Rapid onset pembrolizumab-induced inflammatory arthritis diagnosed using musculoskeletal ultrasound

Abstract

Immune checkpoint inhibitors have revolutionised the treatment of cancer. While very effective, they commonly cause a wide spectrum of immune-related adverse events. These immune-related adverse events can be fatal and often have significant effects on quality of life. They therefore require prompt recognition and management. We report the case of a woman presenting with widespread joint pain and stiffness 6 hours after her first pembrolizumab infusion. She had no joint swelling on physical examination but an ultrasound scan revealed widespread musculoskeletal inflammation, confirming the diagnosis of inflammatory arthritis. To the best of our knowledge, this is the fastest reported inflammatory arthritis onset following immune checkpoint inhibitor treatment. It highlights the importance of timely imaging in patients on immune checkpoint inhibitors who present with new non-specific musculoskeletal pain. Her symptoms improved dramatically with intramuscular triamcinolone injection.

Background

Over the past decade, immune checkpoint inhibitors (ICIs) have revolutionised the management of metastatic melanoma and other malignancies. These monoclonal antibodies improve cancer outcomes by blocking T-cell immune checkpoint receptors. Pembrolizumab blocks the programmed death 1 (PD-1) receptor. Other therapies target its ligand (PD-L1) or cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). As ICI use has increased, so has the occurrence of a wide spectrum of immune-related adverse events (irAEs). Specifically, 90% of patients treated with a CTLA-4 agent and 70% of patients treated with a PD-1/PD-L1 agent will experience an irAE.¹ IrAEs can affect multiple organ systems and their onset is often acute and severe. Musculoskeletal toxicities are the fourth most common organ system affected and ICI-induced inflammatory arthritis (IA) seems to occur more frequently with anti PD-1 and anti PD-L1 agents.^{1 2} IA occurs in around 7% of patients whereas arthralgia is more common and reported in up to 43% of patients.³ Interestingly, cancer outcomes are better in patients who experience irAEs.⁴ Given the increasing frequency of irAEs, our understanding of how to recognise and manage the symptoms that herald their onset is of vital importance. Here, we report a case of rapid onset pembrolizumab-induced IA presenting with severe, non-specific musculoskeletal pain 6 hours post-initial therapy.

Case presentation

A Caucasian woman in her early 50s was diagnosed with malignant melanoma following a skin biopsy of a mole on her upper back. She had a medical history of hypertension and normal body mass index. There was no history of autoimmunity or joint pain. Her regular medications included amlodipine, perindopril, omeprazole and vitamin D. She was a non-smoker and did not drink alcohol. She underwent wide local excision of the mole and an initial staging CT showed no metastases. A sentinel lymph node biopsy was then performed which revealed a positive lymph node in her left axilla and she was diagnosed with stage IIIC melanoma. Her oncologist commenced adjuvant pembrolizumab as a 400 mg infusion every 6 weeks to reduce the cancer recurrence rate. Approximately 6 hours after she completed her first pembrolizumab infusion she noticed abrupt stiffness and pain which started in her legs and rapidly spread throughout her body. She denied joint swelling but struggled to walk the next day. Co-codamol (8 mg codeine and 500 mg paracetamol per tablet) two tablets four times a day did not relieve her symptoms and she was reviewed by oncology 4 weeks later. Further pembrolizumab infusions were withheld and she was commenced on naproxen 500 mg two times a day, which also did not alleviate her symptoms. By the time of rheumatological assessment 6 weeks after the first pembrolizumab infusion the pain had increased in severity and most affected her knees, wrists, hips, shoulders and elbows. She had 1 hour of early morning stiffness but still had not noticed any joint swelling. She did not have psoriasis or connective tissue disease symptoms. There was no family history of autoimmune disease. On examination, she had widespread tenderness with seven tender joints in her metacarpophalangeal (MCP), wrist, elbow and shoulder joints but no clinical synovitis or tenosynovitis. Her Health Assessment Questionnaire Disability Index score was 1.5.

Investigations

Her inflammatory markers were raised, with a C reactive protein (CRP) of 45 mg/L (normal <10), erythrocyte sedimentation rate of 40 mm/hour (normal 1–15) and platelet count of 459 10×9/L. Rheumatoid factor (RF), anti-cyclic citrullinated peptide antibodies (anti-CCP) and anti-nuclear antibodies were all negative and she was human leukocyte antigen (HLA)-B27 negative. Her serum urate and creatine kinase levels were within normal reference range. Given the lack of specific examination findings, a same day musculoskeletal ultrasound (US) assessment was performed in clinic, which revealed active inflammation in several joints, tendons and entheses (see figure 1).

Figure 1.

(A) Tenosynovitis of the right flexor carpi radialis (FCR) seen using a longitudinal approach. Note the presence of a small area of synovial hypertrophy in the tendon sheath of the FCR (asterisk) with intratendinous and peritendinous power Doppler signal (red spots). (B) Longitudinal view of the third left finger flexor tendons showing a small effusion in the tendon sheath (rounded dot) and mild intratendinous power Doppler signal (red spots) indicating ‘active’ tenosynovitis. (C) Longitudinal view of the right second metacarpophalangeal joint showing intra-articular synovial hypertrophy and moderate power Doppler signal (red spots) as per ‘active’ synovitis. (D) Thickened insertion of the common extensor tendons into the lateral epicondyle with entheseal and peri-entheseal power Doppler signal (red spots). (E and F) Moderately large area of joint effusion (rounded dots) in the right suprapatellar recess and right tibiotalar joint, respectively. Contralateral ultrasound assessment of the left knee and ankle revealed unremarkable sonographic findings (ie, no joint effusion).

