Article, Rheumatology

Utility of point-of-care synovial lactate to identify septic arthritis in the emergency department

a b s t r a c t

Background: Synovial lactate is a promising biomarker to distinguish septic from aSeptic arthritis. If available as a point-of care test, synovial lactate would be rapidly available to aid the emergency provider in clinical decision making. This study assesses the test characteristics of synovial lactate obtained using an EPOC(C) point-of-care (POC) analyzer to rapidly distinguish septic from aseptic arthritis in the emergency department.

Methods: We enrolled a convenience sample of patients with possible septic arthritis presenting to the emer- gency department at a large urban academic center between October 2016 and April 2018. Enrolled patients underwent arthrocentesis based on the clinical judgment of the treating provider. We obtained synovial lactate levels (SLL) from the POC device. Standard Laboratory analysis, synovial fluid culture, emergency and hospital course, operative procedures, antibiotics, and discharge diagnosis were abstracted from the electronic medical record.

Results: Thirty-nine patients undergoing forty separate arthrocentesis procedures were enrolled in this study over the two-year period. The sensitivity and specificity of SLL >= 5 mmol/L was 0.55 and 0.76 respectively, with +LR 2.3 and -LR 0.6. The sensitivity and specificity of SLL >= 10 mmol/L was 0.27 and 0.97 respectively, with +LR 7.9 and -LR 0.8; SLL >= 10 mmol/L performed similarly to overall synovial WBC >= 50,000/uL by conven- tional laboratory testing.

Conclusion: It is feasible to obtain a synovial lactate level using the EPOC(C) POC device. In our study, POC SLL per- forms similarly to other markers used to diagnose septic arthritis. Further study with larger sample sizes is warranted.

(C) 2018

Introduction

Septic arthritis remains a relatively rare diagnosis, with roughly 16,000 cases seen in the Emergency Department (ED) per year, or 0.01% of annual ED visits in the United States [1]. It is difficult to distin- guish septic from aseptic arthritis, since both frequently present with a similar clinical picture. Prior research suggests that physicians are able to accurately diagnose the etiology of acute non-traumatic monoarticular arthritis at approximately 3 days [2]. ED physicians do not have the luxury of this timeline, and inadequate or delayed treat- ment of septic arthritis can lead not only to increased morbidity from Joint destruction in up to 50% of cases [3], but also to mortality in N10% of cases [4].

* Corresponding author at: Department of Emergency Medicine, UCSF Fresno Center for Medical Education and Research, 155 N Fresno St, Fresno, CA 93701, United States of America.

E-mail addresses: [email protected] (L. Farshidpour), [email protected] (M. Young), [email protected] (M. Darracq), [email protected] (C. Ives Tallman).

Classic signs and symptoms of septic arthritis, such as non-traumatic acute monoarticular joint pain, swelling, erythema, or micromotion ten- derness to palpation, are not sufficiently sensitive nor specific for the di- agnosis of septic arthritis [5]. Adding to the complexity of making an accurate diagnosis, there is not a clear diagnostic gold standard [4,6,7], with case definitions based on one of four criteria: (1) isolation of path- ogenic organism from affected joint, (2) isolation of pathogenic organ- ism from another source (i.e. blood) in the context of a suspicion of joint infection, (3) typical clinical features and turbid joint fluid in joint previously treated with antibiotics, or (4) postmortem features suspicious of septic arthritis [4]. Because synovial fluid culture is only 75-95% sensitive for septic arthritis [8], and time to culture limits time- liness of diagnosis, accurate and fast diagnostic laboratory markers are necessary to identify septic arthritis.

