|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 12 (April 20), 2007: pp. 1498-1504 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.06.2539 High Body Mass Index Increases the Risk for Osteonecrosis in Children With Acute Lymphoblastic Leukemia
From the Departments of Pediatrics and Radiology, Oulu University Hospital, Oulu; and the Departments of Radiology and Pediatrics, Kuopio University Hospital, Kuopio, Finland Address reprint requests to Riitta A. Niinimäki, MD, Department of Pediatrics, Oulu University Hospital, Box 23, 90029 Oulu, Finland; e-mail: rsalonen{at}paju.oulu.fi
Purpose: The aim of the study was to determine the incidence of and clinical risk factors for radiographic osteonecrosis (ON) in children treated for acute lymphoblastic leukemia (ALL) using the Nordic ALL protocols. Patients and Methods: Ninety-seven consecutive patients with childhood ALL were studied prospectively by magnetic resonance imaging (MRI) of the lower extremities at the end of the treatment. Results: Twenty-three (24%) of the 97 patients had ON. Seven of the patients (30%) were symptomatic, and three patients (13%) required surgical interventions. Multiple logistic regression analysis showed that high body mass index (BMI; P = .04), female sex (P = .01), older age at diagnosis (P < .001), and higher cumulative dexamethasone dose (P = .03) were independent risk factors for radiographic ON. The cumulative prednisone dose did not differ significantly between the patients with and without ON. The incidence of radiographic ON decreased significantly, from 36% to 7%, when the duration of dexamethasone exposure during the delayed-intensification phase was shortened from 3 to 4 weeks to 2 weeks with a taper (P = .001). Conclusion: ON as determined by MRI was found to be a common complication in children and adolescents after treatment with the Nordic ALL protocols. Revision of the ALL protocols by shortening the single exposure to dexamethasone has diminished the risk for ON remarkably. High BMI was identified as a new significant risk factor for ON.
Approximately 80% of the patients with acute lymphoblastic leukemia (ALL) become long-term survivors.1,2 Along with the improved cure rate, the importance of treatment-related late effects has increased. Osteonecrosis (ON) has been recognized as an important complication of treatment for ALL.3-7 ON may cause severe pain, loss of function, and even joint damage and articular collapse.8 However, ON may remain asymptomatic and cause no disability.9 The proportion of patients needing to undergo orthopedic surgery for a variety of procedures and even total joint replacement in early adulthood is not known. The prevalence of symptomatic ON has varied from 1.1% to 9.3% in patients with ALL.3,4,10,11 The great variation in incidence may be a result of differences in treatment, but it could also be a result of variable reporting activity and imaging methods. Magnetic resonance image (MRI) has been established as the most sensitive method in the early diagnosis of ON.12-14 Only a few prospective studies have used MRI and included asymptomatic patients. Those studies showed the incidence of ON to range from 15.5% to 38%.6,9,15 The studies included patients examined during or immediately after treatment6,9,15 or up to 5.6 years after the diagnosis15 as well as relapsed patients.15 A prospective MRI study showed that ON developing during treatment mostly resolves during follow-up.16 This highlights the importance of uniform timing of imaging to find out the true prevalence. The etiology of ON is not well understood, and the risk factors are unclear. Corticosteroids, especially dexamethasone, are considered one main cause.17 Long continuous exposure to dexamethasone has been reported to increase the incidence of ON.18 Older age at diagnosis has been reported as a risk factor, but the role of female sex has been controversial.3,4 Body mass index (BMI) was not found to be a significant risk factor for ON in one study.15 The aim of this prospective MRI study was to define the prevalence of ON identified by end-of-therapy MRI screening among consecutive patients with ALL and to assess the clinical risk factors for radiographic ON. Better understanding of the risk factors will contribute to the development of ALL therapy and possibly allow the development of preventive interventions.
