Many classification systems have been proposed but the most widely used one is that of Salter and Harris [24] based on joint congruity and physeal involvement and alignment.

Treatment varies with fracture type and site, but in general, best practice guidelines are based on case series (LoE IV) [2, 24–27]. Care must be taken with manipulations, surgical approaches and the use of fixation devices to ensure that no additional, iatrogenic damage takes place.

In Salter-Harris type I and II injuries permanent growth disturbance is rare and healing is usually uneventful unless damage to the blood supply co-exists e.g. in intra-articular physeal injury (proximal femur, humerus and radius) (LoE IV) [24, 28]. Treatment varies with the specific site of the fracture. In general, reduction should be gentle and anatomical: closed reduction and cast immobilisation is considered adequate treatment but, if anatomic closed reduction cannot be achieved then open reduction should be considered. If the reduction is unstable, fixation is required (LoE IV) [24, 26, 29].

In Salter-Harris type III and IV injuries both the physis and the articular cartilage are involved; both must be reduced anatomically and open reduction and fracture stabilisation is common [29].

In Salter-Harris fracture types I-IV, fixation, when necessary, is usually achieved with non-threaded wires or cannulated screws in the metaphysis or epiphysis parallel to the physis, or non-threaded wires across the growth plate supplemented by cast immobilisation [26, 29]. Fixation methods have not been compared. Experimental and clinical studies reported the use of biodegradable rods or screws across the growth plate with no evidence of growth arrest at 2 years (LoE IV) [30, 31]. However Bostman described a high incidence of discharging inflammatory foreign-body reaction and osteolytic foci in patients with displaced malleolar fractures treated by open reduction and internal fixation using absorbable polyglycolide rods [32]. Moreover, pin fracture did occur and the technique was not recommended for supracondylar fractures [33]. The technique has fallen out of favour over recent years.

The Salter-Harris type VI injury was not included in the original Salter-Harris classification. It was first described by Rang in 1969 as a compression injury to the periosteum or perichondrial ring resulting in a bony bridge, and subsequent angular deformity [27].

Growth arrest can affect any injured physis, however it is more likely to affect some physes than others. Growth disturbance has been reported after injuries to the triradiate cartilage [34, 35], the physes of the spine [36], the clavicle, [37] and those of the hands [38], and feet [39], but these are relatively rare events.

Treatment aims to prevent and/or manage the length inequality. Options consist of no intervention, compensatory orthoses for lower limb discrepancies, epiphyseodesis or shortening of the contralateral or paired bone, ipsilateral bone lengthening, or a combination of these (LoE IV/ V) [2].

Bright [4] classified partial growth arrest into three types: peripheral (type I), central (type II), and combined (type III). Partial arrests often result in some loss of length in addition to angular deformity and both problems need to be considered.

Angular deformity can be treated by acute (osteotomy) or gradual correction (distraction osteogenesis) in combination with a formal physeal closure and /or contralateral epiphyseodesis; or by physeal bar excision with interposition of an inert material. In the past, angular deformities were also treated by distraction of the growth plate and bar in children close to skeletal maturity (Hemichondrodiastasis) [2] (Fig. 45.7).

Physeal bar excision can be considered if less than 30–50 % of growth plate is involved (LoE IV/V) [2]. Younger children tend to have a better prognosis and less than 2 years of remaining growth is a relative contraindication for bone bridge resection (LoE IV) [63, 64]. It has also been noted that central bars are more amenable to resection than peripheral ones: probably related to the periosteal stripping required to resect peripheral lesions. Ischemic or septic-related bone bars have a poorer prognosis (LoE IV) [2]. The forces associated with normal growth can overcome a small physeal bar and although correction up to 30° has been reported, the degree of correction is variable and inconsistent (LoE IV) [4]. Angular deformities greater than 20° are unlikely to correct following bar resection and thus a corrective osteotomy is required. (LoE IV) [2, 62, 65]. Several interposition materials have been used: fat [53, 66, 67], muscle [67], polymeric silicone [68–70], bone wax [71], and bone cement, [63] and none has been shown to be superior. Current literature on the Langenskiold procedure – for the treatment of partial growth arrest is summarized in Table 45.1 [5, 51–58, 60, 69–71]. Recent experimental studies have explored the use of autogenous chondrocytes to fill the defect [72]. The main problem is to find a suitable donor site with cells retaining a certain growth potential. These experimental studies are still in the preclinical evaluation phase.

The purpose of epiphysiodesis is to produce either a permanent or temporary growth arrest. Its primary indication is a mild to moderate, (2–5 cm), actual or predicted leg length discrepancy in a skeletally immature child with adequate growth remaining and an acceptable final height at skeletal maturity [75–79].

Epiphysiodesis must be timed accurately so that leg lengths becomes equal at skeletal maturity: the timing of surgery is the most challenging aspect of this procedure. [75–80]. Different methods have been proposed to determine the amount of growth remaining at any individual time point: each has advantages and disadvantages [75, 80–83].

Several surgical techniques for epiphysiodesis are available. Some destroy the physis permanently [84, 85] while others, implant-mediated, are reversible and produce a temporary growth arrest [86–91]. Current literature on epiphysiodesis consists mainly of retrospective reviews, case series and expert opinion with few comparative studies (LoE III, IV and V).

Green and Anderson reported preliminary results of 77 procedures in 50 patients using several epiphysiodesis techniques, including their modified Phemister technique [93]. Later on they evaluated the results on 237 patients, 173 had permanent epiphysiodeses and 83 staplings [94]. With few exceptions, results were good or excellent in the 173 epiphysiodesis patients. Two had overcorrection greater than ½ inch, and 5 others underwent a contralateral epiphysiodesis to prevent overcorrection. Five patients developed angular deformity necessitating a corrective osteotomy in 4 and a repeat epiphysiodesis in 1. One patient developed osteomyelitis. Their overall rate of complications (including slow fusion) was 9.3 % (LoE III) [94].

