Published online May 1, 2006
PEDIATRICS Vol. 117 No. 5 May 2006, pp. e840-e844 (doi:10.1542/peds.2005-1515)
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow P3Rs: View responses
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jones, O. Y.
Right arrow Articles by Rabinovich, C. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jones, O. Y.
Right arrow Articles by Rabinovich, C. E.
Related Collections
Right arrow Musculoskeletal System

A Multicenter Case-Control Study on Predictive Factors Distinguishing Childhood Leukemia From Juvenile Rheumatoid Arthritis

Olcay Y. Jones, MD, PhDa, Charles H. Spencer, MDb, Suzanne L. Bowyer, MDc, Peter B. Dent, MDd, Beth S. Gottlieb, MD, MSe and C. Egla Rabinovich, MD, MPHf

a Department of Pediatric Rheumatology, Children's National Medical Center and George Washington University, Washington, DC
b Department of Pediatric Rheumatology, La Rabida Children's Hospital and University of Chicago, Chicago, Illinois
c Department of Pediatric Rheumatology, James Whitcomb Riley Hospital, Indiana University, Indianapolis, Indiana
d Division of Rheumatology, McMaster University, Hamilton, Ontario, Canada
e Department of Pediatric Rheumatology, LIJ Schneider Children's Hospital, New Hyde Park, NY
f Department of Pediatric Rheumatology, Duke University, Durham, North Carolina


   ABSTRACT
TOP
 ABSTRACT
METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
OBJECTIVE. Acute lymphocytic leukemia (ALL) often presents with musculoskeletal concerns such as pain or swelling, even before appearance of blasts in the peripheral blood. Such presentation may lead to misdiagnosis of a child with juvenile rheumatoid arthritis (JRA). This study was designed to identify the predictive factors for leukemia using basic clinical and laboratory information.

METHODS. A retrospective chart review was performed using a simple questionnaire to compare the clinical and laboratory findings present during the initial visit to a pediatric rheumatology clinic for 277 children who were ultimately diagnosed with either JRA (n = 206) or ALL (n = 71). Sensitivity and specificity analysis of a variety of parameters, both singly and in combination, was performed to identify predictive value for ALL.

RESULTS. The majority (75%) of children with ALL did not have blasts in the peripheral blood at the time of evaluation by pediatric rheumatologists. In children presenting with unexplained musculoskeletal complaints, the 3 most important factors that predicted a diagnosis of ALL were low white blood cell count (<4 x 109/L), low-normal platelet count (150–250 x 109/L), and history of nighttime pain. In the presence of all 3, the sensitivity and specificity for a diagnosis of ALL were 100% and 85%, respectively. Other findings, including antinuclear antibody, rash, and objective signs of arthritis, were not helpful in differentiating between these diagnoses because they occurred at similar rates in both groups.

CONCLUSIONS. When a child develops new-onset bone-joint complaints, the presence of subtle complete blood count changes combined with nighttime pain should lead to consideration of leukemia as the underlying cause.

Key Words: arthritis • leukemia • musculoskeletal pain

Abbreviations: JRA—juvenile rheumatoid arthritis • ALL—acute lymphocytic leukemia • ANA—antinuclear antibody • WBC—white blood cell • LDH—lactic dehydrogenase • CI—confidence interval

Juvenile rheumatoid arthritis (JRA) is a chronic idiopathic autoimmune disease with an annual incidence of 2 to 20 per 100000 children.1 Clinically, JRA is stratified into 3 subtypes2: pauciarticular (≤4 joints involved), polyarticular (≥5 joints), and systemic (characterized by fever, rash, and arthritis). The diagnosis of JRA is guided by published criteria from the American College of Rheumatology and requires exclusion of other potential etiologies of arthropathy, including infectious agents, orthopedic conditions, and malignancies.3 There is no single test for diagnosis of JRA; rather, physicians must determine if the clinical pattern of presentation is consistent with the diagnosis. This determination is based on an integrative analysis of data, including history, examination findings, and results of simple laboratory tests such as blood counts, radiographs, or other imaging studies.

