`
`C
`
`of,
`
`C
`
`t (cid:9)
`
`lots
`
`C
`
`C
`C
`
`-
`
`C.) (cid:9)
`
`A
`
`1O
`
`BL
`
`12 Months (cid:9)
`
`Exit
`
`6 Months
`Post-Treatment
`
`Total Fusobacterium counts recovered
`
`106
`
`2
`00
`0
`
`20 q
`---0-- Placebo
`20 bid
`10 qd
`
`E (cid:9)
`
`105
`
`00 I 10
`
`B
`
`BL (cid:9)
`
`12 Months (cid:9)
`
`Exit (cid:9)
`
`6 Months
`Post Treatment
`
`Total Actinomyces counts recovered
`ri
`
`(cid:151)0-------
`
`Placebo
`20 bid
`10 q
`
`106
`
`10
`
`2 (cid:9)
`00
`C
`
`of,
`
`. (cid:9)
`
`00 I (cid:9)
`
`BL
`
`12 Months (cid:9)
`
`Exit
`
`6 Months
`Post-Treatment
`
`Figure I.
`A. Total cultivable anaerobic counts recovered for each treatment at
`each sample period. B. Total cultivable Fusobacterium spp counts
`recovered for each treatment at each sample period. C. Total
`cultivable Actinomyces spp counts recovered for each treatment at
`each sample period.
`
`(cid:9)
`

`

`Periodontol September 2000 (cid:9)
`
`Thomas, Walker, Bradshaw
`
`Table 2.
`ANOVA Analysis of Differences in Total Anaerobic
`Bacteria, Actinomyces spp, and Fusobacterium spp
`Counts Among Treatment Groups (study 1)
`
`Colony Count of
`Bacterial Group
`
`Baseline
`(P Value)
`
`12 Months
`(P Value)
`
`15 to 18 Months
`(P Value)
`
`Sample Period
`
`6 Months Post-
`Treatment (P Value)
`
`Total anaerobic bacteria
`
`0.8584
`
`0.8775
`
`Actinomyces isolates
`
`0.4870
`
`0.5007
`
`Fusobocterium isolates (cid:9)
`
`0.3216 (cid:9)
`
`0.4274 (cid:9)
`
`0.8018
`
`0,5209
`
`0.5422 (cid:9)
`
`0. 19 18
`
`0.5380
`
`0.098 1
`
`(21 to 24 months post-baseline), no
`differences in M10 50 or MIC90 values
`were found for any of the four treat-
`ment groups relative to each other
`or to baseline for Actinomyces spp
`isolates (Table 3).
`The MIC 50 values for Fusobac-
`terium spp isolates were essentially
`identical for all treatment groups at
`all sample periods (Fig. 213). In-
`creases were noted in the doxycy-
`dine M1C90 of Fusobacterium spp
`isolates in all treatment groups,
`including placebo, at 15 to 18
`months compared with baseline
`(data not shown). There was no dif-
`ference, however, among the treatment
`groups in the doxycycline M1C 90 at the 12-
`month assessment. At assessments taken 6
`months after treatment ended (21 to 24
`months post-baseline), no differences in
`M1050 or MIC 90 values were found for any of
`the four treatment groups relative to each
`other or to baseline for Fusobacterium spp
`isolates.
`
`35
`
`30 (cid:9)
`
`CL
`a, 25
`
`. 20 (cid:9)
`C.)
`x 15
`
`to
`
`A
`
`30
`
`a’ 25
`a-
`a,
`
`20
`
`X 15 (cid:9)
`0
`
`10
`
`0
`
`B
`
`.. -...
`
`- (cid:9)
`
`- (cid:9)
`
`-.----------
`
`o2omgqd
`020 mg bid
`alomgqd
`c3 placebo
`
`Baseline (cid:9)
`
`12 months (cid:9)
`
`Study exit (cid:9)
`
`6 months post
`
`Sample Period
`
`Baseline (cid:9)
`
`12 months (cid:9)
`
`Study exit 6 months post
`
`Sample Period
`
`[20 mg qd 1
`El
`020mg bid
`010 mgqd
`o placebo
`
`Study 2 Results
`Patterns and cross-resistance. Among the
`Actinomyces spp and Fusobacterium spp
`isolates collected in study 2, there were no
`changes in NCCLS antibiotic patterns (Sus-
`ceptible [i.e., MIC !!~4 Ig/ml], Intermediate
`[i.e., MIC 5 to 15 ig/ml], or Resistant [i.e.,
`MIC ~!16 tg/ml]) nor significant changes in
`antibiotic susceptibility to the six antibiotics
`tested (tetracycline, erythromycin, penicillin,
`ampicillin, cefoxitin, and metronidazole) in
`any of the sample periods for any of the
`treatment groups. The MIC 50 for each antibi-
`otic tested was well below the NCCLS cut-
`off levels that determine antimicrobial resis-
`tance for each of the sample periods (Table
`4). Antibiotic profiles and cumulative M10 50
`remained stable regardless of doxycycline
`dosage, indicating the absence of multiple
`antibiotic resistance.
`Studies 3 and 4 Results
`Distribution of resistance. At baseline and
`at 3, 6, and 9 months, five taxa accounted
`for 57% to 80% of doxycycline-resistant iso-
`lates (Bacteroides coagulans, Campylobacter concisus,
`Fusobacterium spp, Prevotella spp, and Streptococcus
`spp). Four of these taxa accounted for 60% to 80% of
`doxycycline -resistant isolates at 12 months (C. con-
`cisus, Fusobacterium Spp, Prevotella spp, and Strep-
`tococcus spp). These resistant taxa were consistent
`
`A. Doxvcvcltne MiCro values for Actinomyces spp isolates for each sample period by
`treatment group. B. Doxycycline M10 50 values for Fusobacterium sop isolates for
`each sample period by treatment group. (cid:9)
`
`’ (cid:9)
`
`line. However, 15 to 18 months post-baseline, the
`M1050 and MIC90 values for the 20 mg bid treatment
`group were similar to the placebo, although the M10 50
`and M1C 90 values for the 20 mg qd treatment group
`remained elevated compared with other treatments.
`At assessments taken 6 months after treatment ended
`
`1477
`
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`

