ORIGINAL ARTICLE

Sequelae of Septic Arthritis of the Hip in Children

A New Classification and a Review of 41 Hips
Edilson Forlin, MD, MSc, PhD*Þ and Carlo Milani, MD, MSc, PhDþ§

Conclusions: The classification we proposed showed to be simpler

Background: The sequelae of septic arthritis of the hip in children
and more reliable than the classification proposed by Choi et al. It
are diverse and represent a complex problem. A classification of the
may be useful to guide treatment and prognosis. Longer follow-up is
deformities is important to plan treatment and to give prognostic
needed to assess the real benefit of the reconstruction for unstable hips.
information. The radiographic main classification used provides de-
Level of Evidence: Therapeutic study, level IV (case series [no
tailed information of the deformities, but it is complicated and not
control, or historic, group]).
feasible to apply in young children. The authors developed a simple
and reliable method of classification and present the results of a Key Words: hip, septic joint, deformities, neonates, children
large series of patients.
(J Pediatr Orthop 2008;28:524Y528)
Methods: The authors reviewed 37 children with 41 hips with
sequelae of septic arthritis of the hip. Twenty-four were male subjects,
and 13 were female subjects. Twenty-five had the acute infection at
the neonatal period, and 16 were aged between 1 month and 3 years.
None of them received adequate treatment at the acute process:
20 hips were not treated, and the other 21 had delayed drainage (at
D elay of the diagnosis and treatment of septic arthritis of
the hip (SAH) leads to involvement and damage of struc-
tures of the joint such as articular cartilage, epiphyses, ace-
least 5 days after the onset of the infection). The hips were classified tabulum, proximal femur metaphysic, and growth epiphysis.
into 2 groups on the basis of radiographic appearance. Hips with Consequently, it may result in loss of motion, pain, leg-length
sequelae grade 1 were reduced hips, with femoral head preserved discrepancy, consequent limp, and early osteoarthritis.1Y13
(1A) or absent (1B); hips grade 2 were dislocated hips, with the Surgical treatment of the sequelae may be indicated to
femoral head preserved (2A) or absent (2B). There were 10 hips with improve the joint condition and aspect, providing the best
grade 1A, 11 with grade 1B, 3 with grade 2A, and 17 were grade 2B. function possible at the adulthood. To guide treatment, some
We compared our material and results with those presented by Choi radiographic classifications have been proposed. The most
et al and also compared the reliability of both classification. referred is the Choi et al,14 a modification of the Hunka et al15
The criteria for satisfactory results were stability, range of motion system. It classified the hips in 4 types and 8 subtypes, but
of 50 degrees or more of flexion, a flexion contraction below 20 the classification is complex and difficult to apply in young
degrees, and a painless hip. children when treatment may be desirable. We believe that
Results: Our classification presented a greater concordance (25 of the 2 most important aspects for treatment and prognosis
41 hips against 7 in 41 for the Choi et al classification; P = 0.01). are the instability and the grade of the destruction of the
Thirty-seven hips were pain-free, and 4 had occasional pain. Thirty- proximal femur.
seven hips had a satisfactory range of motion (90.2%). Stability was In this study, we present a series of patients with se-
present in 29 hips (70.7%). Final results showed 24 hips classified as quelae of SAH and a simple new classification for sequelae.
satisfactory and 17 as unsatisfactory. Our material and end results We compared it with the classification and results published
were comparable with those reported by Choi et al. by Choi et al.14
Septic arthritis of the hip developed at the newborn period
produced more severe sequelae. Hips with sequelae grade 1 have a METHODS
good prognosis. Surgical procedures in these grades were necessary We retrospectively reviewed the medical records and
to correct acetabular dysplasia and improve congruency. Fifteen of 20 radiographs of all patients who had been treated at the Pequeno
hips with sequelae grade 2 (unstable) underwent reconstruction. At a Principe Hospital between 1985 and 1997 for SAH. Most
mean follow-up of 6.9 years (range, 2.7Y11.3 years), satisfactory patients in this study were treated at other medical centers at
results were obtained in 8 of them. the acute phase and were referred to our hospital to treat
complications. The diagnosis was made with typical history
From the *Hospital Pequeno Principe, Curitiba, Pr; †Hospital de Clinicas da of SAH and characteristic signs of the sequelae. Confirma-
Universidade Federal do Parana, Curitiba, Pr; ‡University Federal of São tion by culture was not necessary.
Paulo; and §Medicine School of Foundation of ABC, Sao Bernardo, Sao
Paulo, Brazil. The criteria to be included in this study were the
Study performed at Hospital Pequeno Prı́ncipe, Curitiba, Paraná, Brazil, and following:
Universidade Federal of São PauloYEscola Paulista de Medicina, São 1. children with history of developed septic arthritis before
Paulo, Brazil. age of 3 years,
None of the authors received financial support for this study.
Reprints: Edilson Forlin, MD, MSc, PhD, Rua Buenos Aires, 1020, Curitiba
2. minimal age of last clinical and radiographic evaluation
Pr, 80250-070 Brazil. E-mail: ediforlin@hotmail.com. of 4 years or at least 2 years of follow-up for the patients
Copyright * 2008 by Lippincott Williams & Wilkins who underwent reconstruction, and

