Journal of Orthopedic Surgery and Techniques

The Use of Preoperative CRP and ESR as Predictive Markers of Prosthetic Joint Infection in Primary Total Hip and Knee Arthroplasty

Abstract

Background : C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) are inflammatory markers commonly used during the diagnosis and treatment of periprosthetic joint infection (PJI) in arthroplasty patients. The purpose of this study was to assess for a correlation between pre-operative CRP/ESR and post-operative development of PJI in patients undergoing primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) and to assess modifiable and non-modifiable risk factors in patients with elevated preoperative inflammatory markers.

Methods : Retrospective review of 806 patients at a single institution who underwent THA (n = 291) and TKA (n = 515). CRP and ESR were obtained on all patients during the preoperative workup. Patient demographics, medical comorbidities, and occurrence of PJI were extracted. Positive CRP was set at any value > 0.3 mg/dL and positive ESR as any value > 30 mm/hr.

Results:  In our study, we found no statistically significant correlation between pre-operative CRP or ESR and PJI. Although notably, we found that a higher proportion of patients with PJI had elevated preoperative CRP (70.6%) vs. those with normal CRP (29.4%).

Conclusions:  This study is unable to validate the use of preoperative CRP/ESR as a predictive measure of PJI in primary THA/TKA. However, it does provide quantitative insight into the prevalence of elevated preoperative CRP/ESR in all patients prior to THA/TKA with a considerable proportion having modifiable risk factors. Since a large proportion of patients with elevated CRP/ESR didn’t develop PJI, we do not recommend cancellation of THA/TKA unless there is an obvious modifiable risk factor that confers substantially heightened risk of PJI.

Keywords

Arthroplasty, Periprosthetic joint infection, ESR, CRP, Complications, Risk factors

Introduction

Total hip (THA) and total knee (TKA) arthroplasty are two of the most performed orthopaedic surgeries each year in the United States. Some of the more common complications include prosthetic joint infection (PJI), aseptic loosening, instability, dislocation, periprosthetic fracture, stiffness, chronic pain, and DVT/PE [1-4]. Currently, more than 400,000 THA and more than 700,000 TKA are performed each year with recent projections to increase to an estimated 635,000 THAs and 1,260,000 TKAs by year 2030 [5].

Despite advances, PJI remains one of the most common serious complications among all patients undergoing total joint arthroplasty (TJA) [6-10]. The overall incidence of PJI in THA and TKA is 1-2% and 2-3% respectively and it is increasing proportionately with the overall volume of TJA being performed nationwide [11]. It is estimated that PJI is responsible for 20% of all revision THA and 25% of all revision TKA [10]. In addition to the health burden of PJI to patients, the projected annual economic burden by 2030 is estimated to be $753.4 million for THA PJI and $1.1 billion for TKA PJI [12].

With the shift to value-based care and alternative payment models, it has become increasingly important to optimize patients prior to surgery to decrease the risk of postoperative complications [13-16]. Typical components of an optimization panel include history and physical exam, complete blood count, comprehensive metabolic panel, coagulation studies (PT/INR), electrocardiogram, body mass index (BMI), mental health screen, dental screen, nicotine screen, and alcohol and illicit drug use screen [15,16].

C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) are commonly used, but nonspecific, markers for inflammation and infection in a wide variety of contexts throughout all fields of medicine [17-22]. Within the field of orthopaedic surgery, CRP and ESR are widely used to diagnose and track infection, including THA and TKA PJI. There has been a significant amount of literature on the use of CRP and ESR in the field of TJA, specifically for the diagnosis of PJI [23-27]. However, there is much less data focused on the use of preoperative CRP and ESR in the setting of primary TJA. Previous studies have found a correlation with elevated CRP and PJI, while others have found no correlation [6,28-31]. The purpose of this study is to further investigate any correlation that exists between preoperative CRP/ESR and PJI in primary THA and TKA. It also serves to assess risk factors among patients with elevated preoperative inflammatory markers to better understand overall incidence in this population of patients.

