Acute hematogenous osteomyelitis (AHO) and septic arthritis (SA) are severe and not uncommon problems of the neonate. A classic definition of the neonatal period is from birth to 28 days; however, when considering the unusual features of skeletal sepsis, that period can be extended to 8 or 10 weeks of age. Death is a rare sequela of neonatal AHO and SA in developed countries, but severe complications such as joint destruction, limb shortening, and deformity are common. Treating low birthweight and high risk babies in neonatal intensive care units saves many lives, but these babies are at even greater risk for skeletal infections. The diagnosis of bone and joint sepsis in the neonate is more difficult, the causative factors are different, and the complications are worse than in the older child.
The pathogenesis of skeletal infection in the neonate differs from that involved in the older child in both its origin and spread. For discussion, these aspects are best divided into physiologic and anatomic differences.
In physiologic terms, the neonate is immunologically impaired because both the humoral and the cellular defense mechanisms are immature. 1,2 For this reason, the neonate is more vulnerable to and less aefficient in fighting infection. The lack of immune defense permits organisms other than the usual pathogens to affect the neonate. In addition, the neonate's local and systemic inflammatory response to infection is weaker than that of the older child.3'6 Consequently, the local signs are often minimal, making both the diagnosis and the ability to locate the infection much more difficult. The very common finding of multifocal bone and joint involvement is due to this weak immune response, as is, in part, the rapid local spread. 1,6,8
Anatomically, the metaphyseal and epiphyseal areas of the neonate are different from those of the older child. In the neonate, there is a small or often no ossifie nucleus. In this situation, the metaphyseal vessels communicate with the epiphyseal vessels in the cartilagenous precursor of the ossifie nucleus (Figure I).8 This allows early and rapid spread of infection from the metaphysis into the cartilagenous epiphysis. The consequence is rapid destruction of the epiphysis and early spread into the joint, accounting for the frequent association of joint infection with bone infection. 3,6'9 As the ossifie nucleus and the epiphyseal plate form, separate epiphyseal arteries arise to supply the epiphysis and the communication with the metaphyseal vessels disappear, providing temporary protection for both the joint and the epiphysis (Figure 2).
Figure 1. Femoral head from a 3-day-ofd infant Note the vessels traversing the femoral neck directly into the cartilaginous epiphyseal precursor. This provides a direct route of spread for bacterial invasion of the cartilaginous anlage of the femoral head (From Chung SMK: Arterial supply of the developing proximal end of the human femur. J Bone Joint Surg 1976; 58A:996).
The destruction of the epiphysis and the physeal plate in the neonatal period have profound consequences on future growth. The two most severe complications of musculoskeletal sepsis in the neonate are shortening and angular deformity. 3·6'7 Joint destruction is also common. However, the neonate's tremendous remodeling potential permits significant adaptation in the affected joints, often preserving a reasonable amount of congruity (Figure 3).
In children, there are four areas where the metaphysis - the usual site where infection begins - is intracapsular: the hip, proximal humerus, proximal radius, and distal- lateral tibia. However, in the neonate, the capsule and ligaments of the joint are closer to and often surround the metaphysis. Thus, the early spread of pus through the thin maze- like cortex may track beneath these structures into the joint (JA Ogden, personal communication).8 This anatomic fact coupled with immature defense mechanisms may help to explain why bone and joint involvement are seen together in over one half of the cases of musculoskeletal sepsis in the neonate.7,8
The elevation of the periosteum by pus has a second effect. Because the periosteum is thick and well vascularized from its external surface and is also loosely attached to the bone, it is easily elevated from the cortex. The bone becomes devascularized forming a sequestrum. At the same time the periosteum, which remains healthy, begins to lay down bone. This new bone surrounds the dead sequestrum and is called the involucrum.
The manner in which bacteria destroy bone and cartilage is similar in both the neonate and the older child. Although a complete discussion of the complex mechanisms involved in bone and cartilage destruction is beyond the scope of this article, some factors will be mentioned as they are important in understanding the role of surgical drainage.
