Synovial Fluid pH as an Indicator of Infected Joint Arthroplasty

Synovial Fluid pH as an Indicator of Infected Joint Arthroplasty

Giles Stafford, Mohammed Akmal, Catriona Mitchell-Hynd, John Skinner, George Bentley.
Royal National Orthopaedic Hospital, Brockley Hill, Stanmore, Middlesex HA7 4LP, U.K.

Corresponding Author:
Giles Stafford FRCS, 12 Beaver Close, Crowborough, East Sussex, TN6 2NZ, U.K.

E-mail: gstafford@doctors.org.uk

Abstract

Background
The synovial fluid pH of patients with septic arthritis has been shown to be lower than normal subjects. We hypothesise that this may be true of patients with infected arthroplasty.

Methods
We collected samples of synovial fluid from 47 osteoarthritic patients undergoing either primary or revision hip and knee arthroplasty. Of these, 8 revisions were proven to be infected. Synovial fluid pH and lactate parameters were measured using a blood gas analyser perioperatively.

Results
Our results demonstrate no significant statistical difference in the synovial fluid pH between the primary and non-infected revision groups (Primary group mean pH 7.30 ±0.19, Uninfected revision group mean pH 7.23 ±0.09, p=0.727). Conversely, the synovial fluid pH of the infected revision group was lower (mean pH 7.06 ±0.12), with a significant statistical difference between the infected and non-infected groups (p=0.006).

Conclusions
Synovial fluid pH could become a useful tool in the diagnosis of infection prior to revision arthroplasty.

Background
The cause of failure of joint arthroplasty can pose a difficult question, especially when differentiating between septic and aseptic loosening. It is important to define the cause of prosthetic failure, as this can influence the decision between a single or two stage revision procedure. A significant proportion of prosthetic infections are not definitively apparent by clinical, biochemical or radiological methods (1).
Previous research has shown that the biochemical properties of synovial fluid are altered in the presence of infection, with increasing lactate levels and a subsequent lowering of pH (2,3,4). This has been demonstrated in both humans and animals (5).
However no work has been done to correlate the biochemical properties of synovial fluid in the prosthetic joint. If the trend that has been shown in septic arthritis is also true of prosthetic joints, the synovial fluid biochemistry could be used as an additional modality to aid the diagnosis of infected arthroplasty.

Methods
Synovial fluid samples were taken perioperatively from 45 osteoarthritic patients undergoing either primary or revision hip and knee arthroplasty. Samples were taken using a standard 5ml heparinised blood gas syringe (Becton-Dickson Drihep™ Plus) with a 19G needle intraoperatively before the joint capsule was opened. The synovial fluid pH and lactate were measured using a blood gas analyser (Radiometer ABL 700 series, Radiometer, Copenhagen, Denmark) within one hour of the sample being taken. Previous work has shown that the pH in aspirated synovial fluid rises slowly with time, but not significantly within six hours (6). 
Biopsies were also sent perioperatively for both microbiological and histological analysis if appropriate. A pro forma was filled out for each patient recording their personal details, pathology, operation, pH and lactate plus any further comment. From the results of pre, peri and postoperative microbiology and histology results, the patients were allocated to one of three groups – a) uninfected primary arthroplasty, b) uninfected revision arthroplasty, c) infected revision arthroplasty.

Results
Demographics
The uninfected primary group consisted of 18 patients. The mean age was 65.33 years (SD=8.63) with a male to female ratio of 4:5. Of this group, 14 were undergoing knee and 4 total hip replacements. The uninfected revision group consisted of 21 patients with a mean age of 69.38 years (SD=8.06) and M/F ratio of 3:4. Of these, 13 underwent revision knee and 8 revision hip arthroplasty. The infected revision group had 8 patients, mean age 65.5 (SD=8.67) and M/F ratio of 3:5. Seven of the 8 patients had infected knee prostheses.
There was no significant difference in the synovial fluid pH and lactate values between the hips and knees in each group.