Materials and Methods

All patients who underwent primary total hip or total knee arthroplasty between 2016 and 2020 at John Peter Smith Hospital were identified and studied via a retrospective chart review. Patient characteristics, acute infections, labs, and past medical history were extracted from their medical records. Exclusion criteria included THA for femoral neck fracture, follow-up < 1 year, and federal inmates (due to minimal follow-up documentation). After excluding those patients, 806 patients were included in the final cohort for analysis. Minimum follow-up was 1 year. All surgeries were performed at a single institution (John Peter Smith Hospital, Fort Worth, TX) by one of two attending orthopaedic surgeons with fellowship training in adult reconstruction.

Baseline patient demographics, medical history, preoperative labs (CRP, ESR, urinalysis), and outcomes (prosthetic joint infection) were extracted from their medical records. Patient past medical history was used to assess the impact of documented obesity, smoking status, rheumatoid arthritis, liver disease, renal disease, HIV, and other conditions (DVT/PE, coagulopathy, stroke, transient ischemic attacks, coronary artery disease). For the purpose of an objective cutoff, positive CRP was set at any value > 0.3 mg/dL and ESR as any value > 30 mm/hr. PJI was defined using MSIS criteria.

A logistic regression was performed to assess the impact of ESR on positive CRP labs. To establish a full model, all items abstracted from the medical record underwent univariate pre-filtering. All available items were analyzed via univariate regression, and items with a p-value of 0.2 or lower were analyzed in our multivariate models. Backwards stepwise regression was used to construct parsimonious models. All testing across surgery type (i.e. Total Hip Arthroplasty and Total Knee Arthroplasty) was performed based on the data’s distribution; where appropriate Wilcoxon-Mann-Whitney U test, chi square, or Fisher’s exact test was used.

Approval through the North Texas Regional Institutional Review Board (IRB) was obtained for this chart review and thus informed consent was waivered by the IRB, approval number 1354130-2. All review of PHI in this study was HIPAA-compliant.

Results

Demographics and THA vs. TKA

A total of 806 patients underwent primary total hip (291) and total knee arthroplasty (515). Table 1 provides an overview of patient demographics and medical comorbidities distributed by surgery performed. There was a significant difference when comparing patient age (p-value < 0.0001) across the two surgery types, the median age in the THA group was 55 (IQR: 49 to 60), whereas the median age was 62 for patients undergoing a TKA (56-69). When assessing BMI across the two surgery types, there was a significant difference in proportion (p-value = 0.0001) and there were a larger proportion of patients who were obese undergoing a TKA (70.1%) when compared to those who had a THA (49.1%). Likewise, there was a significant difference (p-value < 0.0001) in patients who suffered avascular necrosis; there were zero patients suffering from avascular necrosis who underwent a TKA, and 22.3% of those who underwent THA had avascular necrosis. In total, there were 65 patients that developed avascular necrosis, two of whom had PJI. Similarly, there was a significantly lower (p-value = 0.0487) proportion of patients suffering from a PJIs who underwent a TKA (1.4%) when compared to those who underwent a THA (3.4%).

Patients who had a prior surgery make up a significantly greater proportion (p-value = 0.0001) of patients undergoing a TKA (22.1%) when compared to those undergoing a THA (11.4%). When assessing smoking status there was a significantly higher (p-value < 0.0001) proportion of patients who never smoked undergoing a TKA (61.9%) when compared to patients undergoing a THA (44.8%). There were significantly (p-value = 0.0009) more patients who suffered from diabetes mellitus undergoing a TKA (28.0%) when compared to those undergoing a THA (17.5%).

CRP and ESR distribution

The distribution of patient CRP and ESR results are detailed in Table 2. Overall, 49.9% of patients had no elevation in CRP (< 0.3 mg/dL) and 50.1% of patients had elevation in CRP (> 0.3 mg/dL). 82.8% of patients had no elevation in ESR (30 mm/hr or less) and 17.3% of patients had elevated ESR (> 30 mm/hr). 53.7% of patients had elevation of either CRP or ESR, 46.3% of patients had both normal CRP and ESR, and 13.7% of patients had elevation of both CRP and ESR. When assessing CRP and ESR elevation across the different lab results, there were no significant differences in PJI infection. Furthermore, when combining ESR and CRP results there were no significant differences in PJI status across the combination labs.

