Table of Contents
Children often present to the ED with a limp. When there is a history of trauma, the management is often straight forward. Without a history of trauma, the differential can be overwhelming.
The causes of limping or leg pain in a child can be as benign as "growing pains" or as malignant as a tumor. To avoid a "shot gun" approach in the evaluation of a child with a non traumatic limp, the emergency physician should be knowledgeable in the history and physical examination findings of the common etiologies of the limping child to avoid unnecessary tests and radiographs. In this issue of Pediatric Emergency Medicine Practice, we will review the available evidence on the evaluation of a child with a non traumatic limp.
Abbreviations Used in this Article
AP - Antero-posterior
CBC - Complete Blood Count
CRP - C Reactive Protein
CT - Computed Tomography
ED - Emergency Department
ESR - Erythrocyte Sedimentation Rate
HSP - Henoch Schonlein Purpura
JRA - Juvenile Rheumatoid Arthritis
LCPD - Legg-Calve-Perthes disease
MRI - Magnetic Resonance Imaging
NSAID - Nonsteroidal anti-inflammatory drug
ROC - Receiver Operator Characteristic
SCFE - Slipped Capital Femoral Epiphysis
WBC - White Blood Cell
Critical Appraisal of the Literature
While limping has been reported and studied for over 100 years,1 the past 10 years have yielded very little new literature in support of the ED approach to the diagnosis and treatment of limping. While there are several review articles discussing the evaluation and management of the child with a limp,2-12 there is a lack of methodologically sound studies examining this broad topic, specifically regarding the overall approach to the limping child. Most studies relate to the management of specific disorders once the diagnosis is already made. The key, of course, is getting to the correct diagnosis. In general, the history and physical examination will help narrow the focus in the extensive differential diagnosis and help guide the ED physician in the initial work-up and management.
Epidemiology, Etiology, and Pathophysiology
The literature on the epidemiology of the child with a limp presenting to the ED is quite limited. One study reported up to 4% of pediatric ED visits are for evaluation of limping or refusal to walk.13
Limping was more common in males than females (1.7:1) with a median age of 4.3 years.14
Toddlers are very active but have immature gaits leading to frequent falls.
Infections are common in this age group since the bony cortex is developing and there is little resistance to bacterial invasion. School-aged children are more ambulatory and rambunctious which increases their risk of injuries. Jumping off objects, such as bunk beds, trampolines, and trees can lead to injuries such as fractures, dislocations, and ligamentous injuries. In this age group, the bony architecture is more mature and resilient. Muscle strength has also increased dramatically. A SCFE is an example of how bone maturation, strength, and weight mismatches can result in problems.
Abnormal gait in children may be due to a wide variety of causes including infectious, inflammatory, musculoskeletal, and malignant. In a prospective study conducted in a pediatric ED during a 6 month period, Fischer et al found an incidence of acute atraumatic limp of 1.8/1000. Of the 243 subjects studied, 193 (80%) presented with pain; the hips and knees were the most common locations of pain, occurring in 33.7% and 19.3% of the subjects, respectively. The most common diagnosis was transient synovitis or "irritable hip," comprising approximately 40% of the subjects. Of note, 63 of the 243 subjects (26%) had no definitive diagnosis.14
Studies on septic arthritis and osteomyelitis have shown a tremendous decline in Haemophilus influenzab as a pathogen in pediatric septic arthritis in the late 1990's compared to the 1980's and earlier, largely due to widespread immunization programs.15,16 Most recent studies report Staphylococcus aureus as the predominant organism isolated in bacterial septic arthritis and osteomyelitis in all age groups, accounting for up to 53% of the cases.15-17 In Moumile's study, Kingella kingae was the second most common organism, occurring in 14% of the total isolates.17 This was followed by Streptococcus pyogenes and Streptococcus pneumoniae. In Luhmann's study, Staphylococcus aureus, Streptococcus pyogenes, and Enterobacter were the common organisms, followed by Kingella, Neisseria meningitides, Streptococcus pneumoniae, Neisseria gonnorhoeae, Candida, and Staphylococcus epidermidis.16 The pathogens involved may also exhibit some geographic variability. In a study conducted in Israel from 1988-1993, Kingella and Haemophilus influenza b were the two most common organisms isolated in patients less than 2 years of age with septic arthritis.18 Among neonates, Staphylococcus aureus remains the most common pathogen, followed by Escherichia coli and Group B Streptococcus.19 Salmonella was found to be the predominant organism causing osteomyelitis in patients with sickle cell disease, 20,21 though Staphylococcus aureus is also a common pathogen.
Walking is comprised of two distinct phases, the stance and the swing. In order for gait to be smooth and fluid, joint flexibility, pelvic rotation, pelvic tilt, balance, and strength all have to be unimpaired. Any aberration of the cycle will be noticed as a limp.
The type of limp can help determine the cause. Limps have been divided into three types: antalgic, trendelenburg, and short leg. The antalgic or "quick step" gait is a painful limp with a shorter stance on the affected or painful leg. It is commonly seen with traumatic injuries (fractures, sprains, or strains), tumors, or infectious etiologies. The Trendelenburg or "lurch gait" is a painless limp primarily due to musculoskeletal weakness. The affected hip drops down during the swing phase of the contralateral leg. The pelvis tilts into the affected side when standing. This may be seen with LCPD, SCFE, developmental dysplasia of the hip, and neuromuscular diseases (poliomyelitis). The developmental status of the child must be taken into consideration when assessing a gait disorder. Limping cannot be diagnosed until the infant can stand. Infants generally pull to a stand by nine months of age and start to "cruise" around while holding onto furniture or other items. Most children older than one year can walk unassisted. Toddlers initially have a wide-based gait. Intrinsic hip abductor weakness leads to a mild Trendelenburg gait and a noticeably shorter stance phase. By age three, children have assumed adult gait characteristics.
The differential diagnosis of limping is extensive. (See Table 1
). There may also be some variations in the common diagnoses based on the age of the patient. (See Table 2
). The following is a description of the various etiologies of limping in children.
Hip dysplasia has previously been called "congenital hip dislocation." The current accepted term of developmental dysplasia of the hip illustrates that the infant hip may be abnormally developed but not actually dislocated. The diagnosis is not always made in the nursery. It may not come to attention until the child becomes ambulatory, at which time a Trendelenburg gait may be noted. An ultrasound is used for diagnosis in early infancy, while plain films are more useful in the older child. Orthopedic referral is needed for further treatment.
Numerous rheumatologic illnesses have arthritis as a significant symptom. Therefore, they can present with limping as the primary complaint. Initially, it may be difficult to make the diagnosis, particularly if only one joint is involved. In these cases, it may be confused with a septic arthritis. Other systemic signs and symptoms, and chronicity of the arthritis may help make the correct diagnosis. For instance, an associated rash, eye abnormalities, recurrent fever, or other joint pain in the past may lead the physician toward a diagnosis of JRA. Psoriasis and inflammatory bowel disease may have arthritis associated with their other findings. Ironically, the arthritis can precede the skin changes in psoriatic arthritis. HSP may also present with joint pain and limping before the typical purpuric rash in 15% of the cases.22 Historical and examination findings with HSP include abdominal pain, palpable purpura (usually localized to the lower extremities and buttocks), joint pain, hematuria, hematochezia, hematemesis, headaches, and peripheral edema. Approximately 40% of patients may have had a preceding illness such as an upper respiratory tract infection.23, 24
Many infectious diseases include a component of arthritis or arthralgias either as an acute finding, or as a post-infectious or reactive arthritis. These include post-streptococcal arthritis, as well as viral illnesses such as hepatitis and Coxsackie virus.
A more serious bone and joint infection can be easily missed and result in permanent sequelae if the diagnosis is delayed. Findings include swelling, erythema, and tenderness near the area of infection. Osteomyelitis in the pediatric population is commonly spread hematogenously. Staphylococcus aureus and Streptococcus pyogenes (group A beta-hemolytic streptococcus) constitute the major pathogens. In some circumstances, other pathogens need to be considered. Neonates are susceptible to group B Streptococcus and Escherichia coli, adolescents to Neisseria gonorrhea, and sickle cell patients to Salmonella.15-17,19-21 In addition to examining the involved extremity, a complete evaluation should include the spine and pelvis to rule out a diskitis or sacroiliac disease.
Transient synovitis may present in a similar manner to septic arthritis. It is a self-limited inflammation of the synovial lining, usually resolving within 3 to 10 days of the onset of symptoms. It is the most common cause of painful limp in childhood, accounting for up to 40% of non traumatic limps.14 Although the cause is unknown, it is hypothesized that transient synovitis is due to a post-infectious phenomenon. It often presents following a viral illness or upper respiratory infection.