Synovial Fluid pH
The mean synovial fluid pH values of the uninfected primary group was 7.30 (SD=0.19). This was higher than those of the uninfected revision group which had a mean of 7.23 (SD=0.09). In turn, this was again higher than the infected revision group, mean pH 7.06 (SD=0.12). (Fig.1).

Figure 1. Histogram of mean pH

Mathematical analysis of the synovial fluid pH in the three groups shows no significant statistical difference between the primary and uninfected revision groups (Kruskal – Wallis pairwise comparisons, P=0.727). However, the difference in pH between the uninfected revision group and the infected revision group was significantly different (P=0.006). This difference was also present for comparisons between the primary and infected revision groups.  (P=0.0048). After pooling the primary and uninfected revision data (not significantly different) and comparing with the infected revision group, a very significant difference is observed (one sided student’s t-test of unequal variance, P=0.0009).

Synovial Fluid Lactate
The mean synovial fluid lactate levels of the uninfected primary group was 2.95 mmol/L (SD=1.22). The mean values of the uninfected revisions and infected revisions were 4.6 mmol/L (SD=1.15) and 6.03 mmol/L (SD=2.26) respectively (Fig. 2).

Figure 2. Histogram of mean Lactate

Figure 2. Histogram of mean Lactate

Statistical analyses of synovial fluid lactate levels show differences between the groups. There was significant difference between the uninfected revision and infected revision groups (P=0.0486). However, unlike the values for synovial fluid pH, the lactate values from the primary and uninfected revision groups were also shown to be different (P=0.002).

Figure 3. Scatter graph of data from all three groups with linear correlation illustrating trend.

Figure 4. Table of individual results

Microbiology
Breakdown of the positive microbiological cultures obtained included three patients with staphylococcus epidermidis, one infected with staphylococcus aureus, one with streptococcus viridans and one further patient with peptostreptococcus. Diagnosis of infection in the remaining two patients was made on histological grounds perioperatively, but culture of specimens did not yield significant growth.
We had no detrimental effects to the blood gas analyser by using it in this fashion.

Discussion
The trends demonstrated in the results are consistent with previous studies looking at synovial fluid biochemistry in septic arthritis. The synovial fluid pH of the infected revision arthroplasty group was significantly more acidotic than the pH of the primary and non-infected revision arthroplasty groups. Our results also indicate that this may be due to factors other than the build up of lactate alone.
Our original hypothesis was that the metabolic products of infective bacteria and subsequent inflammatory response may cause a decrease in synovial fluid pH. We also attempted to determine any difference in synovial fluid pH between uninfected primary joints and uninfected revision joints. The inflammatory response to particulate wear debris in joint replacement prostheses does have an effect on synovial fluid biochemistry (7,8). Despite this, our data does not demonstrate statistically significant differences in synovial fluid pH between uninfected primary joints and uninfected revisions. 
In addition to this, studies specifically investigating the effect of effusions (which are commonly seen in prosthetic infection) on synovial fluid biochemistry have shown that high intra-articular pressure is thought to cause acidosis independently (9,10). The mechanism described is via the shunting of blood away from the subsynovial capillaries, resulting in local hypoxia and the build up of lactate.  This does not detract from synovial fluid pH being a useful diagnostic aid, but one needs to be aware of the possibility of a false positive indicator of infection in a sterile joint if a large traumatic effusion is present.
Another correlation which has been documented is between synovial fluid leukocyte count and pH (11). The authors of this study postulate that a decrease in synovial fluid pH may not only be due to an increase in lactate, but also the result of the hexose monophosphate metabolism of the increased numbers of leukocytes found in infected joints. It is conceivable that the drop in synovial fluid pH associated with infection is caused by a combination of all of the above factors. 
Studies into the synovial fluid biochemistry of patients with various inflammatory and non-inflammatory arthritidies have shown that the pH and corresponding lactate levels are altered in those with rheumatoid arthritis when compared to other arthritic conditions (3,8,12,13). This has been shown to be due in part to acid secretion by synovial fibroblasts in patients with rheumatoid arthritis (13). Due to low numbers of rheumatoid patients available at the time of the study, we selected only osteoarthritic subjects. These patients have been shown to have similar synovial fluid biochemical profiles to normal subjects (12). A separate study is needed to clarify the relationship of synovial fluid biochemistry in rheumatoid patients and those with prosthetic sepsis.
We are also as yet to ascertain the effect of gonococcal sepsis in the prosthetic joint. Previous studies have demonstrated that in the presence of gonococcal infection of the virgin joint, synovial fluid lactate is not significantly raised above normal values although pH may be lower (2,3,4).