Modifiable and non-modifiable risk factors

The distribution of modifiable and non-modifiable causative risk factors with associated elevations in CRP and/or ESR is detailed in Table 3. When assessing all potential causative risk factors (increased BMI, urinalysis, current smoking status, diabetes, rheumatoid arthritis, HIV, renal, liver, CHF, or coagulopathy), there were no significant differences across any combination of ESR and CRP labs. Elevations of ESR (p- value < 0.0001), and patients with positive CRP and negative ESR negative (p-value = 0.0132), were significantly associated with increased likelihood of having an unmodifiable risk factor (diabetes, rheumatoid arthritis, HIV, renal, liver, CHF, or coagulopathy). Similarly, elevations of ESR (p-value = 0.0039), along with positive CRP and negative ESR (p-value = 0.0051) were significantly associated with increased likelihood of having a modifiable risk factor (increased BMI, positive urinalysis, current smoking status).

Odds of PJI with elevated CRP and/or ESR

The full logistic model is presented in Table 4, and we found that a positive preoperative ESR significantly increases the odds of a positive preoperative CRP test (p-value < 0.0001). For each unit increase in ESR the odds of a positive CRP lab increase by a factor of 1.056 (95% CI: 1.040, 10.072). Additionally, the odds of a positive CRP was 1.973 times higher in patients who had a total hip arthroplasty when compared to patients who had a total knee arthroplasty (95% CI: 1.389, 2.803; p-value = 0.0001). The odds of a positive CRP was 46% lower in patients who are former smokers, when compared to patients who currently smoke (95% CI: 0.328, 0.888; p-value = 0.0116). We also found that patients who were obese had a significantly (p-value < 0.0001) higher odds of having a positive CRP lab. The odds of an obese patient having a positive CRP lab were 4.102 times higher when compared to patients with a normal BMI (95% CI: 2.442, 6.89).

The results of the reduced model resulting from a stepwise regression are summarized by Table 5. The most parsimonious model consisted of ESR, procedure type, smoking status, and BMI. The odds of a positive CRP lab significantly (p-value < 0.0001) increases by 1.061 for each unit of ESR (95% CI: 1.046, 1.076). The odds of a positive CRP lab is 2.014 times higher (95% CI: 1.425, 2.845; p-value < 0.0001) in patients who underwent total hip arthroplasty, when compared to patients who underwent total knee arthroplasty. Furthermore, the odds of a positive CRP was 44.8% lower in patients who are former smokers, when compared to patients who currently smoke (95% CI: 0.336, 0.904; p-value = 0.0137). We also found that patients who were obese had a significantly (p-value < 0.0001) higher odds of having a positive CRP lab. The odds of an obese patient having a positive CRP lab were 4.203 times higher when compared to patients with a normal BMI (95% CI: 2.502, 7.059).

Discussion

To date, only a handful of studies have investigated preoperative ESR or CRP and the potential correlation with PJI. Results have been mixed; some studies have found a positive correlation while others have found no correlation [6,28-32]. Xu, et al. looked specifically at TKA in osteoarthritis patients. In their retrospective review of 3,376 cases, they found the overall prevalence of elevated preoperative inflammatory markers to be 4.1%. The rate of PJI was higher in patients with elevation of both CRP and ESR (12.5%) compared to either high (0.9%) or both normal groups (1.4%) [28]. In another retrospective review of 50 matched patients, Pfitzner, et al. found that the average preoperative CRP in the PJI group was 1.3 mg/dL vs. 0.4 mg/dL in the non-infected group. They recommend performing CRP on all patients before THA/TKA and suggest a threshold of 0.5 mg/dL in which you should perform further investigation as to a possible cause [29]. Although not specific to PJI, two additional studies have assessed preoperative CRP and outcomes in orthopaedic patients. In a study by Ghosh, et al. they found that patients with high preoperative CRP (> 3 mg/dL) may be at high risk of developing complications after postoperative day 14 and increased operative time [31]. Similarly, in a study by Ackland, et al. they found that in patients with higher preoperative CRP, there was an increase in delayed post-operative complications and longer length of stay [32].

In a retrospective review of 351 TKAs, Godroy, et al. found no statistical difference between CRP or ESR and any complications. The overall number of infections in this group was 8, only 2 of which were considered deep infections [6]. Another study looked specifically at preoperative CRP on patients with femoral neck fracture before they underwent hemiarthroplasty. The overall infection rate was 4.85%. Their study did not validate the use of CRP levels or suggest a threshold that was predictive of pre-existing infection before performing a hemiarthroplasty for femoral neck fracture [30].