524 J Pediatr Orthop & Volume 28, Number 5, July/August 2008

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Pediatr Orthop & Volume 28, Number 5, July/August 2008 Sequelae of Pediatric Septic Arthritis of the Hip

3. patients seen at our institution without previous treatment

of the sequelae.

We excluded all patients with minimal sequelae (type 1

by Choi et al classification), tuberculosis, and neurological
involvement and the ones whose diagnosis of bacterial
infection was not clear.
Forty-one hips in 37 patients met the criteria to be
included in this study. There were 24 boys and 13 girls. The
right side was involved in 16 patients; the left side, in 17;
and bilateral, in 4. Twenty-five patients presented the acute
disease when newborn; and 16, between ages of 1 month
and 3 years. Twenty-one patients had late drainage (at least
5 days after onset), and 20 did not undergo drainage.
Fifteen hips underwent a reconstruction procedure: open
reduction (OR) in 6, OR associated with Dega osteotomy in
5, OR associated with Dega and varus proximal femoral
osteotomy in 2, and OR associated with Salter osteotomy in
2 hips.
Seventeen hips were submitted to other modalities of
surgeries: valgus proximal femur osteotomy in 9, shelf

FIGURE 2. Our classification.

procedure in 3, distal transfer of the great trochanter in 3, and

Steel osteotomy in 1 hip. Three of these patients also had an
epiphysiodesis of the contralateral distal femur associated.
Other 9 hips had no surgery.

Radiographic Classification
We classified the hips using 2 different classifications:
1. The method proposed by Choi et al14 in their original
article (Fig. 1). There were 18 hips classified as type 2A, 4
as 2B, 1 as 3A, 1 as 3B, 7 as 4A, and 10 as 4B.
2. Our classification (Fig. 2). It was based on the relation
between the femur and the acetabulum (1 or 2) and at the
severity of the lesion of the proximal femur (A or B).
Therefore, the classification has 2 groups and 2 subgroups:
& grade 1, hips with the head or the femoral neck within
the acetabulum. It was subdivided in A when the femoral
head (total or part of) was present and B when it was
absent.
& grade 2, hips are dislocated. It was also subdivided in A
if the femoral head was present and B when it was
absent.

There were 10 hips classified as 1A, 11 as 1B, 3 as 2A,

and 17 as 2B.
Comparing the Two Systems of Classification
To assess the concordance of both classifications, 3 in-
dependent senior pediatric orthopaedic surgeons evaluated the
initial radiography of each case. They were not aware of the
patient condition, treatment, or outcome. We asked them to
classify the hips using both systems.
FIGURE 1. Classification as published by Choi et al. (From Choi The surgeons were given Choi et al14 original drawing
et al. Sequelae and reconstruction after septic arthritis of the hip and explanation as published. Our classification consisted of
in infants. J Bone Joint Surg [Am]. 1990;72-A:1150Y1164. an explanation through text and illustrations as shown here.