`

`Subantimicrobial Doxycycline and Microbial Susceptibility (cid:9)
`
`Volume 71
`
`(cid:149) Number 9
`
`Table 3.
`Doxycycline Susceptibilities for Representative Gram-
`Positive (Actinomyces) and Gram-Negative
`(Fusobacterium) Bacteria
`
`Treatment
`
`N Subjects
`
`Taxa
`
`N Strains
`
`MIC Range
`
`-
`
`Baseline
`20 mg qd
`
`20 mg bid
`
`0 mg qd
`
`Placebo
`
`12 months
`20 mg qd
`
`20 mg bid
`
`10 mg qd
`
`Placebo
`
`10
`
`10
`
`10
`
`10
`
`8
`
`7
`
`7
`
`7
`
`Actinomyces
`Fusobacterium
`
`Acinemyces
`Fusobacterium
`
`Actinomyces
`Fusobacterium
`
`Actinomyces
`Fusobacterium
`
`Actinomyces
`Fusobacterium
`
`Actinomyces
`Fuse be cterium
`
`Actinomyces
`Fusobacter;um
`
`Actinomyces
`Fusobacterium
`
`6
`6
`
`3
`4
`
`8
`6
`
`6
`10
`
`6
`5
`
`3
`0
`
`2
`5
`
`!~*25 - 4
`!~0.25 - 0.5
`
`<025 -
`0.25 - 0.5
`
`0.5 __ 8
`<0.25 -
`
`<0.25
`0.25
`
`8
`32
`
`>32
`2 (cid:9)
`:~*25 (cid:9)
`4
`
`0.5 - 32
`N/A
`
`2.0 - 8
`<0.25 - 8
`
`!~O.25
`
`!~0.25 (cid:9)
`
`0,25
`
`32 (cid:9)
`!~0.25 (cid:9)
`
`32 (cid:9)
`N/A (cid:9)
`2 (cid:9)
`!~0.25 (cid:9)
`
`4 (cid:9)
`:~0.25 (cid:9)
`
`In assessing the distribution of doxycycline
`resistance, there were no statistically signifi-
`cant differences between treatment groups.
`The distribution of doxycycline-resistant iso-
`lates was consistent across NCCLS resistance
`categories at all sample periods after the b