In our study, which is the second largest cohort to date that investigates preoperative inflammatory markers before primary TJA, we found no statistically significant correlation between pre-operative CRP or ESR and PJI. In total, 806 primary TJA cases were reviewed (515 TKA, 291 THA). The rate of PJI was 2.1%. We did find that a higher proportion of patients with PJI had elevated preoperative CRP (70.6%) versus those with normal CRP (29.4%); however, among patients without PJI, the preoperative CRP was still found to be elevated > 0.3 mg/dL in half of all patients (49.7%). ESR followed a different trend with only 29.4% of patients with PJI having an elevated preoperative ESR. Looking at the overall cohort, just over half of all patients with or without PJI had elevated preoperative CRP (50.1%), while only 17.3% of all patients had elevated ESR.

While our results regarding the correlation of preoperative CRP and ESR with PJI do not meet statistical significance, they do highlight an important finding. There is a large proportion of patients with elevated preoperative CRP and ESR that undergo primary TJA and most of them do not go on to develop a PJI. Given these findings, we investigated risk factors among our cohort of patients to see if there was a correlation. No study to date on the topic of TJA and preoperative CRP/ESR has investigated this. Among all patients with elevated preoperative CRP or ESR, we identified a potential risk factor in 95.6% of cases, with risk factors including increased BMI, positive urinalysis, current smoking status, diabetes mellitus, rheumatoid arthritis, HIV, renal disease, liver disease, CHF, and coagulopathy. Furthermore, 94.0% of the time, at least one modifiable risk factor was present, which we defined as increased BMI, positive urinalysis, and current smoking status. This finding can be explained by the fact that CRP/ESR are both non-specific inflammatory markers and, as such, are often elevated in patients with chronic disease which involves a systemic inflammatory response. A publication by Watson, et al. looked specifically at idiopathically elevated CRP and ESR in patients undergoing primary TKA. They found that higher BMI was correlated with higher levels of preoperative CRP and ESR [33]. Our study supports this same finding; patients with BMI in the obese category (> 30), were more likely to have elevated preoperative CRP level (OR 4.2, p < 0.0001). Lastly, data in Table 2 displays an interesting finding. Patients with elevation in CRP or ESR (Table 2) begins to closely approach statistical significance (p = 0.0557), indicating that pre-operative elevation in either CRP or ESR is sensitive for potential correlation with PJI. This highlights that there is value in obtaining these pre-operative labs despite the results not being statistically significant.

Even after identifying numerous potential inflammatory risk factors, there were still nearly 5% of cases in which there were no identifiable risk factors. This is potentially explained by the disease process of osteoarthritis itself inherently leading to an inflammatory response. A study by Takahashi investigated this, showing elevation of CRP in patients with generalized osteoarthritis and a positive correlation of elevated inflammatory markers with Kellgren-Lawrence scale of arthritis severity [34]. Similar findings were seen in a study that looked at CRP and ESR in patients with and without knee osteoarthritis. They found a positive correlation of elevated ESR and CRP with higher Kellgren-Lawrence grade in patients with knee osteoarthritis [35]. No literature, to our knowledge, has investigated this concept in hip osteoarthritis as both aforementioned studies were specific to knee osteoarthritis. In our study, we saw a slightly higher proportion of hip arthritis patients with elevated preoperative labs than knee arthritis patients, as 54.6% of all patients who underwent THA had elevation in preoperative CRP, versus 47.6% of all TKA patients.

Weaknesses of our study include its retrospective nature and relatively low amount of total PJI (n = 17). To definitively identify a correlation with PJI, we would need thousands of patients in our cohort to investigate a highly substantial number of total PJIs. Despite this, our study does highlight the fact that there is a substantial number of patients that demonstrate elevated pre-op inflammatory markers, and most of them do not go on to develop PJI after primary TJA. Improved documentation in future studies could be helpful for increasing sample size and thus achieving a high statistical power. This study was unable to address the percentage of the time that a modifiable risk factor could have been addressed and resolved prior to surgery. This information was unable to be analyzed due to inconsistent documentation, which could be in part due to the retrospective nature of this study. A prospective analysis may have allowed the participating physicians to create a dataset as they operated for which patients’ risk factors were able to be addressed. This information could be of clinical significance when compared to postoperative outcomes in this patient cohort versus those that were not able to have modifiable risk factors resolved prior to surgery. Future studies should aim at providing a high-powered study seeking a more definitive answer on this topic since the results thus far have been mixed, in addition to longer term follow-up and investigation into other complications (e.g., thromboembolic disease, stiffness, instability, and loosening).

