897
Endotoxin Levels in Periodontally Healthy and Diseased Sites: Correlation With Levels of Gram-Negative Bacteria Daniel H. Fine, * Carlos Mendieta, * Michael L. Naini, * and Jack W. Vincent*
Barnett,f David Furgang, * Ali
between endotoxin levels and the perin sites and in Periodontitis sites. Twelve of bacteria healthy Gram-negative centage Each 3 adults participated. subject provided Periodontitis sites with 5 to 8 mm healthy that bled on probing depths gentle probing and 3 healthy sites with sulcus depths of 1 to 3 mm that did not bleed. Clinical examinations and sterile paper point sampling of all study sites were conducted on days 0, 7, and 14, and site-specific endotoxin levels and percentage of Gram-negative bacteria were determined. There were significant differences in both endotoxin levels and percentage Gram-negative bacteria between healthy and Periodontitis sites across all 3 sampling periods, but no difference across sampling periods in the healthy sites and the Periodontitis sites, respectively. Correlation coefficients revealed a high degree of correlation between site-specific endotoxin levels and percentage of Gram-negative organisms. Using a sample dilution of 1 x 104, endotoxin levels differentiated healthy from Periodontitis sites with a specificity of approximately 91% and a sensitivity of approximately 90%. / Periodontol 1992; 63:897-901. This
study investigated the correlation
Key Words: Endotoxin; periodontitis/microbiology; periodontitis/diagnosis.
The onset and progression of periodontal diseases have been shown to be associated with both quantitative and qualitative changes in the dental plaque microbial flora. The majority of putative periodontopathogens identified to date have been Gram-negative organisms, and the proportion of plaque organisms found to be Gram-negative increases markedly as sites progress from health to disease.17 It is not surprising, therefore, that crevicular levels of endotoxin, a lipopolysaccharide component of Gram-negative bacterial cell walls, have been shown to correlate with increasing severities of gingivitis811 since determinations of endotoxin levels may serve as a surrogate for enumeration of Gramnegative bacteria. Moreover, endotoxins possess biologic activities which suggest a role in the pathogenesis of periodontal lesions.12 20 Taken together, these findings suggest that endotoxin levels could potentially be used in periodontal diagnosis and/or the monitoring of periodontal therapy.21,22 The primary objective of this study was to determine the correlation between endotoxin levels and the percentage of Gram-negative bacteria in healthy sites and in Periodontitis sites. The study also investigated the consistency of results "Division of Oral Infectious Diseases, School of Dental and Oral Surgery, Columbia University, New York, NY. 'Consumer Products Research and Research and Development, WarnerLambert Company, Morris Plains, NJ.
obtained with multiple sampling of untreated sites over time and the ability of endotoxin levels to distinguish between healthy sites and Periodontitis sites. MATERIALS AND METHODS Twelve healthy adults (7 female, 5 male; ages 29 to 62) participated in this study. Eight of the subjects were Caucasian and four were Black. Subjects had chronic adult Periodontitis, as determined by periodontal probing and radiographie examination, with at least 3 sites with probing depths of 5 to 8 mm that bled on gentle probing and 3 healthy sites with sulcus depths of 1 to 3 mm that did not bleed on gentle probing. Sites selected for study were separated by at least one tooth. Third molars, teeth with orthodontic appliances, and teeth with endodontic lesions were not utilized. To qualify, subjects could not have received any periodontal therapy, including scaling, root planing, or prophylaxis, or have had any antibiotic therapy in the previous 6 months. Additional exclusion criteria included a history of diabetes, blood dyscrasias, hepatic or renal disease, immunosuppression, rheumatic fever, or any other condition that would require prophylactic antibiotic therapy prior to invasive dental procedures. Clinical examinations and sampling of all study sites were conducted by the same investigator (CM.) on days 0 (baseline), 7, and 14. Clinical parameters recorded included the
898
gingival index23 and a modified Quigley-Hein plaque index,24 as well as probing depth and attachment level measurements using a calibrated periodontal probe.* Subgingival samples from each study site were used to determine endotoxin levels by means of a standard limulus amoebocyte lysate assay (LAL),25 as well as the percentage of Gram-negative bacteria at each site using both a direct Modified
and
an
"indirect" method
as
described below.
Subgingival samples were collected from the mesial-buccal crevice of each of the 3 periodontally healthy teeth and from the selected proximal pocket of each of the 3 periodontally involved teeth in each subject. Prior to sampling, each area was isolated with cotton rolls, supragingival plaque
removed with a dental sealer, and the tooth surface dried with a sterile cotton pledget. One sterile #3 absorbent endodontic paper point8 was placed in each crevice or pocket for 10 seconds, dried for 1 minute with a 1500 watt space heater, placed in 500 µ of sterile pyrogen-free distilled water in a sterile 2 ml Eppendorf tube, disrupted by agitation on a vortex mixer for 30 seconds, and frozen 20°C for endotoxin assay at a later date. at The tooth surface was again dried with a cotton pledget and 2 additional sterile paper points were consecutively placed in the crevice or pocket for 10 seconds each. These points were placed in 500 µ of reduced transport fluid26 in a sterile 2 ml Eppendorf tube and sent to the laboratory for culture and microscopic determination of Gram-negative was was
—
organisms.