Though Laboratory markers such as C-reactive protein (CRP) and pe- ripheral white blood cell count (pWBC) have been used to distinguish between the two types of arthritis, several studies have shown that these markers do not differ significantly between septic and aseptic joints [9,10]. The classically taught synovial WBC (sWBC) >= 50,000/uL also does not adequately distinguish between these two disease pro- cesses [4,11]. Other biomarkers, such as laboratory-measured synovial

https://doi.org/10.1016/j.ajem.2018.12.030 0735-6757/(C) 2018

E. Shu et al. / American Journal of Emergency Medicine 37 (2019) 502505 503

glucose and Uric acid were also considered but found to be inferior in ac- curacy to synovial lactate level (SLL) [9]. A few studies have been done on the potential of laboratory-measured SLL to serve as a diagnostic marker to identify bacterial etiologies of arthritis, but the overall quality of evidence is relatively low with the majority of studies being retro- spective [2,9].

A limitation in the potential usefulness of SLL in the emergency de- partment is the ability of the clinical laboratory to process the synovial sample and obtain rapid lactic acid values for the clinician’s use. Many emergency departments, including the authors’, may lack the ability to quickly obtain a SLL. In such a setting, a point of care test is clearly desir- able. Given how promising laboratory-measured SLL is in the rapid identification of septic joints, this study was designed to assess the use- fulness of EPOC(C) point of care (bedside) lactate values as a tool to rap- idly distinguish between septic and aseptic joints.

Methods

We prospectively enrolled patients undergoing arthrocentesis for the evaluation of a swollen or painful joint in the ED at the University of California San Francisco Fresno campus, a large urban academic cen- ter, between October 2016-April 2018. Arthrocentesis was performed at the discretion of treating providers according to standard practices. Synovial fluid was sent to the main laboratory for analysis and desired fluid studies were ordered by the treating physicians. The study proto- col did not affect whether or not joint aspiration occurred, nor did it dic- tate which laboratory studies each treatment provider ordered. If the provider completing the aspiration was able to obtain at least 1 mL of additional synovial fluid in a separate sterile syringe, the patient was able to be enrolled in our study. The study sample was then analyzed promptly using the point of care EPOC(C) blood gas analysis system by Alere using the standard test card. Point of care (POC) SLL was recorded, in addition to joint location, fluid characteristics, and patient identifiers. From the electronic medical record, investigators then abstracted data including sWBC and cell count with differential, synovial Gram stain and culture. Investigators also reviewed the patient’s ED and hos- pital course, any operative procedures done, antibiotics given, and final discharge diagnosis. Since synovial fluid culture itself is an imperfect di- agnostic measure [8], we defined septic arthritis as patients with (1) sy- novial fluid culture positive or (2) septic arthritis diagnosed by our orthopedists with surgical intervention and IV antibiotics given during the hospital stay, even if cultures were ultimately negative. We also

included in our final analysis a small number of patients who had fluid samples from septic bursitis and were treated with antibiotics.

Statistical analysis was done using Excel(R) (Microsoft Office 2016) and SPSS(C) (IBM, version 24). Sensitivity and specificity for each diag- nostic test was analyzed using the chi-squared analysis, Fisher’s exact test, and linear Regression modeling. ROC curves were calculated using both parametric and non-parametric analyses. The C-statistic (area under the ROC curve) for each test was obtained using the parametric analysis. The positive and negative likelihood ratio for each test was cal- culated using the chi-squared table, and values were confirmed with the non-parametric ROC analysis.

Results

During the study period, 39 patients were enrolled with 40 arthrocenteses performed (one patient had the procedure on bilateral knees). Eleven (28%) of the 39 patients were female. Mean patient age was 51 years, ranging from 16 to 85 years. There was no significant dif- ference in diagnosis of septic joint by gender (p = 0.82) or age (p = 0.60).

Of the 40 separate joints sampled, there were 29 knees (73%), 6 el- bows (15%), 3 ankles (8%), 1 wrist (2%) and 1 shoulder (2%). Eleven (27.5%) joints were ultimately diagnosed as septic arthritis or infected bursitis, with 6 (15%) undergoing operative intervention [Table 1]. The diagnoses for patients without septic joints were simple effusion (n = 15, 37%), crystal arthropathy including gout or pseudogout (n = 9, 22%), and inflammatory arthritis (n = 6, 15%).