Lyme disease, a tick-borne infection due to Borrelia burgdorferi, also includes arthritis as a characteristic feature. Lyme arthritis can be classified according to the number of joints involved and duration. Episodic arthritis involves 1 to 4 joints for a duration of <1 week with recurrence at least 2 weeks later. Acute and chronic pauciarticular arthritis involves 1 to 4 joints with chronic form lasting >4 weeks. The disease includes the typical expanding target lesion (erythema migrans), involvement of central and peripheral nervous systems, and migratory arthritis. Bilateral Bell's Palsy can be seen with Lyme disease.
When a clear history of trauma precedes the limp, the diagnosis is often much easier to make. However, toddlers can present without a history for injury, or with a trauma or fall that is thought to be insignificant. Such a scenario is often found in children who have a fracture of the tibia, known as the toddler's fracture. (Please see Figure 1
The toddler's fracture is a spiral, oblique, non displaced fracture of the distal tibia, typically seen in children <3 years of age. Sometimes the fracture is only picked up on oblique views of the tibia, followup radiographs done weeks later, or by bone scan. Halsey et al found the most common symptoms of a toddler's fracture to be point tenderness and refusal to bear weight. The sensitivity and specificity of point tenderness (59% and 53%, respectively) and refusal to bear weight (82% and 30%, respectively) were found to be poor.25 In the same study, patients with a presumptive diagnosis of toddler's fracture were placed in a long leg cast or splint. However, 21 of the 59 patients did not have evidence of fracture even on follow-up x-rays. The final diagnosis in these patients is not known, and the inconvenience of taking care of a toddler in a long leg cast or splint is not discussed. Also, the question of what would have happened if these children were not splinted still remains unanswered.
Lower extremity fractures in non-ambulatory children should also raise suspicion for non-accidental trauma. Look for bucket handle fractures or corner fractures that are suggestive of child abuse. While spiral fractures are traditionally thought of as being suspicious for non-accidental trauma and are more likely to be investigated, they are not pathognomonic for abuse. Scherl et al found equal numbers of transverse and spiral femur fractures among cases with positive results of investigations for abuse.26 Mellick also found that isolated spiral tibial fractures are most commonly accidental.27 The history surrounding the event and other evidence of injury or neglect help determine whether child protective services investigation is warranted.
Toxic/metabolic causes include drugs, rickets, scurvy, and hyperparathyroidism. Surprisingly, some drugs can cause limb pain and present as limping. Vitamin A intoxication and carbamazepine have been implicated. Corticosteroids, ergotamine, phenothiazines and thiazides may cause muscle cramps. Hypercalcemia can cause bone pain. Rickets, the most common form of which in the U.S. is vitamin D deficient rickets, may also be associated with fractures and pain. Other metabolic diseases that affect the skeletal system include the mucopolysaccharidoses. All of these disorders usually involve some type of spinal deformity, with kyphosis being the most common.
Neoplastic causes of limping include benign bone tumors (unicameral and aneurysmal bone cysts), malignant bone tumors (Ewing's sarcoma, osteosarcoma), metastatic bone disease, and leukemia or lymphoma. Several case studies are reported with children presenting with limping and bony pain in which a subsequent diagnosis of a malignancy is made.28-31
Concurrent systemic signs and symptoms of fever, weight loss, lymphadenopathy, or hepatosplenomegaly should prompt further investigation with laboratory studies, particularly a CBC to evaluate cell lines, and radiographs as indicated by the examination.
Primary malignant bony tumors such as osteosarcoma or Ewing's sarcoma commonly involve the long bones, and thus may be seen in the lower extremity. Growing children are affected most often, with the peak incidence occurring in the pre-teen and early teen years. Spinal tumors, such as sacrococcygeal teratoma or bony tumors of the axial skeleton may also cause lower extremity weakness, back pain, and limping.
Benign bone lesions such as osteochondromas and osteoid osteomas may be found in the limping child. Osteoid osteomas are a common benign bone tumor that may occur in the lower extremity or back, causing limping and pain. It is common in early adolescence, with pain characteristically worse at night, and relieved by NSAIDs. Radiographs usually confirm the diagnosis.
Degenerative causes of limping include avascular necrosis of the femoral head, LCPD. LCPD is caused by osteonecrosis of the proximal epiphysis of the femoral head. LCPD is most likely the result of repeated trauma to the hip in active boys which may impair the blood supply to this area. It is most commonly seen in children ages 4 to 10 years. Patients typically present with chronic limping. While it is classically known as a "painless limp," there may be associated hip pain and referred pain to the ipsilateral knee or groin. Swelling and tenderness on examination are rare. Patients may also keep the hip externally rotated with limited abduction and internal rotation. Avascular necrosis of the femoral head has also been reported to be increased in children with renal failure. In one study, it was noted to occur in 1 out of 15 children with renal failure, which is a significantly higher rate than the general pediatric population.32
Children who are being treated with corticosteroids may be at a higher risk, although avascular necrosis has been noted to occur in the absence of steroid use.32
Patients with sickle cell disease may also be at increased risk for developing avascular necrosis of the femoral head. Laboratory tests are generally normal, and plain radiographs of the hips make the diagnosis once significant necrosis has occurred.
Osgood Schlatter's disease is characterized by pain over the tibial tuberosity. This disease is felt to be a result of repeated microtrauma to the insertion of the patellar tendon, similar to Sever's disease, which is pain in the area of the insertion of the Achilles tendon to the calcaneus. It is more common in adolescent boys, and may be bilateral 25-50% of the time.33 The diagnosis is made based on the presence of tenderness to palpation over the tibial tuberosity. In addition, pain is exacerbated with repeated activity, particularly jumping or knee extension against resistance. Resolution occurs when the secondary ossification center fuses to the proximal tibia, which occurs with maturation. Treatment consists primarily of limiting activity, immobilization, and NSAIDs.
SCFE is a medial or posterior slipping of the femoral capital epiphysis. The typical patient with SCFE is an overweight adolescent male. Manoff et al studied the association between body mass index and SCFE. They retrospectively reviewed 106 subjects with radiographically confirmed SCFE to 46 normal controls. In the SCFE group, 81.1% had a body mass index greater than the 95th percentile, compared to 41.3% in the control group (p< 0.0001).34 However, it is also seen in tall, thin adolescents who have undergone a recent growth spurt, resulting in shearing stress on the weakened epiphysis. SCFE will often present with a painful limp, hip or groin pain, or referred knee pain. Knee pain was found to be the primary presenting symptom in 15% of the patients with SCFE.35 Physical examination finds the hip externally rotated with painful range of motion, especially internal rotation, abduction, and flexion. Plain radiographs are generally the initial step in the diagnosis. Up to 60% of patients may develop bilateral SCFE, with about 23% having bilateral slips at the time of initial presentation. 36 There is also an association of SCFE with endocrine abnormalities such as hypothyroidism, panhypopituitarism, and hypogonadism. Such abnormalities should be suspected and evaluated in a younger child with SCFE, short stature, and hypogonadism. These children are also more likely to have bilateral SCFE.37 Adelay in diagnosis of SCFE results in increased slip severity and potentially higher risk of long-term complications.35 Treatment involves emergent orthopedic consultation for internal fixation.
Osteochondritis dissecans is a disease in which a small island of bone dies and is then sloughed. Typically, the child complains of poorly localized knee pain. Osteochondritis dessicans commonly presents in the preteen or early adolescent period. Pain with full flexion is usually found. If there is a piece of sloughed bone, one may find an effusion. Locking of the knee can occur while in flexion. Typically involving the distal femur, osteochondritis dessicans can be easily identified on plain AP radiographs of the femur. Treatment includes immobilization, isometric exercises to retain quadriceps tone, and pain control. Arthroscopic surgery is indicated for continued pain. (Please see the May 2006 issue of Pediatric Emergency Medicine Practice of pain management.)
Neuromuscular conditions which cause weakness in the hip girdle, lower spine, or extremities can present as limping. These include muscular dystrophies, myelitis or myositis, neuropathies, demyelinating disorders, or disorders involving the neuromuscular junction. A thorough discussion of these disorders is beyond the scope of this review article.
The role for pre-hospital care for non traumatic limping is limited. In cases of an obvious traumatic injury, the affected extremity should be immobilized and splinted in a position of comfort until further evaluation and management can be completed in the ED.
Emergency Department Evaluation
Given the broad differential diagnosis of the child with a limp, the emergency physician must rely a great deal on the history and physical examination findings to help narrow the possible range of diagnoses, and thus guide the appropriate evaluation and management of the patient.
