Chapter 32
Adult Septic Arthritis

Ian T. Ferguson and Christopher R. Carpenter

Department of Emergency Medicine, Washington University School of Medicine, St. Louis, MO, USA

Highlights

The accuracy of history, physical exam, and labs differ for gonococcal and nongonococcal septic arthritis.
Similarly, native joint septic arthritis exhibits different patient demographics, serum, and synovial features than prosthetic joint septic arthritis.
For the diagnosis of a nongonococcal septic joint, the most useful findings on history and physical exam are cellulitis overlying a prosthetic joint (LR+ 15) and recent (<3 months) joint surgery (LR+ 7).
Peripheral WBC does not accurately distinguish nongonococcal native joint septic arthritis from other etiologies of monoarticular arthritis, but procalcitonin may rule in septic arthritis.
Synovial white blood cell counts >100,000 (LR+ 28) or >50,000 (LR+ 8) and Gram stain (LR+ 21) are the most useful tests to rule in the diagnosis of nongonococcal native joint septic arthritis.

Background

Acute monoarticular arthritis presenting to the emergency department (ED) can be due to a number of etiologies as summarized in Table 32.1.¹ Septic (bacterial) arthritis has an annual incidence of 10 per 100,000 individuals but is more common among those with rheumatoid arthritis or a prosthetic joint (up to 70 cases per 100,000).^2,3 Among patients in whom clinically concerning monoarticular arthritis is evaluated, the prevalence of septic arthritis across ED studies ranges from 7% to 27%.^4,5 Septic arthritis most commonly affects the knees, which account for 50% of cases. In decreasing order of frequency, septic arthritis also occurs in the hip, shoulder, and elbow. However, any articular surface can become infected.⁶ Most cases result from hematogenous spread since bacterial organisms can easily enter the synovial fluid through synovial tissue, which lacks a basement membrane. Although native joint surgery for septic arthritis need not occur emergently, prompt diagnosis, and appropriate antibiotic management are essential in the ED since cartilage can be destroyed within days and the in‐hospital mortality of treated infections can reach 15%.^7,8 The incidence of prosthetic joint infections is increasing.⁹ Important differences exist in the diagnosis and management of native joint versus prosthetic joint septic arthritis (Table 32.2).^10,11

Table 32.1 Differential diagnosis of monoarticular arthritis

Source: Data from [1].

Infection (bacterial, fungal, mycobacterial, viral, and spirochete)

Rheumatoid arthritis

Gout/pseudogout

Lupus

Lyme disease

Sickle cell disease

Dialysis‐related amyloidosis

Transient synovitis

Plant thorn synovitis

Metastatic carcinoma

Hemarthrosis

Pigmented villonodular synovitis

Neuropathic arthropathy

Intra‐articular injury

Staphylococcal and streptococcal species produce 70% of nongonococcal (GC) septic arthritis cases. These Gram‐positive species are isolated in both blood and synovial fluid in only 48% of cases.^6,12 Methicillin‐resistant Staphylococcus aureus septic arthritis is being increasingly reported, particularly in older adults and in those with intravenous drug abuse.^13–15 Gram‐negative septic arthritis more often involves the knee with pseudomonas species and Escherichia coli the most common organisms.¹⁶ Patients with human immunodeficiency virus (HIV) are not at increased risk for septic arthritis.¹⁷ Less common causes of septic arthritis include mycobacterium tuberculosis, brucella from unvaccinated livestock or unpasteurized milk, penetrating injuries from thorny plants, and Whipple disease.¹⁸

Table 32.2 Differences between native joint and prosthetic joint septic arthritis

Source: Data from [10].

Risk factor	Septic arthritis	Prosthetic joint infection
Incidence (%)	0.006–0.010	1–3
Common age (years)	<3 and >55	>65
Synovial WBC cutoff	>50,000 cells/μL	Acute >10,000 cells/μL
Synovial WBC cutoff	>50,000 cells/μL	Chronic >3,000 cells/μL
Management recommendation	Needle lavage, arthroscopic lavage, or open arthrotomy lavage	Open surgical debridement or implant removal and 1‐ or 2‐stage reimplantation
Antibiotic duration	6 weeks	Up to 3 months

Table 32.3 Risk factors for gonococcal arthritis

Source: Data from [19].