Patients Patients with primary ALL treated according to the Nordic ALL protocols were eligible for the prospective MRI study in two centers in Finland during the study period, which was from September 1992 to December 2005 in Oulu University Hospital, and from September 1994 to December 1999 in Kuopio University Hospital. Eligible patients had to be in continuous complete remission at the end of the treatment and to give informed consent. Patients referred for bone marrow transplantation and patients with Down syndrome were excluded from the study. A total of 122 patients with ALL were diagnosed and treated according to the Nordic ALL protocols during the study period. Reasons for exclusion from the study were death during the treatment or referral to a bone marrow transplantation (n = 9), Down syndrome (n = 4), and relapse before the end of the treatment (n = 3). Of the remaining 106 eligible patients, nine patients refused to give consent, which resulted in a participation rate of 91.5%. Ninety-seven patients were included in the analysis. There were 51 boys and 46 girls, and their mean age at diagnosis was 6.4 years (range, 1.2 to 15.3 years). MRI scanning covering the lower extremities from the femoral head to the ankle was performed at the cessation of therapy. The MRI results of 34 patients from Oulu obtained either during or after therapy have been reported earlier in two publications.6,9 Patient files were carefully reviewed manually by two of the authors (R.A.N. and A.H.H.-S.). Age at diagnosis, sex, height, weight at diagnosis and at the end of the treatment, any symptoms or procedures for ON, and disease and treatment details were recorded.
Childhood BMI values were computed and converted into percentiles for age and sex by using graphs based on the growth data of 2,514 Finnish children born between 1954 and 1972.19,20 Overweight was defined as a BMI between the 85th and 95th percentiles, obesity was defined as a BMI greater than the 95th percentile, underweight was defined as a BMI less than the fifth percentile, and normal weight was defined as a BMI between the fifth and 85th percentiles, all by sex.21 We also calculated the weight for height percentage, which is in clinical use in Finland.22 The diagnosis of being overweight and obese in childhood is based on the child's weight for height expressed as a percent deviation of weight from the mean weight for height and sex. Overweight is defined as a 10% to 20% higher weight than the mean in children younger than 7 years and a 20% to 40% higher weight than the mean in children age 7 years or older. Obesity is defined as a more than 20% higher weight or a more than 40% higher weight than the mean in children age younger than 7 years or Informed consent was obtained from the patients and/or the guardians. The research protocol was approved by the ethical committees of the Oulu and Kuopio University Hospitals.
Leukemia Therapy and Corticosteroid Dose The total cumulative doses of prednisone and dexamethasone on each therapeutic protocol are listed in Table 1. The IR 86 protocol included a delayed-intensification phase with dexamethasone 10 mg/m2/d for 4 weeks followed by a 10-day taper, and the HR and IR 92 protocols included dexamethasone for 3 weeks followed by a 7- to 10-day taper. The latest protocols, IR and HR 00, include an intensification phase with dexamethasone 6 to 10 mg/m2/d only for 2 weeks followed by a 2-week taper. In the other phases of treatment, dexamethasone is administered in short 5-day pulses in all protocols, and the minimum interval between the start points of these pulses is 4 weeks.
Some patients had also received remarkable cumulative doses of dexamethasone (up to 375 mg/m2) for the treatment of nausea or as prophylaxis during cranial radiation therapy. The actual received doses of prednisone and dexamethasone were carefully calculated from the patient files. The cumulative dose of corticosteroids per square meter was calculated as prednisone equivalents. Dexamethasone doses were converted to prednisone equivalents by multiplying the dexamethasone dose by 6.67 based on the anti-inflammatory properties of the two drugs.4,26 The cumulative doses of high-dose methotrexate were also calculated from the patient files.
Imaging Circumscribed lesions with a rim of low signal (T1-weighted images) or a double-line sign (T2-weighted images) were considered typical of ON. Patchy low-intensity lesions without a typical rim were also classified as ON. The involved sites of ON were first determined exactly and then classified according to the closest joints.
Statistical Analysis
Incidence and Influence of Sex, Age, and Corticosteroid and Methotrexate Doses ON was detected by MRI in 23 (24%) of 97 patients. ON was less frequent in the patients treated with the standard-risk protocol (6%) compared with patients treated with the IR (30%) or HR protocols (35%). The patients with ON were older at diagnosis than the patients without ON, and they had received significantly more dexamethasone compared with patients without ON. However, there was no difference in the prednisone equivalents. The incidence of ON was not associated with higher cumulative doses of methotrexate (Table 2).