When evaluating a child with a limp, a complete and comprehensive history is first and foremost in helping to make the correct diagnosis. Some typical characteristics of the history or examination in certain diseases can help narrow the differential diagnosis. (See Table 3
.) A history of associated symptoms such as joint edema, erythema, and pain may be suggestive of an infectious etiology such as a septic joint or osteomyelitis, or of a rheumatologic condition if the symptoms are recurrent or chronic.
Constitutional symptoms such as fever, pallor, easy bruising, chills, and weight loss with limping or bony pain should prompt an investigation for malignant causes of limping. Chronic limping or pain with a lack of other associated symptoms are more likely to occur with degenerative causes such as LCPD or SCFE. Also, a preceding illness such as an upper respiratory tract infection, pharyngitis, or scarlet fever is often noted with transient synovitis or post-infectious (post-strep) reactive arthritis. A history of abdominal pain, rash, travel history, or tick bite may also be useful in determining the etiology of the limp.
The physical examination of the child with a limp should be done in a systematic, head to toe manner. Specific examination findings may be suggestive of the etiology of the limp. Fever noted on the vital signs should lead one to suspect infectious etiologies. Certain rashes or skin findings can also help lead to the diagnosis. A palpable purpuric rash over the lower extremities with joint pain or limping suggests HSP. Erythema nodosum can be seen with inflammatory bowel disease which occasionally has associated arthritis causing limping. In addition, a target lesion (erythema migrans) is seen with Lyme disease. Localized erythema, warmth, edema, and pain are indicative of infectious causes such as osteomyelitis or septic arthritis. Ophthalmic abnormalities, such as iritis, in conjunction with arthritis can be seen with rheumatologic diseases. Also, lymphadenopathy or hepatosplenomegaly in the limping child is certainly concerning for malignancy as a possible etiology. Back pain, or pain with palpation over the spine may be elicited in discitis.
Evaluating lower extremity muscle bulk, strength, sensation, and deep tendon reflexes are also key aspects of the physical examination. Lack of deep tendon reflexes suggests Guillain-Barre syndrome, with limping due to weakness. Calf muscle hypertrophy may be seen with certain muscular dystrophies. The position that the legs are held in may also help localize the abnormality. For instance, patients with SCFE or LCPD often keep the affected hip externally rotated and slightly flexed. A positive Galleazzi test, with asymmetric knee heights with the patient supine, hips and knees flexed, is suggestive of developmental dysplasia of the hips. This test is more appropriate in an older child, as opposed to the Ortolani-Barlow maneuvers done on the neonate.
Local tenderness, masses, or swelling can be seen with osteomyelitis or bony tumors. Tenderness over the tibial tuberosity in an otherwise well child is diagnostic for Osgood Schlatter's disease. An evaluation of range of motion of the hips, knees, and back can also help localize the source of limping. Of course, observing the child's gait is a key part of the exam. The physician should determine if there is an antalgic gait, a Trendelenburg gait, or if it is actually ataxia, which would lead to an entirely different set of diagnostic considerations, such as cerebellar tumors. The child's shoes, socks, and, if possible, pants should be removed for the examination of gait.
Generally, laboratory studies are not a key component in the evaluation of the limping child. Laboratory tests may help to support a diagnosis already made on clinical grounds. Imaging studies can be more useful, depending on the etiology of the limp. Much of the literature on the usefulness of laboratory studies in the evaluation of the child with a limp centers around the WBC count, ESR, and CRP, and their use with septic arthritis, osteomyelitis, and transient synovitis. In general, studies have shown a high sensitivity but low specificity for the use of CRP and ESR. However, many of the studies have small sample sizes, poorly designed methodology, and varying standards for what are considered normal values.
Complete Blood Count
The total WBC count has been studied to determine its usefulness in predicting septic arthritis versus transient synovitis. Eich et al found that leukocytosis (defined by age ranges) was not significantly different between patients with septic arthritis and transient synovitis.38
Del Beccaro and Zawin found similar overlap in the WBC count between these groups.39,40
Kocher et al found a significant difference in total WBC count between patients with septic arthritis and transient synovitis in a study aimed at developing a clinical prediction algorithm.41
However, this algorithm has not been adequately validated.
The CBC may be useful when malignancy is high on the list of possible diagnoses for the patient. Thrombocytopenia, anemia, neutropenia, and/or evidence of blast cells on the peripheral smear certainly support the diagnosis of a malignancy. In addition, in a child in whom HSP is suspected, a CBC should be done to rule out thrombocytopenia as a cause of the purpura.
The ESR has consistently been found to be higher among patients with septic arthritis versus those with transient synovitis.38, 39, 41, 42
However, there is no agreement as to a cut-off value above which one can make the diagnosis of septic arthritis. For instance, Eich et al found a mean ESR of 103 mm/h among patients with septic arthritis.38
However, they recommend using a value >20 mm/h to suggest septic arthritis. Yet, their own data shows that 25% of the patients with transient synovitis also have an ESR >20 mm/h. This illustrates that while patients with septic arthritis have a statistically significantly higher ESR than those with transient synovitis, the value is not very specific but can be highly sensitive depending on the cut-off value chosen.
Several studies found a mean ESR among patients with septic arthritis to be approximately 50 mm/h.41, 43, 44 In a study of 26 patients, Klein reported a sensitivity of 91% for an ESR >30 mm/h, and 95% for an ESR >20 mm/h.44 However, specificities were not given. Del Beccaro et al showed an increased relative risk of 4.96 and 5.52 for having septic arthritis versus transient synovitis with an ESR of >20 mm/h and >30 mm/h, respectively. When calculating the relative risk of septic arthritis versus transient synovitis with an ESR of >20 mm/h and/or temperature >37.5 °C, the relative risk was found to be 22.48.39 However, in all these cases, the 95% confidence intervals were wide, likely due to the small sample sizes. In addition, Levine et al calculated an area under the ROC curve of 0.61 for ESR.45 Likelihood ratios for having septic arthritis did not have significant 95% confidence intervals with different ESR values, even with an ESR >75 mm/h. The ESR alone cannot make the diagnosis of septic arthritis, but may be useful in conjunction with other historical and examination findings that are suggestive of septic arthritis.
CRP is an acute phase protein that generally rises rapidly, reaching a peak value around the second day of illness, then normalizing by 7-10 days with appropriate therapy.43,46
Among patients with septic arthritis, Kallio found a mean CRP of 8.5 mg/dL.43
Eich's study showed a mean CRP of 9.9 mg/dL.38
In that study, while all patients with septic arthritis had a CRP of >2.0 mg/dL, 14% of patients with transient synovitis also had a CRP of >2.0 mg/dL.
In a retrospective study on 278 children with multiple different etiologies of arthritis, Kunnamo found that a CRP of >2.0 mg/dL had a sensitivity of 94% and a specificity of 92% for the 18 patients with septic arthritis.47 Yet the positive predictive value was only 57%, with a negative predictive value of 99%. A higher CRP of >4.0 mg/dL had a lower sensitivity of 71%, but higher specificity of 98%. Levine et al found sensitivities of 41-90%, with specificities of 29-85%, depending on the cut-off value chosen.45 The area under the ROC curve for CRP was 0.72, compared with 0.61 for the ESR. This study also concluded that the CRP may be a better negative predictor than a positive predictor of septic arthritis, with a NPV of 87% if the CRP is <1.0 mg/dL.
If septic arthritis is suspected, then fluid from the appropriate joint should be aspirated for cell count and culture. While synovial fluid culture is the current accepted gold standard for diagnosing septic arthritis, studies have shown that 30-70% of cases have no growth on culture.48,49 In general, nucleated cells of >50,000/mm3 in the joint aspirate is concerning for septic arthritis.
A urinalysis is recommended as part of the evaluation for patients suspected of having HSP. Renal involvement can persist long term and is a major cause of long-term morbidity among patients with HSP.50 Evidence of nephritis may not be evident until several weeks after initial presentation.51 Urinalysis may also be done as part of the evaluation for abdominal pain that may be causing a change in gait, as in appendicitis, testicular pain, urinary tract infections, or nephrolithiasis.
Additional serum testing may be indicated based on the history and physical findings of the patient. Among patients with HSP, checking stool for occult blood may be warranted, particularly in cases with severe abdominal pain. Gastrointestinal bleeding and intussusception are known complications of HSP. If positive, the patient should be admitted, or at least observed for a longer period of time for serial examinations and further management, if needed. In addition, a chronic history of recurrent arthritis with other systemic features such as fever, anemia, serositis, and/or rash, may prompt investigation for rheumatologic causes of limping. These tests include antinuclear antibodies (ANA), rheumatoid factor (RF), and anti-double stranded DNA. Also, evidence of proximal muscle weakness with other characteristic findings of rash may warrant checking creatinine kinase (CK) and aldolase levels for dermatomyositis. In areas endemic for Lyme disease, with a history of a tick bite, Lyme titers may be indicated as well. These tests might best be done by the primary care physician or with the guidance of a rheumatologist, as results generally will not be readily available to the emergency physician and follow-up will be indicated.
