Vertebral Compression Fracture as a Presenting Feature of Acute Lymphoblastic Leukemia in Children

RAUL C. RIBEIRO, MD, CHING-HON

PUI, MD.*t AND MICHAEL J. SCHELL, PHDS

Twenty-four (1.6%) of 1466 children with acute lymphoblastic leukemia (ALL) treated at St. Jude Childrens Research Hospital had vertebral compression fractures at diagnosis. When compared with patients without this complication, they were more likely to have good prognostic features, including a leukocyte count of greater than 25 X 109/1, a leukemic cell DNA index of greater than 1.15, and hyperdiploidy (>50 chromosomes). Complete remission of ALL was induced in all patients, and symptoms of vertebral compression fractures abated following antileukemia therapy. Although the diagnosis of ALL was delayed for some patients because this unusual presenting complication was not recognized as such, their treatment outcome was as good as that for other children with standard-risk ALL. Cumer 61589-592,1988.

rare complication of childhood acute lymphoblastic leukemia (ALL).- Because the presenting symptoms in childhood ALL patients with this complication are generally atypical, the diagnosis of leukemia can be delayed, resulting in increased complications. Based on case reports, Blatt et al. have suggested that these cases represent a unique subset of ALL characterized by low leukemic cell burden. Otherwise, little is known about the presenting features or outcome of these patients. We report the clinical and biologic features of 24 consecutive children with vertebral compression fractures associated with ALL. We found that these children usually have favorable presenting features for leukemia and that they have a relatively good treatment outcome, comparable to that of others with standard-risk leukemia.

ERTEBRAL COMPRESSZON FRACTURE is a

Patients and Methods

From 1970 to 1986, 1466 consecutive, previously untreated children with ALL were admitted to St. Jude Childrens Research Hospital. Routine chest roentgenograms were performed for all patients. Children with vertebral compression fractures were identified by both computer-assisted review of case material from the medical records and radiology departments and by chart review of each case. Patients were enrolled in one of five successive clinical trials afier informed consent was obtained. Details of the treatment plans have been described previously.*-*

Leukemic Cell Phenotype Studies

The diagnosis of ALL was based on morphologic and cytochemical staining properties and, more recently, cases were classified according to French-AmericanBritish (FAB) criteria. l 3 Leukemic cell immunophenotype was determined from reactivity with monoclonal antibodies to lymphoid-associated antigens, including J5 (common ALL antigen, CD-lo), TI 1 (CD-2), TI01 (CD-5) and T3 (CD-3).I4 Blast cells were also tested for surface and cytoplasmic immunoglobulin and rosette formation with sheep erythrocytes.

From the Departments of *Hematology/Oncology, and SBiostatistics and Information Systems, St. Jude Childrens Research Hospital, Memphis, and the ?Division of Hematology and Oncology, Department of Pediatrics, University of Tennessee at Memphis, College of Medicine, Memphis, Tennessee. Supported in part by G a t CA-20180 and CA-21765 from the rns National Cancer Institute, and by the American Lebanese Syrian Associated Chanties. The authors thank D .A. Thomas Look for DNA content detennir nations, Drs. Sue Melvin and Frederick Behm for immunologic studies, Drs. Dorothy L. Williams and Susana C. Raimondi for cytogenetic analyses, Ms. Marlene Jacks for typing and Ms. Linda Daniels for manuscript editing. Address for reprints: Ching-Hon h i , MD,St. Jude Childrens Research Hospital, 332 North Lauderdale, P.O. Box 318, Memphis, TN 38101. Accepted for publication August 25, 1987.

DNA Content Determination

Leukemic bone marrow samples were stained with the DNA-specific dye, propidium iodide, and analyzed lw by f o cytometry, as previously described. The DNA index (ratio of DNA content in leukemic GO/G1 cells to

589

590

CANCER February I 1988

TABLE. Presenting Clinical and Laboratory Characteristics of Patients With Vertebral Compression Fracture 1

No. 3

Patient no.
3. .