The sensitivity and specificity of a SLL >= 5 mmol/L was 55% [95% CI

32-94%] and 76% [95% CI 62-93%] respectively, with LR+ 2.3 and LR-

0.6. The sensitivity and specificity of a SLL >= 10 mmol/L was 27% [95% CI 10-72%] and 97% [95% CI 79-99%] respectively, with LR+ 7.9 and LR

– 0.8. The sensitivity of the commonly used physical examination find- ings and laboratory tests to evaluate septic joints was poor in our study [Table 2]. The specificity of sWBC >=50,000/uL, sWBC >=100,000/uL, and WBC differential of polymorphonuclear neutrophils >=90% all performed well in ruling in septic arthritis, though they were all poorly sensitive in ruling it out.

The C-statistic for SLL >= 10 mmol/L was 0.69 [Fig. 1]. (A C-statistic of

0.5 represents a tool no better than chance, N0.7 a good model, N0.8 a strong model, and 1.0 is a model with perfect prediction.) Similar or poorer C-statistic values were obtained for all of the commonly used

Table 1

Patients diagnosed with septic arthritis

Synovial culture Operative

Patient number

Age (years)

Gender

Joint

Synovial lactate (mmol/L)

Synovial WBC (leukocytes/ml)

PMNs (%)

Gram stain

1

63

M

Knee

1.27

750

79

Neg

2

57

M

Knee

16.87

40,000

84

Pos

3

16

F

Knee

1.17

5622

86

Neg

4

53

M

Ankle

5.76

100,800

86

Neg

5

52

M

Elbow

4.41

1650

87

Neg

6

58

M

Knee

7.28

1935

88

Neg

7

33

M

Knee

4.25

15,390

93

Neg

8

55

F

Elbow

4.74

80,750

94

Neg

9

55

M

Elbow

11.79

133,750

95

Neg

10

85

F

Shoulder

11.61

18,117

96

Neg

11

64

M

Knee

9.27

47,520

97

Pos

intervention

Blood culture Diagnosis

Staphylococcus aureus

No Micrococcus Species

Septic joint

Staphylococcus aureus

Yes No Growth Septic joint

No growth Yes No Growth Septic joint and

osteomyelitis

No growth Yes N/A Septic joint and gout

Staphylococcus aureus Staphylococcus aureus

No N/A Infected bursitis

Yes No Growth Septic joint with cellulitis

aureus

Staphylococcus No

aureus

Staphylococcus

aureus

Multiple septic joints,

bacteremia

No growth Yes

N/A

Septic joint and gout

Staphylococcus No

N/A

Infected bursitis

Escherichia coli Yes

No Growth

Septic joint

Group B No

N/A

Infected bursitis

Streptococcus

504 E. Shu et al. / American Journal of Emergency Medicine 37 (2019) 502505

Table 2

Test Characteristics of synovial lactate and fluid characteristics used in the diagnosis of septic joints

Test

p value

Sensitivity

Specificity

C-Statistic

LR+

LR-

Synovial lactate >=5

0.075

0.55 [0.32, 0.94]

0.76 [0.62, 0.93]

0.68

2.3

0.6

Synovial lactate >=10

0.056

0.27 [0.10, 0.72]

0.97 [0.79, 0.99]

0.69

7.9

0.8

Cloudy synovial fluid

0.578

0.45 [0.24, 0.87]

0.52 [0.36, 0.74]

0.48

0.9

1.1

Warmth or erythema

0.183

0.64 [0.40, 0.99]

0.59 [0.43, 0.80]

0.62

1.5

0.6

WBC >= 50 K

0.056

0.27 [0.10, 0.71]

0.97 [0.79, 0.99]

0.69

7.9

0.8

WBC >= 100 K

0.071

0.18 [0.05, 0.64]

1

n/a

10

0.8

PMN >= 90%

0.467

0.45 [0.24, 0.87]

0.62 [0.47, 0.82]

0.54

1.2

0.9

Positive gram stain

0.071

0.18 [0.05, 0.64]

1

n/a

10

0.8

Micromotion tenderness

0.196

0.27 [0.10, 0.72]

0.9 [0.72, 0.97]

0.62

2.6

0.8

tests and Physical exam findings used to distinguish septic from non- septic joints [Table 2].

