Reproducibility of microscopic and cultural data in repeated subgingival plaque samples

Publisher Website
Google Scholar
Abstract
In association studies, micro-organisms can only be recognized as suspects for playing a major role in the development of a pathological environment, if their destructive action goes along with a marked proportional increase of their numbers or if their first detection can be related to the clinical onset of the disease. Limitations in the reproducibility of repeated samples have to be taken into account, when changes of the microbial composition of subgingival environments are to be studied, and when local clinical changes are to be related to shifts in the composition of the pertaining microbiological compartment. To study reproducibility, a total of 109 sites was sampled repeatedly with sterile paperpoints at an interval of 7 to 10 days in 24 patients suffering from periodontal disease and 12 edentulous patients wearing successful and failing osseointegrated titanium implants. Using continuous anaerobic techniques, the samples were cultured on nonselective and selective media and were studied by darkfield microscopy. Both the intertest-agreements of frequencies of detection (.vkappa.-statistics) as well as the discrepancies of proportions of bacterial groups and selected bacterial species were determined. The standard deviation of proportional differences between first and second samples ranged between 6.4% (fusiform organisms) and 17.2% (coccoid cells) for darkfield parameters, between 4.3% (B. melaninogenicus on ETSA/Kana). and 14.0% (B. gingivalis on ETSA/Kana.) for selected bacterial species and between 6.9% (gram-negative anaerobic cocci) and 24.0% (gram-positive facultative cocci) for bacterial groups classified according to gram stain characteristics and atmospheric growth conditions. As the magnitude of deviation depended upon the mean value, coefficients of variation (cv) were also determined. The lowest cv values (0.25) were found for the proportions of spirochetes in the darkfield microscope, for the proportions of the gram-positive facultative rods (0.29) for gross bacterial groups and for Fusobacterium spp. (0.60) for selected bacterial species. The intertest-agreement of the frequencies of detection were best for spirochetes in the microscope and for B. intermedius with selected cultivated species. The results did not indicate a general bias of the results of second samples due to the previous sampling (McNemar''s .chi.2). Methodological proposals for estimation and comparison of proportions of bacteria are made and discussed in an appendix.