Age (yr) 5.3 4.4 4. I 16.2 5. I 18.3 2.4 6.1 9.5 6.4 7.5 9.8 9.3 5.6 2.1 3.9 2.8 7.4 4.2 4.0 4. I 6.7 12.9 10.1

Race
W W W

Sex

Leukocyte count (X 109/1)

24.0 6.9 6.4 I .8 10.5 4.0 6. I 6.7 4.5 3.4 2.3 4.2 9.1 16.0 6. I 9.8 85.0 22.5 7.8 5.4 5. I 8. I 2.4 4.1

Hemoglobin level (ddU

7.9 5.3
7.1 10.0

Platelet count ( X I 09/1)

22 NA 9 I96 18 150 13 106 260 53 206 280 48 30 400 8 19 82 57 46 106 562 I73 72

Leukemic cell lmmunophenotype NA NA NA NA NA NA NA NA Common CALLANA Common Common Common CALLACommon Common Common NA Common NA Pre-B* Common Common DNA index NA NA NA NA NA NA NA NA NA NA NA 1.17 1.17 1 .oo NA 1.18 1.21 1.17 1.23 1.18 NA
I .oo 1.15 1.12

Ploidy NA NA NA NA NA NA NA NA Pseudodiploid Diploid Diploid Hyperdiploid Hyperdiploid Hypodiploid Hyperdiploid Hyperdiploid Hyperdiploid Hyperdiploid NA Hyperdiploid NA Pseudodiploid Hyperdiploid Hvrxrdiuloid

M
M M F M M F F F F M F F M F F F F F F M

3 4 5 6
7

W W
W W

8 9 10
11

W
W

W W
W W

I? 13 14 15 16 17 18 19 20 21 22 23 24

W W W W
W

W W W W W W

M
F

2.9 7.3 6.6 8.5 7.8 9.6 7.6 10.3 9.2 5.9 11.5 5.8 6.5 5.2 2.4 5.4 8.5 10.7 7.9 11.4

NA: not available; CALLA: negative for common acute lyrnphoblastic leukemia antigen.

Presence of cytoplasmic immunoglobulin.

that of normal diploid GO/Gl cells) was determined. This measure correlates closely with chromosome number (ploidy); hence, leukemic cells with a normal chromosome number have a DNA index of 1 .O.

induction and relapse or death due to any cause. Patients who did not enter complete remission were assigned a failure time of zero.

Cytogenetic Studies
Bone marrow samples were prepared for cytogenetic analysis by a direct technique developed for ALL and a modified trypsin-Giemsa technique for chromosome banding.I6 Cases were classified according to the International System of Human Cytogenetic Nomenclature (1985)''

Results
Of 1466 children with ALL, 24 ( 1.6%)presented with vertebral compression fractures. The initial clinical and laboratory characteristics of these patients are shown in Table 1. At diagnosis, the 15 girls' and nine boys' ages ranged from 2.4 to 18.3 years (median, 5.8 years), leukocyte counts from 1.8 to 85 X lo9/] (median, 6.2 X 109/1),hemoglobin levels from 2.4 to 11.5 g/dl (median, 7.7 g/dl) and platelet counts from 8 to 562 X 109/1 (median, 72 X 109/l).No patient had a mediastinal mass or central nervous system leukemia at diagnosis. The liver or spleen was palpable more than 5 cm below the costal margin in eight patients. Cell morphology was L1 in 13 patients and L2 in two patients tested. Of the 13 cases tested for leukemic cell immunophenotype, ten had common and one pre-B ALL; the remaining two cases did not express any lineage-specific markers or common ALL antigen. Notably, there were no cases of T-cell ALL. Of the 1 1 cases tested, seven had a blast cell DNA index of greater than 1.1 5 . Karyotype analysis of leukemic cells from 14 patients disclosed that nine cases were hyperdiploid with more than 50 chromosomes,

Statistical Analysis
Fisher's exact test for contingency tables was used to compare differences in the distribution of presenting clinical features and the remission induction rates for patients with or without vertebral compression fracture. Bonferoni correction was used to adjust for multiple comparisons." Since 1 1 features were compared, any P value less than 0.05/11 (0.0045) was considered to be statistically significant. Time-to-failure curves were constructed by the Kaplan-Meier procedure, and differences were analyzed by the Cox-Mantel test. Time to failure was defined as the interval between remission

Vol. 61

VERTEBRAL COMPRESSION FRACTURE ALL IN

Ribeiro ct a/.