Influence of the Duration of Dexamethasone Treatment The incidence of ON was significantly greater in the patients who had received dexamethasone for more than 3 weeks with a taper during the delayed-intensification phase compared with patients who had received dexamethasone for  2 weeks with a taper (Table 2).
Role of Obesity
Details of Patients With ON
Risk Factors in Logistic Regression Analysis In the logistic regression analysis in the univariate context, when ON was used as the dependent variable, older age at diagnosis, higher dose of dexamethasone, and higher BMI both at diagnosis and at the end of the treatment emerged as significant risk factors for radiographic ON. In addition, weight for height both at diagnosis and at the end of the treatment was identified as a risk factor for radiographic ON. Sex did not show a significant association with ON in the univariate analysis. The multiple logistic regression analysis demonstrated that older age at diagnosis, higher dose of dexamethasone, higher BMI at the end of the treatment, and female sex were independent risk factors for ON (Table 4).
In this prospective MRI study, which included consecutive patients with childhood ALL, ON was detected frequently (ie, in 24% of the MRI scans obtained at the end of the treatment). Most of the patients with ON were asymptomatic, but orthopedic procedures were performed frequently in patients with symptomatic ON (in three of seven patients). High BMI and high weight for height emerged as new risk factors for radiographic ON. Long continuous exposure to dexamethasone was found to increase the risk for ON significantly. In a previous large prospective MRI study including asymptomatic ALL and lymphoma patients, ON was found in 17 (15%) of 116 patients,15 which is less than in the present study. The time point of MRI varied from on therapy to off therapy for some years.15 It has been shown that some changes may already resolve during a short follow-up.9 In the study by Ribeiro et al,15 the protocol included only prednisone, and only 13 of 110 patients in the study had received any dexamethasone; however, the cumulative total dose of corticosteroids was markedly higher than in our patients. The slightly higher frequency of ON in our patients may be explained by the more frequent use and longer continuous exposure to dexamethasone.
Three large retrospective studies have reported symptomatic ON in 1.1% to 9.3% of children with ALL in remission,3,4,10,11 whereas the incidence of symptomatic ON in our study was 7%. In a Children's Cancer Group study of 1,409 patients up to 20 years of age with HR ALL, the incidence of ON was highest, namely 9.3%.3 ON was more common in patients who had received two 21-day dexamethasone courses versus one course, which parallels our finding that long continuous exposure to dexamethasone increases the risk for ON. The high ON rate may also be explained by the higher age of the patients compared with our study. In an Italian study, only 1.1% of the patients developed symptomatic ON.10 All patients had received Berlin-Frankfurt-Munster–type intensive chemotherapy for non–B-cell ALL, but only a small minority of patients belonged to the HR group. The authors suggest that reporting bias should be taken into account. A total of 176 ALL patients, who were High BMI and high weight for height percentage both at diagnosis and at the end of treatment were found to be significant risk factors for radiographic ON, indicating that obesity may notably contribute to the development of ON. There is only one previous study concerning the role of BMI in the development of ON in patients with leukemia, and that study failed to find an association.15 The phase of treatment at the time of the BMI measurement was not known, and the exact way of analysis was not reported. According to earlier studies, BMI is a good measure for adiposity.27,28 High BMI could cause ON by increasing the fat content of bone marrow or through the development of fat embolism in the marrow. Some studies suggest that abnormal lipid metabolism may be associated with the development of ON. In an autopsy study on corticosteroid-induced ON, the size of bone marrow fat cells was