Endotoxin Assay A 100 µ aliquot of the sample collected for endotoxin assay was diluted with pyrogen-free water to produce serial dilutions of from 1:10 to 1:500,000. A 100 µ aliquot of each dilution was placed in a 75 x 10 mm pyrogen-free borosilicate tube to which an equal volume of limulus lysate11 was added. Samples were incubated at 37°C for 1 hour followed by refrigeration for 15 minutes at 4°C. The tubes were then gently inverted and results were recorded as positive if a firm adherent clot remained at the bottom of the tube, or negative if the combined sample remained liquid. Samples were tested concurrently with both a negative control (water) and a positive control (E. coli endotoxin standards).11 The final end point of limulus activity was considered to be the highest sample dilution producing a positive reaction. Endotoxin levels were expressed as ng/ml and calculated by comparison of the dilution end point with the end point of the E. coli standards.
Gram-Negative Organism Determinations Microscopy counts (direct Gram-stain). After the tubes were
J Periodontol November 1992
ENDOTOXIN LEVELS AND GRAM-NEGATIVE BACTERIA
vigorously agitated for 30 seconds using a Vortex dislodge microorganisms from the paper points,
mixer to
•Williams probe; Hu-Friedy Co., Chicago, IL. •Johnson & Johnson Co., New Brunswick, NJ. Cape Cod Associates, Woods Hole, MA.
Table 1. Clinical Parameters at Initial Examination
Healthy Sites 1.03 (±0.69)* Plaque index 0.79 (±0.21) Gingival index 2.66 ( ± 2.99) Probing depth (mm) 2.54 (±1.98) Attachment loss (mm) *Mean (± standard deviation). Parameter
20 air
Periodontitis Sites 1.54 1.25 5.83 6.16
(±1.08) (±0.64) ( ± 1.29) (±1.66)
µ aliquote were placed on duplicate glass slides, then dried, heat fixed, and Gram-stained for counting by
microscopic examination at 100 x. The proportion of Gramnegative organisms in the sample was determined by counting 200 microorganisms and calculating the ratio of Gramnegative to total number counted. Since spirochetes contain endotoxin but are not visualized by Gram-stain, this ratio was adjusted to account for the percent of spirochetes which was determined by examining duplicate wet mounts of each sample using darkfield microscopy according to the method of Listgarten et al.27
Cultural counts ("indirect" Gram-stain). Serial 10-fold dilutions from 102 to 106 were made from 100 µ of the original sample, and a 100 µ aliquot of each dilution was placed in duplicate on enriched trypticase soy agar,28 streaked with an L-shaped glass rod, and incubated anaerobically at 37°C for 48 to 114 hours. Plates containing suitable numbers of colonies were selected and 25 colonies per duplicate plate were randomly selected for Gram staining. Results were adjusted for percent spirochetes and recorded as the proportion of Gram-negative microorganisms per sample. Statistical Analyses A repeated measures analysis of variance (ANOVA) was used to test for differences between healthy and diseased sites as well as for significant differences within a given site across sampling periods. A simple one-way ANOVA was used to test for differences between healthy and diseased sites at each of the sampling times. Data from each sampling interval were tested separately and results applied only within a given sampling period. An alpha of .05 was
significant differences within or across sampling periods. Additionally, correlation coefficients between endotoxin levels and percent Gram-negative organisms as determined by both direct and "indirect" Gram considered to indicate
stain
were
calculated.
RESULTS Twelve subjects completed this study with each subject providing 3 healthy sites and 3 Periodontitis sites. Results of the baseline clinical examinations are contained in Table 1. There were no statistically significant differences in sitespecific clinical parameters between the baseline examination and either of the subsequent 2 weekly examinations
(data not shown).