Conclusions
Extrapolating our results into the clinical setting, if one uses pH 7.2 at which under is deemed likely to be infected, the returned sensitivity is 88% and specificity 67%.
Synovial fluid pH may not be an appropriate tool upon which to solely diagnose infection, but it is cheap, simple, fast and readily available.  Our results indicate that synovial fluid pH is a useful addition to the diagnosis of septic prosthetic failure. Further investigation should yield clinically useful results.

Competing Interests
The authors declare that they have no competing interests.

Authors Contributions
GH collected the majority of the data, collated and interpreted the results and drafted the manuscript. MA provided the original premise for the study and collected data. CM-H collected background data. JS and GB supervised the study.

Acknowledgements
We would like to thank to Mr. Bernie Higgins, Senior Lecturer in Statistics at University of Portsmouth, for his invaluable help with the analysis of the data.

References
1. Spangehl MJ, Masri BA, O’Connell JX, Duncan CP. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg [Am] 1999;81-A:672-83.
2. Brook IK, Reza MJ, Bricknell KS, Bluestone R, Finegold SM. Synovial fluid lactic acid – a diagnostic aid in septic arthritis. Arthritis Rheum 1978;21-7:774-79.
3. Treuhaft PS, McCarty DJ. Synovial fluid pH, lactate, oxygen and carbon dioxide partial pressure in various joint diseases. Arthritis Rheum 1971;14-4:475-84.
4. Gobelet C, Gerster JC. Synovial fluid lactate levels in septic and non-septic arthritidies. Ann Rheum Dis 1984;43(5):742-45.
5. Tulamo RM, Bramlage LR, Gabel AA. Sequential clinical and synovial fluid changes associated with acute infectious arthritis in the horse. Equine Vet J 1989;21(5):325-31.
6. Falchuk KH, Goetzl EJ, Kulka JP. Respiratory gases of synovial fluids.  Am J Med 1970;49:223-31.
7. Kovacik MW, Gradisar IA, Haprian JJ, Alexander TS. Osteolytic indicators found in total knee arthroplasty synovial fluid aspirates. Clin Orthop 2000;379:186-94.
8. Chen FS, Scher DM, Clancy RM, Vera-Yu A, Di Cesare PE. In vitro and in vivo activation of polymorphonuclear leukocytes in response to particulate wear debris. J Biomed Mater Res 1999;48(6):904-12.
9. Richman AI, Su EY, Ho Jr G. Reciprocal relationship of synovial fluid volume and oxygen tension. Arthritis Rheum 1981;24(5):701-5.
10. James MJ, Cleland LG, Rofe AM. Determinants of synovial fluid lactate concentration. J Rheumatol 1992;19(7):1107-10.
11. Ward TT, Steigbigel RT. Acidosis of synovial fluid correlates with synovial fluid leukocytosis.  Am J Med 1978;64:933-36.
12. Cummings NA, Nordby GL. Measurement of synovial fluid pH in normal and arthritic knees. Arthritis  Rheum 1966;9(1):47-56.
13. Parak WJ, Dannöhl S, Geaorge M, Schuler MK, Schaumburger J, Gaub HE, Müller O, Aicher WK. Metabolic activation stimulates acid production in synovial fibroblasts. J Rheumatol 2000;27(10):2312-22.