The two routes by which bacteria gain access to the site of infection are local and hematogenous. Common procedures that have been shown to result in local bone or joint infection are heel puncture, femoral artery or vein puncture, and electrode insertion for fetal monitoring during birth.3·10'12 In most cases, however, bacteria gain access to the site of infection by the hematogenous route. Bacteremia can be secondary to colonization or other infection in the body. Procedures such as exchange transfusion, intravenous therapy, umbilical artery catheterization, and surgery are known to be related to neonatal sepsis.3,4,7,13,14
Figure 2. Femoral head from a Doy 3 years 4 months of age. Note the epiphyseal vessels supplying the distinct ossifie nucleus. Note persistence of a few vessels traversing the epiphyseal plate. (From Chung SMK: Arterial supply of the developing proximal end of the human femur. J Bone Joint Surg 1976; 58A:996).
During the last 10 to 15 years, group B Streptococcus, a type of bacteria frequently found in the birth canal, has become more common and in many reports is now the most frequent agent causing bone and joint infection in the neonate.3·15·16 Staphylococcus aureus remains a frequent cause of infection, and the incidence of this bacteria has increased since the discontinuation of hexachlorophene in bathing neonates. The gram-negative bacteria are also common. Other frequent pathogens in the neonate are Escherichia coli, Serratia marcescens, and Proteus mirabilis. Due to the immune compromised status of the neonate, a large variety of other bacteria rarely pathogenic under normal circumstances are found to cause skeletal infeaction. Candida albicans, a fungus, is not infrequently the cause of skeletal infection in the debilitated neonate with a central vein catheter, hyperalimentation, or prolonged antibiotic treatment.4,17
The classic signs of local or systemic bacterial infection are mainly a result of the inflammatory response. As mentioned before, the neonate's immune system is not mature. Consequently the local and systemic signs and symptoms are often minimal.
Systemic signs may include irritability or lethargy, poor feeding, and failure to gain weight. These signs are not specific and are common to other neonatal pathologies as is sepsis in any other site. Fever might be present, but normal temperature is found in more than one half of the cases.4,6,7 A septic neonate may present with hypothermia.18
Local signs of bone and joint infection include swelling of the involved area and decreased mobility. Less reliable signs are erythema, warm skin, or tenderness. All of these signs can be absent or minimal due to the lack of an inflammatory response. In addition, presence of these signs may be obscured by the fatty and ill defined contour of the neonate's limbs as well as the normal joint contractures during and soon after birth. This is especially true for the hip joint, which is difficult to examine at any age because of its complex motion and location, buried beneath tat and muscle. For all of these reasons, the orthopedic manifestations of sepsis in the neonate are often not suspected and the diagnosis is missed (Figure 3).
Figure 3. This child was admitted at I week of age with sepsis. The cause was never diagnosed but she remained on intravenous antibiotics for 4 weeks. After walking began, the parents noticed a limp 3A. X-rays of a female 4 years 4 months of age. The acetabulum is well formed indicating that cartilagenous remnant remains. The good acetabular development indicates the presence of a sizable remnant of the femoral head and neck which is still present and functioning within the acetabulum.
Laboratory tests such as white blood count, differential count, and erythrocyte sedimentation rate are less helpful in the neonate than in the older child. The normal values for full term and premature newborns are different and change during the first week of life. l9 For the same reasons that systemic and local inflammatory signs do not correspond with the severity of the infection, laboratory tests can also be deceiving. It is common to see a neonate with severe infection and a white blood count within normal limits. *'6 Erythrocyte sedimentation rate is somewhat better as a nonspecific test, although about 25% of the neonates with bone and joint infection present with erythrocyte sedimentation rate under 40 mm per hour.
Bony changes appear on radiographs about 7 to 10 days after infection begins. Radiography is very important in the early stages of bone and joint infection because it demonstrates the soft tissue changes that appear in most cases. 4,6,20,21 Changes in soft tissue that result from infection include joint swelling, deep soft tissue swelling, and changes in or disappearance of the fat planes between muscle layers. In the hip, lateral subluxation of the femoral head is a late but important sign. It is essential to have radiographs showing both limbs symmetrically positioned, so that the soft tissues can be compared (Figure 4). The exposure should show soft tissue, and a skeletal survey is recommended because of the multifocal nature of musculoskeletal sepsis in the neonate.