Concerns about the cosmetic appearance of the scars and joint stiffness with the open technique, prompted interest in percutaneous modifications.

In 1984, Bowen and Johnson published the results of their percutaneous technique of epiphysiodesis which consisted of curetting the physis from either side and inserting an osteotome to a depth of 5 mm. They observed complete physeal arrest within 4 months in all 12 patients treated: the only complication was a keloid scar (LoE IV) [95]. In 1986, both Canale and Ogilvie, independently, reported that percutaneous techniques using a combination of drills and high-speed burrs effectively produced the desired epiphysiodesis effect in an animal model [96, 97]. These experimental findings have since been confirmed by many authors in clinical series using a variety of instruments (alone or in combination) such as cannulated tube saws or reamers, drills, burrs and curettes. Complete growth arrest was observed within 3–6 months from surgery in all series (LoE III and IV) [96, 98–107].

Haas described the first method of temporary, reversible epiphysiodesis, observing the retardation of bone growth when using a wire loop around the physis in an animal model. He used the same technique in five patients: in two the wire loop broke and had to be replaced but he noted growth retardation while the wires were in place, and subsequent resumption of growth when wires broke (LoE IV) [86]. Subsequently, he used staples in an animal model and obtained similar effects on growth inhibition but growth resumption after staple removal was not normal and not universal [87]. Blount and Clarke published the first clinical results of the use staples in 13 patients, 7 treated for leg length inequality and 6 for angular deformity. They noted that inserting three staples on either side of the physis, stopped growth immediately and almost completely (LoE IV) [88]. Green and Anderson evaluated 83 stapling procedures in 61 patients at skeletal maturity and found stapling to be effective, although the distal femur grew an average of 6 mm after stapling (LoE III) [94]. Similar experience was reported by other authors (LoE IV) [89, 108–110]. Bylander observed a gradual cessation of growth across the stapled physis over 6 months; moreover the incidence of staple extrusion was lower if staples had been bent prior to insertion (LoE IV) [108, 109]. Blount reported that 426 operations were necessary in 185 patients, but only 2 patients required osteotomy for final correction (LoE IV) [89]. Sengupta et al. found that 71 % of 503 patients treated with stapling had less than 1 cm of discrepancy at skeletal maturity and concluded that stapling is a simple, effective procedure in developing countries (LoE IV) [110].

The main complications reported for epiphyseal stapling were slow arrest, asymmetric growth, overgrowth or reduced growth after staple removal, staple extrusion and a need for surgery to correct residual deformities [89, 94, 109, 111–113].

In 1998, Metaizeau described his technique of temporary percutaneous epiphysiodesis using transphyseal screws (PETS). He reported results on 32 patients with limb length inequality concluding that the screws began to cause detectable growth inhibition within 6 months of insertion, slowing down the distal femoral and proximal tibial physes by 68% and 56 %, respectively. Maximum growth retardation was achieved at 12 months, when the distal femoral physis was inhibited by 89 % and the proximal tibial physis by 95 %. At skeletal maturity, total femoral growth was 45 % and total tibial growth was 52 % of the normal side (LoE IV) [90].

Several reports have documented the success of PETS in the management of limb length difference (LLD). Nouth reported on nine patients, showing an average reduction in LLD from 3.33 to 1.36 cm over an average 2.2 year follow-up: 56 % reached an LLD of less than 1 cm (LoE IV) [114]. Khoury followed 30 patients to maturity: PETS was successful in all cases and the average final femoral and tibial lengths were 0.15 cm and 0.05 cm, respectively, from predicted length (LoE IV). Ilharreborde observed a mean efficacy of the femoral epiphysiodesis of 35 % at 6 months, and 66 % at maturity and of the tibial epiphysiodesis, 46 % at 6 months, and 66 % at maturity when looking at 45 patients. The revision rate was 18 % (8 patients): 7 of 8 revisions (87.5 %) involved the tibia and the authors advised caution using the technique at this site. (LoE IV) [115]. In a retrospective series of 59 patients by Song, PETS showed an average LLD correction of 75.5 % in the distal femur and 78.9 % at the proximal tibia: the authors recommended that PETS should be performed at least 1 year earlier than the estimated epiphysiodesis timing in order to achieve equal leg length (LoE IV) [116]. Monier reported similar results in 16 PETS patients and concluded that results would have been better if the procedure had been performed at an earlier skeletal age [117].

The most recent implant introduced on the marked for growth modulation is the 8-plate (LoE IV) [91, 118]. Early experience was on the correction of angular deformities and only recently has the implant been used to correct leg length discrepancy. Pendleton reported his results on 34 patients with either congenital or developmental LLD followed until screw removal or skeletal maturity. The LLD change in patients who had guided growth of the tibia was minimal, but in those patients who had femoral or combined femoral and tibial guided growth the change was an average 10 mm. Twenty of 33 patients had a LLD of less than 1 cm at maturity or screw removal. One patient required treatment for angular deformity after guided growth for LLD (LoE IV) [91]. Stewart reported 11 patients who had dual 8-plate for LLD and observed a mean 4 mm improvement after 18 months, compared toa mean 15.5 mm improvement in those patients who had physeal ablation (LoE) [119]. Gaumetou reported similar results with a mean efficacy for femoral epiphysiodesis of 23 % at 6 months and 68 % at 18 months. The procedure was even less effective on the proximal tibia, with only 42 % of the expected growth arrest at 18 months (LoE IV) [120].

Current literature [94, 98, 102, 119, 121–125] is summarized in Table 45.2.