Limitations

The EPOC(C) point of care analyzer and cartridges were not developed for use in analyzing synovial fluid, nor are they approved for this indica- tion. Furthermore, we were unable to compare POC SLL to a laboratory value as our hospital does not process these samples. We did not mea- sure time from sample collection to analysis but did request that sam- ples be analyzed as soon as possible after collection. This was a small, single-center study, and our data did not reach statistical significance, likely secondary to the small sample size and low incidence of disease within our population. Given our small sample size, our sensitivity and specificity data have wide confidence intervals. A large multi- center, prospective study would better elucidate whether POC SLL pro- vides adequate sensitivity and specificity in the diagnosis of septic arthritis.

Discussion

In our present study, POC SLL >= 10 mmol/L performed similarly with the classic diagnostic marker of sWBC >=50,000/uL with poor sensitivity at 27% and near perfect specificity at 97%. Our results reflect similar sen- sitivity and specificity as prior studies measuring sWBC for the diagnosis of septic arthritis, with wide ranges for both measures (sensitivity: 31-70% with cutoff at 50,000/uL and 6-31% at 100,000/uL; specificity 74-97% with cutoff at 50,000/uL and 94-100% at 100,000/uL) [2,7,11,12]. Though POC SLL and sWBC had similar test characteristics in our small study, the advantage of bedside POC SLL is a shorter time to diagnosis. The additional cost of the test was also not burdensome,

Fig. 1. ROC curve for synovial lactate level (SLL) >= 10 mmol/L. C-Statistic for Lactate >=10 = 0.689704.

at approximately $3.40. This is in comparison to a cost of approximately

$6.80 for the formal laboratory cell count and differential [13].

Of the eleven patients ultimately diagnosed with septic joints, two had positive blood cultures, though only 54% of these patients had blood cul- tures drawn at all. The same organism grew from the blood and synovial fluid in only one patient. Since one of our metrics for the diagnosis of sep- tic arthritis or bursitis was synovial fluid culture growth, there may be some concern that improper cleansing of the skin led to the contamina- tion of the sample with methicillin-resistant Staphylococcus aureus (MRSA), staph aureus or another cutaneous bacteria and therefore a false positive. In our study population of 40 joints, only one patient ulti- mately grew MRSA as part of their synovial fluid culture; an additional 5 patients had Staphylococcus aureus cultured from synovial fluid. How- ever, all of these patients had additional findings to suggest a septic joint independent of the synovial fluid culture result alone. Future studies with larger sample size will need to address this clinical question.

The measurement of SLL is not a new idea and has been previously shown to correlate well with response to Antimicrobial treatment [14- 16], however, explicit guidelines and diagnostic cut off values still have not been clearly established. Mean SLL was 15.1 (+-5.7 SD) mmol/L among non-gonococcal septic arthritis joints in one study and reportedly markedly elevated compared to gonococcal septic arthritis and non-septic arthritides [16]. Another paper reported mean SLL of

24.4 mmol/L for culture-positive and 17.3 mmol/L for culture-negative septic arthritis, whereas mean SLL for rheumatoid arthritis, crystal ar- thropathy, and osteoarthritis were 5.9, 2.7, and 1.8 mmol/L, respectively [15]. Neither of these early groundbreaking studies reported sensitivity, specificity, or positive/negative likelihood ratios.