Several childhood malignancies, including acute lymphocytic leukemia (ALL), the most common childhood malignancy,4 may initially present with musculoskeletal complaints that mimic JRA.5–10 Such presentation can occur in 15% to 30% of ALL cases at disease onset, when peripheral blood changes are subtle or even absent.5–7,11–15 Thus, there is concern that a malignancy could be misdiagnosed, causing a delay in proper management.16–18

The literature on comparing such children with ALL with those with JRA for the pattern of initial findings is limited and involves small numbers of case series. Most of those reports reiterate the importance of bone marrow examination for correct diagnosis before emergence of blasts in the peripheral blood.6–9,11–13,18 Our purpose was to examine the value of diagnostic markers derived from initial basic clinical and laboratory data in predicting a potential malignancy using a large cohort of patients. Our results may aid in the decision of whether to obtain an oncology consultation in children with recent-onset arthropathy.


   METHODS
TOP
 ABSTRACT
 METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
This is a multicenter, case-control chart review study using a simple questionnaire examining the data from the initial visit to a pediatric rheumatology clinic before initiation of treatment. All the children included in the study had an established diagnosis of ALL or JRA at the time of data collection. Children with ALL were identified by using the Pediatric Rheumatology Disease Registry that encompassed the years 1992 to 199519 and by direct communication with pediatric rheumatology colleagues through the Pediatric Rheumatology list server. The control subjects, children with JRA, were randomly selected from the established patient populations followed at 7 different medical centers in North America. The content of the questionnaire was based on the available literature and clinical experience among the authors. The study design was compatible with the Health Insurance Portability and Accountability Act (HIPAA) regulations and approved by the institutional review boards of the University of Chicago, Illinois, and All Children's Hospital, St Petersburg, Florida.

Physicians were asked to complete a simple questionnaire for each subject by circling a yes or no option for the diagnostic markers listed in Table 1: "Nighttime pain" signified complaint of limb or joint pain at night; "early disability" described a child's inability to walk freely and need to be carried; "<5 joints on examination" was to assess the numbers of joints with signs of arthritis (ie, a pauciarticular presentation); arthritis was defined as a joint with swelling or with minimum of 2 of the following: pain, limitation, tenderness, and warmth2; antinuclear antibody (ANA) was regarded as positive if titer was 1:80 or above; "low normal platelets" queried if the platelet count was 150 to 250 x 109/L to target a subtle change rather then a thrombocytopenia that would readily alert the physician to a bone marrow disorder; "low white blood cell (WBC) count" queried if the WBC count was <4 x 109/L, an unusual finding for JRA; "low Hgb" defined a hemoglobin <11 g/dL; "high LDH" and "high uric acid" signified elevated levels above the established reference range for that clinic. In addition, the questionnaire queried actual values of abnormal lactic dehydrogenase (LDH) and uric acid levels as well as details of abnormal radiograph findings. The fold increase in LDH and uric acid levels were calculated relative to the upper limits of the norm of the given institution.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Reported Results of Diagnostic Markers Among Children With Blast-Negative ALL (n = 53) in Comparison to JRA (n = 206)

 
All statistical analysis was performed by using Systat software (Systat, Evanston, IL) using a 2-sided Student's t test, a 2-sided Pearson {chi}2, or Fisher's exact test to determine if a presenting sign, symptom, or diagnostic test was different between the JRA and ALL groups. Significance level was adjusted using the Bonferroni correction for multiple comparisons. The value of given diagnostic test as a predictive factor for leukemia was assessed by calculating its specificity and sensitivity for leukemia in comparison to JRA along with the 95% confidence interval (CI). The differences in the denominators were the result of missing information on some patients for the given question asked.


   RESULTS
TOP
 ABSTRACT
 METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
We examined data on 71 children with ALL and 206 with JRA. The mean age of children with leukemia was 88.6 months (range: 12–222 months) and that for JRA was 76.6 months (range: 29–205 months) (P = .06). A female predominance was evident only among the children with JRA (75% vs 51%). The ethnic breakdown was similar for the 2 diagnoses and included 60% white, 10% black, 7% Hispanic, and 23% other. Subtype information was available on 146 subjects with JRA and consisted of 42% pauciarticular, 38% polyarticular, and 20% systemic.

Only 25% (18 of 71) of children with leukemia had blast cells in the initial blood smear (patients with blast-positive leukemia), and the remaining 75% (53 of 71) did not (patients who had blast-negative leukemia). Table 1 summarizes the results from the children with blast-negative leukemia compared with those with JRA. Among the children with ALL, the proportion positive for a given diagnostic test was the same whether the patients who were blast-positive were included or excluded in the analysis (data not shown). The data from the JRA group combine the results from all 3 subtypes (pauciarticular, polyarticular, and systemic) because a subtype-specific statistical analysis of the data did not significantly alter the results (data not shown).