Plain radiographs are often obtained as the initial evaluation of a child with a limp. When an area of pain or swelling is localized, both AP and lateral films should be obtained. When the pain cannot be localized, as in a nonverbal child, plain radiographs from the hips to feet can identify a fracture in 1/5 of patients presenting with a limp.52
When considering a diagnosis of SCFE, LCPD, or fracture, plain radiographs are usually adequate in the initial ED evaluation of the child. If a child presents with knee pain as a possible site for the limp, obtain films above and below the joint in question (i.e. the ankle and the hip).7
If the plain films are inconclusive, it may be beneficial to repeat the films in 10 - 14 days. Followup films may reveal a healing fracture that was not evident on initial presentation.53
LCPD is usually diagnosed by plain radiographs with AP and frog leg lateral views. Typically the radiograph shows sclerosis, flattening, or fragmentation of the femoral head. Early in the illness, a small femoral head with a possible widened medial joint space is seen, especially in comparison with the contralateral side. As the disease progresses, a crescentshaped radiolucent line, the crescent sign, appears along the proximal femoral head. Later in the diseases progression, the femoral head can become more radiopaque with subsequent fragmentation and collapse of the epiphysis. (See Figure 2). No recent studies have evaluated the diagnosis of LCPD. Instead, most studies have focused on outcome and management of LCPD by different radiographic methods including plain films, bone scan, ultrasound, CT, and MRI.54-60
The key to the diagnosis of SCFE is obtaining the appropriate radiographs and interpreting them correctly. Obtaining an AP view and a special frog leg lateral view is usually recommended because the AP view may fail to show the displacement in 14% of cases.61 (See Figure 3). Also, radiographs of bilateral hips can aid in the diagnosis with the unaffected hip serving as a comparison view. On an AP view, the "Klein line" is drawn along the lateral cortex of the femoral neck and should intersect a portion of the lateral femoral head. If it does not, then a medial slip of the femoral epiphysis is indicated. 62,63 (See Figure 4). Comparison to the contralateral hip may also reveal a smaller portion of the femoral head above the "Klein line" that can also suggest an early slip of the involved hip. On a frog leg lateral view, the slip may be more obvious. On a lateral view, if a line passing through the center of the femoral neck does not intersect the center of the femoral head, a posterior slip is suggested. A subtle finding of an early SCFE may be a physis that is widened or blurred (Bloomberg's sign) compared to the contralateral hip.64,65
Plain radiographs have a limited role in the emergent evaluation of septic arthritis. It has also been found to be unreliable for the diagnosis of a septic hip.66 In the presence of a widened joint space, the likelihood of septic arthritis is raised, but most plain radiographs are found to be normal even in the presence of a joint effusion. In a small retrospective study of children diagnosed with septic arthritis, Gandini showed that 12% of plain radiographs were found to be abnormal, in contrast to 83% of ultrasonographic studies showing an effusion. 67 Marchal also found that, of 21 children with transient synovitis, a joint effusion was detected on 20 patients with ultrasonography. But only 8 (42%) were found to have an increased joint space on plain radiographs.68
Ultrasonography examination of the hip has been shown to identify most cases of painful hips. In Germany, ultrasonography was able to make a diagnosis in 90% of children; 53 children with transient synovitis, 47 with LCPD, 28 with SCFE, and 3 with septic arthritis. They concluded that ultrasonography is a useful tool for the diagnosis of a child with a painful hip.69
No studies have been conducted comparing ultrasonography to plain radiographs in the ED evaluation of LCPD. Few studies have compared other radiographic modalities to plain radiographs in the diagnosis of SCFE. Billing et al believed that ultrasonography had not been proven to add much to the definitive diagnosis of SCFE.70
In contrast, in a small study of 21 symptomatic overweight children with a possible diagnosis of SCFE, Magnano et al showed a sensitivity rate of 66% for the diagnosis of SCFE with plain AP radiographs. With the addition of a frog leg lateral view, the sensitivity rate increased to 80%. Ultrasonography and MRI revealed a higher sensitivity in the detection of SCFE, 95% and 88%, respectively. They recommended that ultrasonography be considered for the early detection of slipped capital femoral
Most studies on ultrasonography have been focused on detecting an effusion in children with hip pain and septic arthritis. In a study conducted by Miralles, 500 children were prospectively evaluated by plain films and ultrasonography. Only 58 plain films were found to be abnormal when ultrasonography detected a hip effusion in 235 children.72 Gordon et al carried out ultrasonography in 132 children with hip pain. Follow up to determine the absence or presence of septic arthritis was conducted with 73 of the patients. Only four patients were found to have no effusion on initial ultrasonography but were later determined to have septic arthritis. Two had inadequate initial studies, and the other two had symptoms for <24 hours.73 Tien et al found 31 of 40 patients with suspected septic arthritis had an ultrasound confirmed joint effusion. Of these patients, 22 were found to have confirmed septic arthritis by needle aspiration of the joint.74
Another cause of hip effusions is transient synovitis. Marchal found that, of 21 children with transient synovitis, a joint effusion was detected on 20 patients with ultrasonography.68 Distinguishing between septic arthritis and transient synovitis is difficult as the two share many clinical and laboratory features. Several studies investigating the echogenicity of the effusion to differentiate septic arthritis from transient synovitis have had conflicting results. Dorr and colleagues found that 13 patients with confirmed septic arthritis had effusions on ultrasonography that were non-echofree. Of 58 patients with transient synovitis, 42 were echofree, 12 had a low level of echogenicity, and four had a very small amount of effusion that could not be classified confidently.75 In a prospective controlled study, Zieger also found that transient synovitis had effusions to be echofree, whereas septic arthritis showed non-echofree effusions. 76 In contrast, Marchal found increased echogenicity in the effusion of patients with transient synovitis, but no echogenicity in the one patient with a final diagnosis of septic arthritis.68 In 235 patients with a hip effusion, Miralles also found that no sonographic signs served to differentiate sterile, purulent, or hemorrhagic effusions.72 In an animal study of rabbits, Strouse et al demonstrated increased synovial vascularity in approximately 50% of septic arthritis cases by Doppler ultrasonography.77 In clinical practice, hypervascularity proved to be less useful showing only 1 of 11 patients with a positive finding on Doppler ultrasonography.78 Thus, while ultrasonography may be useful in detecting joint effusion, the specifics of the findings are conflicting and would be operator and viewer dependent.
When the history and physical examination do not localize a source of a child's limp and plain radiographs do not show any evidence of pathology, a bone scan may help to localize the site in question. A bone scan uses IV technetium 99m-labeled methylene diphosphonate that accumulates in areas of increased cellular activity and blood flow. Choban and Killian reviewed the records of 60 children aged less than five years with limp. 35 bone scans were performed that led to 18 definitive diagnoses including synovitis, osteomyelitis, LCPD, JRA, soft tissue infection, and discitis.89 In another prospective series of 50 limping toddlers that had no clear diagnosis, a bone scan localized the lesion in 27 patients (54%) with no infections being missed.80 Goergens showed that 88% of bone scans were found to be positive in a series of patients with a diagnosis of septic arthritis.81
CT is rarely needed in the emergency evaluation of a child with a non traumatic limp. It can be useful for a more detailed evaluation of a possible bony etiology for the limp. Most studies on CT scans and limping children have been centered on the severity and long term management of children.82-85 No studies were directed at the ED evaluation of a child with a non traumatic limp.
Magnetic Resonance Imaging
MRI has become increasingly important in the evaluation of possible infectious etiologies in children with a limp. The utility of MRI in the ED evaluation of a child with a limp has a limited role, as further diagnostic evaluation can be completed non-emergently either on an inpatient or outpatient basis. The MRI is useful for the detection of bone marrow and soft tissue lesions as well as joint effusions. The sensitivity and specificity of detecting a joint effusion was found to be 100% and 77%.86 Lee et al attempted to describe any differential findings on MRI between septic arthritis and transient synovitis. Nine patients with septic arthritis and 14 patients with transient synovitis were evaluated and there was significantly increased signal intensity in patients with septic arthritis (p = < 0.001).87 In a prospective study of 45 children with acute hip pain, Ranner compared plain radiographs, bone scan, and MRI. The study found that the MRI provided more morphologic information than the other techniques.88 In a small study of 13 children with SCFE, Umans et al compared radiographs and CT with MRI in the diagnosis of early physeal lesions for a "pre slip." Of the one child with a pre slip of the femoral head, only the MRI clearly delineated the physeal changes of the pre slip.89 Pinto et al noted that, in two patients suspected of having LCPD, the plain radiograph and bone scan were found to be normal. An MRI was obtained and the diagnosis of LCPD was made by showing definite avascular necrosis of the affected hip.90
Treatment for the child who presents with a limp is, of course, dependent on the ultimate etiology of the limp. Some primary considerations are pain control, antibiotic therapy if indicated (i.e. for septic arthritis or osteomyelitis), splinting, or need for further subspecialty consultation, such as with orthopedics, rheumatology, or oncology.