increased significantly at the early stages of ON development.29 A significant increase in bone marrow fat cell size has been reported in cortisone-treated rabbits,30 and the diameter of marrow fat cells was significantly larger in rabbits with ON compared with rabbits without ON.31 Obesity is a very common occurrence after ALL treatment and may cause significant psychosocial harm, which is why its prevention is thought to be important. Whether interventions to prevent obesity could decrease the risk for ON is not known. The dose of dexamethasone and the duration of dexamethasone treatment, but not the total cumulative dose of corticosteroids, were found to be risk factors for ON in this study. Most reports of children with leukemia and ON suggest that corticosteroids are the main pathogenetic factor.3,9,10,32 However, there are case reports on patients with malignancies in whom ON has developed after chemotherapy protocols without corticosteroids.33-35 In the only prospective follow-up study, ON was found to develop during treatment after delayed-intensification therapy including dexamethasone.9 Dexamethasone has been found to be more effective against CNS leukemia,36-38 but it seems to carry a risk for adverse effects such as ON, especially when continued for several weeks. Our study patients treated with older protocols had more ON than the patients treated with the latest protocols, which include a larger total dose of dexamethasone but administered for a shorter time. This indicates that the duration of dexamethasone treatment might be as equally important as the total dose. The cumulative dose of additional dexamethasone used for nausea was notable in some patients and may have contributed to the development of ON. The realization that high cumulative doses of dexamethasone are used against nausea in some patients has led us to avoid the routine use of dexamethasone for that indication. Methotrexate use is a known risk factor for osteoporosis.39 It has also been suggested to be involved in the development of ON. In our study, however, the higher cumulative dose of methotrexate was associated with a decreased risk of ON. This is probably because patients who received more methotrexate received lower cumulative doses of dexamethasone. Ten of the 23 patients with ON underwent follow-up MRI scanning 1 to 8 years after the end of the treatment. In the group of asymptomatic patients, the ON lesion had disappeared totally in one patient, whereas in eight patients, the lesions had remained unchanged or regressed in size. ON had progressed in one symptomatic patient, who thus underwent bilateral hip replacement. The clinical importance of these asymptomatic findings is uncertain, and MRI examination seems to be of benefit only in symptomatic patients. Only ON affecting the femoral head seems to progress and require surgical treatment. Remarkable progress has been achieved in the last few years in the survival of ALL patients. According to our study, the development of the Nordic treatment protocols has also decreased the debilitating adverse effect of ON. The role of obesity in the development of ON and the possible benefit of interventions against obesity for the prevention or treatment of ON require further studies.
The authors indicated no potential conflicts of interest.
Conception and design: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Raija M. Seuri, Pekka V. Riikonen, Eija L. Pääkkö Financial support: Marjatta Lanning Administrative support: Marjatta Lanning Provision of study materials or patients: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Raija M. Seuri, Pekka V. Riikonen, Merja I. Möttönen, Marjatta Lanning Collection and assembly of data: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Raija M. Seuri, Eija L. Pääkkö Data analysis and interpretation: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Eija L. Pääkkö Manuscript writing: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Raija M. Seuri, Pekka V. Riikonen, Eija L. Pääkkö, Merja I. Möttönen, Marjatta Lanning Final approval of manuscript: Riitta A. Niinimäki, Arja H. Harila-Saari, Airi E. Jartti, Raija M. Seuri, Pekka V. Riikonen, Eija L. Pääkkö, Merja I. Möttönen, Marjatta Lanning
