Surgical Versus Ponseti Approach for the Managementof CTEV: A Comparative Study

Catherine M. Duy, MD, FRCS (Tr & Orth), Jose J. Salazar, PhD, Lee Humphreys, MCSP,and Brona C. McDowell, PhD, MSCP

Background: Results from a comparative study of Ponseti versus

surgical management for congenital talipes equino varus

(CTEV), using historically managed patients, are presented.

No bias existed in terms of management choice or participants

recruited.

Methods: Twenty-three surgically treated children (31 club feet;

mean age 9.1 y) and 29 treated by the Ponseti technique (42 club

feet; mean age 6.5 y) agreed to participate in the study. Twenty-

six typically developing children (mean age 7.9 y) were also

recruited as a control group. A physical examination and

3-dimensional gait analyses were carried out on all participants,

and each child and his/her parent also, independently, com-

pleted the Oxford Ankle Foot Questionnaire (OxAFQ).

Results: The Ponseti group underwent fewer joint-invasive

procedures than the surgical group. Passive range of dorsiexion

and plantarexion were signicantly less in the CTEV groups

when compared with the control group (P

METHODSEthical approval was obtained from ORECNI

(reference number 08/NIR02/84).For this study a sample size of 40 children, 20 per

group, was chosen. A chart search was used to identifythose children with idiopathic CTEV treated between1999 and 2001, by serial casting and surgery, and thosetreated between 2001 and 2003, by the Ponseti technique.To allow for lack of consent, 80 children were identied,40 in each group, and invited, by letter, to take part. Allchildren who started in the Ponseti program were con-sidered as part of that program, regardless of subsequentintervention. Twenty-six children, from among friendsor family, without impairment were also recruited as acontrol group. Informed consent was obtained fromparents and children.

All participants underwent physical examinationand 3-dimensional gait analysis. All data were capturedby a senior physiotherapist and bioengineer.

Kinematic and kinetic data were collected using a 6camera Vicon 612 motion analysis system with Work-station software (Oxford Metrics, Oxford, United King-dom) and 2 AMTI force plates (Advanced MedicalTechnology Inc., Watertown, MA). A full description ofthe data capture technique is detailed elsewhere.20 Par-ticipants walked barefoot at a self-selected speed along aat 8m walk-way. A minimum of 5 trials of valid kine-matic and kinetic data for each limb were captured.

Representative gait cycles were chosen for analysis by thebioengineer. The following variables were preselected foranalysis; sagittal plane knee and ankle kinematics,transverse plane hip and foot kinematics, and ankle ki-netics. The GDI15 was calculated from the kinematic dataand used to provide an overall score of the walking pat-tern. Visual inspection of kinematic graphs was also usedto identify evidence of foot drop, equinus, calcaneus, andprolonged stance time. The latter 2 were detected byvalues >1 SD from normal mean values.

Children and parents completed the OxAFQ.19 Thisis a nondisease specific instrument developed to considerthe impact of foot problems from the point of view of achild (childs questionnaire) and their care-giver (parentsquestionnaire). It considers 4 domains; physical, schooland play, emotional, and footwear. Items are rated on a5-point scale indicating how frequently they affect thechild (from never to always). These scores are trans-formed to a percentage scale (0 to 100), in which a higherscore represents higher function. Parents filled in theirown form, whereas the physiotherapist used the interviewtechnique to complete the childs questionnaire.

For those with bilateral CTEV data were collectedfrom both limbs, and for those with unilateral CTEV datawere collected from the aected side only. Control datawere taken from both limbs. All data were analyzed usingSPSS version 13 (SPSS Inc., Chicago, IL). The analysis wasnot blind to the treatment modality that had been used.Data were tested for normality using the Kolmogorov-Smirnov test. For data normally distributed an analysis ofvariance was carried out. If signicant dierences wereobserved between the groups, a test for homogeneity ofvariances was used to determine which post hoc test wasappropriate for pair-wise comparisons: the Bonferroni testif equal variances were assumed or the Dunnett T3 test ifequal variances could not be assumed. For data that werenot normally distributed the Kruskal-Wallis test and sub-sequent Mann-Whitney U tests, as appropriate, were usedto test for dierences between groups. The Pearson corre-lation test was also used to assess the relationship betweenrotational measures.