The 3 strongest predictive factors for ALL with narrow CIs were the presence of a low WBC count, low-normal platelet count, and the history of nighttime awakenings because of pain. As shown in Table 2, the combination of any low hematologic value, including anemia, in conjunction with nighttime pain provided increased sensitivity; presence of 2 low hematologic values plus nighttime pain yielded 100% sensitivity and 85% specificity. Among the children with leukemia, presence of 2 low hematologic values was more frequent in the blast-negative population compared with those who were blast-positive (46% vs 28%), but the difference only approached significance (P = .06). There were no significant differences between the blast-positive and -negative groups in the proportion with nighttime pain, or with 1 or 2 low hematologic values in combination with nighttime pain (data not shown).


View this table:
[in this window]
[in a new window]
  		TABLE 2 Predictive Value of Complete Blood Count Changes and Nighttime Pain for ALL

 
Both JRA and ALL groups were similar regarding a history of early disability and number of affected joints. Presence of a rash was similar among children with ALL or JRA. A positive ANA also did not discriminate between the 2 groups. As expected, approximately one third of patients with JRA were ANA-positive. Of those children with leukemia who were tested (n = 47), 17% also had a positive ANA, yielding limited specificity of this test. The levels of LDH were increased in 77% of leukemia patients (n = 58), but also in 38% (n = 104) for the JRA population, reducing the specificity of an elevated LDH for leukemia. The absolute values of elevated LDH were available for 32 patients with leukemia (25 blast-negative) and tended to be higher (range: two- to fivefold greater than normal) than those of patients with JRA (range: less than twofold greater than normal). Elevated uric acid levels were found in 19% of the patients with leukemia. The small number of subjects who had a uric acid examined, particularly in the JRA group, limited meaningful comparison. Of note, all the patients with leukemia with increased uric acid levels (n = 8) also had elevated LDH values.

A radiographic examination of affected joints was included in the initial workup of 80% of children with leukemia and 69% of those with JRA. The frequency of abnormal radiograph findings was similar between the 2 populations. Although the radiograph findings of radiolucent bands, sclerosis, lytic lesions, and periosteal elevation were seen only in children with ALL, osteopenia and joint effusions were noted in both groups.


   DISCUSSION
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Literature over the last 50 years describes small numbers of children who presented to rheumatologists for arthritis who were later diagnosed with ALL. Among those cases, severe nighttime pain, mild to moderate blood smear abnormalities, and increased LDH were often reported.6–9 Our results support previous observations and provide statistical validity on the usefulness of history, physical, and simple serologic and radiographic tests in decision-making. Physicians evaluating a child for musculoskeletal complaints should remain vigilant for that unusual child with an underlying malignancy, and as shown in our study, one cannot rely on finding blasts on the peripheral smear to identify the child with ALL. The presence of subtle changes in the blood count along with complaints of nighttime pain provided the highest predictive value for distinguishing ALL from JRA. In particular, presence of the combination of blood count changes that may or may not be associated with night pain was almost exclusively a finding for ALL with a sensitivity and specificity of ≥85%.

Rheumatologists usually anticipate increased, not decreased, WBC and platelet counts when considering a diagnosis of systemic JRA, and increased platelet counts are not unusual for any type of JRA. Anemia from chronic illness does occur among children with JRA and thus it is not surprising to find a low specificity for leukemia. There seems to be a unique subgroup of leukemic children presenting initially with a low WBC count.6–8,11,12,18,20 These children can have leukemic expansion in the peripheral organs without blastic changes in the blood smear for several months. Their disease may mimic systemic JRA with fever, chronic arthritis, hepatosplenomegaly, and anemia, delaying a correct diagnosis. In our study, 14 children with ALL had a low WBC count; among those, this was the only abnormal diagnostic test for 3.

Night pain has been observed frequently in children with leukemia.12,13 Children with JRA usually describe achiness, stiffness, and a dull discomfort over the joint(s) that is worse in the mornings. Pain that is excruciating and wakes the child from sleep is not typical for JRA.2,6,7,21 We found 17% of children with JRA reported night pain compared with 65% with leukemia. Children with malignancies may develop a spectrum rather than a unique type of bone and joint complaints, including pain.5–7,11,13,22 Unlike JRA, pain from leukemia may, but not always, be point tenderness over the diaphysis rather than diffusely over the joint(s).5,7,11–15