* 2008 Lippincott Williams & Wilkins 525

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Forlin and Milani J Pediatr Orthop & Volume 28, Number 5, July/August 2008

FIGURE 3. A, Radiograph of a patient who presents a septic arthritis of the right hip as a newborn. B, At age of 8 years, there is
some degree of subluxation and dysplasia (type 1A sequelae). C, As the subluxation progresses and the patient presented pain, a
shelf acetabuloplasty was performed at age of 10 years. At the age of 13 years and 5 months, the radiograph shows a good joint
coverage.

No other explanation was provided. The analysis of the 3 Study of the Author’s Classification
surgeons and the authors was used for statistical evaluation. and Variables
We also compared our series of patients and results with Using the classification proposed by the author, we
those reported by Choi et al.14 To do that, we divided their consider the possibility of a relationship between sequelae
patients in 2 groups by the stability criteria. Those classified as with sex, side, age at time of the acute infection, performance
types 1, 2, and 3 were considered stable, and the type 4 was of drainage or not, reconstruction surgery, and results.
considered unstable. For our classification, grade 1 was con-
sidered stable; and grade 2, unstable.
RESULTS
Classification of the Results The average age at the last visit was 9.9 years (range,
The results were classified clinically by Hunka et al15 3.9Y19 years). The average length of follow-up from the first
criteria. Satisfactory results had the following conditions: surgery for the treatment of the sequelae was 5.2 years (range, 3
1. the joint was stable, months to 11.3 years). For the patients who underwent reconstruc-
2. range of motion of 50 degrees or more of flexion and a tion, the mean follow-up was 6.9 years (range, 2.7Y11.3 years).
flexion contraction below 20 degrees, and Thirty-seven hips were pain-free, and 4 presented
3. a painless hip for the daily activities. pain. Thirty-seven hips had a satisfactory range of motion
(90.2%). Stability was present in 29 hips (70.7%), and 12 were
The hip was classified as unsatisfactory when one or unstable (29.3%).
more criteria above were present. Final results showed 24 hips classified as satisfactory
and 17 as unsatisfactory.
Statistical Analysis In the group of reconstructed hips, 14 of them were
We considered full concordance when all 4 surgeons painless. A satisfactory range of motion was present in 13
came up with the same grade or type for the analyzed hip. hips. Eleven hips were stable, and 4 were unstable. At the end,
A W2 test and a Fisher exact test were used for no- 8 were classified as satisfactory; and 7, as unsatisfactory.
temporal variable and t-student for temporal. Of the 26 hips without a reconstruction surgery, 23
The decisions for statistic test were based on 5% or were pain-free, and 3 presented pain; 24 had a satisfactory
0.05 significance. range of motion, and 2 had not; 18 were stable, and 8 were

FIGURE 4. A, Radiograph of a 13-year-old boy with a sequelae type 1A at the left side showing a subluxation that was
progressive. B, After a Steel triple osteotomy, there is improvement at the joint relationship, but some incongruence remains.
C, Important narrowing of the joint space was noticed, and a valgus osteotomy was performed. The result was unsatisfactory with
ankylosis of the hip.