TABLE 2. Surgical Procedures Carried Out in the Surgical and Ponseti Groups

Surgical Group Primary Procedure Secondary Procedure (n=27 Patients)

n=31 Turco21 17 (52%) Redo soft tissue release 12Crawford et al22 7 (21%) Tibialis anterior transfer 17Posterolateral release 5 (15%) Plantar fasciotomy 8Other 4 (12%) Cuboid osteotomy 2

Dilwyn Evans procedure 2Ilizarov frame 4Other tendon transfer 2

Ponseti Group Ponseti Procedures Adjuvant Procedures (n=4 Patients)

n=42 Achilles tenotomy 33 (79%) Posteromedial release 2Redo tenotomy 7 (17%) Cuboid osteotomy 2Tibialis anteriorTendon transfer

14 (33%) Split tibialis posterior transfer 1

Posterior release 3Ilizarov frame 2Plantar fasciotomy 1

TABLE 1. Participants Demographics

Control Groupn=26

Surgical Groupn=23

Ponseti Groupn=29

Male/female 17/9 20/3 20/9Mean age (range) 7.9 (5.2-10.8) 9.1 (7.5-10.3) 6.5 (5.0-8.0)Club feet: unilateral NA 15 (65%) 16 (55%)Club feet: bilateral NA 8 (35%) 13 (45%)Club feet: total NA 31 42

NA indicates not applicable.

J Pediatr Orthop Volume 33, Number 3, April/May 2013 Surgical Versus Ponseti Approach for Management of CTEV

r 2013 Lippincott Williams & Wilkins www.pedorthopaedics.com | 327

RESULTSTwenty-three patients of the surgical group and 30

of the Ponseti group responded to the invitation. One ofthose in the Ponseti group was found to have an under-lying neurological condition and was excluded from thestudy, leaving 29 children in this group.

The demographics for the participants are givenin Table 1. The surgical procedures carried out for bothgroups are outlined in Table 2. For the surgical group,27 of 31 feet (87%) required secondary surgery resultingin 47 remedial procedures. For the Ponseti group the in-itial success rate was 40 of 42 (95.3%) feet, with both feetin 1 patient failing to correct. This was due to the severityof the deformity. This patient, aged 1 year, went forwardto bilateral posteromedial releases and, later, cuboid os-teotomies. Thirty-three of 42 (78.6%) feet underwentAchilles tenotomy. Of the Ponseti group, 17 of 40 feet(40%) required 28 remedial procedures.

All clinical examination data followed a normaldistribution. In all measurements there were signicantdierences between groups (Table 3). Both of the CTEVgroups had signicant restriction of dorsiexion com-pared with the control group (P

Ponseti group was external and those in the surgicalgroup internal. The mean external hip rotation angle wassignicantly greater in both CTEV groups than in thecontrol group (P

DISCUSSIONThis was a retrospective trial in which the results for

the rst cohort of patients receiving Ponseti treatmentwere compared with those for the last cohort treatedsurgically. As there was no overlap time during whichboth treatment modalities were used, there was no bias inthe choice of management for either group. However,this also means that it is unavoidable that the childrentaking part in the study are necessarily of dierent ages,as has been the case in other studies5 in which sequential,rather than concurrent, treatment practices have beencompared.

In the interests of preserving all available data, andwith statistical advice that, as has been noted in otherstudies,7,23 it makes no dierence to the ndings, incommon with many other studies,3,57,16,2431 we includeddata from both limbs in cases of bilateral CTEV.

Initial correction of the club foot was achievedby the Ponseti technique in 95%, which is similar toother studies,4,23,32 as is the recurrence rate of40%4,23,24,27,30,32,33; most (90.4%) did not undergo anyprocedure other than those prescribed as part of thetechnique1 (Table 2). However, it was considered im-portant to include those children for whom other proce-dures were performed so that the entire treatment groupcould be assessed. We found, as others have,7 the need forremedial surgery to be less in those treated by the Ponsetitechnique, although it should be remembered that theirfollow-up time is also shorter. The requirement for sec-ondary procedures in the surgical cohort in this studymay appear quite high at 87%. Other studies quote muchlower revision rates14,25,28,31,34 after surgery, after a sim-ilar follow-up period, although Dobbs et al27 reported arevision rate of 87% in older patients who had beensimilarly treated. Cohen-Sobel29 reported a revision rateof 34% after surgery, but also reported marked residualfoot deformities and gait anomalies indicating that thethreshold for reoperation is a subjective phenomenon.Thirty-six percent of the secondary procedures done inthe surgical group were tibialis anterior transfers used totreat dynamic supination, rather than residual or re-current deformity, and to a certain extent this cohort mayhave been self-selecting: those who had more recent ormost frequent contact with the orthopaedic departmentperhaps being more likely to respond to the invitation toattend for the trial.