Some diagnostic tests were not useful in differentiating between ALL and JRA, including ANA. It is not unusual for clinicians to consider an autoimmune disorder if ANA is positive. Previous studies cautioned the use of ANA, because it can be found among patients with leukemia.9,21,22 In our study, the frequency of a positive ANA among the ALL group was similar to that reported for age-matched healthy control subjects,23 and furthermore, the difference between the ALL and JRA groups was not statistically significant. We avoided investigation of fever, lymphadenopathy, organomegaly, or increased sedimentation rate, because they have been shown to be of limited value in the distinguishing leukemia from systemic JRA.6,8–11,13,15–17

In our study, LDH and uric acid were not found to be useful as predictive factors for ALL; LDH showed low sensitivity for leukemia (53%), similar to a previous report.9 However, we found a twofold or above increase of LDH almost exclusively among children with ALL, as suggested previously.24 The small sample size of those with uric acid level determination limited our ability to determine its potential use. In the literature, development of gout among children with leukemia is extremely rare,25,26 and uric acid levels were usually reported as being normal8,9 or mildly increased.22 Radiographic studies of the affected joints are strongly encouraged to rule out trauma or any bone abnormalities. Although the frequency of finding an abnormal radiograph was similar between ALL and JRA, presence of radiolucent bands, lytic lesions, and sclerotic lesions should alert the physician to consider malignancy until proven otherwise.14,15,22 A prospective cohort study of these factors may clarify their diagnostic value.


   CONCLUSIONS
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The blood count may not have blasts on the peripheral smear at early stages of leukemia, an important consideration when evaluating a child with recent-onset musculoskeletal complaints. Our results suggest that peripheral blood smear abnormalities, especially a low-normal platelet count and low WBC count, along with presence of nighttime pain are the most sensitive and specific factors for the diagnosis of ALL. For such children, we suggest primary care physicians seek oncology consultation before rheumatology consultation for consideration of additional workup, including examination of the blood smear, bone scan, and bone marrow examination. Our results may also provide a platform to expand predictive factors to distinguish ALL from JRA in future prospective studies.


   ACKNOWLEDGMENTS
 
We thank Drs G.D. Cawkwell, C.P.T. Chao, A. Gedalia, A. Gillies, D.P. Goldsmith, S.I. Goodman, K. Gross, K.A. Haines, M. Henrickson, N.T. Ilowite, C.B. Lindsley, D.J. Lovell, D.K. McCurdy, T. Nix, K.N. Schikler, J. Sullivan, I.S. Szer, Y. Uziel, L. Vogler, D. Wright, L. Wagner-Weiner, and L.S. Zemel and nursing staff members M. Belvedere, S. Hunter, M. Kipta, J. Norris, A. Sharpko, and P. Towell for their contributions to data collection.


   FOOTNOTES
 
Accepted Nov 4, 2005.

Address correspondence to Olcay Y. Jones, MD, PhD, Children's National Medical Center and George Washington University, Division of Pediatric Rheumatology, 111 Michigan Ave NW, WW 3.5, Suite 300, Washington, DC 20010. E-mail: oyjones{at}cnmc.org

The authors have indicated they have no financial relationships relevant to this article to disclose.