Analgesic medication may include acetaminophen, NSAIDs, and narcotics, depending on the severity of the pain. In a randomized, blinded, placebo controlled study, Kermond et al found that NSAIDs shortened the duration of symptoms in children with a clinical diagnosis of transient synovitis of the hip.91
If a diagnosis of SCFE or LCPD is suspected, the patient should be non-weight bearing, and an urgent orthopedic referral should be made. Goals of therapy for LCPD include resolving the hip joint inflammation and positioning the femoral head in such a way as to promote healing. If 50% or more of the femoral head can be seen, therapy is aimed at maintaining the range of motion until healing occurs. When more than 50% of the head is involved, orthopedists will place the hip in an abduction brace or hip spica cast, or perform an osteotomy of the proximal femur to position the femoral head well into the acetabulum. Usually the brace is used for children under seven, and osteotomy reserved for the older patient. Ultimately, the outcome for those who present with less than 50% of their femoral head involved is good. Children under six do very well, but those over 10 or those with flattening of the femoral head do not. Treatment for SCFE involves in situ pinning to prevent further slippage.
The use of glucocorticoids has been studied for patients with HSP. In general, the routine use of prednisone in HSP is not recommended. Patientsmay be treated with steroids to relieve abdominal pain. However, Rosenblum and Winter showed that abdominal pain in patients with HSP is largely selflimited. 92 By 72 hours, there was no difference in the number of patients still complaining of abdominal pain between the group treated with prednisone and the group not treated with steroids. Huber et al conducted a randomized, placebo-controlled trial of prednisone in early HSP.93 They found no significant reduction in the risk of renal involvement or gastrointestinal complications with early prednisone therapy when compared to placebo. However, the total study population of 40 patients may be too small to detect relatively rare events such as intussusception. Mollica et al did find a significant difference in reduction of the development of nephropathy among patients with HSP.94 However, this study was not randomized or blinded. Most other studies are retrospective studies, with mixed results. Other studies have shown improvement in the progression of nephritis in more severe cases of HSP, where there is already evidence of significant nephropathy among patients treated with corticosteroids.95-97 The studies on the prevention of intussusception with the use of corticosteroids are generally limited by their retrospective nature and small number of patients.
Antibiotic therapy for suspected or proven infectious causes of limping such as osteomyelitis or septic arthritis should be guided by the age group, the common pathogens in the geographic area for that age group, and susceptibility patterns of the area. In general, therapy should include anti-staphylococcal coverage such as nafcillin, oxacillin, clindamycin, or vancomycin. Broader, gram-negative coverage with, for instance, a third-generation cephalosporin, may also be appropriate pending culture results.
In June 2003, an interdisciplinary expert committee in Boston set a clinical practice guideline for the treatment of septic arthritis of the hip in children to try to improve the process of care and the outcome of these children. They retrospectively reviewed the medical records of 30 children with septic arthritis as a control group and prospectively applied the clinical guideline in 30 consecutive patients seen at their hospital. The clinical guideline recommended that if a previously healthy patient between 6 months and 18 years of age is found to have a history and physical examination suggestive of a septic arthritis, the patient was to follow a set clinical guideline. Initial labs included CRP, ESR, CBC with differential, blood culture, throat culture and anti-streptolysin O antibody titers, along with radiographs of the hip. If the labs or radiographs suggested septic arthritis, then aspiration of the joint was done. However, the authors failed to define their laboratory or radiographic definitions for septic arthritis. If the joint aspirate showed WBC >50,000 /mm3 or a positive gram stain, then the patient was admitted for possible operative drainage or intravenous antibiotics. Once patients showed clinical improvement after 72 hours of treatment, they could be discharged home if able to tolerate oral antibiotics. Although this study showed a lower rate of bone scanning, lower rate of presumptive drainage, and a shorter hospital stay, there was no difference with outcome, readmission, or recurrent infection.98
Children with special needs and chronic illnesses present additional challenges to the emergency physician. For instance, patients with sickle cell disease may have limping due to pain from vaso-occlusive crisis. They are also at risk for avascular necrosis. If the pain is not a usual site for their pain crisis, radiographs may be indicated. In addition, as noted earlier, the pathogens involved in septic arthritis for sickle cell patients may be different. Hemophiliacs with hemarthrosis may also present with acute limping. These patients need recombinant factor VII replacement. In some cases, families can do this therapy in an outpatient basis. However, a discussion with the hematologist may be helpful in determining which cases should be admitted for closer observation and therapy. Finally, wheelchair bound patients often develop osteopenia. Caretakers may report crying with movement or difficulty with physical therapy. Diagnosis in this case is particularly difficult, as the etiology can range from constipation to urinary tract infection to fracture.
In recent years, procalcitonin has been studied as a marker to differentiate between inflammatory conditions and more serious bacterial infections. One prospective study has found elevated procalcitonin levels >0.5ng/ml to have a sensitivity of 43.5% and a specificity of 100% in identifying patients with skeletal infections (osteomyelitis and septic arthritis) from those with other causes of fever and limping (reactive arthritis, JRA, soft tissue infection).99 None of the patients with a diagnosis other than skeletal infection had an elevated procalcitonin level. In this study, there was no significant difference between the CRP, ESR, and WBC count among patients with skeletal infection vs. those with "other" diagnoses for their fever and limp. Another study found similarly low sensitivity with high specificity for procalcitonin in identifying patients with bacterial septic arthritis (55% and 94%, respectively).100 This study also showed that a CRP value of >5.0 mg/dL had a 100% sensitivity, but only 40% specificity. These are early studies with small sample sizes. Nonetheless, they do represent an interesting area of possible future research.
Clinical Practice Guidelines
An evidence based clinical prediction algorithm was developed by Kocher et al in an attempt to differentiate septic arthritis from transient synovitis. They retrospectively reviewed the medical records of 82 patients with septic arthritis and 86 patients with transient synovitis. Multiple logistic regression identified history of fever, non weight-bearing, an ESR >40mm/h, and a serum WBC of >12,000 /mm3
to be independent predictors for differentiating septic arthritis from transient synovitis. An algorithm based on the number of predictors found that having all four predicted 99.6% of those with septic arthritis, three predicted 93.1%, two predicted 40.0%, one predicted 3.0%, and having none of the four predicted <0.2%. When all four parameters were met, they also found excellent diagnostic performance of this algorithm on a receiver operating characteristic curve with an area under the curve of 0.96.41
Follow up studies of this clinical prediction rule have shown mixed results. In a follow up validation study by Kocher, applying the prediction rule prospectively to 51 patients with septic arthritis and 103 patients with transient synovitis revealed the same four predictors on multivariate analysis. Applying the algorithm with the same four predictors found the predicted probability of septic arthritis was 93.0% with four predictors, 72.85% with three predictors, 35% with two predictors, 9.5% with one predictor, and 2% with no predictor. The area under the receiver operating characteristic curve was 0.86. Although the clinical prediction rule did not perform as well as the initial study, they concluded that it maintained very good diagnostic performance in a new patient population.101
When applying the clinical algorithm to 163 patients, another validation study on a different study population found the four predictors proposed by Kocher had a predicted probability of a patient having septic arthritis of 59%, compared to the initial 99.6% published by Kocher.41,42
The majority of children who present to the ED with a chief complaint of limping have a minor, self-limited disease. These children can be safely discharged home with instructions for pain control if needed, and instructions to follow-up with their primary care physicians. Discharge instructions should include recommendations to return for re-evaluation if symptoms worsen, or new symptoms develop in addition to limping that may be concerning for a more serious etiology. These include fever, joint edema and erythema, new rash, weight loss, or hematemesis and/or hematochezia.
Some cases will require admission for further therapy and consultation. Among patients with suspected septic arthritis or osteomyelitis, the patient should be admitted for IV antibiotics, and, if needed, surgical drainage and debridement. In cases where pain is not adequately controlled with standard oral analgesics, admission for pain control and serial exams may be indicated. Suspected cases of neoplasm also warrant admission for further diagnostic evaluation, such as bone marrow biopsy or bony tumor biopsy. In addition, in traumatic cases where the history is not consistent with the injuries, and abuse is suspected, the child should be admitted for further investigation.
