A Pain in the Foot: Delayed Diagnosis of Primary Septic Arthritis of Naviculocuneiform and Second/Third Tarsometatarsal Joints

Background

Acute painful joints are frequent presenting symptoms in the Emergency Department. The differential diagnoses for monoarticular arthritis are diverse, and it can be challenging for physicians to immediately identify the cause. Septic arthritis (SA) occurs when an infectious agent affects the synovium. The routes of infection include hematogenous spread, direct inoculation, and extension from an infected contiguous site [1–3]. The condition needs to be recognized early because delays in the treatment can lead to significant morbidity and mortality. Gout is clinically suspected in middle-aged men presenting with acute monoarthritis involving the foot or ankle [4], while osteoarthritis (OA) of the foot is a significant cause of foot pain among older adults [5,6]. We present a case of a middle-aged man who was treated for acute gout and OA. He presented multiple times to outpatient clinics and Emergency Departments (EDs) with worsening foot pain, and the lack of specific physical findings and laboratory and radiological results led to delays in diagnosis and appropriate management.

Case Report

A middle-aged man with a past medical history of hypertension and hyperlipidemia presented to our ED owing to severe left foot pain. He had developed sudden-onset, atraumatic mild pain in his left ankle that was worse with movement and ambulation. After 3 days, the pain localized to his left foot. The pain was associated with swelling and redness, which prompted the patient to consult his family physician. There was no fever or other associated symptoms. A provisional diagnosis of gout was made because of the patient’s demographic profile and overall clinical presentation. He was prescribed nonsteroidal anti-inflammatory drugs (NSAIDs) and colchicine, which did not improve his symptoms. After 4 more days, the patient consulted an orthopedic specialist. Magnetic resonance imaging (MRI) of the ankle was done, which showed medial talar and midfoot osteoarthritic changes (Figure 1). Analgesics and observation were again advised. Two weeks from the onset of his symptoms, the patient presented to the ED reporting severe, constant left foot pain that affected his sleep and prevented him from bearing weight on that foot. On examination, his vital signs were normal. The left foot was noted to have generalized swelling, with erythema and significant tenderness over the midfoot. Examination of the ankle was normal, and the plain radiograph of the foot was un-remarkable. The clinical impression was OA, and the patient was discharged with NSAIDs. Two days after, he consulted the orthopedic specialist again, and investigations revealed C-reactive protein (CRP) of 60.4 mg/L and a white blood cell count (WBC) of 7.97×109/L without neutrophilia. The erythrocyte sedimentation rate (ESR), uric acid, and rheumatoid factor results were normal. A repeat radiograph of the left foot showed mild degenerative changes at the first metatarsophalangeal (MTP) joint (Figure 2). The clinical impression was OA, and the patient was again discharged with analgesics. After 5 more days, he re-presented to our ED owing to persistent severe pain despite medication compliance and was admitted to the Orthopedic Department.

Upon admission, further investigations were performed, which showed WBC 12.92×109/L, CRP 136 mg/L, and ESR 62 mm/h. Procalcitonin (PCT), antinuclear antibody, anti-double-stranded DNA antibody, and cyclic citrullinated peptide antibody findings were all normal. No predisposing factors were identified in the patient. His hemoglobin A1c was 5.7%, and the HIV screen was nonreactive. His blood chemistry showed normal renal and liver function. He did not have constitutional symptoms such as weight loss, night sweats, and anorexia. He was a healthy man with no history of recurrent infections and frequent admissions.