   REFERENCES
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 

  1. Manners PJ, Bower C. Worldwide prevalence of juvenile arthritis why does it vary so much? J Rheumatol. 2002;29 :1520 –1530[ISI][Medline]
  2. Cassidy JT, Petty RE. Juvenile rheumatoid arthritis. In: Cassidy JT, Petty RE, eds. Textbook of Pediatric Rheumatology. Philadelphia, PA: WB Saunders; 2001:218–321
  3. Cassidy JT, Levinson JE, Bass JC, et al. A study of classification criteria for a diagnosis of juvenile rheumatoid arthritis. Arthritis Rheum. 1986;29 :274 –281[ISI][Medline]
  4. Margolin JF, Poplack DG. Acute lymphoblastic leukemia. In: Pizzo PA, Poplack DG, eds. Principles and Practice of Pediatric Oncology. Philadelphia, PA: Lippincott-Raven Publishers; 1997:409–482
  5. Cabral DA, Tucker JB. Malignancies in children who initially present with rheumatic complaints. J Pediatr. 1999;134 :53 –57[CrossRef][ISI][Medline]
  6. Schaller J. Arthritis as a presenting manifestation of malignancy in children. J Pediatr. 1972;81 :793 –797[CrossRef][ISI][Medline]
  7. Bowyer SL, Hollister JR. Limb pain in childhood. Pediatr Clin North Am. 1984;31 :1053 –1081[ISI][Medline]
  8. Fink CW, Windmiller J, Sartain P. Arthritis as the presenting feature of childhood leukemia. Arthritis Rheum. 1972;15 :347 –349[ISI][Medline]
  9. Ostrov BE, Goldsmith DP, Athreya BH. Differentiation of systemic juvenile rheumatoid arthritis from acute leukemia near the onset of disease. J Pediatr. 1993;122 :595 –598[ISI][Medline]
  10. Sills EM. Errors in diagnosis of juvenile rheumatoid arthritis. Johns Hopkins Med J. 1973;133 :88 –95[ISI][Medline]
  11. Jonsson OG, Sartain P, Ducore JM, Buchanan GR. Bone pain as an initial symptom of childhood acute lymphoblastic leukemia: association with nearly normal hematologic indexes. J Pediatr. 1990;117 :233 –237[CrossRef][ISI][Medline]
  12. Silverstein MN, Kelly PJ. Leukemia with osteoarticular symptoms and signs. Ann Intern Med. 1963;59 :637 –645[ISI][Medline]
  13. Bichel J. Arthralgic leukemia in children. Acta Haematol. 1948;1 :154 –164
  14. Thomas LB, Forkner CE, Frei E, Besse BE, Stabenau JR. The skeletal lesions of acute leukemia. Cancer. 1961;14 :608 –621[CrossRef][ISI][Medline]
  15. Rogalsky RJ, Black B, Reed MH. Orthopaedic manifestations of leukemia in children. J Bone Joint Surg [Am]. 1986;68 :494 –501[Abstract/Free Full Text]
  16. Case records of the Massachusetts General Hospital (case 36-1981). N Engl J Med. 1981;305 :567 –574[ISI][Medline]
  17. Case records of the Massachusetts General Hospital (case 14-1986). N Engl J Med. 1981;314 :973 –981
  18. Bradlow A, Barton C. Arthritic presentation of childhood leukemia. Postgrad Med J. 1991;67 :562 –564[Abstract]
  19. Bowyer SL, Roettcher PA, Higgins GC, et al. Health status of patients with juvenile rheumatoid arthritis at 1 and 5 years after diagnosis. J Rheumatol. 2003;30 :394 –400[ISI][Medline]
  20. Needleman M. Childhood leukemia mimicking arthritis. J Am Board Fam Pract. 1996;9 :56[Medline]
  21. Costello PB, Brecher ML, Starr JI, Freeman AI, Green FA. A prospective analysis of the frequency, course and possible prognostic significance of the joint manifestations of childhood leukemia. J Rheumatol. 1983;10 :753 –757[ISI][Medline]
  22. Spilberg I, Meyer GJ. The arthritis of leukemia. Arthritis Rheum. 1972;15 :630 –635[ISI][Medline]
  23. Deane PMG, Liard G, Siegel DM, Baum J. The outcome of children referred to a pediatric rheumatology clinic with a positive antinuclear antibody test but without an autoimmune disease. Pediatrics. 1995;95 :892 –895[Abstract/Free Full Text]
  24. Wallendal M, Stark L, Hollister JR. The discriminating value of serum LDH in children with malignancy presenting as limb pain. Arthritis Rheum. 1992;35(suppl 9) :S56
  25. Whitaker JA, Shaheddy M, Baum J, James J, Flume JB. Gout in childhood leukemia. J Pediatr. 1963;63 :961 –966[CrossRef][ISI][Medline]
  26. Vining CW, Thomson JG. Gout and aleukemic leukemia in a boy aged five. Arch Dis Child. 1934;9 :277 –284
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics



This article has been cited by other articles:


Home page
 Arch. Dis. Child.Home page
Journal Watch Pediatrics & Adolescent Medicine
Arch. Dis. Child., November 1, 2006; 91(11): 949 - 950.
[Full Text] [PDF]

Home page
 JWatch PediatricsHome page
Acute Lymphocytic Leukemia or Juvenile Rheumatoid Arthritis?
Journal Watch Pediatrics and Adolescent Medicine, August 2, 2006; 2006(802): 5 - 5.
[Full Text]

P3Rs:

Read all P3Rs

Need for statistical review
Tari J Turner, et al.
Pediatrics Online, 29 Apr 2007 [Full text]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow P3Rs: View responses
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jones, O. Y.
Right arrow Articles by Rabinovich, C. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jones, O. Y.
Right arrow Articles by Rabinovich, C. E.
Related Collections
Right arrow Musculoskeletal System