The overall mean endotoxin levels (SD) for healthy and Periodontitis sites were 0.367 (.613) and 3.600 (5.246) ng/
Volume 63 Number 11
FINE, MENDIETA, BARNETT, KURGAN G, NAINI,
Table 2: Mean Endotoxin Levels (S.D.) at Each
Sampling Period
(ng/ml)
Healthy
Periodontitis
Sampling Day
14
.394 3.562
(.650) (5.310)*
.400 4.250
Significantly different from healthy sites,
(.648) (6.366)*
.306 2.988
(.549) (3.817)*
Endotoxin Levels in Periodontally Healthy and Diseased Sites: Correlation With Levels of Gram-Negative Bacteria Daniel H. Fine, * Carlos Mendieta, * Michael L. Naini, * and Jack W. Vincent*
Barnett,f David Furgang, * Ali
between endotoxin levels and the perin sites and in Periodontitis sites. Twelve of bacteria healthy Gram-negative centage Each 3 adults participated. subject provided Periodontitis sites with 5 to 8 mm healthy that bled on probing depths gentle probing and 3 healthy sites with sulcus depths of 1 to 3 mm that did not bleed. Clinical examinations and sterile paper point sampling of all study sites were conducted on days 0, 7, and 14, and site-specific endotoxin levels and percentage of Gram-negative bacteria were determined. There were significant differences in both endotoxin levels and percentage Gram-negative bacteria between healthy and Periodontitis sites across all 3 sampling periods, but no difference across sampling periods in the healthy sites and the Periodontitis sites, respectively. Correlation coefficients revealed a high degree of correlation between site-specific endotoxin levels and percentage of Gram-negative organisms. Using a sample dilution of 1 x 104, endotoxin levels differentiated healthy from Periodontitis sites with a specificity of approximately 91% and a sensitivity of approximately 90%. / Periodontol 1992; 63:897-901. This
study investigated the correlation
Key Words: Endotoxin; periodontitis/microbiology; periodontitis/diagnosis.
The onset and progression of periodontal diseases have been shown to be associated with both quantitative and qualitative changes in the dental plaque microbial flora. The majority of putative periodontopathogens identified to date have been Gram-negative organisms, and the proportion of plaque organisms found to be Gram-negative increases markedly as sites progress from health to disease.17 It is not surprising, therefore, that crevicular levels of endotoxin, a lipopolysaccharide component of Gram-negative bacterial cell walls, have been shown to correlate with increasing severities of gingivitis811 since determinations of endotoxin levels may serve as a surrogate for enumeration of Gramnegative bacteria. Moreover, endotoxins possess biologic activities which suggest a role in the pathogenesis of periodontal lesions.12 20 Taken together, these findings suggest that endotoxin levels could potentially be used in periodontal diagnosis and/or the monitoring of periodontal therapy.21,22 The primary objective of this study was to determine the correlation between endotoxin levels and the percentage of Gram-negative bacteria in healthy sites and in Periodontitis sites. The study also investigated the consistency of results "Division of Oral Infectious Diseases, School of Dental and Oral Surgery, Columbia University, New York, NY. 'Consumer Products Research and Research and Development, WarnerLambert Company, Morris Plains, NJ.
obtained with multiple sampling of untreated sites over time and the ability of endotoxin levels to distinguish between healthy sites and Periodontitis sites. MATERIALS AND METHODS Twelve healthy adults (7 female, 5 male; ages 29 to 62) participated in this study. Eight of the subjects were Caucasian and four were Black. Subjects had chronic adult Periodontitis, as determined by periodontal probing and radiographie examination, with at least 3 sites with probing depths of 5 to 8 mm that bled on gentle probing and 3 healthy sites with sulcus depths of 1 to 3 mm that did not bleed on gentle probing. Sites selected for study were separated by at least one tooth. Third molars, teeth with orthodontic appliances, and teeth with endodontic lesions were not utilized. To qualify, subjects could not have received any periodontal therapy, including scaling, root planing, or prophylaxis, or have had any antibiotic therapy in the previous 6 months. Additional exclusion criteria included a history of diabetes, blood dyscrasias, hepatic or renal disease, immunosuppression, rheumatic fever, or any other condition that would require prophylactic antibiotic therapy prior to invasive dental procedures. Clinical examinations and sampling of all study sites were conducted by the same investigator (CM.) on days 0 (baseline), 7, and 14. Clinical parameters recorded included the
898
gingival index23 and a modified Quigley-Hein plaque index,24 as well as probing depth and attachment level measurements using a calibrated periodontal probe.* Subgingival samples from each study site were used to determine endotoxin levels by means of a standard limulus amoebocyte lysate assay (LAL),25 as well as the percentage of Gram-negative bacteria at each site using both a direct Modified
and
an
"indirect" method
as
described below.
Subgingival samples were collected from the mesial-buccal crevice of each of the 3 periodontally healthy teeth and from the selected proximal pocket of each of the 3 periodontally involved teeth in each subject. Prior to sampling, each area was isolated with cotton rolls, supragingival plaque
removed with a dental sealer, and the tooth surface dried with a sterile cotton pledget. One sterile #3 absorbent endodontic paper point8 was placed in each crevice or pocket for 10 seconds, dried for 1 minute with a 1500 watt space heater, placed in 500 µ of sterile pyrogen-free distilled water in a sterile 2 ml Eppendorf tube, disrupted by agitation on a vortex mixer for 30 seconds, and frozen 20°C for endotoxin assay at a later date. at The tooth surface was again dried with a cotton pledget and 2 additional sterile paper points were consecutively placed in the crevice or pocket for 10 seconds each. These points were placed in 500 µ of reduced transport fluid26 in a sterile 2 ml Eppendorf tube and sent to the laboratory for culture and microscopic determination of Gram-negative was was
—
organisms.