Two types of bone scan may be useful in the neonate: the technetium phosphate and gallium scans. The technetium scan shows increased uptake in areas of increased blood flow and new bone formation, whereas the gallium scan shows increased uptake in areas of white blood cell accumulation. Gallium scans require a waiting period of 48 to 72 hours after injection before isotope accumulation is sufficient for accurate interpretation. The gallium scan also delivers a high radiation dose to the patient. For these reasons, it is usually reserved for complex situations in which the technetium phosphate scan has not provided the necessary information.22
Figure 3B. A 7 years 6 months a small amount of ossification has occurred. The trochanter continues to grow. This patients function remains good as a child, although leg length discrepancy is developing and arthritis can be anticipated. She represents a case of failure to diagnose a septic hip.
In older children with skeletal infection, the technetium scan has been shown to have a 70% to 95% accuracy rate in different series.23,24 However, in the neonate several problems lower this level of accuracy. First, the neonate lacks an adequate inflammatory response. Second, the very active epiphyses of the neonate concentrates a large amount of isotope, which makes the increased uptake in the adjacent metaphysis more difficult to see. Third, the small size of the bones and joints coupled with the difficulty of eliminating motion compound the problems of bone scanning. These difficulties can be overcome to some extent by using pinhole imaging. In one study of 10 neonates with 20 proven sites of osteomyelitis, the scan was positive in only 6 sites, normal in 11, and equivocal in 3.25 Despite these difficulties, technetium scans should be obtained in all neonates with suspected skeletal sepsis to aid in locating other silent sites of sepsis.
However, bone scans should never delay other diagnostic and therapeutic measures, eg, aspiration of the bone or joint. These important diagnostic tests have been shown not to cause false positive readings of a later bone scan and so can be performed as soon as a site is suspected.26,27
Unlike a radiograph which reflects anatomic changes, the bone scan reflects localized physiologic changes in the skeleton. Therefore, it is important to realize that there may be a point in the pathologic process where the infected area may be avascular. In such a circumstance the isotope cannot gain access to the area and the bone scan will reveal decreased uptake or a "cold scan."24,28
STEPS IN DIAGNOSIS
A high index of suspicion is most important in any neonate who is septic. Frequently, the neonate will be evaluated and treated for sepsis and the musculoskeleton focus will be missed.
Figure 4. This infant was admitted to the hospital with a history of failure to move the arm and fever for the past 4 to 5 days. Note the swelling of the deep soft tissue layer on the left shoulder as compared with the right shoulder. This indicates septic arthritis or osteomyelitis. The superficial soft tissues remain of equal width, indicating no cellulitis. Symmetrically positioned comparison views taken with soft tissue technique are valuable in the diagnosis of bone and joint sepsis.
Important points in the history include problems during pregnancy or delivery (eg, prematurity and low birthweight); infection in any other site that could cause bacteremia; and invasive procedures (eg, umbilical vessel catheterization, femoral punctures, exchange transfusion, intravenous therapy, or intramuscular injections). The time and duration of observed symptoms are also important. Pseudoparalysis first noted at 4 days of age is probably not due to Erb's palsy or birth trauma; infection is a more likely etiology (Figure 4).
When examining the neonate, every joint should be tested accurately for the range of motion and every bone should be palpated for swelling and tenderness. The classic signs of infection should be looked for but in the neonate those signs are often minimal or absent, even in severe infection. A great deal of experience in the examination of newborns is necessary since as little as 10° less internal rotation in one hip as compared with the other is significant.
Blood cultures and Gram's stains of any puni lent material should be obtained early. Blood cultures will be positive and identify the causative organism in 50% to 70% of bone and joint infections.4'6 Radiographic survey of the entire skeleton with symmetric positioning of the limbs and attention to soft tissue detail should be obtained. Technetium bone scan should be ordered in all cases of neonatal skeletal sepsis to search for multiple sites.