Rheumatology was consulted to rule out crystal arthropathy. Dual-energy computed tomography (CT) of the left foot showed multiple periarticular erosions in the tarsal bones and metatarsal bases with adjacent soft tissue swelling, suspicious for an inflammatory etiology (Figure 3). There was no evidence of monosodium urate (MSU) crystal deposition and soft tissue calcification. MRI of the foot revealed extensive joint effusion and synovitis involving the second to fifth tarsometatarsal (TMT) joints and intercuneiform joints, with associated extensive periarticular marrow edema and erosions (Figure 4). The patient underwent drainage, washout, debridement, and biopsy of his left second/third TMT joints on day 3 of admission. Intraoperative findings revealed purulent fluid and clumps of debris within the joint. Tissue culture was positive for Staphylococcus aureus sensitive to cefazolin, and he was advised to complete a 6-week course of this antibiotic. On day 5 of admission, the patient underwent left foot wound debridement and washout of the second/third TMT joint and the medial naviculocuneiform (NC) joint. Intraoperative findings revealed pus within the second TMT joint and medial NC joint and unhealthy synovium and connective tissue. A wound drain was inserted and removed after few days. The patient was eventually transferred to a community hospital, where he completed his intravenous antibiotics and continued his rehabilitation for a month. His inflammatory markers were all normal on discharge. He remained well but returned to the ED after 2 weeks for recurrence of persistent left foot pain. He was admitted again and repeat investigations were done, revealing CRP 54.3 mg/L (baseline 2), WBC 11×109/L (baseline 5), and ESR 12 mm/h (baseline 8). A repeat plain radiograph of the foot showed worsening cortical irregularity and lucencies in the anterosuperior aspect of the navicular bone and superior cuneiform (Figure 5). MRI was also repeated and showed interval progression of extensive periarticular marrow edema involving all tarsal bones and TMT joints, with increased bony erosions, especially along the dorsal aspect of the tarsal bones. Other findings included the presence of joint effusion and synovitis of the midfoot and the second to fifth TMT joints, with increased surrounding reactive sub-cutaneous soft tissue, plantar, and interosseous muscular enhancement (Figure 6). An infectious disease specialist was consulted and advised to continuation of intravenous cefazolin for another 6 weeks. Inflammatory markers were regularly monitored and were noted to be in a downward trend after 2 days. The patient was subsequently transferred to a community hospital to complete his intravenous antibiotics. He had an uneventful recovery and was discharged after few weeks. At present, he continues to have intermittent foot pain, and he is still on active follow-up with an orthopedic specialist and the Rehabilitation Department.

Discussion

Early recognition of SA is important to prevent significant morbidity and mortality in affected patients. The reported incidence varies between studies and ranges between 2 and 60 cases per 100 000 person-years [1,7–9]. Delays in diagnosis and treatment lead to irreversible joint destruction in up to 25–50% of cases, and functional disability has been noted in 25–75% of the survivors [9,10]. Mortality rates of 3–40% have been described [1,7–8]. Patients most commonly present with acute joint pain, swelling, erythema, pain on movement, and decreased range of motion [3,8]. Reported affected areas are often the knee, hip, shoulder, and elbow [3,7,8]. Fever appears to be present in only half of the cases [8,9]. Based on the literature, the most important predisposing factor in native joints is a pre-existing joint disease, and almost half of the patients diagnosed with SA have previous joint conditions [1,3,8]. Immunocompromised states such as diabetes mellitus, cirrhosis, intravenous drug use, end-stage renal disease, and immunosuppressant use are also known risk factors [3,7]. Interestingly, another study showed that 22% of patients diagnosed with SA had no known risk factors [1,3].

SA of the foot is very uncommon, and it comprises only 3–7% of all SA cases reported [11]. Tibiotalar and first MTP joints are typically affected. A case report of primary SA involving the talonavicular joint in a healthy adult was previously published [11]. Our case is unusual because the main joints involved were the intercuneiform and NC joints. To the best of our knowledge, there is no published case report of primary SA involving these joints in an immunocompetent adult. Although the first MRI showed osteoarthritic changes in the midfoot, the patient denied having similar symptoms in the past and he was never diagnosed with OA.

The criterion standard for diagnosing SA is culture-positive synovial fluid or pus aspiration from the joint [1,2,12]. However, synovial gram stain sensitivity has been reported to be as low as 29% [1]. Synovial culture sensitivities vary from 40% to 60% and results require a few days for return. Hence, these tests may not be very helpful in the ED [2]. Serum inflammatory markers such as total WBC, CRP, and ESR are routinely tested in most cases of arthritis, but unfortunately, studies show that these are not reliable in distinguishing infectious from noninfectious causes [3,7]. Recently, PCT has gained much attention as a biomarker for bacterial processes. A study by Maharajan et al [10] shows that PCT levels of more than 0.4 ng/mL represent a sensitive (85.2%) and specific (87.3%) marker for SA and osteomyelitis. A meta-analysis by Zhao et al [13] involving 10 studies demonstrated PCT has a sensitivity of 54% and a specificity of 95%, a positive likelihood ratio (LR) of 10.7, and a negative LR of 0.45. They concluded that PCT was a better diagnostic marker than CRP in differentiating SA from non-SA. Of note, our patient had a normal PCT level, which further complicated the diagnosis of SA and makes this case even more intriguing. It must be remembered that PCT remains low if the cause of SA is local, rather than systemic [7].