Conclusions

This study is unable to validate the use of preoperative CRP/ESR as a predictive measure of future risk of PJI in primary TJA. However, it does provide quantitative insight into the prevalence of elevated preoperative inflammatory markers in all patients prior to TJA with a considerable proportion of patients having modifiable risk factors. Even then, there are still patients with elevated inflammatory markers and no known cause, which may be attributed to osteoarthritis itself. Because almost half of all patients studied were found to have elevated CRP/ESR, we cannot provide a firm recommendation in favor of or against the utility of obtaining these preoperative labs. Since there is such a large proportion of patients with elevated inflammatory markers that do not go on to develop PJI, we do not recommend cancellation of TJA unless there is an obvious identified modifiable risk factor that puts them at increased risk of PJI. Interestingly, we found that once we removed obesity as a modifiable risk factor, much of our data that was previously statistically significant became no longer significant. This shines important light onto the effect of obesity on elevation of inflammatory markers, especially CRP when ESR was negative. Lastly, we found that elevation in either CRP or ESR closely approached statistical significance for correlation with PJI, indicating perhaps some sensitivity for development of PJI. Our study does not refute the use of CRP/ESR in the setting of diagnosis and treatment of PJI after TJA, as this remains one of the gold standard biochemical markers.

Contributors

James-Michael Blackwell, MPH, ^­ Andrew Merwin, MD.