As reported in other papers,8,9,11,16,24,27,28,30,3538

restriction of passive ankle movement was found for bothCTEV groups compared with normal (Table 3), but wasstill within an acceptable range, as dened by previousstudies.18,35,39,40 However, as also found by others13 thisrestriction of passive ankle movement did not result incorresponding reduction of dorsiexion in stance, whichdid not dier between the 3 groups. Indeed, in commonwith other studies,4,11,13,14 it was noted that there weremore children in calcaneus in the CTEV groups, partic-ularly the surgical one. This, together with prolongedstance time, suggests that the triceps surae in childrenwith treated CTEV are unable to control the forward

movement of the tibia in stance, thus allowing the ankleto sag into dorsiexion.

Both CTEV groups exhibited poor plantarexion atpush-o compared with the control group. This, togetherwith the reduction in ankle power, noted also in severalother studies,8,1012,32,34,35,38 further indicates in-suciency of the triceps surae. The actual work done atthe ankle was also signicantly less in both of the treat-ment groups than in the control group, as found byAlkjaer et al.10 It is not possible to conrm whether thesendings are the result of primary pathology or if they arecontributed to by treatment.3

In common with other authors,4,12,35 foot drop wasfound in a signicant proportion (one third) of both theCTEV groups (Table 6), which is reected in the reducedmean dorsiexion in swing (Table 4). The nding of back-kneeing in 67% of the Ponseti group (Table 6) also tiesin with the signicantly greater mean knee extensionin stance found in this group (Table 4). This is almostcertainly a reection of proximal tightness of the tricepssurae.

In-toeing is a common feature of CTEV,1 whichmay relate to the initial severity of the deformity and theinvasiveness of treatment required.14 It has been attrib-uted to a combination of metatarsus adductus,11,12,14,16,29

internal tibial torsion,11,14,16 overactivity of tibialis ante-rior,11 talo-calcaneal malalignment,9,11 posterolateraltether,9 hind foot equinus,11 medial rotation of the talarneck,11 and internal rotation of the hind foot with respectto the tibia.3,12,13 Karol et al9 found the bimalleolar axis,as an indicator of tibial torsion, to be less external on thenonaected than the aected side in children with uni-lateral CTEV, and Laaveg and Ponseti16 found internaltibial torsion to be greater in those with CTEV who in-toed than in those with CTEV who did not in-toe.

The greater bimalleolar axis measurements found inthe Ponseti group (Table 3) indicate that this group hadless internal tibial torsion, compared with the surgicalgroup, and although the dierence did not achievestatistical signicance, the mean foot progression anglewas internal in the surgical group only. All the childrentreated by the Ponseti method wore abduction braces,1

which is considered inuential in overcoming in-toeing.3,4

Bracing was not used as part of the surgical technique.The lesser degree of in-toeing noted in the Ponseti groupmay also have been contributed to by the improvement inthe inclination of the talar neck, and hence the alignmentof the fore foot on the hind foot, brought about by thePonseti technique.41,42

Not all children with CTEV and internal tibialtorsion, metatarsus adductus, or medial spin of the hindfoot, in-toe,3,9 and these eects may be compensated, asdemonstrated by the children in the surgical group in thisstudy (Table 4) by greater external rotation12,15 at the hip.

GDI15 (Table 4) is a measure of the quality of gait,assessed using a scoring system that derives from kine-matic variables. The scores for those treated by thePonseti technique (90.6) were signicantly closer to nor-mal (97.3) compared with those recorded in the surgical

Duy et al J Pediatr Orthop Volume 33, Number 3, April/May 2013

330 | www.pedorthopaedics.com r 2013 Lippincott Williams & Wilkins

group (83.5). To our knowledge, no other paper hasdemonstrated this, although Karol14 did nd that thosefeet which had undergone the least surgery were asso-ciated with the most normal gait.