The differential diagnosis of patients who present to the ED with a complaint of a painful hip or a limp is extensive. Evaluating a child with a limp may be quite challenging if one does not use a systematic approach. The approach to narrowing the differental should begin with a thorough history and physical exam. When more serious conditions are suspected or cannot be ruled out, then further evaluation such as laboratory testing or further consultation is warranted. Selection of laboratory studies and imaging should be guided by the history and physical examination. Ultimately, the goal of the clinician is to exclude the more serious, life-threatening diseases and to prevent permanent impairment in function.
A thorough history and examination of a child with a limp will help focus and guide the ED evaluation.
Consider referred pain from the hip in a child with complaints of limping with knee pain.
Plain radiographs are generally the first step in the evaluation of a limping child without a fever.
Plain radiographs are generally adequate in determining the etiology of limping in a child without a fever.
The ESR and CRP are not specific to the diagnosis of septic arthritis, but can be useful in conjunction with the historical and physical examination findings.
In a non-ambulatory child with a lower extremity fracture, consider non accidental trauma.
The gold standard for the diagnosis of septic arthritis remains joint aspirate.
In a well-appearing child without a definitive diagnosis for limping and without concern for an infectious etiology, outpatient follow-up and re-evaluation can be appropriate therapy.
1. "I didn't think that a febrile child with a normal CBC could have a serious cause for his limp, so I discharged him home."
Studies to date have not shown the CBC alone to be a reliable predictor for septic arthritis. Although not adequately validated, a history of fever, non weight-bearing, an ESR >40mm/h, and a serum WBC of >12,000 /mm3 can be predictive of septic arthritis.
2. "Although the child with fever and pallor didn't have a source for his limp, an ESR was found to be elevated so I started him on steroids for a presumptive rheumatologic disease."
In the presence of concurrent systemic signs and symptoms of fever, pallor, weight loss, lymphadenopathy, or hepatosplenomegaly, a CBC should be ordered to evaluate cell lines to rule out leukemia. Steroid use prior to a thorough evaluation by a hematologist/oncologist can alter management of a child with leukemia.
3. "The patient complained of knee pain and had a normal plain radiograph of the knee. Since he walked with a limp, I placed him in a knee immobilizer and discharged him home with instructions to follow up with his primary doctor within a week."
Caution should be taken when evaluating a child with complaints of knee pain. If a patient presents with knee pain with a normal examination of the knee, consider further evaluation of the hip as a source of the limp.
4. "The history was concerning for a hip disorder but the AP and lateral radiograph of the hips werefound to be normal."
The addition of a frog leg lateral view is usually recommended. In the possible diagnosis of SCFE, a sensitivity rate of 66% was found with plain AP radiographs. The addition of a frog leg lateral view increased the sensitivity rate to 80%.
5. "I did not want to get social work involved as the parents seemed genuinely concerned for their child. But in the presence of a spiral fracture to the tibia without a clear history, I called child protective services."
This scenario is often found in children who have a non-displaced fracture of the tibia, known as the toddler's fracture. The toddler's fracture is a spiral, oblique non displaced fracture of the distal tibia, typically seen in children <3 years of age.
6. "I was uncertain whether this patient had transient synovitis or septic arthritis until an ESR wasfound to be 26 mm/h. Now I know he has septic arthritis."
No study to date has clearly shown a cut off value to predict septic arthritis in children. Using a cut off value of 20 mm/h has shown a high sensitivity for the diagnosis of septic arthritis, but a low specificity. Currently, the gold standard remains synovial fluid culture.
7. "The child has sickle cell disease and fever with a limp. No obvious swelling was noted on examination, so I treated him with narcotics to manage his pain."
If a patient with sickle cell disease presents with symptoms that are not consistent with their pain crisis, caution must be taken to ensure that they do not have osteomyelitis or a septic joint. A common pathogen to consider in a child with sickle cell disease is Salmonella or Staphylococcus aureus.
8. "I obtained a plain radiograph of the right hip due to pain and limping in an overweight male. I wassurprised to find it normal as I was highly suspicious of SCFE."
Obtaining bilateral hip films can be useful in the diagnosis of SCFE. A "Klein line" may appear to be normal, but in comparison to the contralateral hip, it may show a mild medial slip as evidenced by a difference in the "Klein line" intersecting the femoral head.
9. "The child had a limp with a normal examination. Since there was no focal area to radiograph, I discharged the patient home."
If no clear source is found for the limp, obtaining a radiograph of the hips to the feet can identify a fracture in 1/5 of children.
10. "I was considering a diagnosis of a septic hip in this patient, but since the plain radiographs showed no evidence of an effusion, I diagnosed her with transient synovitis."
Plain radiographs rarely have evidence of effusion in septic arthritis. Most effusions can be detected by ultrasonography; however, effusions do not differentiate septic arthritis from transient synovitis. A joint aspirate should be obtained in all cases of suspected septic arthritis.
Clinical Pathway Approach To The Acutely Limping Child With Fever
Clinical Pathway: Approach To The Acutely Limping Child Without Fever
Tables and Figures
Lovett R, Morse J. A transient or ephemeral form of hip disease with a report of cases. Boston Med Surg J. 1892;127:161. (Retrospective; 156 cases)
Barkin RM, Barkin SZ, Barkin AZ. The limping child. J Emerg Med. 2000;18(3):331-9. (Review)
Bowyer SL, Hollister JR. Limb pain in childhood. Pediatr Clin North Am. 1984;31:1053-81. (Review)
Dabney KW, Lipton G. Evaluation of limp in children. Curr Opin Pediatr. 1995;7:88-94. (Review)
Hensinger RN. Limp. Pediatr Clin North Am. 1986;33:1355-64. (Review)
Lawrence LL. The limping child. Emerg Med Clin North Am. 1998;16(4):911-29. (Review)
Leet AI, Skaggs DL. Evaluation of the Acutely Limping Child. Am Fam Physician. 2000;61(4):1011-8. (Review)
MacEwen GD, Dehne R. The limping child. Pediatr Rev. 1991;12(9):268-74. (Review)
Myers MT, Thompson GH. Imaging the child with a limp. Pediatr Clin North Am. 1997;44:637-58. (Review)
Nogi J. Common pediatric musculoskeletal emergencies. Emerg Med Clin North Am. 1984;2(2):409-423. (Review)
Phillips WA. The child with a limp. Orthop Clin north Am. 1987;18:489-501.) (Review)
Renshaw TS. The child who has a limp. Pediatr Rev. 1995;16:458-65. (Review)
Singer JI. The cause of gait disturbance in 425 pediatric patients. Pediatr Emerg Care. 1985;1(1):7-10.
