Differences between leukemic arthritis and juvenile idiopathic arthritis

Acute lymphoblastic leukemia (ALL) is the most common type of cancer in children, and juvenile idiopathic arthritis (JIA) is the most common chronic rheumatic disease in childhood. Most patients with leukemia present with symptoms of bone marrow failure (anemia, leukopenia, thrombocytopenia), however; musculoskeletal manifestations occur in approximately 30% of cases, manifesting as bone pain, arthralgia, myalgia, limp or arthritis. Sometimes these joint symptoms are the only ones present, leading to a misdiagnosis of juvenile idiopathic arthritis, with the consequent use of steroids or disease-modifying drugs, which increase morbidity and mortality. We present a study that shows the clinical and laboratory differences between these 2 diseases, and we propose criteria to help the clinician to easily differentiate between them [1‐4].

Patients under 16 years of age, who initially presented to the pediatric rheumatology service of the General Hospital of the La Raza National Medical Center, from January 2013 to December 2022, with arthritis, who did not show blasts in peripheral blood, in whom the final diagnosis was juvenile idiopathic arthritis or acute lymphoblastic leukemia. Arthritis is defined as the presence of joint inflammation and effusion. The clinical manifestations evaluated were: age, gender, time of onset of symptoms, pain intensity, presence of nocturnal pain, remission with analgesics, fever, weight loss, nocturnal diaphoresis, lymph node enlargement, and hepatosplenomegaly. The laboratory variables evaluated were: hemoglobin, leukocytes, neutrophils, lymphocytes, platelets, erythrocyte sedimentation rate and C-reactive protein. The clinical and laboratory variables of both groups were compared. Leukemias were classified according to the French-American-British classification of acute lymphoblastic leukemia (FAB) classification, and juvenile idiopathic arthritis according to the International League of Associations for Rheumatology (ILAR) criteria [5, 6].

Data from a total of 76 patients were analyzed, 14 with leukemic arthritis and 62 with JIA (Table 1). The mean age in the leukemic arthritis (LA) group was 8.1 years (3–14 years) and in the JIA group, 10.3 years (2–15 years) (p = 0.084), the female gender predominated in the JIA group (p = 0.026). The pattern of arthritis in the group of patients with leukemia was polyarticular in 43% and oligoarticular in 50% (migratory in 71% of these). In the JIA group, the polyarticular pattern was the most frequent (85%) (Table 2).

Comparing the leukemic arthritis group with the JIA group, we found that the onset time of symptoms was 4.2 months vs. 9.1 months (p = 0.017) and the pain intensity was 10 vs. 7 on the pain analogue scale (p = 0.0001), the presence of nocturnal pain that awakens the patient 79% vs 10% (p = 0.0001, RR 12.7), pain remission with analgesics 0% vs 74% (p = 0.0001, RR 1.8), fever 64% vs 6% (p = 0.0001, RR 8.7), weight loss 86% vs 19% (p = 0001, RR 13), nocturnal sweating 50% vs 0% (p = 0.0001, RR 9.8), lymph nodes 86% vs 3% (p = 0.0001, RR 26), hepato-splenomegaly 57% vs 3% (p = 0.0001, RR 8.8) respectively. Anemia was observed in 42% vs 8% (p = 0.001, RR 4.4), leukopenia in 71% vs 2% (p = 0.0001, RR 14.7), neutropenia 92% vs 0% n (p = 0.0001, RR 63), thrombocytopenia in 43% vs 0% (p = 0.0001, RR 8.7), mean hemoglobin value was 11.5 g/dL vs 13 g/dL (p = 0.009), leukocytes 3508 mm3 vs 9111 mm3 (p = 0.0001), neutrophils 695 mm3 vs 5143 mm3 (p = 0.0001), platelets 191,000 mm3 vs 378,000 mm3 (p = 0.0001) in leukemic arthritis and JIA group respectively. No difference was found in the values of lymphocytes, erythrocyte sedimentation rate and C-reactive protein (Table 3).

A binary logistic regression model was developed to differentiate leukemic arthritis from JIA. The probability associated with the statistic (Chi-square) was 0.000, and the Cox and Snell R2 and Nagelkerke R2 values were 0.615 and 1, respectively. The developed model correctly classified 100% of the cases (Tables 4 and 5).

Musculoskeletal pain is a common symptom, affecting 10–20% of school-age children. Most of these complaints are benign in nature. However, some hematological diseases have been associated with rheumatic symptoms. These diseases can be benign (sickle cell disease, thalassemia, hemochromatosis, hemophilic arthropathy) or malignant (leukemia, lymphoma, multiple myeloma, neuroblastoma) [1].

Leukemia is the most common childhood cancer (30–40% of cases) and acute lymphoblastic leukemia is the most common subtype (80%), with a peak incidence between 2–6 years of age. Most patients with leukemia present with pallor, fever, lymphadenopathy, visceromegaly, and a variety of hematologic abnormalities (anemia, thrombocytopenia, leukopenia, leukocytosis, and peripheral blood blasts). However, one-third of pediatric patients with acute lymphoblastic leukemia initially present with symptoms associated with the musculoskeletal system [2]. Localized or diffuse joint pain, claudication, monoarthritis or polyarthritis, mialgia, or limp are the most common musculoskeletal symptoms. Occasionally, these symptoms are the only ones at the time of presentation, with subtle or absent peripheral blood changes [2, 3].