Whereas some scoring systems16 categorize theoutcome of club foot management as poor to excellent,the OxAFQ produces a score whose proximity to that of acontrol group is used as an outcome. In this instance, theparents and children of the Ponseti group produced me-dian scores in all domains of >85%; only the parentreport on the footwear item was lower than this (75%).The parents for the surgical group produced medianscores in all 4 domains of

22. Crawford AH, Marxen JL, Osterfeld DL. The cincinnati incision: acomprehensive approach for surgical procedures of the foot andankle in childhood. J Bone Joint Surg Am. 1982;64:13551358.

23. Richards S, Faulks S, Rathjen K, et al. A comparison of two non-operative methods of idiopathic clubfoot correction: the Ponsetimethod and the French functional (Physiotherapy) method. J BoneJoint Surg Am. 2008;90:23132321.

24. Cooper DM, Dietz FR. Treatment of idiopathic clubfoot. A thirtyyear follow-up note. J Bone Joint Surg Am. 1995;77:14771489.

25. Reichel H, Lebek S, Milikic L, et al. Posteroplantar release forcongenital clubfoot in children younger than 1 year. Clin OrthopRelat Res. 2001;387:183190.

26. Huang Y-T, Lei W, Zhao L, et al. The treatment of congenital clubfoot by operation to correct deformity and achieve dynamic musclebalance. J Bone Joint Surg Br. 1999;81:858862.

27. Dobbs MB, Nunley R, Schoenecker PL. Long term follow upof patients with clubfeet treated with extensive soft tissue release.J Bone Joint Surg Am. 2006;88:986996.

28. Hutchins PM, Foster BK, Paterson DC, et al. The long term results ofearly surgical release in club feet. J Bone Joint Surg Br. 1985;67:791799.

29. Cohen-Sobel E, Caselli M, Giorgini R, et al. Longterm followupof clubfoot surgery: analysis of 44 patients. J Foot Ankle Surg.1993;32:411423.

30. Changulani M, Garg NK, Bass A, et al. Treatment of idiopathicclub foot using the Ponseti method. initial experience. J Bone JointSurg Br. 2006;88:13851387.

31. Uglow and Clarke. Relapse in staged surgery for congenital talipesequino varus. J Bone Joint Surg Br. 2000;82:739743.

32. Dobbs MB, Rudzki JR, Purcell DB, et al. Factors predictive ofoutcome after use of the Ponseti method for the treatment ofidiopathic clubfeet. J Bone Joint Surg Am. 2004;86:2227.

33. Haft GF, Walker CG, Crawford HA. Early clubfoot recurrenceafter use of the Ponseti method in a New Zealand population.

J Bone Joint Surg Am. 2007;89:487493.

34. Aaronson J, Puskarich CL. Deformity and disability from treatedclubfoot. J Pediatr Orthop. 1990;10:109119.

35. Bor N, Coplan J, Herzenburg JE. Ponseti treatment for idiopathicclubfoot: minimum 5-year follow up. Clin Orthop Relat Res.

2009;467:12631270.

36. Gelder JH, van Ruiten AGP, Visser JD, et al. Long term resultsof the postero-medial release in the treatment of idiopathic clubfoot.

J Paediatr Orthop. 2010;30:700704.

37. Bach CM, Wachter R, Stockl B, et al. Significance of talar distortionfor ankle mobility in idiopathic clubfoot. Clin Orthop Relat Res.

2002;398:196202.

38. Sanker WN, Rethlefsen SA, Weiss J, et al. The recurrent club foot:Can gait analysis help us make better pre-operative decisions? Clin

Orthop Relat Res. 2009;467:12141222.

39. Sinclair MF, Bosch K, Rosenbaum D, et al. Pedobarographicanalysis following Ponseti treatment for congenital clubfoot. Clin

Orthop Relat Res. 2009;467:12231230.

40. Simons GW. Complete subtalar release in club feet: part II.Comparison with less extensive procedures. J Bone Joint Surg Am.

1985;67:10561065.

41. Pirani S, Zenznik L, Hodges D. Magnetic resonance imaging studyof the congenital club foot treated by the Ponseti method. J Pediatr

Orthop. 2001;21:719726.

42. Ippolito E, Fraracci L, Farsetti P, et al. The influence of treatmenton the pathology of club foot: CT study at maturity. J Bone Joint

Surg Br. 2004;86:574580.

Duy et al J Pediatr Orthop Volume 33, Number 3, April/May 2013

332 | www.pedorthopaedics.com r 2013 Lippincott Williams & Wilkins