Plain radiographs are often performed and considered essential in most cases of arthritis, yet results are usually normal in the early stages [12,14,15]. In addition, they are unable to distinguish SA from other inflammatory conditions. Hence, radiography is not useful in diagnosing SA [2].

CT and MRI generally provide a better joint assessment. While CT offers better visualization of bony details [14], MRI is more sensitive and is the preferred diagnostic imaging procedure for suspected bone and soft tissue infections [12,14–16]. The earliest signs of SA in MRI include synovial inflammation and effusion, while marrow signal abnormalities and bony erosions develop later [8,12,15]. A study by Karchevsky et al [17] showed that among 50 patients with SA, synovial enhancement and perisynovial edema were seen consistently, while joint effusion was observed in only 70% of the patients. In addition, effusion was not evident in half of the small joints; hence, the absence of effusion should not rule out the presence of infection. In our case, the small size of the TMT and NC joints might explain the negative findings in the first MRI.

Furthermore, these findings are also nonspecific and can be seen in noninfectious arthritis [3,12,14,17]. Although abnormalities can be seen as early as 2 days of symptom onset, MRI should be repeated if the clinical suspicion of infection remains high. A recent review by Lim et al [16] explained that MRI findings may lag behind symptoms by 2 weeks. In our case, the initial MRI findings performed about a week after the onset of symptoms were nondiagnostic, while the repeat scan revealed significant joint effusions, bone marrow edema, and periarticular erosions.

Clinical judgment appears to be the most reliable factor in differentiating SA from non-SA joints [3]. In this scenario, the main differential diagnoses are acute gout and OA. An acute gout episode commonly presents with rapid onset monoarticular arthritis, classically involving the first MTP joint. The symptoms are maximal within a day, last for about a week, and in most cases are self-limiting [4,18]. Its prevalence is higher in men and often associated with other comorbidities [4,19,20]. A retrospective study done locally by Koh et al [20] showed that in Singapore, gout mainly affects middle-aged men with chronic conditions such as hypertension, hyperlipidemia, and ischemic heart disease. Experts advise that gout should be considered in the diagnosis of any acute arthritis presenting with monoarticular foot or ankle involvement in an adult who is male and has cardiovascular risk factors [4]. Given the patient’s demographic profile and overall clinical picture, gout was a reasonable primary impression by the family physician. The patient has hypertension and hyperlipidemia, is a middle-aged man, and presented with acute onset monoarthritis. It should be emphasized that although these features are highly suggestive, they are not specific for gout [4]. During hospital admission, dual-energy CT was performed to aid in identifying MSU crystals. This recently developed technology uses 2 different energy rays to demonstrate 2 different color-coded lesions. Urate deposits appear green, while calcium appears purple. It has high sensitivity and high specificity, making it a good screening and diagnostic tool in gout [21]. As shown in Figure 5, there was an absence of green lesions, ruling out gout. In addition, the patient’s clinical course was not consistent with gout. His pain was worsening despite appropriate treatment for weeks. Most treatment guidelines recommend early use of colchicine and NSAIDs [22], which the patient received. This should have prompted the attending physicians to consider other causes earlier.

Because of the first MRI and plain radiograph results, the subsequent physicians were likely led to believe that the diagnosis was OA. It is the most common arthritis affecting adults [23–25], yet few studies discuss the involvement of the foot [5,6]. Based on a review by Roddy and Menz [5], the most commonly affected joint in foot OA is the first MTP, followed by the second cuneometatarsal. Symptomatic midfoot OA is known to be present in only about 12% of people aged more than 50 years old and is more common among elderly women [6]. Given the patient’s demographic profile, it is doubtful that he had pre-existing mid-foot OA. Although the first MRI showed degenerative changes in the midfoot, the patient denied having similar symptoms in the past and he was never diagnosed with OA. MRI features suggestive of OA typically include the presence of joint space narrowing, osteophytes, effusion/synovitis, bony erosions, and bone marrow lesions [26]. While there is no finding specific for OA, Halstead et al [26] proposed a scoring system using these MRI features to evaluate foot OA. They reported excellent inter-reader reliability scores for joint space narrowing, osteophytes, and cysts, but lower scores for bony erosions, bone marrow lesions, and effusion. This highlights the difficulty in interpreting MRI of the foot because of its inherent anatomical complexity [14,26].