526 * 2008 Lippincott Williams & Wilkins

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Pediatr Orthop & Volume 28, Number 5, July/August 2008 Sequelae of Pediatric Septic Arthritis of the Hip

FIGURE 5. A, Radiograph of a patient at age 1 year and 3 months who developed an SAH as a newborn presenting type 2B sequelae
submitted to an OR and a Dega osteotomy. B, At age of 5 years and 2 months, a well-centered hip is seen. It was classified as a
satisfactory result.

unstable. The final result was satisfactory in 16 hips and un- sis to develop sequelae; however, if they do, sequelae will be
satisfactory in 10. more severe.
Full concordance was found in 25 of 41 hips for the Sequelae may vary from very subtle alteration to severe
classification proposed by the authors and in 7 of 41 hips using destruction of the proximal femur and dislocation of the joint.
the Choi et al system (P = 0.01). It makes important to have a classification useful to treatment
Of the 25 hips in patients who had the acute phase at decision and prognosis. We realized that the Choi et al14
neonatal age, 7 had sequelae 1A or 1B, and 18 presented classification describes in details the anatomical alterations,
sequelae 2A or 2B. Of the 16 who had the infection late, only and thus, it became the referred method. However, the num-
2 had a grade 2A or 2B sequelae (significant difference). ber of types and details of each one may make it difficult to
Of the 21 classified as sequelae grade 1A or 1B, 16 hips reach an agreement by different observers and do not allow
(77%) presented a satisfactory result. Of the 20 sequelae drawing significant conclusions.
grade 2A or 2B, only 8 (40%) presented a satisfactory result We agree that the evaluation of an immature hip, mostly
(P = 0.00). formed by cartilaginous tissue, may not represent the real anato-
There were no statistical relations between grades of se- mical situation.9 Uses of arthrography or magnetic resonance
quelae and the gender, side, and delay for drainage or no drainage. imaging are effective to demonstrate the nonosseous structure.
There were no differences, based on stability criteria However, these examinations may not be available, have high
before treatment between the group of patients from our study costs, and/or require anesthesia. In addition, we do not have
and the one published by Choi et al (P = 0.36). Also, there studies that provide a reliable correlation between their findings
were no differences in the results for both studies (P = 0.73). and the final outcome. Roentgenograms are still the standard for
the evaluation.
Our proposal of classification intends to be simple, ana-
DISCUSSION lyzing factors that could be important for the decision of
The diagnosis of SAH in infants and especially in treatment: instability and the presence of the femoral head. A
newborns may be difficult.6,9,11,14 Many cases have delayed study of concordance showed that it has less interobserver
diagnosis and treatment causing devastating complications variability than the one proposed by Choi et al14 (full
leading to decrease of life quality, multiple medical procedures, concordance of 25 in 41 hips against 7 in 41). Only on 5
and hospitalizations with a high social and economic cost. hips, there was no concordance for the stability criteria (eg, for
Some factors are related with the occurrence of the the same hip, 2 surgeons classified as grade 1 and 1 classified
sequelae: age at the time of infection, prematurity, newborns in grade 2).
that needed intensive care, type of bacteria, and delay of the Comparing the patients and results presented by Choi
treatment.1,2,4Y6,8,9,11,13 Most of the authors agree that a delay of et al14 and those presented to us, no differences were found.
4 days or more increases the chance of bad prognosis.1,5,16,17 There were 13 unstable hips (types 4A and 4B) and 21 stable
Only 21 hips of this series underwent surgical drainage at the hips (types 1, 2, and 3) in the Choi et al14 study and 20
acute phase. All of them had a delay of 5 days or more. unstable and 21 stable hips for our classification (P = 0.37).
Two anatomical factors may contribute for the poor re- The results obtained also did not present significant differ-
sults in young patients. The presence of intraarticular meta- ences either for stable (P = 0.73) or unstable hips (P = 0.72).
physis allows dissemination from bone to joint. Also, if the For these reasons, it seems more appropriate to use the sim-
main blood supply of the femoral head is intraarticular, the plest classification.
increase of pressure may cause avascular necrosis.18Y21 In our Hips with type 1 sequelae had a much better progno-
series, the newborns presented a worse result. Of the 25 hips sis, and the treatment goal was diverse. The procedures in
that had the acute disease as a newborn, 18 had more severe these hips are indicated to correct acetabular dysplasia or to
sequelae. Because we did not study children without se- improve congruence. For this, we prefer the shelf procedure
quelae, we cannot affirm that newborns have worse progno- or a proximal femoral valgus osteotomy. Both are safe and