Female

Pregnancy

Menses

Multiple sexual partners

Low socioeconomic status

Intravenous drug abuse

Mucosal infection (asymptomatic)

HIV

Lupus

Complement deficiencies

Gonococcal organism factors

With an annual incidence of 3 per 100,000, GC arthritis is a less common cause of septic arthritis in the United States. Unlike non‐GC septic arthritis, GC joint infections respond rapidly to appropriate antimicrobial management. GC septic arthritis is frequently accompanied by a tenosynovitis and, in the absence of pelvic symptoms, can be difficult to distinguish from non‐GC septic arthritis. GC septic arthritis is more common in younger adults, with a 4:1 female‐to‐male predominance. Risk factors for GC arthritis are listed in Table 32.3. Gram stain (<50%), blood (<50%), and synovial cultures (∼50%) are less often positive in GC septic arthritis, but genitourinary specimens will be positive in up about 90% of cases.¹⁹

Table 32.4 Historical risk factors for nongonococcal septic arthritis

Source: Data from [5].

Risk factor	Positive likelihood ratio	Negative likelihood ratio
Age > 80 years	3.5 (1.7–6.4)	0.86 (0.70–0.96)
Diabetes mellitus	2.7 (1.1–6.2)	0.93 (0.79–1.0)
Rheumatoid arthritis	2.5 (1.9–2.9)	0.45 (0.27–0.67)
Recent joint surgery	6.9 (3.7–11.6)	0.78 (0.63–0.90)
Prior same joint surgery	2.3 (1.0–5.1)	0.55 (0.18–1.6)
Prior septic arthritis	0 (0–5.9)	1.1 (0.55–1.4)
Hip or knee prosthesis	3.1 (1.9–4.5)	0.73 (0.55–0.88)
Skin infection	2.8 (1.7–4.2)	0.76 (0.58–0.91)
Prosthesis and skin infection	15.0 (8.0–26)	0.77 (0.62–0.88)
HIV infection	1.7 (0.76–1.5)	0.64 (0.23–1.37)

Clinical question

Can history or physical exam accurately and reliably distinguish non‐GC septic arthritis from other causes of monoarticular symptoms?

History and physical exam alone are inadequate to rule in or rule out septic arthritis.⁵ The prevalence of non‐GC septic arthritis in ED patients with concerning monoarticular symptoms ranges from 7% to 27%.^4,20,21 Table 32.4 summarizes septic arthritis risk factors from the medical history, along with estimates of their diagnostic accuracy.^1,3,5 Notably, none of these findings, when absent, significantly reduced the probability of septic arthritis. A single‐center prospective study found chills to have the highest LR+ among all clinical features evaluated (2.2, CI 1.2–7.3) although given a pretest probability of 27%, this would only change the posttest probability to 45%.²¹ Objective fever has a reported LR+/LR− of 1.13/0.88 in one study and 0.67/1.7 in another, making it unhelpful to rule in or out the diagnosis of septic arthritis.^21,22

Table 32.5 summarizes physical exam accuracy from one small single‐center prospective study.⁴ Global clinician intuition prior to lab testing most accurately rules‐in septic arthritis, but the confidence intervals from this study are too wide to be convincing. Interrater reliability for physical exam findings was good for erythema (κ = 0.83), fair for palpation pain (κ = 0.41), and poor for palpable effusion (κ = 0).^4,23

Clinical question

Can serum labs accurately and reliably distinguish septic arthritis from other causes of monoarticular symptoms?

White blood cell (WBC) over 10,000 cells is not useful acutely in ruling in or ruling out the diagnosis of septic arthritis (Table 32.6).⁵ On the other hand, some studies demonstrate a role for erythrocyte sedimentation rate (ESR) and C‐reactive protein (CRP) in adults with monoarticular arthritis. Li et al.^24,25 assessed only cases of confirmed septic arthritis and demonstrated a sensitivity of 96% for ESR > 30 mm/hour. Given the possibility of selection bias, this sensitivity may be falsely elevated. In contrast, Li et al.²⁶ demonstrated a ESR sensitivity of 75% and a specificity of 11% using the same definition of abnormal and Carpenter et al.⁴ demonstrated a sensitivity of 75% with specificity of 78% for ESR > 49 mm/hour. Hariharan²⁷