The question remains as to whether this case was primary osteomyelitis that was missed in the first MRI and subsequently complicated by secondary arthritis. We speculate that this is unlikely because MRI is highly sensitive and can identify bone marrow edema as early as 1–2 days after the onset of infection [16]. Negative findings can essentially rule out the presence of osteomyelitis [15]. However, we recognize that a delay in radiologic findings is still a small possibility. In this regard, nuclear medicine could have been considered as an adjunct diagnostic test. While other imaging demonstrates anatomical changes, nuclear imaging shows the physiologic state and bone metabolism. As early as 2–3 days from the onset of infection, a triple-phase bone scan can show the presence of increased focal uptake indicating osteomyelitis [15,27]. In this case, if the patient had primary osteomyelitis, nuclear imaging would have identified an early reactive or infectious activity. Nevertheless, findings are also not specific for infection and it is more useful in ruling out osteomyelitis when results are negative [15,27]. If results are unclear, a gallium scan can be utilized because it offers higher specificity [28]. Gallium 67 citrate is a traditional radiotracer that easily associates with transferrin. In areas of infection, blood flow and vascular permeability are increased, which leads to higher gallium deposition. Its application has been replaced by labeled leucocyte scintigraphy, which is presently considered to be the radionucleotide criterion standard for diagnosis of osteomyelitis [27–29]. Patients’ WBCs are labeled in vitro with indium-111 or 99mTc-exametazime (99mTc-HMPAO) and returned intravenously. Increased uptake is observed in the presence of infection. Physiologic uptake is also expected in bone marrow; hence, this is commonly complemented with Tc-sulfur colloid imaging [28,29]. Disadvantages include limited spatial resolution, labor-intensive procedures, and prolonged results [27–29]. The enhanced images obtained from single photon emission computed tomography (SPECT)/ CT hybrid augments the diagnostic accuracy and refines the anatomic definition and localization [27,30,31]. A study by Ar€can et al [31] showed that SPECT/CT changed the evaluation of osteomyelitis in 16.5% of their patients. Currently, nuclear medicine is recommended when MRI is contraindicated or infection is multifocal [15,16,27,28]. Moreover, nuclear imaging has limited value in diagnosing SA. Triple-phase bone scan, gallium, and indium scintigraphy are all unreliable in distinguishing infectious from inflammatory arthritis [15,27].

Although the CRP level was increased at the second orthopedic consult, this test is not specific for infection. Multiple studies now support the “inflammatory” theory as driving the OA process, and increased CRP is considered by some studies to be associated with OA [23–25]. Clinically, OA presents with gradual onset of pain with stiffness; it is associated with remitting and relapsing symptoms. Its pain is generally worse on activity, especially following a period of rest [32]. Given the patient’s acute presentation with no previous episodes of foot pain, the diagnosis of OA should have been questioned.

Ideal management of SA primarily involves broad-spectrum antibiotics as soon as synovial fluid samples are collected, and early orthopedic consult for surgical management [1–3,8,12]. Open and arthroscopic techniques can be utilized to drain the infected joint and eliminate the purulent material [1,8]. Aside from decompression, drainage also helps in improving blood flow and removing toxins [2]. Intravenous antibiotics are generally continued for 3–4 weeks [1,8,12]. It must be emphasized that the first step is rapid recognition of SA in any patient presenting with acutely inflamed joint

Overall, a few factors contributed to the delay in diagnosing SA in this case. First, SA of the midfoot is extremely rare even for immunocompromised patients. In addition, the nondiagnostic initial plain radiograph and MRI findings, coupled with seemingly acceptable serum markers, may have given a false reassurance for ruling out an infectious cause.

This case serves as a reminder that a high index of suspicion is required in the early diagnosis of SA of small joints. Laboratory tests and imaging studies are now readily available to aid in diagnostic uncertainties, but proper clinical evaluation continues to play a vital role in clinching the correct diagnosis. Symptoms such as unremitting pain for 1 week despite analgesics, inability to bear weight, and pain affecting sleep are concerning for SA; these patients warrant a close follow-up regardless of unremarkable laboratory and imaging results. It is important to acknowledge that radiologic findings may lag behind the onset of symptoms and repeat diagnostic tests are needed in the presence of concerning symptoms.

Conclusions

SA of the midfoot is rare but needs to be considered in any patient presenting with worsening symptoms after more than 1 week of anti-inflammatory medications. Laboratory and radiological investigations have limitations and should be guided by appropriate clinical findings and judgment. It is important to maintain a high index of suspicion for these cases to prevent morbidity and mortality in patients.