Endotoxin Assay A 100 µ aliquot of the sample collected for endotoxin assay was diluted with pyrogen-free water to produce serial dilutions of from 1:10 to 1:500,000. A 100 µ aliquot of each dilution was placed in a 75 x 10 mm pyrogen-free borosilicate tube to which an equal volume of limulus lysate11 was added. Samples were incubated at 37°C for 1 hour followed by refrigeration for 15 minutes at 4°C. The tubes were then gently inverted and results were recorded as positive if a firm adherent clot remained at the bottom of the tube, or negative if the combined sample remained liquid. Samples were tested concurrently with both a negative control (water) and a positive control (E. coli endotoxin standards).11 The final end point of limulus activity was considered to be the highest sample dilution producing a positive reaction. Endotoxin levels were expressed as ng/ml and calculated by comparison of the dilution end point with the end point of the E. coli standards.
Gram-Negative Organism Determinations Microscopy counts (direct Gram-stain). After the tubes were
J Periodontol November 1992
ENDOTOXIN LEVELS AND GRAM-NEGATIVE BACTERIA
vigorously agitated for 30 seconds using a Vortex dislodge microorganisms from the paper points,
mixer to
•Williams probe; Hu-Friedy Co., Chicago, IL. •Johnson & Johnson Co., New Brunswick, NJ. Cape Cod Associates, Woods Hole, MA.
Table 1. Clinical Parameters at Initial Examination
Healthy Sites 1.03 (±0.69)* Plaque index 0.79 (±0.21) Gingival index 2.66 ( ± 2.99) Probing depth (mm) 2.54 (±1.98) Attachment loss (mm) *Mean (± standard deviation). Parameter
20 air
Periodontitis Sites 1.54 1.25 5.83 6.16
(±1.08) (±0.64) ( ± 1.29) (±1.66)
µ aliquote were placed on duplicate glass slides, then dried, heat fixed, and Gram-stained for counting by
microscopic examination at 100 x. The proportion of Gramnegative organisms in the sample was determined by counting 200 microorganisms and calculating the ratio of Gramnegative to total number counted. Since spirochetes contain endotoxin but are not visualized by Gram-stain, this ratio was adjusted to account for the percent of spirochetes which was determined by examining duplicate wet mounts of each sample using darkfield microscopy according to the method of Listgarten et al.27
Cultural counts ("indirect" Gram-stain). Serial 10-fold dilutions from 102 to 106 were made from 100 µ of the original sample, and a 100 µ aliquot of each dilution was placed in duplicate on enriched trypticase soy agar,28 streaked with an L-shaped glass rod, and incubated anaerobically at 37°C for 48 to 114 hours. Plates containing suitable numbers of colonies were selected and 25 colonies per duplicate plate were randomly selected for Gram staining. Results were adjusted for percent spirochetes and recorded as the proportion of Gram-negative microorganisms per sample. Statistical Analyses A repeated measures analysis of variance (ANOVA) was used to test for differences between healthy and diseased sites as well as for significant differences within a given site across sampling periods. A simple one-way ANOVA was used to test for differences between healthy and diseased sites at each of the sampling times. Data from each sampling interval were tested separately and results applied only within a given sampling period. An alpha of .05 was
significant differences within or across sampling periods. Additionally, correlation coefficients between endotoxin levels and percent Gram-negative organisms as determined by both direct and "indirect" Gram considered to indicate
stain
were
calculated.
RESULTS Twelve subjects completed this study with each subject providing 3 healthy sites and 3 Periodontitis sites. Results of the baseline clinical examinations are contained in Table 1. There were no statistically significant differences in sitespecific clinical parameters between the baseline examination and either of the subsequent 2 weekly examinations
(data not shown).
The overall mean endotoxin levels (SD) for healthy and Periodontitis sites were 0.367 (.613) and 3.600 (5.246) ng/
Volume 63 Number 11
FINE, MENDIETA, BARNETT, KURGAN G, NAINI,
Table 2: Mean Endotoxin Levels (S.D.) at Each
Sampling Period
(ng/ml)
Healthy
Periodontitis
Sampling Day
14
.394 3.562
(.650) (5.310)*
.400 4.250
Significantly different from healthy sites,
(.648) (6.366)*
.306 2.988
(.549) (3.817)*