Supported by grants from the Nona and Kullervo Väre Foundation, Finland. Presented at the 23rd Annual Meeting of the Nordic Society for Paediatric Haematology and Oncology, May 8-11, 2005, Lillehammer, Norway, and at the 38th Congress of the International Society of Paediatric Oncology, September 17-21, 2006, Geneva, Switzerland. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Brenner H, Kaatsch P, Burkhardt-Hammer T, et al: Long-term survival of children with leukemia achieved by the end of the second millennium. Cancer 92:1977-1983, 2001[CrossRef][Medline] 2. Pui CH, Evans WE: Acute lymphoblastic leukemia. N Engl J Med 339:605-615, 1998 3. Mattano LA Jr, Sather HN, Trigg ME, et al: Osteonecrosis as a complication of treating acute lymphoblastic leukemia in children: A report from the Children's Cancer Group. J Clin Oncol 18:3262-3272, 2000 4. Strauss AJ, Su JT, Dalton VM, et al: Bony morbidity in children treated for acute lymphoblastic leukemia. J Clin Oncol 19:3066-3072, 2001 5. Murphy RG, Greenberg ML: Osteonecrosis in pediatric patients with acute lymphoblastic leukemia. Cancer 65:1717-1721, 1990[CrossRef][Medline] 6. Ojala AE, Lanning FP, Paakko E, et al: Osteonecrosis in children treated for acute lymphoblastic leukemia: A magnetic resonance imaging study after treatment. Med Pediatr Oncol 29:260-265, 1997[CrossRef][Medline] 7. Korholz D, Bruder M, Engelbrecht V, et al: Aseptic osteonecrosis in children with acute lymphoblastic leukemia. Pediatr Hematol Oncol 15:307-315, 1998[Medline] 8. Mankin HJ: Nontraumatic necrosis of bone (osteonecrosis). N Engl J Med 326:1473-1479, 1992[Medline] 9. Ojala AE, Paakko E, Lanning FP, et al: Osteonecrosis during the treatment of childhood acute lymphoblastic leukemia: A prospective MRI study. Med Pediatr Oncol 32:11-17, 1999[CrossRef][Medline] 10. Arico M, Boccalatte MF, Silvestri D, et al: Osteonecrosis: An emerging complication of intensive chemotherapy for childhood acute lymphoblastic leukemia. Haematologica 88:747-753, 2003[Medline] 11. Burger B, Beier R, Zimmermann M, et al: Osteonecrosis: A treatment related toxicity in childhood acute lymphoblastic leukemia (ALL): Experiences from trial ALL-BFM 95. Pediatr Blood Cancer 44:220-225, 2005[CrossRef][Medline] 12. Robinson HJ Jr, Hartleben PD, Lund G, et al: Evaluation of magnetic resonance imaging in the diagnosis of osteonecrosis of the femoral head: Accuracy compared with radiographs, core biopsy, and intraosseous pressure measurements. J Bone Joint Surg Am 71:650-663, 1989 13. Saini A, Saifuddin A: MRI of osteonecrosis. Clin Radiol 59:1079-1093, 2004[CrossRef][Medline] 14. Mitchell DG, Rao VM, Dalinka MK, et al: Femoral head avascular necrosis: Correlation of MR imaging, radiographic staging, radionuclide imaging, and clinical findings. Radiology 162:709-715, 1987 15. Ribeiro RC, Fletcher BD, Kennedy W, et al: Magnetic resonance imaging detection of avascular necrosis of the bone in children receiving intensive prednisone therapy for acute lymphoblastic leukemia or non-Hodgkin lymphoma. Leukemia 15:891-897, 2001[CrossRef][Medline] 16. Ojala AE, Paakko E, Lanning FP, et al: Bone marrow changes on MRI in children with acute lymphoblastic leukaemia 5 years after treatment. Clin Radiol 53:131-136, 1998[CrossRef][Medline] 17. Gaynon PS, Lustig RH: The use of glucocorticoids in acute lymphoblastic leukemia of childhood: Molecular, cellular, and clinical considerations. J Pediatr Hematol Oncol 17:1-12, 1995[Medline] 18. Mattano L: The skeletal remains: Porosis and necrosis of bone in the marrow transplantation setting. Pediatr Transplant 7:71-75, 2003 (suppl 3)[CrossRef][Medline] 19. Sorva R, Lankinen S, Tolppanen EM, et al: Variation of growth in height and weight of children: II. After infancy. Acta Paediatr Scand 79:498-506, 1990[Medline] 20. Wei Y, Pere A, Koenker R, et al: Quantile regression methods for reference growth charts. Stat Med 25:1369-1382, 2006[CrossRef][Medline] 21. Whitaker RC, Wright JA, Pepe MS, et al: Predicting obesity in young adulthood from childhood and parental obesity. N Engl J Med 337:869-873, 1997 22. Sorva R, Perheentupa J, Tolppanen EM: A novel format for a growth chart. Acta Paediatr