In the few available studies since the 1980s, laboratory-measured SLL

has shown to be nearly 100% specific ata threshold of SLL >= 10 mmol/L [2,9]. The only study with sensitivity, specificity, and likelihood ratios for SLL analysis that we identified showed mean SLL 11.7 and 3.5 mmol/L in septic and gouty arthritis, respectively (p = 0.0003) [9]. This same 2014 retrospective study found that a cutoff threshold to maximize sensi- tivity (89.5%) and specificity (77.3%) of SLL was >=4.3 mmol/L, with AUC 0.901 [9]. Our study with bedside POC SLL testing did not perform as well as the 2014 paper which utilized main Laboratory measurements, but this may be secondary to a very small sample size.

To our knowledge, this is the first study to prospectively assess bed- side POC SLL in the ED diagnosis of septic arthritis. Our study was a proof-of-concept study with promising results. We were able to quickly obtain reasonable SLL values from the POC device. In our study, the test characteristics of POC SLL N 10 mmol/L were similar to those of sWBC N 50,000. However, due to the small sample size of our study there is not enough evidence at this time to conclusively state that POC SLL is bet- ter than classic laboratory testing and more research is necessary prior to recommending its routine usage in the diagnosis of septic arthritis.

Acknowledgments

The EPOC point of care analyzer and test cartridges were purchased with a faculty research grant from Central California Faculty medical group (CCFMG)/University of California San Francisco (UCSF).

E. Shu et al. / American Journal of Emergency Medicine 37 (2019) 502505 505

References

  1. Singh JA, Yu S. Septic arthritis in emergency departments in the US: a national study of health care utilization and time trends. Am Coll Rheumatol 2018;70(2):320-6.
  2. Carpenter CR, Schuur JD, Everett WW, Pines JM. Evidence-based diagnostics: adult septic arthritis. Acad Emerg Med 2011;18(8):781-96.
  3. Nade S. Septic arthritis. Best Pract Res Clin Rheumatol 2003;17(2):183-200.
  4. Matthews CJ, Weston VC, Jones A, Field M, Coakley G. Bacterial septic arthritis in adults. Lancet 2010;375:840-55.
  5. Margaretten ME, Kohlwes J, Moore D, Bent S. Does this adult patient have septic ar- thritis? JAMA 2007;297(13):1478-88.
  6. Newman JH. Review of septic arthritis throughout the antibiotic era. Ann Rheum Dis

    1976;35:198-205.

    Raukar NP, Zink BJ. Chapter 136: bone and joint infections. In: Marx A J, Hockberger RS, editors. Rosen’s Emergency Medicine: Concepts and Clinical Practice. 8th edition. Walls RMElsevier; 2013.

  7. Schmerling RH, Delbanco TL, Tosteson AN, Trentham DE. Synovial fluid tests: what should be ordered? JAMA 1990;264:1009-14.
  8. Lenski M, Scherer MA. Analysis of synovial inflammatory markers to differ infectious from gouty arthritis. Clin Biochem 2014;47(1-2):49-55.
  9. Soderquist B, Jones I, Fredlund H, Vikerfors T. Bacterial or crystal-associated arthri- tis? Discriminating ability of serum inflammatory markers. Scand J Infect Dis 1998;30(6):591-6.
  10. McGillicuddy DC, Shah KH, Friedberg RP, Nathanson LA, Edlow JA. How sensitive is the synovial fluid white blood cell count in diagnosing septic arthritis? Am J Emerg Med 2007;25:749-52.
  11. Ross JJ. Septic arthritis of native joints. Infect Dis Clin N Am 2017;31:204-18.
  12. Clinical Diagnostic Laboratory Fee Schedule. https://www.cms.gov; 2018.
  13. Brook I, Reza MJ, Bricknell KS, Bluestone R, Finegold SM. Synovial fluid lactic acid: a diagnostic aid in septic arthritis. Arthritis Rheum 1978;21:774-9.
  14. Riordan T, Doyle D, Tabaqchali. Synovial fluid lactic acid measurement in the diagno- sis and management of septic arthritis. J Clin Pathol 1982;35:390-4.
  15. Gobelet C, Gerster JC. Synovial fluid lactate levels in septic and non-septic arthritides. Ann Rheum Dis 1984;43:742-5.