Fischer SU, Beattie TF. The limping child: epidemiology, assessment and outcome. J Bone Joint Surg Br. 1999;81(6):1029-34. (Prospective; 242 children)
Fink CW, Nelson JD. Septic arthritis and osteomyelitis in children. Clin Rheum Dis. 1986; 12(2): 423-35. (Retrospective)
Luhmann JD, Luhmann SJ. Etiology of septic arthritis in children: an update for the 1990's. Pediatr Emerg Care. 1999; 15(1): 40-2. (Retrospective; 64 children)
Moumile K, et al. Bacterial aetiology of acute osteoarticular infectiouns in children. Acta Paediatrica. 2005; 94: 419-22. (Retrospective; 407 cases)
Yagupsky P, et al. Epidemiology, etiology, and clinical features of septic arthritis in children younger than 24 months. Arch Pediatr Adolesc Med. 1995; 149(5): 537-40. (Retrospective; 40 children)
Wong M, et al. Clinical and diagnostic features of osteomyelitis occurring in the first three months of life. Pediatr Infect Dis J. 1995; 14(12): 1047-53. (Retrospective; 30 infants)
Burnett MW, et al. Etiology of osteomyelitis complicating sickle cell disease. Pediatrics. 1998; 101(2): 296-7. (Review)
Chambers JB, et al. Retrospective review of osteoarticular infections in a pediatric sickle cell age group. J Pediatr Orthop. 2000; 20(5): 682-5. (Retrospective; 14 cases)
Trapani S, et al. Henoch Schonlein purpura in childhood: epidemiological and clinical analysis of 150 cases over a 5-year period and review of literature. Semin Arthritis Rheum. 2005; 35(3) 35-7. (Retrospective; 150 children)
Chao HC, Kong MS, Lin SJ, et al. Gastrointestinal manifestation and outcome of Henoch-Schonlein purpura in children. Chang Gung Med J. 2000; 23(3): 135-41. (Retrospective; 158 children)
Gonzalez-Gay MA, Calvino MC, Vazquez-Lopez ME, et al. Implications of upper respiratory tract infections and drugs in the clinical spectrum of Henoch- Schonlein purpura in children. Clin Exp Rheumatol. 2004; 22(6): 781-4. (Retrospective; 78 children)
Halsey MF, Finzel KC, Carrion WV, et al. Toddler's fracture: presumptive diagnosis and treatment. J Pediatr Orthop. 2001; 21(2): 152-6. (Retrospective; 39 children)
Scherl SA, Miller L, Lively N, et al. Accidental and nonaccidental femur fractures in children. Clin Orthop Relat Res. 2000; 376: 96-105. (Retrospective; 207 children)
Mellick LB, Milker L, Egsieker E. Childhood accidental spiral tibial (CAST) fractures. Pediatr Emerg Care. 1999;15(5): 307-9. (Retrospective; 55 cases)
Fadilah SA, Cheong SK, Shahdan S, et al. The pale and limping child. Postgrad Med J. 2000; 76(901): 725-6. (Case report)
Postovsky S, Bialik V, Keider Z, et al. Large cell lymphoma of bone presented by limp. J Pediatr Orthop B. 2001; 10(1): 81-4. (Case report)
Tuten HR, Gabos PG, Kumar SJ, et al. The limping child: a manifestation of acute leukemia. J Pediatr Orthop. 1998; 18(5): 625-9. (Case reports)
Wong M, Chung CH, Ngai WK. Hip pain and childhood malignancy. Hong Kong Med J. 2002; 8(6): 461-3. (Case report)
Boechat MI, Winters WD, Hogg RJ, et al. Avascular necrosis of the femoral head in children with chronic renal disease. Radiology. 2001; 218: 411-3. (Prospective, randomized, longitudinal; 205 children)
Rosenberg ZS, Kawelblum M, Cheung YY, et al. Osgood-Schlatter lesion: fracture or tendonitis? Scintigraphic, CT, and MR imaging features. Radiology. 1992; 185: 853-8. (Retrospective; 28 cases)
Manoff EM, Banffy MB, Winell JJ. Relationship between Body Mass Index and slipped capital femoral epiphysis. J Pediatr Orthop. 2005;25(6):744-6. (Retrospective; 106 children)
Matava MJ, Patton CM, Luhmann S, et al. Knee pain as the initial symptom of slipped capital femoral epiphysis: an analysis of initial presentation and treatment. J Pediatr Orthop. 1999; 19(4): 455-60. (Retrospective; 106 children)
Jerre R, Billing L, Hansson G, et al. Bilaterality of slipped capital femoral epiphysis: importance of reliable radiographic method. J Pediatr Orthop B. 1996; 5(2): 80-4. (Retrospective; 100 children)
Wells D, King JD, Roe TF, et al. Review of slipped capital femoral epiphysis associated with endocrine disease. J Pediatr Orthop. 1993; 13(5): 610-4. (Retrospective; 131 children)
Eich GF, Superti-Furga A, Umbricht FS, et al. The painful hip: evaluation of criteria for clinical decisionmaking. Eur J Pediatr. 1999; 158: 923-8. (Retrospective; 89 children)
Del Beccaro MA, Champoux AN, Bockers T, et al. Septic arthritis versus transient synovitis of the hip: the value of screening laboratory tests. Ann Emerg Med. 1992; 21(12): 1418-22. (Retrospective; 132 children)
Zawin JK, Hoffer FA, Rand FF, et al. Joint effusion in children with an irritable hip. US diagnosis and aspiration. Radiology. 1993; 187(2): 459-63. (Prospective; 96 children)
Kocher MS, Zurakowski D, Kasser JR. Differentiating between septic arthritis and transient synovitis of the hip in children: an evidence-based clinical prediction algorithm. J Bone Joint Surg Am. 1999; 81(12): 1662-70. (Retrospective; 168 children)
Luhmann SJ, Jones A, Schootman M, et al. Differentiation between septic arthritis and transient synovitis of the hip in children with clinical prediction algorithms. J Bone Joint Surg Am. 2004;86(5):956-962. (Retrospective; 163 children)
Kallio MJT, Unkila-Kallio L, Aalto K, et al. Serum Creactive protein, erythrocyte sedimentation rate and white blood cell count in septic arthritis of children. Pediatr Infect Dis J. 1997; 16(4): 411-3. (Prospective; 100 children)
Klein DM, Barbera C, Gray ST, et al. Sensitivity of objective parameters in the diagnosis of pediatric septic hips. Clin Orthop Relat Res. 1997; 338: 153-9. (Retrospective; 26 children)
Levine MJ, McGuire KJ, McGowan KL, et al. Assessment of the test characteristics of C-reactive protein for septic arthritis in children. J Pediatr Orthop. 2003; 23(3): 373-7. (Retrospective; 133 children)
Pepys MB. C-reactive protein fifty years on. Lancet. 1982; 1(6): 53-7. (Review)
Kunnamo I, Kallio P, Pelkonen P, et al. Clinical signs and laboratory tests in the differential diagnosis of arthritis in children. Am J Dis Child. 1987; 141(1): 34-40. (Prospective; 278 children)
Deshpande SS, Taral N, Modi N, et al. Changing epidemiology of neonatal septic arthritis. J Orthop Surg. 2004; 12(1): 10-3. (Retrospective; 15 neonates)
Lyon RM, Evanich JD. Culture negative septic arthritis in children. J Pediatr Orthop. 1999; 19(5): 655-9. Retrospective; 105 children)
Chang WL, Yang YH, Wang LC, et al. Renal manifestations in Henoch-Schonlein purpura: a 10-year clinical study. Pediatr Nephrol. 2005; 20(9): 1269-72. (Retrospective; 261 patients)
Narchi H. Risk of long term renal impairment and duration of follow up recommended for Henoch- Schonlein purpura with normal or minimal urinary findings: a systematic review. Arch Dis Child. 2005; 90(9): 916-20. (Review)
Oudjhane K, Newman B, Oh KS. Occult fractures in preschool children. J Trauma. 1998;28:858-60. (Retrospective, 500 consecutive patients)
Fordham L, Auringer ST, Frush DP. Pediatric imaging perspective: acute limp. J Pediatr. 1998;132(5):906-8. (Case reports)
Akgun R, Yazici M, Aksoy MC, et al. The accuracy and reliability of estimation of lateral pillar height in determining the herring grade in Legg-Calve-Perthes disease. J Pediatr Orthop. 2004;24(6):651-653. (Prospective; 50 patients)
Gigante C, Frizziero P, Turra S. Prognostic Value of Catterall and Herring Classification in Legg-Calve- Perthes Disease: Follow-Up to Skeletal Maturity of 32 Patients. J Pediatr Orthop. 2002;22(3):345-349. (Retrospective; 32 patients)
Kaniklides C, Sahlstedt B, Lonnerholm T, et al. Conventional radiography and bone scintigraphy in the prognostic evaluation of Legg-Calve-Perthes disease. Acta Radiol. 1996;37(4):561-6. (Retrospective; 75 patients)
Sales de Gauzy J, Kerdiles N, Baunin C, et al. Imaging evaluation of subluxation in Legg-Calve-Perthes disease: magnetic resonance imaging compared with the plain radiograph. J Pediatr Orthop B. 1997;6(4):235-8. (Retrospective; 33 blinded measurements)
Stucker MHF, Buthmann J, Meiss AL. Evaluation of hip containment in Legg-Calve-Perthes Disease: A comparison of ultrasound and magnetic resonance imaging. Ultraschall in Med. 2005;26:406-410. (Retrospective; 26 patients)
Van Campenhout A, Moens P, Fabry G. Serial bone scintigraphy in Legg-Calve-Perthes disease: correlation with the Catterall and Herring classification. J Pediatr Orthop. 2006;15(1):6-10. (Prospective; 86 patients)
Wirth T, LeQuesne GW, Paterson DC. Ultrasonography in Legg-Calve-Perthes disease. Pediatr Radiol. 1992;22(7):498-504. (Retrospective; 23 patients)
Cowell HR. The significance of early diagnosis and treatment of slipping of the capital femoral epiphysis. Clin Orthop. 1966;48:89-94. (Retrospective; 66 children)
Klein A, Joplin RJ, Reidy JA, et al. Roentgenographic features of slipped capital femoral epiphysis. AJR. 1951;66:361-374.