The most important diagnostic tool is aspiration of suspected joints and bones. When sepsis is suspected, any local, clinical, or radiographic sign is an indication for an early aspiration. Positive aspiration will confirm the diagnosis, help to identify the organism, and aid in determining the treatment. Gram's stain and cultures should be obtained from any aspirated material, even if it does not appear to be pus. If no material can be aspirated, sterile saline without preservative can be injected and reaspirated for culture. When no fluid is found in a suspected hip joint, needle placement should be confirmed by injecting a small amount of radiopaque fluid into the joint under fluoroscopy.
It is wise to aspirate both hips of any neonate who has a bone or joint infection at another site. There are three reasons for this recommendation: (1) the hip is the most common joint to be infected in the neonate; (2) it is the most difficult joint to examine; and (3) its destruction is so rapid and the sequelae so severe that no risk of missing an infection in this joint should be permitted.
Intravenous antibiotic treatment should start as soon as local blood cultures are taken and aspiration of suspected areas is completed. Gram's stain findings or previous cultures from other sites of infection can be very helpful in choosing antibiotics. Knowing the flora in the nursery or the neonatal intensive care unit will also help in selecting antibiotics. Without other guidelines, the broad spectrum of common causative organisms should be covered. In the neonate, those organisms are: Streptococcus group B, Staphylococcus aureus, and gram-negative bacteria. Cefotaxime sodium is a good choice, as it has the needed broad spectrum activity and can penetrate the blood-brain barrier. Another good choice is the combination of cloxacillin and an aminoglycoside. Great care should be used in the administration of these antibiotics because the metabolisms of prematures and neonates are so varied. Expert management by an infectious disease specialist is strongly recommended. Antibiotic treatment should be reviewed once the organism and its sensitivity are known.
In the older child, intravenous antibiotics can be changed to oral treatment after about 1 week if certain conditions have been met. 29 This is not recommended for the neonate with bone and joint sepsis. Infection in the neonate is usually more severe, involves multiple sites, and has a worse outcome. The gastrointestinal side effects of oral antibiotics are also more common in the neonate. Parenteral treatment for at least 3 weeks is the treatment of choice. A central line is helpful in many cases.
The decision to stop treatment is based on physical examination and laboratory tests. The presenting signs such as lethargia, poor feeding, swelling, tenderness, and limited range of motion should disappear. Erythrocyte sedimentation rate should be under 20 mm per hour. Radiographs should show a healing phase with no new foci.
Local antibiotic treatment has been suggested for septic joints. Local administration of antibiotic to a septic joint is not needed and can be harmful. With intravenous administration, antibiotic concentration in the joint fluid will be at least equal to the concentration in the serum.29,30 Given locally the antibiotic can be very irritating, which perpetuates the synovitis. In addition, the antibiotic may be absorbed from the joint with adverse consequences for the patient.
The main goal of surgery is debridement. Articular cartilage is destroyed by the inflammatory response which is caused by a number of factors in addition to the live bacteria. The dead bacterial cell wall fragments, the products liberated by the bacteria, the products of tissue destruction, and the enzymes from the synovium and the white blood cells are all responsible for continuation of the inflammatory process and consequent tissue destruction. 31,32 Although the host may eventually remove these substances that perpetuate the inflammatory response and interfere with antibiotic action, the progress takes days to weeks during which cartilage destruction continues. Surgical removal of these substances is an augmentation of the host's defense mechanisms. The same holds true for the bone.
Because of the fibrinous exudate in a septic joint, adequate debridement is impossible by aspiration. In addition, the necessary repeated aspirations do not make for happy patients. A small arthrotomy which allows thorough irrigation of the joint is the best method of achieving joint debridement. Every bone and joint in which pus is found should be debrided if the child's general condition permits. A closed system of suction-irrigation for 24 hours provides a means of continuous debridement and seems particularly well suited to joint infections. The joint capsule is left open but the skin is closed. The hip joint is best approached anteriorly. The anterior approach does not endanger the blood supply to the femoral head and does not leave an opening in the posterior capsule which might predispose to posterior dislocation.