References

1. Ethgen O, Bruyère O, Richy F, et al. (2004) Health-related quality of life in total hip and total knee arthroplasty. J Bone Joint Surg Am 86: 963-974.
2. Singh JA, Vessely MB, Harmsen WS, et al. (2010) A population-based study of trends in the use of total hip and total knee arthroplasty, 1969-2008. Mayo Clin Proc 85: 898-904.
3. Zhang Y, Jordan JM (2010) Epidemiology of osteoarthritis. Clin Geriatr Med 26: 355-369.
4. Bumpass DB, Nunley RM (2012) Assessing the value of a total joint replacement. Curr Rev Musculoskelet Med 5: 274-282.
5. Sloan M, Premkumar A, Sheth NP (2018) Projected volume of primary total joint arthroplasty in the U.S., 2014 to 2030. J Bone Joint Surg Am 100: 1455-1460.
6. Godoy G, Sumarriva G, Ochsner JL, et al. (2016) Preoperative acute inflammatory markers as predictors for postoperative complications in primary total knee arthroplasty. Ochsner J 16: 481-485.
7. Kunutsor SK, Beswick AD, Whitehouse MR, et al. (2018) Debridement, antibiotics and implant retention for periprosthetic joint infections: A systematic review and meta-analysis of treatment outcomes. J Infect 77: 479-488.
8. Bozic KJ, Kurtz SM, Lau E, et al. (2009) The epidemiology of revision total hip arthroplasty in the united states. J Bone Joint Surg Am 91: 128-133.
9. Pitta M, Esposito CI, Li Z, et al. (2018) Failure after modern total knee arthroplasty: A prospective study of 18,065 knees. J Arthroplasty 33: 407-414.
10. AAOS (2020) Download the AJRR 2020 Annual report.
11. Kurtz SM, Lau EC, Son MS, et al. (2018) Are we winning or losing the battle with periprosthetic joint infection: Trends in periprosthetic joint infection and mortality risk for the medicare population. J Arthroplasty 33: 3238-3245.
12. Premkumar A, Kolin DA, Farley KX, et al. (2021) Projected economic burden of periprosthetic joint infection of the hip and knee in the United States. J Arthroplasty 36: 1484-1489.e3.
13. Porter ME (2010) What is value in health care? N Engl J Med 363: 2477-2481.
14. Porter ME, Larsson S, Lee TH (2016) Standardizing patient outcomes measurement. N Engl J Med 374: 504-506.
15. Volanakis J (2001) Human C-reactive protein: Expression, structure, and function. Mol Immunol 38: 189-197.
16. Pfitzner T, Krocker D, Perka C, et al. (2008) Das C-reaktive Protein. Der Orthopäde 37: 1116-1120.
17. Bernstein DN, Liu TC, Winegar AL, et al. (2018) Evaluation of a preoperative optimization protocol for primary hip and knee arthroplasty patients. J Arthroplasty 33: 3642-3648.
18. Dlott CC, Moore A, Nelson C, et al. (2020) Preoperative risk factor optimization lowers hospital length of stay and postoperative emergency department visits in primary total hip and knee arthroplasty patients. J Arthroplasty 35: 1508-1515.e2.
19. Battistelli S, Fortina M, Carta S, et al. (2014) Serum C-reactive protein and procalcitonin kinetics in patients undergoing elective total hip arthroplasty. Bio Med Research International 2014: 1-6.
20. Tsimikas S, Willerson JT, Ridker PM (2006) C-reactive protein and other emerging blood biomarkers to optimize risk stratification of vulnerable patients. J Am Coll Cardiol 47: C19-31.
21. Ferri C, Croce G, Cofini V, et al. (2007) C-reactive protein: Interaction with the vascular endothelium and possible role in human atherosclerosis. Curr Pharm Des 13: 1631-1645.
22. Bray C, Bell LN, Liang H, et al. (2016) Erythrocyte sedimentation rate and c-reactive protein measurements and their relevance in clinical medicine. WMJ 115: 317-321.
23. Ridker PM (2007) C-reactive protein and the prediction of cardiovascular events among those at intermediate risk. J Am Coll Cardiol 49: 2129-2138.
24. Della Valle C, Parvizi J, Bauer TW, et al. (2011) American academy of orthopaedic surgeons clinical practice guideline on. J Bone Joint Surg Am 93: 1355-1357.
25. Piper KE, Fernandez Sampedro M, Steckelberg KE, et al. (2010) C-reactive protein, erythrocyte sedimentation rate and orthopedic implant infection. Bereswill S, editor. PLoS One 5: e9358.
26. Parvizi J, Zmistowski B, Berbari EF, et al. (2011) New definition for periprosthetic joint infection: From the workgroup of the musculoskeletal infection society. Clin Orthop Relat Res 469: 2992-2994.
27. McArthur BA, Abdel MP, Taunton MJ, et al. (2015) Seronegative infections in hip and knee arthroplasty. Bone Joint J 97: 939-944.
28. Parvizi J, Tan TL, Goswami K, et al. (2018) The 2018 definition of periprosthetic hip and knee infection: an evidence-based and validated criteria. J Arthroplasty 33: 1309-1314.e2.
29. Xu C, Guo H, Qu P, et al. (2018) Preoperatively elevated serum inflammatory markers increase the risk of periprosthetic joint infection following total knee arthroplasty in patients with osteoarthritis. Ther Clin Risk Manag 14: 1719-1724.
30. Buchheit J, Uhring J, Sergent P, et al. (2014) Can preoperative CRP levels predict infections of bipolar hemiarthroplasty performed for femoral neck fracture? A retrospective, multicenter study. Eur J Orthop Surg Traumatol 25: 117-121.
31. Ghosh S, Paul S, Bhattacharjee DP, et al. (2009) Prognostic value of baseline high-sensitivity C-reactive protein in patients undergoing replacement arthroplasty. JNMA J Nepal Med Assoc 48: 144-148.
32. Ackland GL, Scollay JM, Parks RW, et al. (2007) Pre-operative high sensitivity C-reactive protein and postoperative outcome in patients undergoing elective orthopaedic surgery. Anaesthesia 62: 888-894.
33. Probasco WV, Cefalu C, Lee R, et al. (2020) Prevalence of idiopathically elevated ESR and CRP in patients undergoing primary total knee arthroplasty as a function of body mass index. J Clin Orthop Trauma 11: S722-S728.
34. Takahashi M, Naito K, Abe M, et al. (2004) Relationship between radiographic grading of osteoarthritis and the biochemical markers for arthritis in knee osteoarthritis. Arthritis Res Ther 6: 208-212.
35. Hanada M, Takahashi M, Furuhashi H, et al. (2016) Elevated erythrocyte sedimentation rate and high-sensitivity C-reactive protein in osteoarthritis of the knee: Relationship with clinical findings and radiographic severity. Ann Clin Biochem 53: 548-553.