59 1

two pseudodiploid (46 chromosomes with structural abnormalities), two diploid, and one hypodiploid (less than 46 chromosomes). Comparisons of the presenting clinical and biologic features disclosed that patients with vertebral compression fracture were significantly more likely to have leukocyte counts < 25 X 109/1and to have leukemic cell DNA index > 1.15 than were those without this complication (Table 2). They also tended to have leukemic cell hyperdiploidy. There were no significant differences between the two groups in sex, age, race, liver and spleen size, hemoglobin level, platelet count, and FAB morphology. The duration of initial symptoms ranged from 1 to 22 weeks (median, 7 weeks). Nineteen patients had back and leg pain as their initial chief complaint. Five of the patients were unable to walk because of intense pain. Despite seeking medical attention, the diagnosis of leukemia was delayed (5 to 13 weeks) for six patients because initial findings from physical examination, radiographic examination of the spine (three patients), and complete blood counts (two patients) were normal. Twenty-two patients had multiple vertebral compression fractures and two had only one vertebra involved. In addition to vertebral compression fractures, 1 1 patients had extensive spinal osteoporosis. Midthoracic and upper lumbar vertebrae were most frequently involved (17 and 14 cases, respectively). Seven patients had involvement of both thoracic and lumbar vertebrae. In no instance were cervical vertebrae affected. Two of the six patients tested had elevated serum calcium levels at diagnosis (1 l .7 and 12.7 mg/dl); serum parathormone level was elevated ( 1350 ng/l; normal < 885 ng/l) in one of the two. The hypercalcemia was corrected after antileukemic therapy but the serum calcium level was again elevated in one patient at relapse of leukemia. All patients completely recovered from vertebral compression fractures. Twenty-one patients were free of pain and ambulatory within 4 weeks from the beginning of antileukemic therapy. Only three patients required an orthopedic brace and physical therapy for 2 to 3 months. Complete remission was induced in all 24 patients and 16 remained free of leukemia for 2 months to 15.2 years (median, 4.3 years). When the 23 patients with vertebral compression fracture and presenting leukocyte counts of less than 25 X 109/1were compared to the 943 patients who also had leukocyte counts of less than 25 X 109/1 but had no such complication, there was no difference in the treatment outcome. Among those patients, 53% with vertebral compression fractures compared with 5 1% without the complication were estimated to be in complete remission 4 years after diagnosis.

TABLE Comparisons of Clinical and Biologic Features Between 2. Children With or Without Vertebral Compression Fracture at Diagnosis of ALL
Vertebral compression fracture Feature Leukocytes Category
<25 225 >].I5
51.15

Present
23 1 7 4 9 5 15 9 20 4

Absent
943 499 107 436 157 404 62 1 82 I 956 486

PValue
0.001
0.002 0.006 0.06 0.08

(x 109/1)
DNA index* KarYOtYpe ploidy*
Sex

Age (Yr)

H yperdiploidyt Others Girls Boys 2-10 <2 or >I0

A L L acute lymphoblastic leukemia. * Data not available for some patients. t >50 chromosomes.

Discussion

Children with vertebral compression fractures at diagnosis of ALL were demonstrated to be more likely to have leukocyte counts of less than 25 X 109/1, leukemic cell DNA indexes of greater than 1.15 and hyperdiploidy-favorable presenting features known to be associated with standard-risk Although vertebral compression fractures are associated with those classic indicators of good prognosis, results of our analysis disclosed that the presence of this complication has no independent prognostic value. Bone pain and radiographically detected skeletal changes are common in children with ALL. The radiographic abnormalities include radiolucent metaphyseal bands, periosteal reaction, osteolysis, and osteosclerThe vertebral column is involved less often than are the long bones. Spinal os?eoporosisleading to vertebral compression fractures has been reported rarely.- Baty and V0gt3 reported vertebral involvement f in two of 43 children with leukemia. Blatt e a/. described this complication in two of 350 patients with ALL. Leheup et al. noticed this complication in one of 250 patients. None of the 137 patients in a series studied by Rajantie et al. had vertebral compression fractures.20 Our study confirms the rarity of this axial skeletal lesion, occumng in fewer than 2% of our newly diagnosed patients with childhood ALL. The mechanisms that cause osteoporosis in children with ALL are uncertain. Release of parathormone or an osteoclast activating factor by leukemic cells has been implicated in the demineralization p r o ~ e s s . *In * ~ ~ ~ this study, serum calcium was increased in two of six patients tested at diagnosis, and serum parathormone was