Immobilization has been a classic modality in the treatment of bone and joint infections. Recently there has been some enthusiasm for the use of continuous motion.33 However, in the neonate this may pose formidable problems. In the case of the hip joint which is subluxated, immobilization is advisable to allow capsular tightening and healing.
Death is a rare outcome of bone and joint infection in the neonate today. But even with modern antibiotics and good neonatal intensive care units, the morbidity is high.
Chronic or recurrent osteomyelitis is rare7 and can be eliminated with proper treatment. Most morbidity is from growth plate damage with consequent shortening and angular deformity. In the long term, however, all joint destruction that results in architectural alteration of the joint will result in arthritis. One half to two thirds of infected joints will later show evidence of permanent damage. Damage can range from mild destrucrion and narrowing of the joint space to total destruction and an ankylosed joint.3,34 The hip, which is the joint most frequently involved, has the poorest outcome. Bad prognostic factors include dislocation of the joint, delay in diagnosis and treatment, and a young age. 34 Bone deformity is found in more than 50% of the cases. 3·6 Bone involvement can range from a minimal length discrepancy to severe deformity or complete destruction of the epiphysis.
The only way to minimize the severe sequelae of bone and joint infection in the neonate is to have a very high index of suspicion and the ability to institute correct treatment, including a low threshold for aspiration, adequate radiologic facilities, knowledge of antibiotic usage in the neonate, and the ability to do adequate surgical debridement.
Bone and joint infections during the neonatal period differ from those in the older child. Neonates are relatively immunocompromised and their growth plate is pierced by blood vessels. Special characteristics of skeletal infection in the neonate are: (1) multiple foci of infection; (2) simultaneous involvement of the bone and the adjacent joint; and (3) limited systemic and local inflammatory response.
The more common causative organisms are group B Streptococcus, Staphylococcus aureus, and gram-negative bacteria. Clinical findings can be minimal or absent. The more common findings include local swelling and limited motion in a joint or a limb. Early soft tissue radiographic changes are found in most cases and should be looked for. Every suspected bone or joint should be aspirated. Hip joint aspiration is recommended in the presence of other skeletal infection, even without local signs. Intravenous antibiotic treatment is started according to Gram's stain results or the best "educated guess." Surgical debridement is indicated for every site of pus in the extremities. Joint destruction and bone deformity and shortening are a common outcome. The only way to minimize early and late complications is high index of suspicion, aggressive workup, and adequate early treatment.
1. Cutoff SP: Neonatal immunity. J Pediatr 1974: 85(21:149-154.
2. Kuo KN, Uoyd-Roberrs UC. Orme IM. et al: Immunodetkiencv and infantile hone and pint infection. Arch Da Chad 1975; 50(11:50-56.
3. Fox L. Sprunt K: Neonatal osteomyelitis. Pidkttrics 1978; 62(41:535-542.
4. Pirrard WB. Thullen JD. FanarorY AA: Neonatal septic arrhriris. I Rsfiarr 1976; 88(41:621-624.
5. Squire E, Favata B, ToJdJ: Diagnosis of neonatal bacterial infection: Hematologic and pathologie findings in fetal and nonfatal cases. rVthemcs 1979; 64(11:60-64.
6. Weissberg ED, Smith AL, Smith DH: Clinical features of neonatal osteomyelitis. Pediatrics 1974; 53(4):505-509.
7. Mok PM. Reilly BJ. Ash JM: Osteomyelitis in the neonate. Radiology 1982: 145:677-682.
8. Ogden JA: Pediatric Osteomyelitis and septic arthritis: The pathology of neonatal Disease. Yale J Biol Med 1979; 52:423-448.
9. Alderson M, Speers D. Emslie K. et al: Acute haematogenous osteomyelitis and septic arrhriris- A single disease. J Bone Joint Surg 1986; 68B(21:268-274.
10. Barry DMJ. Reeve AW: Increased incidence of gram-negative neonatal sepsis with intramuscular iron administration, Pediatric 1977: 60(6):909-912.
11. Nelson DL, HaMe KA, Matsen JM: Proteus mirabilis osteomyelitis in nu neonates following needle puncture. Am J Dis Child 1973: 125:109-110.