* 2008 Lippincott Williams & Wilkins 527

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Forlin and Milani J Pediatr Orthop & Volume 28, Number 5, July/August 2008

can be done with an irregular femoral head (Figs. 3AYC). The 9. Mitchell GP. Management of acquired dislocation of the hip in septic
redirectional osteotomies like Salter or Steel may increase arthritis. Orthop Clin North Am. 1980;11:51Y64.
10. Morrey BF, Bianco AJ, Rhodes KH. Suppurative arthritis of the hip in
the pressure leading do stiffness (Figs. 4AYC). children. J Bone Joint Surg Am. 1976;58-A(3):388Y392.
Treatment of type 2 sequelae is controversial. Some 11. Morrissy RT. Bone and joint infection in the neonate. Pediatr Ann.
authors (Betz et al3 and Wopperer et al22) do not recommend 1989;18:36Y44.
treatment because of the good functional results in patients 12. Obletz BE. Acute suppurative arthritis of the hip in the neonatal period.
J Bone Joint Surg Am. 1960;42-A(1):23Y30.
who did not undergo reconstruction. They only recom- 13. Peters W, Irving J, Letts M. Long-term effects of neonatal bone and joint
mended contralateral epiphysiodesis to address the leg-length infection on adjacent growth plates. J Pediatr Orthop. 1992;
discrepancy. Many others recommended reconstruction, but 12:806Y810.
there is no consensus about the procedure to be used.14,23Y31 14. Choi IH, Pizzutillo PD, Bowen JR, et al. Sequelae and reconstruction after
septic arthritis of the hip in infants. J Bone Joint Surg Am. 1990;72:
Because of the risk of the stiffness, a proximal femoral valgus
1150Y1165.
osteotomy concomitant with lengthening of distal femur and 15. Hunka L, Said SE, MacKenzie DA, et al. Classification and surgical
external fixation was proposed.31,32 management of the severe sequelae of septic hips in children. Clin Orthop.
Like many authors, we prefer to perform the reconstruc- 1982;171:30Y36.
tion in a young child.9,14,15,22 Better results were obtained 16. Lunseth PA, Heiple KG. Prognosis in septic arthritis of the hip in children.
Clin Orthop. 1979;139:81Y85.
with the association of concomitant OR, proximal femoral 17. Parsch K, Savvidis E. Coxitis in the newborn and infant: diagnosis and
varus osteotomy, and Dega acetabuloplasty (Figs. 5A, B). therapy. Orthopaedic. 1997;26(10):838Y847.
Although treatment can transform the unstable hip (type 2) 18. Chung SMK. The arterial supply of the developing proximal end of the
in a stable one (type 1), theoretically improving the prognosis, human femur. J Bone Joint Surg Am. 1976;58-A(7):961Y970.
19. Kemp HBS, Lloyd-Roberts GC. Avascular necrosis of the capital
we are aware that the results may be deteriorated after a epiphysis following osteomyelitis of the proximal femoral metaphysis.
longer follow-up. J Bone Joint Surg Br. 1974;56-B(4):688Y697.
Leg-length discrepancy should be expected in all pa- 20. Trueta J. The normal vascular anatomy of the human femoral head during
tients with lesion of the proximal epiphysis. In the stable hips, growth. J Bone Joint Surg Br. 1957;39-B(2):358Y394.
the discrepancy varies from 2 to 5 cm, and for the unstable 21. Vidigal EC Jr, Vidigal EC, Fernandes JL. Avascular necrosis as a
complication of septic arthritis of the hip in children. Int Orthop.
hips, it reaches up to 14 cm. In our series, the maximum dis- 1997;1(6):389Y392.
crepancy in the centered hip (with or without reconstruction) 22. Wopperer JM, White JJ, Gillespie R, et al. Long-term follow-up of