Scand 73:527-529, 1984[Medline] 23. Salo M, Anglé S, Kaukua J, et al: [Lasten lihavuus] Current care guidelines in Finland 2005: Childhood obesity. Duodecim 121:2015-2024, 2005 24. Gustafsson G, Schmiegelow K, Forestier E, et al: Improving outcome through two decades in childhood ALL in the Nordic countries: The impact of high-dose methotrexate in the reduction of CNS irradiation—Nordic Society of Pediatric Haematology and Oncology (NOPHO). Leukemia 14:2267-2275, 2000[CrossRef][Medline] 25. Gustafsson G, Kreuger A, Clausen N, et al: Intensified treatment of acute childhood lymphoblastic leukaemia has improved prognosis, especially in non-high-risk patients: The Nordic experience of 2648 patients diagnosed between 1981 and 1996—Nordic Society of Paediatric Haematology and Oncology (NOPHO). Acta Paediatr 87:1151-1161, 1998[CrossRef][Medline] 26. Haynes RC Jr, Murad F: Adrenocorticotropic hormone; adrenocortical steroids and their analogs; inhibitors of adrenocortical steroid biosynthesis, in Gilman AG, Goodman LS, Rall TW, et al (eds): The Pharmacological Basis of Therapeutics (ed 7). New York, NY, Macmillan Publishing Company, 1985, pp 1459-1489 27. Lindsay RS, Hanson RL, Roumain J, et al: Body mass index as a measure of adiposity in children and adolescents: Relationship to adiposity by dual energy x-ray absorptiometry and to cardiovascular risk factors. J Clin Endocrinol Metab 86:4061-4067, 2001 28. Pietrobelli A, Faith MS, Allison DB, et al: Body mass index as a measure of adiposity among children and adolescents: A validation study. J Pediatr 132:204-210, 1998[CrossRef][Medline] 29. Motomura G, Yamamoto T, Miyanishi K, et al: Bone marrow fat-cell enlargement in early steroid-induced osteonecrosis: A histomorphometric study of autopsy cases. Pathol Res Pract 200:807-811, 2005[CrossRef][Medline] 30. Wang GJ, Sweet DE, Reger SI, et al: Fat-cell changes as a mechanism of avascular necrosis of the femoral head in cortisone-treated rabbits. J Bone Joint Surg Am 59:729-735, 1977 31. Miyanishi K, Yamamoto T, Irisa T, et al: Bone marrow fat cell enlargement and a rise in intraosseous pressure in steroid-treated rabbits with osteonecrosis. Bone 30:185-190, 2002[Medline] 32. Vegna ML: Avascular necrosis of the femoral head in acute leukemia patients treated with combination therapy including steroids. Eur J Haematol 42:209-211, 1989[Medline] 33. Ishii E, Yoshida N, Miyazaki S: Avascular necrosis of bone in neuroblastoma treated with combination chemotherapy. Eur J Pediatr 143:152-153, 1984[CrossRef][Medline] 34. Harper PG, Trask C, Souhami RL: Avascular necrosis of bone caused by combination chemotherapy without corticosteroids. Br Med J (Clin Res Ed) 288:267-268, 1984[Medline] 35. Obrist R, Hartmann D, Obrecht JP: Osteonecrosis after chemotherapy. Lancet 1:1316, 1978[Medline] 36. Mitchell CD, Richards SM, Kinsey SE, et al: Benefit of dexamethasone compared with prednisolone for childhood acute lymphoblastic leukaemia: Results of the UK Medical Research Council ALL97 randomized trial. Br J Haematol 129:734-745, 2005[CrossRef][Medline] 37. Jones B, Freeman AI, Shuster JJ, et al: Lower incidence of meningeal leukemia when prednisone is replaced by dexamethasone in the treatment of acute lymphocytic leukemia. Med Pediatr Oncol 19:269-275, 1991[Medline] 38. Bostrom BC, Sensel MR, Sather HN, et al: Dexamethasone versus prednisone and daily oral versus weekly intravenous mercaptopurine for patients with standard-risk acute lymphoblastic leukemia: A report from the Children's Cancer Group. Blood 101:3809-3817, 2003 39. Pfeilschifter J, Diel IJ: Osteoporosis due to cancer treatment: Pathogenesis and management. J Clin Oncol 18:1570-1593, 2000 Submitted February 20, 2006; accepted October 23, 2006. This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
7 years, respectively.
2 test or Fisher's exact test was used to compare proportions. Univariate and multivariate logistic regression analyses were used to determine whether prospectively defined factors were useful in predicting a patient's ON risk. Data were analyzed using the Windows-based SPSS statistical package (version 11.0; SPSS Inc, Chicago, IL), and two-tailed P < .05 was considered significant. 
