Papers

Relation between mouth and haematogenous infection in total joint replacements

BMJ 1994; 309 doi: http://dx.doi.org/10.1136/bmj.309.6953.506 (Published 20 August 1994) Cite this as: BMJ 1994;309:506
  1. C A Bartzokas,
  2. R Johnson,
  3. M Jane,
  4. M V Martin,
  5. P K Pearce,
  6. Y Saw
  1. Wirral Hospital NHS Trust, Wirral, Merseyside L49 5PE Liverpool Dental Hospital, Liverpool L3 5PS
  1. Correspondence to: Dr C A Bartzokas, Department of Medical Microbiology, Wirral Hospital NHS Trust, Clatterbridge Hospital, Bebington, Wirral L63 4JY.
  • Accepted 15 June 1994

Abstract

Objective : To investigate the source of infections associated with orthopaedic prostheses.

Design : Analysis of four infections of prosthetic joints with case records; minimum inhibitory and minimum bactericidal concentrations and sodium dodecylsulphate polyacrylamide gel electrophoresis of the cell wall polypeptides of the Streptococcus sanguis isolates from the mouth and infected prostheses; examination of the patients' mouths for periodontal disease and caries.

Subjects : Four adults (three men) aged 58-83.

Results : For each patient the strain of S sanguis isolated from the mouth was indistinguishable from that isolated from the prosthesis. All patients had severe periodontal disease and caries.

Conclusions : The mouth was probably the source of bacterial infection in the prosthetic joints of these patients; the route of infection was possibly haematogenous. Incipient oral infection should be treated before joint replacement, and oral health should be maintained indefinitely.

Clinical implications

  • Clinical implications

  • The source of secondary infection of prosthetic joints has been attributed to the mouth

  • No definitive evidence for this association, however, has been reported

  • This study confirms that oral sepsis can be a source of bloodborne infection of prosthetic joints

  • Incipient oral sepsis should be treated before a patient has a joint replacement, and oral health should be maintained indefinitely

Introduction

Deep infection of prosthetic joint replacements is a serious complication.1, 2 Several circumstantial cases of secondary infection of prosthetic joints, presumed to originate in the mouth, have been reported.*RF 3-13* The mouth is a recognised source of bacteraemia, particularly if untreated dental caries or periodontal disease is present. To date, however, no conclusive evidence that the mouth is a definitive source of infection in total joint replacements has been reported. This paper describes four cases of deep infection - three in total knee replacements and one in a total hip replacement - caused by Streptococcus sanguis (an (alpha)-haemolytic streptococcus of the viridans group), a common constituent of the oral microflora.

Patients and methods

Table I shows details of the four patients.

Table I

TABLE I - Details of four patients* with postoperative prosthetic infection caused by Streptococcus sanguis

View this table:

We identified all the strains of S sanguis isolated from the infected prosthesis and mouth using the methods of Beighton and colleagues14 and determined their minimum inhibitory and minimum bactericidal concentrations against penicillin, erythromycin, tetracycline, amoxycillin, and clindamycin according to the methods of Reeves and colleagues.15

Analysis of cell wall polypeptides

We compared the cell wall polypeptides of the four pairs of S sanguis using sodium dodecylsulphate polyacrylamide gel electrophoresis.16 Briefly, the strains were grown in tryptone yeast extract (Lab M, London), supplemented with 2 g glucose, and incubated at 37°C aerobically for 18 hours. The bacterial growth was harvested by centrifugation and suspended in 10% trichloroacetic acid for one hour. The precipitate, collected by centrifugation, was solubilised in sodium dodecylsulphate. About 100 μg protein was subjected to discontinuous gel electrophoresis according to the methods of Laemmli,17 with a 5% stacking gel and a 8.75% resolving gel. All gels were fixed in 12.5% trichloroacetic acid and stained with 2% coomassie blue, before being destained in 10% acetic acid. We photographed the gels on Agfa, Dia Direct black and white transparency film (15 ASA; Agfa, London), thus avoiding any distortion by drying and bubble formation during scanning. We scanned photographs of the gels densitometrically (2202 laser densitometer linked to a 2202 plotting and recording integrator; LKB Pharmacia, Bromma, Sweden) and calculated the molecular weight and number of polypeptides resolved.

We analysed the polypeptide pattern further for qualitative differences by summary methods of measurement using the differences in each gel pattern. We arranged every polypeptide detected from each isolate in a contingency table, recording presence or absence. We compared the differences in the polypeptide pattern of the bands between the isolates from the oral cavity and those from the prosthesis. S sanguis ATCC 10556 was used as a control strain throughout the experiments.

Results

Table II shows the minimum inhibitory and minimum bactericidal concentrations of the isolates of S sanguis derived from the mouth and the prosthesis against five antibiotics. For each patient for each antibiotic the minimum inhibitory and minimum bactericidal concentrations of the isolates from the mouth were the same as for those from the prosthesis and the minimum inhibitory concentrations differed from the minimum bactericidal concentrations.

Table II

Minimum inhibitory and minimum bactericidal concentrations (μg/ml) of Streptococcus sanguis isolated from mouth and prosthesis, against five antibiotics

View this table:

The number of polypeptides detected from the cell wall of the isolates ranged from 26 to 31. In cases 2 and 4 the number of polypeptides in the isolates from the mouth was the same as the number in the isolates from the prosthesis; in cases 1 and 3 the numbers differed by one. None of the isolates completely matched the polypeptide pattern of the control strain: there were between 3 and 6 different bands. The figure shows the differences between the mouth and the prosthesis in the number and distribution of polypeptides from the isolates of S sanguis.

Discussion

We tried to determine whether the strains of S sanguis isolated from the mouth and the infected prosthesis in four patients were identical. As no definitive method of comparing strains of S sanguis has yet been described we compared the isolates by their antibiograms and cell wall polypeptides. The results of these analyses are consistent with the strain of S sanguis isolated from the mouth being the same as that from the infected prosthesis. Although an exogenous source of S sanguis cannot be unequivocally excluded, a bloodborne dissemination of this oral commensal to the prosthesis is the most likely source of the infections in these four patients.

The mouth has been previously implicated as a source of infection in orthopaedic implants.*RF 3-13* The proportion of hip replacements that develop haematogenous infection has been estimated at 0.3%.18, 19 The proportion of hip and knee replacements that are infected from the mouth must be much smaller. The early infections of cases 1 to 3 are the only instances in which S sanguis has been isolated from prostheses at the Wirral hospital in the past four years: an incidence of <0.25%. Despite the unexpected presence of S sanguis in orthopaedic prostheses a neglected mouth can be treated and can be eliminated as a source of such infection. Some studies have suggested a preoperative dental check up before a joint replacement.7 We strongly recommend that before orthopaedic prostheses are implanted incipient oral sepsis should be treated and that patients with implanted prostheses should maintain oral health indefinitely.

Figure1

Profiles from sodium dodecylsulphate polyacrylamide gel electrophoresis of cell wall polypeptides of Streptococcus sanguis isolates from mouth and prosthesis

We thank Mrs Wendy McKeown for the primary microbiological investigations and Mr N P J Geary for allowing us to include the fourth case in our study.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
View Abstract