Articular manifestations can be explained by infiltration of the synovium, periosteum, or bone marrow by malignant cells. They can also be caused by the deposition of uric acid or immune complexes in the joint. A rare cause is hemarthrosis secondary to thrombocytopenia [2‐4]. The joint effusion may be secondary to inflammation of the adjacent metaphysis, rather than direct invasion by blasts. Occasionally, as a result of severe infiltration of the metaphysis, the joint appears to be inflamed, simulating juvenile idiopathic arthritis. Severe nocturnal pain that awakens the child, pain that is out of proportion to clinical findings, or pain that does not respond to analgesics, is highly suggestive of malignancy [2, 3]. For this reason, hematological malignancies must be ruled out in patients with musculoskeletal pain, and for this reason, the physician needs to have tools to help discriminate malignant causes from benign ones, improving early diagnosis, treatment, and patient prognosis.

Based solely on musculoskeletal pain, it is difficult to diagnose acute lymphoblastic leukemia, since musculoskeletal manifestations can mimic JIA, osteomyelitis, septic arthritis, or transient synovitis [2]. Clinicians evaluating children with musculoskeletal symptoms should be aware that some hematologic conditions can present with rheumatologic symptoms and signs, and a high index of suspicion is important for a correct diagnosis. The acute onset of pain, the pattern of joint involvement, periarticular rather than joint pain, predominance at night, presence of hepatomegaly, lymph nodes, petechiae, ecchymosis, abnormal blood counts or with subtle changes, poor response to conventional therapy, and lytic lesions on radiographs warrant further investigation of the causes of musculoskeletal pain [1, 3].

Over the years, several reports have been made on the articular manifestations of hematological malignancies and their difference from JIA. Bone pain, arthralgia, limp, myalgia, true arthritis, night pain, pain out of proportion to clinical findings, elevated DHL, presence of peripheral blood blasts, leukocytosis, anemia, thrombocytopenia or low normal platelet count (150,000–250 000), neutropenia, or even normal leukocyte counts have been reported. In this type of patient in whom leukemia presents with musculoskeletal manifestations, a better prognosis with a longer survival has been observed when compared with those who do not present with musculoskeletal symptoms [2, 4, 7‐9].

In 2005, Jones conducted a multicenter study in 71 children diagnosed with acute lymphoblastic leukemia and 206 JIA patients to determine the factors that help differentiate between the two diseases. Of the patients with leukemia, 75% (n = 53) had no blasts in peripheral blood at the time of evaluation. These patients were compared with those with JIA. The 3 strongest predictors of ALL were leukopenia (< 4000), low normal platelet count (150–250,000), and history of nocturnal pain. The combination of these values showed a sensitivity of 100% and a specificity of 85% for diagnosis of ALL. The number of joints affected in both groups was similar, and uric acid and LDH were not useful as predictors of ALL [10]. In our study, with 2 of 8 clinical variables and 1 of 3 laboratory variables (excluding anemia), it was possible to correctly classify 100% of the patients with leukemia, which makes our study a useful tool for the differential diagnosis of patients with arthritis.

Louvigne in 2020 conducted a study to identify early clinical and laboratory features that help distinguish between JIA and ALL. In a multicenter study, 49 patients with ALL and 98 with JIA were included. Oligoarthritis (43.9%) predominated in the JIA group, followed by polyarthritis (29.6%), arthritis-related enthesitis (18.4%), systemic arthritis (5.1%), and psoriatic arthritis (3.1%). Of the leukemia group, only one patient presented arthritis, the rest presented musculoskeletal pain. Neutrophils, eosinophils, basophils, monocytes, platelets, and hemoglobin were found to be low in patients with ALL. The presence of hepatomegaly, splenomegaly or adenomegaly (p = 0.0001, OR 154), non-articular bone pain (p = 0.001, OR 13), fever (p = 0.01, OR 4.5), neutrophil count less than 2000 (p = 0.002, OR 50), platelets < 300,000 (p = 0.004, OR14) were found to be significantly associated with the presence of leukemia. This article concludes with a decision tree in which if the patient presents with osteoarticular pain for at least 1 month, a bone marrow aspirate should be performed if accompanied by hepatomegaly, splenomegaly, or adenomegaly. If not accompanied by the latter, if associated with fever or increased inflammatory markers (CRP > 6, ESR > 20), plus non-articular bone pain and/or general symptoms (asthenia, adynamia or weight loss), neutrophils < 2,000 or platelets < 300,000, a bone marrow aspirate should be performed to rule out ALL as the cause of the symptoms [11]. In our study, we found similar results, except that all our patients had arthritis, which makes the differential diagnosis more difficult.

The studies mentioned above show us that the clinical manifestations and complete blood count of both diseases are different. We only included patients with true arthritis in our study, which makes it different from previous studies in which most patients had musculoskeletal symptoms. This type of presentation of leukemia with true arthritis makes differential diagnosis more difficult for the pediatric rheumatologist, but with the proposed criteria, patients with suspected leukemia can be appropriately selected and bone marrow aspirate can be performed to establish a diagnosis and appropriate treatment. Prospective studies are needed to help validate our results.