Klein A, Joplin RJ, Reidy JA, et al. Slipped capital femoral epiphysis: early diagnosis and treatment facilitated by "normal" roentgenograms. J Bone Joint Surg (Am). 1952;34:233-9. (Retrospective)
Bloomberg TJ, Nuttall J, Stoker DJ. Radiology in early slipped femoral capital epiphysis. Clin Radiol. 1978;29(6):657-67. (Retrospective; 38 radiographs)
Causey AL, Smith ER, Donaldson JJ, et al. Missed slipped capital femoral epiphysis: illustrative cases and a review. J Emerg Med. 1995;13(2):175-189. (Case reports)
Volberg FM, Sumner TE, Abramson JS, et al. Unreliability of radiographic diagnosis of septic hip in children. Pediatrics. 1984;74:118-120. (Retrospective; 19 patients)
Gandini D. Acute septic arthritis of the hip in children in northern Australia. ANZ J Surg. 2003; 73: 136-139. (Retrospective; 11 patients)
Marchal GJ, Van Holsbeeck MT, Raes M, et al. Transient synovitis of the hip in children: role of US. Radiology. 1987;162(3):825-8. (Retrospective; 46 patients)
Kayser R, Franke J, Mahlfeld K. Value of ultrasound diagnosis in Legg-Calve-Perthes disease. Schweiz Rundsch Med Prax. 2003;92(24):1123-7. (Retrospective; 153 patients)
Billing L, Bogren HG, Wallin J. Reliable X-ray diagnosis of slipped capital femoral epiphysis by combining the conventional and a new simplified geometrical method. Pediatr Radiol. 2002;32:423-430. (Retrospective; 100 SCFE and 95 Control)
Magnano GM, Lucigrai G, De Filippi C, et al. Diagnostic imaging of the early slipped capital femoral epiphysis. Radiol Med (Torino). 1998;95:16-20. (Retrospective; 21 patients)
Miralles M, Gonzalez G, Pulpeiro JR, et al. Sonography of the painful hip in children: 500 consecutive cases. Am J Roentgenol. 1989;152(3):579-82. (Prospective; 500 patients)
Gordon JE, Huang M, Dobbs M, et al. Causes of falsenegative ultrasound scans in the diagnosis of septic arthritis of the hip in children. J Pediatr Orthop. 2002;22(3):312-6. (Prospective; 132 patients)
Tien YC, Chih HW, Lin GT, et al. Clinical application of ultrasonography for detection of septic arthritis in children. Kaohsiung J Med Sci. 1999;15(9):542-9. (Retrospective; 31 patients)
Dorr U, Zieger M, Hauke H. Ultrasonography of the painful hip. Prospective studies in 204 patients. Pediatr Radiol. 1988;19(1):36-40. (Prospective; 204 patients)
Zieger MM, Dorr U, Schulz RD. Ultrasonography of hip joint effusions. Skeletal Radiol. 1987;16(8):607-611. (Prospective; 123 consecutive patients with 20 controls)
Strouse PJ, DiPietro MA, Teo EL, et al. Power Doppler evaluation of joint effusions: investigation in a rabbit model. Pediatr Radiol. 1999;29:617-623. (Animal study)
Strouse PJ, DiPietro MA, Adler RS. Pediatric hip effusions: evaluation with power Doppler sonography. Radiology. 1998;206(3):731-5. (Prospective; 29 patients)
Choban S, Killian JT. Evaluation of acute gait abnormalities in preschool children. J Pediatr Orthop. 1990;10:74-8. (Retrospective; 60 patients)
Aronson J, Garvin K, Seibert J, et al. Efficiency of bone scan for occult limping toddlers. J Pediatr Orthop. 1992;12:38-44. (Prospective; 50 consecutive patients)
Goergens ED, McEvoy A, Watson M, et al. Acute osteomyelitis and septic arthritis in children. J Paediatr Child Health. 2005;41:59-62. (Retrospective; 102 osteomyelitis and 47 septic arthritis)
Guzzanti V, Falciglia F. Slipped capital epiphysis: Comparison of a roentgenographic method and computed tomography in determining slip severity. J Pediatr Orthop. 1991;11:6-12. (Retrospective; 20 radiographs)
Kamegaya M, Saisu T, Ochiai N, et al. Preoperative assessment for intertrochanteric femoral osteotomies in severe chronic slipped capital femoral epiphysis using computed tomography. J Pediatr Orthop B. 2005;14(2):71-8. (Prospective; 22 patients)
Rowe SM, Chung JY, Moon ES, et al. Computed tomographic findings of osteochondritis dissecans following Legg-Calve-Perthes disease. J Pediatr Orthop. 2003;23(3):356-62. (Retrospective; 13 patients)
Stanitski CL, Litts CS, Stanitski DF. Tibial torsion in chronic, stable slipped capital femoral epiphyses: evaluation by CT scan. J Pediatr Orthop. 1997;17(5):657-8. (Retrospective; 44 patients)
Hopkins K, Li K, Bergman G: Gadolinium-DTPAenhanced magnetic resonance imaging of musculoskeletal infectious processes. Skeletal Radiol. 1995;24:325. (Retrospective; 34 radiographs)
Lee SK, Suh KJ, Kim YW, et al. Septic arthritis versus transient synovitis at MR Imaging: Preliminary Assessment with Signal Intensity Alterations in Bone Marrow. Radiology. 1999;211(2):459-65. (Retrospective; 9 septic arthritis, 14 transient synovitis)
Ranner G, Ebner F., Fotter R, et al. Magnetic resonance imaging in children with acute hip pain. Pediatr Radiol. 1989;20:67-71. (Prospective; 45 patients)
Umans H, Leibling MS, Moy L, et al. Slipped capital femoral epiphysis: a physeal lesion diagnosed by MRI, with radiographic and CT correlation. Skeletal Radiology. 1998; 27: 139-44. (Prospective; 15 patients)
Pinto MR, Peterson HA, Berquist TH. Magnetic resonance imaging in early diagnosis of Legg-Calve- Perthes disease. J Pediatr Orthop. 1989;9(1):19-22. (Case report; 2 patients)
Kermond S, Fink M, Graham K, et al. A randomized clinical trial: should the child with transient synovitis of the hip be treated with nonsteroidal anti-inflammatory drugs? Ann Emerg Med. 2002; 40(3): 294-9. (Randomized, blinded, controlled; 40 children)
Rosenblum ND, Winter HS. Steroid effects on the course of abdominal pain in children with Henoch- Schonlein purpura. Pediatrics. 1987; 79(6): 1018-21. (Retrospective; 43 children)
Huber AM, King J, McLaine P, et al. A randomized, placebo-controlled trial of prednisone in early Henoch Schonlein purpura. BMC Med. 2004; 2: 7. (Randomized, blinded, controlled; 40 children)
Mollica F, Li Volti S, Garozzo R, et al. Effectiveness of early prednisone treatment in preventing the development of nephropathy in anaphylactoid purpura. Eur J Pediatr. 1992; 151(2):140-4. (Prospective, controlled; 168 children)
Flynn JT, Smoyer WE, Bunchman TE, et al. Treatment of Henoch-Schonlein purpura glomerulonephritis in children with high-dose corticosteroids plus oral cyclophosphamide. Am J Nephrol. 2001; 21(2): 128-33. (Retrospective; 12 children)
Foster BJ, et al. Effective therapy for severe Henoch- Schonlein purpura with and prednisone azathioprine: a clinical and histopathologic study. J Pediatr. 2000; 136(3): 370-5. (Retrospective; 20 children)
Niaudet P, Habib R. Methylprednisolone pulse therapy in the treatment of severe forms of Schonlein-Henoch purpura nephritis. Pediatr Nephrol. 1998; 12(3): 238-43. (Prospective; 38 children)
Kocher MS, Mandiga R, Murphy JM, et al. A clinical practice guideline for treatment of septic arthritis in children. J Bone Joint Surg Am. 2003;85(6):994-999. (Prospective cohort; 30 children)
Butbul-Aviel Y, Koren A, Halevy R, et al. Procalcitonin as a diagnostic aid in osteomyelitis and septic arthritis. Pediatr Emerg Care. 2005; 21(12): 828-32. (Prospective; 44 children)
Martinot M, Sordet C, Soubrier M, et al. Diagnostic value of serum and synovial procalcitonin in acute arthritis: a prospective study of 42 patiens. Clin Wxp Rheumatol. 2005; 23(3): 303-10. (Prospective; 42 children)
Kocher MS, Mandiga R, Zurakowski D, et al. Validation of a clinical prediction rule for the differentiation between septic arthritis and transient synovitis of the hip in children. J Bone Joint Surg Am. 2004;86(8):1629-1635. (Prospective; 154 children)