was 4.5 cm at the last follow-up. Epiphysiodesis of the contra- infantile hip sepsis. J Pediatr Orthop. 1988;8:322Y325.
lateral distal femur would be the choice for these patients. 23. Forlin E, Cunha LAM, Kamp PA, et al. Tratamento das seqüelas da artrite
In a comparable group of patients, our classification séptica do quadril na crian0a. Rev Bras Ortop. 1992;27:299Y305.
24. Freeland AE, Sullivan DJ, Westin GW. Greater trochanteric hip
showed to be simpler and more reliable than those published arthroplasty in children with loss of the femoral head. J Bone Joint Surg
by Choi et al14 and may be useful for the prognosis and as Am. 1980;62:1351Y1361.
an indicator of treatment. Controversy still remains if the un- 25. Hallel T, Salvati EA. Septic arthritis of the hip in infancy: end result study.
stable hips will benefit from the reconstruction procedures. Clin Orthop. 1978;132:115Y128.
26. Harmon PH. Surgical treatment of the residual deformity from
suppurative arthritis of the hip occurring in young children. J Bone Joint
Surg Am. 1942;24:576Y585.
REFERENCES 27. Wang EB, Ji SJ, Zhao Q, et al. Treatment of severe sequelae of infantile
1. Bennett OM, Namnyak SS. Acute septic arthritis of the hip joint in hip sepsis with trochanteric arthroplasty. J Pediatr Orthop.
infancy and childhood. Clin Orthop. 1992;281:123Y132. 2007;2:165Y170.
2. Berdahl S, Ekengren K, Eriksson M. Neonatal hematogenous osteomye- 28. Rigault P, Lagrange J, Pouliquen JC, et al. La trochantéroplastie.
litis: risk factors for long-term sequelae. J Pediatr Orthop. Opération de sauvetage dans les mutilations de la tête et du col du fémur
1985;5:564Y568. chez l’enfant. Indications, technique et résultats à propos de 10 cas. Revue
3. Betz RR, Cooperman DR, Wopperer JM, et al. Late sequelae of septic Chir Orthop. 1973;59:565Y582.
arthritis of the hip in infancy and childhood. J Pediatr Orthop. 29. Weissman SL. Transplantation of the trochanteric epiphysis into the
1990;10:365Y372. acetabulum after septic arthritis of the hip. Report of a case. J Bone Joint
4. Chacha PB. Suppurative arthritis of the hip joint in infancy: a persistent Surg Am. 1967;49:1647Y1651.
diagnostic problem and possible complication of femoral venipuncture. 30. Krumins M, Kalnins J, Lacis G. Reconstruction of the proximal end of
Bone Joint Surg Am. 1971;53-A(3):538Y544. the femur after hematogenous osteomyelitis. J Pediatr Orthop. 1993;
5. Fabry G, Meire E. Septic orthritis of the hip in children: poor 13:63Y67.
results after late and inadequate treatment. J Pediatr Orthop. 31. Cheng JC, Aguilar J, Leung PC. Hip reconstruction for femoral head loss
1983;3:461Y466. from septic arthritis in children: a preliminary report. Clin Orthop.
6. Gillespie R. Septic arthritis of childhood. Clin Orthop. 1973;96: 1995;314:214Y224.
152Y159. 32. Cheng JCY, Lam TP. Femoral lengthening after type IVB septic arthritis
7. Grill F, Rustler T. Late sequelae of coxitis in infants. Orthopade. of the hip in children. J Pediatr Orthop. 1996;16:533Y539.
1997;26(10):848Y857. 33. Choi IH, Shin YW, Chung CY, et al. Surgical treatment of the severe
8. Howard JB, Highgenboten CL, Nelson JD. Residual effects of septic sequelae of infantile septic arthritis of the hip. Clin Orthop.
arthritis in infancy and childhood. JAMA. 1976;236(8):932Y935. 2005;434:102Y109.

528 * 2008 Lippincott Williams & Wilkins

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.