Grey scale grade 1 was detected in the wrists, right MCP1 joint, left MCP2 joint, left ankle and right second and third flexor tendons. Grey scale grade 2 was detected in the right MCP2 joint, knees, right ankle, left second metatarsophalangeal joint and left flexor digitorum superficialis and profundus tendons. Power doppler grade 1 was detected in the right MCP1 joint, left MCP2 joint and the left flexor digitorum superficialis and profundus tendons and right second and third flexor tendons. Power Doppler grade 2 was detected in the right MCP2 and right ankle joints. There was also hypoechogencity grade 1 and power doppler grade 2 in both epicondyle entheses.

Given the imaging findings, a diagnosis of ICI-induced IA was made.

Differential diagnosis

Musculoskeletal symptoms are common in the general population with a broad differential diagnosis. When there are no objective findings on clinical examination, the clinical history and investigations are critical. Given this patient had a normal serum urate level this makes gout unlikely and pseudogout is rare under the age of 60; with the prevalence doubling in each subsequent decade after 60.⁵ Her CRP was mildly raised but this is a non-specific finding, particularly in the context of cancer. The widespread nature of her pain coupled with the lack of joint swelling initially led to the consideration of fibromyalgia. Mechanical joint pain was another strong possibility; especially as activated osteoarthritis, (worsening joint pains in a typical osteoarthritis distribution and lacking in inflammatory symptoms and signs) secondary to immunotherapy has been reported in the literature.⁶ Although fibromyalgia and osteoarthritis can both occur alongside IA, US power doppler is not typically associated with these conditions.⁷ In this case, the US scan was therefore crucial in confirming the diagnosis of IA. The patient was seronegative for rheumatoid arthritis (RA) autoantibodies with no personal or family history of psoriasis. Given this and the temporal association with her recent pembrolizumab infusion she was diagnosed with ICI-induced IA.

Treatment and outcome

A 120 mg intramuscular methylprednisolone injection was administered. 5 weeks later she received an 80 mg intramuscular triamcinolone injection.

Outcome and follow-up

On review a month later, she reported no improvement in symptoms although her CRP level had decreased to 19 mg/L. After a risk vs benefit discussion with the patient, it was decided to reduce her pembrolizumab dose to 200 mg which she would receive 3 weekly with an aim of reducing ICI toxicity. Her second pembrolizumab infusion was 8 weeks after the first and she reported her musculoskeletal symptoms slightly worsened immediately afterwards. Consequently, her third pembrolziumab dose was withheld. On rheumatology review, no clinical synovitis or tenosynovitis was identified. A repeat US assessment revealed an improvement compared with appearances on the initial scan but she had ongoing tenosynovitis in the wrist extensor tendons. At that stage she received an 80 mg intramuscular triamcinolone injection as anecdotally some patients respond better to this form of steroid. After receiving the triamcinolone injection she reported dramatic symptom resolution. Given this response, no additional immunosuppression with the use of disease-modifying antirheumatic drugs was required.

Discussion

To the best of our knowledge, this case represents the fastest reported arthritis onset following ICI treatment. The majority of irAEs occur in the first 12 weeks after ICI treatment and have a typical chronological onset pattern with dermatological irAEs occurring within the first 2–5 weeks, followed by gastrointestinal, hepatic and endocrine toxicities.⁸ The presentation of musculoskeletal irAEs can vary and tend to occur later with the median onset time of IA after ICI reported between 4 and 5 months.^{2 9 10}

IA after ICIs has been shown to occur more frequently in patients on dual therapy and in those with pre-existing autoantibodies.^{11 12} In their cohort study, Leipe et al included a seropositive (RF and anti-CCP) patient with polyarthritis occurring 1 day after dual ICI therapy.⁹ It has been hypothesised that in cases like this where toxicity develops rapidly after ICI initiation the patient may have been immunologically primed.² In immunologically primed patients, arthritis may have developed anyway regardless of the use of ICI therapy. However, such immunological priming may not be detectable by recognised serum autoantibodies alone. Indeed, the patient in our case had negative RA autoantibodies and 91% of patients with ICI-induced IA in the literature are seronegative for RA autoantibodies.² Rapid onset ICI toxicity has been reported more frequently in dermatological and hepatic toxicities, which may be in part due to these conditions being more clinically or biochemically obvious.⁸

Our case demonstrates the possibility that a significant proportion of patients treated with ICI and labelled with non-specific arthralgia may actually have subclinical inflammation on imaging in keeping with genuine IA. In this case, the US scan was crucial in making the diagnosis of IA. The patient’s initial symptoms were non-specific with limited objective examination findings. These factors lowered our initial index of suspicion for IA. However, the US conducted as part of a ‘one stop’ rheumatological assessment, revealed active inflammation in several joints, tendons and entheses and confirmed the diagnosis of ICI-induced IA. ICI-induced IA has a significant functional and emotional impact on patients hence the benefit of early diagnosis via US.¹³ This case highlights the importance of prompt rheumatology referral, multidisciplinary collaboration between oncologists and rheumatologists and timely imaging in patients who develop any new and significant musculoskeletal symptom after starting ICI therapy.

Patient’s perspective.

I had widespread joint inflammation due to a reaction to my immunotherapy treatment. It was very painful and difficult to move around. The full-body ultrasound indicated joint synovitis. The steroid injections have made a huge difference in my pain management and subsequent mobility.

Learning points.

• Inflammatory arthritis secondary to immune checkpoint inhibitors (ICIs) can develop rapidly after the initial treatment.

• Prompt rheumatology referral and ultrasound assessment should be considered when patients taking ICI therapy present with new non-specific musculoskeletal symptoms.

• Patients with new non-specific arthralgia after ICI therapy may have subclinical inflammation on imaging.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).