12. Oveturf GD. Balfour G: Osteomyelitis and sepsis: Severe complications of fetal monitoring, Pediatrics 1975: 55(2):244-247.
13. Lim MO, Gresharti EL, Franken EA, et al: Osteomyelitis as a complication of umbilical artery catheterization. Am J Dis Child 1977; 131:142-144.
14. Simmon PB, Harris LE, Bianco AJ: Complications of exchange transfusion. Mono Clin Proc 1973; 48:190-193.
15. Edwards MS, Baker CJ. Wagner ML, et al: An etiologic shift in infantile osteomyelitis: The emergence of the group B streptococcus. J Pediarr 1978; 93(4):578-583.
16. Howard JB, McCracken GH: The spectrum of group B streptococcal infections in infancy. Am J Dis Child 1974: 128:815-817.
17. Yousefasdeh DK, Jackson JH: Neonatal and infantile candidal arthritis with or without osteomyelitis: A clinical and radiographicaf review of 21 cases. Skeletal Radiol 1980; 5:77-90.
18. Nelson WD: Nelson Textbook of Pediatrics. Infection of the Newborn General Comsiderations. ed 12. Philadelphia, WB Saunders Co, 1983. pp 399-403.
19. Xanthou M: Leucocyte blood picture in health full-term and premature babies during neonatal period. Arch Dis Child 1970; 45:242-249.
20. Capitanio MA, Kirkpatrkk JA: Early Roentgen Observations in Acute Osteumylitis. Philadelphia, St Christopher's Hospital for Children and Temple University School of Medicine. 1970. pp 488-496.
21. Volberg FM, Sumner TE, Abramson JS, et al: Unreliability of radiographic diagnosis of septic hip in children, Pediatrics 1984; 74(1):118-120.
22. Borman TR, Johnson RA, Sherman FC: Gallium scintigraphy for diagnosis of septic arthritis and osteomyelitis in children. J Pediart Ortho 1986; 6:317-325.
23. Duszynski DO, Kuhn JP, Afshani E, et al: Early radionuclide diagnosis of acute osteomyelitis. Radiology 1975; 117:337-340.
24. Sullivan JA, Vasileff T, Leonard JC: An evaluation of nuclear scanning in orthopaedic infections. J Pediarr Orthop 1981; 1:73-79.
25. Ash JM, Gilday DL; The futility of bone scanning in neonatal osteomyelitis: Concise communication. J Nucl Med 1980: 21:417-420.
26. Canale ST, Harkness RM, Thomas PA, et al: Does aspiration of hones and joints affect results of later bone scanning? J Pediatr Orthop 1985; 5:23-26.
27. McCoy JR, Morissy RT, Siebert J: Clinical experience with the technetium-99 scan in children. CORR 1981; 154:175-180.
28. Trackler RT, Miller KE, Sutherland DH, et al: Childhood pelvic osteomyelitis presenting as a "cold" lesion on bone scan: Case report. J Nucl Med 1976; 17:620-622.
29. Nelson JD, Howard JB, Shekon S: Oral antibiotic therapy for skeletal infections of children. J Pediatr 1978; 92(1)131-134.
30. Nelson JD Antibiotic concentrations in septic joint effusions. N Engl J Med 1971; 284(7):349-353.
31. Daniel D, Akeson W, Amiel D, et al: Lavage of septic joints in rabbits: Effects of chondrolysis. J Bone Joint Surg 1976; 58A(3):393-395.
32. Goldenberg DL, Chisholm PL, Rice PA: Experimental models of bacterial arthritis: A microbiologic and histopathologic characterization of the arthritis after the intraarticular injections of Neisseria gimorrhoeae. Staphylococcus aureus, group A streptococci, and Escherichia coli. J Rheumatol 1983: 10(1): 5-11.
33. Salter RB, Bell RS, Keeley FW: The protective effect of continuous passive motion on living articular cartilage in acute septic arthritis: An experimental investigation in the rabbit. CORR 1981: 159:223-247.
34. Lunseth PA. Heiple KG: Prognosis in septic arthritis of the hip in children. CORR 1979: 139:81-85.