592

CANCER February I 1988

No.3

increased in one patient. Alternatively, this complication is due to a long-standing, slowly proliferative and destructive process of B-cell precursor leukemic cells in bone marrow of the weight-bearing vertebrae. Since Tcell ALL usually involves extramedullary sites and proliferates rapidly, it may not be coincidental that none of our patients with vertebral compression fractures had this immunophenotype. Although many patients in this study had intense back pain that rendered some of them unable to walk, all but three patients had complete resolution of the pain and became ambulatory within 4 weeks from the start of antileukemic therapy. The other three patients required orthopedic braces for periods of 2 to 3 months. Our findings indicate that with antileukemia therapy, the symptoms associated with vertebral compression fractures will resolve rapidly and, except for the few patients with extensive vertebral involvement, special orthopedic care usually is not necessary. We conclude that vertebral compression fracture occurs rarely in children with ALL and is associated with good prognostic features. Complete resolution of the symptoms of vertebral compression fracture can be expected in them with antileukemia therapy. Prompt recognition of this complication could reduce the associated morbidity.
REFERENCES
I . Blatt J, Martini SL, Perchansky L. Characteristics of acute Iymphoblastic leukemia in children with osteopenia and vertebral compression fractures. J Pediatr 1984; 105:280-282. 2. Newman AJ, Melhorn DK. Vertebral compression in childhood leukemia. Am J D s Child 1973; 125:863-865. i 3. Baty JM, Vogt EC. Bone changes of leukemia in children. AJR
1935; 34:310-3 14. 4. Epstein BS. Vertebral changes in childhood leukemia. Radiology 1957; 68:65-69. 5 . Dubovsky D, Jacobs P. Vertebral rarefaction in acute lymphoblastic leukaemia. S Afi Med J 1975; 49:241-242. 6. DeCastro LA, Kuhn JP, Freeman AI, Sinks LF, Mindell ER.

Complete remodeling of the vertebrae in a child successfully treated for acute lymphoblastic leukemia (ALL). Cancer 1977; 40:398-401. 7. Leheup B, Membre H, Gerard H, Olive D. Lymphoblastic leukemia with osteopenia and vertebral compression fractures (Letter). J Pediatr 1985; 106:860. 8. Aur RJA, Hustu HO, Verzosa MS, Wood A, Simone JV. Comparison of two methods of preventing central nervous system leukemia. Blood 1973; 42:349-357. 9. Aur RJA, Simone JV, Verzosa MS er al. Childhood acute lymphocytic leukemia: Study VIII. Cancer 1978; 42:2 123-2 134. 10. Pui C-H, Aur RJA, Bowman WP et al. Failure of late intensification therapy to improve a poor result in childhood lymphoblastic leukemia. Cancer Res 1984; 44:3593-3598. 1 1. Pui C-H, Dahl GV, Bowman WP e a/.Elective testicular biopsy r during chemotherapy for childhood leukaemia is of no clinical value. Lancer 1985; 410-412. 12. Rivera GK, Evans WE, Kalwinsky DK et al. Unexpectedly severe toxicity from intensive early treatment of childhood lymphoblastic leukemia. J Clin Oncol 1985; 3:201-206. 13. Bennett JM, Catovsky D, Daniel MT et al. Proposals for the classification of the acute leukemias: French-American-British (FAB) co-operative group. Br J Haemarol 1976; 33:451-458. Williams DL, Kalwinsky DK et al. Cytogenetic fea14. Pui C-H, tures and serum lactic dehydrogenase. level predict a poor treatment outcome for children with p r e - k l l leukemia. Blood 1986; 67: 16881692. 15. Look AT, Roberson PK, Williams D L er al. Prognostic impor-

tance of blast cell DNA content in childhood acute lymphoblastic leukemia. Blood 1985; 65:1079-1086. 16. Williams DL, Harris A, Williams KJ, Brosius MJ, Lemonds W. A direct bone marrow chromosome technique for acute lymphoblastic leukemia. Cancer Genet Cyfogener 1984; 13:239-257. 17. ISCN: An international system for human cytogenetic nomenclature. In: Harnden DG,Klinger HP, eds. Birth Defects: Original Article Series, vol. 2 I , no. I . New York: March of Dimes Birth Defects Foundation, 1985. 18. Tukey JW. Some thoughts on clinical trials, especially problems of multiplicity. Science 1977; 198:679-684. 19. Pui C-H. Crist WM. High risk lymphoblastic leukemia in children: Prognostic factors and management. Blood Rev 1987; 1:25-33. 20. Rajantie J, Jliiskelainen J, Perkkio M, Siimes MA. Prognostic significance of primary bone changes in children with acute lymphoblastic leukemia. Pediatr Radio1 1985; 15:242-244. 21. Aur RJA, Westbrook HW,Riggs W Jr. Childhood acute lymphocytic leukemia: Initial radiological bone involvement and prognosis. Am J Dis Child 1972; 124653-654. 22. Ramsay NKC, Brown DM, Nesbit ME,Coccia PF, Krivit W, Krubik S. Autonomous production of parathyroid hormone by lymphoblastic leukemia cells in culture. J Pediarr 1979; 94:623-625. 23. Raisz LG, Luben RA, Mundy GR, Dietrich JW, Horton JE, Trummel CL. Effect of osteoclast activating factor from human leuln kocytes on bone metabolism. J C i Invest 1975; 56:408-413.