A ten-year-old limps after training, complains about his heel, and those around him settle it: "it's growth". The word is reassuring, and that is precisely the problem. True growing pains and Sever's disease are not the same thing, are not treated the same way, and are not missed in the same way.

Review article. Every figure carries its source where it is written, with its sample size and method. The evidence base on this condition is thin: ten randomised trials in all, most of low to very low certainty according to the most recent Cochrane review. When a statement has no support, it is flagged as such rather than papered over.

In brief: the clinical synthesis

What the literature says, without the usual rounding. Every point is taken up and sourced in detail in the body of the article.

• "Growing pain" describes neither a disease nor growth. The scoping review that examined 145 sources and two classification systems finds an "extremely poor" consensus: lower limb pain is mentioned by only half the sources, and 93% make no reference to growth¹.
• Two independent cohorts looked for the link with growth, and did not find it. Among 777 Danish preschool children followed by messages every two weeks, the authors conclude that the term is a misnomer² ; among 268 Taiwanese patients, no association with rapid growth in the two years following diagnosis³.
• Sever's disease is an entirely different matter : mechanical pain, on exertion, localised at the calcaneal insertion, in a child whose skeleton is immature. It is the leading cause of heel pain in children, with a measured incidence of 3.7 per 1,000 patients aged 6 to 17 in Dutch general practice⁴.
• It may not even be an apophysitis. MRI of children labelled as having Sever's disease and still in pain despite treatment shows the oedema in the trabecular bone of the metaphysis, not in the apophysis; the authors suggest seeing it as a stress fracture and conclude that the term apophysitis should not be used⁵.
• The diagnosis is clinical. Among 30 symptomatic children and 15 matched controls, single-leg toe standing had a sensitivity of 100%, the medio-lateral squeeze test 97% and palpation 80%, all three with a specificity of 100%⁶.
• The X-ray does not make the diagnosis, but it is not useless for all that. In the same study, all the children, cases as well as pain-free controls, had a dense apophysis, and half the controls had fragmentation⁶. Conversely, among 98 children referred for heel pain, 5.1% had a serious radiographic lesion (cysts, fibroma, stress fractures) and no feature of the history or examination made it possible to spot them⁷.
• The overuse story does not stand up in the way it is usually told. A prospective study concludes that there is no evidence that weight and activity level are risk factors⁸, and a study of 430 young athletes finds a higher risk in the younger and less trained children⁹.
• No treatment has shown itself superior. The 2026 Cochrane review brings together ten randomised trials and 654 children and rates the certainty of the evidence as low to very low for almost everything; the only moderate-certainty result is an absence of difference between foot orthoses and heel lifts¹⁰.
• The most useful trial in practice compares wait-and-see with two active treatments. Among 101 children, waiting, fitting a heel lift or providing physiotherapy all three produced a clinically relevant improvement, with no difference between them at three months¹¹.
• The order of magnitude to announce from the outset is two months, with an enormous spread. In a German football academy, the mean return to play was 60.7 days with a standard deviation of 64.9 days, and 13.6% of cases were recurrences¹².
• What must leave the benign frame : night pain that wakes the child, fever, a limp that persists, deterioration in general condition, non-mechanical pain, pain that no rest relieves. Calcaneal osteomyelitis in children has a median age of 8 years, exactly that of Sever's disease¹³, and limb pain is present in 43% of children at the diagnosis of leukaemia¹⁴.

Who this article is for, and what it does not cover

For physiotherapists who see children and adolescent athletes, in the clinic as well as at the side of the pitch, and who are presented with a complaint the family has already labelled.

It deals with the heel, that is, Sever's disease, and with the term itself, that is, the confusion between benign growing pains and overload apophysitis. It does not go over adolescent knee conditions again, which have their own pages on this site: Osgood-Schlatter disease has its own, and slipped capital femoral epiphysis, Legg-Calve-Perthes disease, Sinding-Larsen-Johansson disease and osteochondritis dissecans are covered together in the article on lower limb growth conditions. The chapter devoted to the knee here does only one thing, but one that none of those pages does: it starts from the complaint as it arrives ("his knee hurts, it's growth") and says which of the conditions we are heading towards.

Why is "growing pain" a bad name?

Because it labels two opposite clinical pictures with the same reassuring tag, and because the only one of the two that really bears the name has almost nothing to do with growth.

The term circulates in three mouths at once. Parents use it to say "it will pass". Coaches use it to say "he can carry on". And some clinicians use it to say "I found nothing worrying". These three uses do not cover the same object, and it is that imprecision that costs weeks of delay: a child who is told he has growing pain does not come back, because the word itself has removed the reason to come back.

What the only review to have counted the definitions found

In 2022, an international team published in Pediatrics a scoping review that does not try to treat growing pains, nor to find their cause: it simply sets out how the literature defines them. Eight databases and six diagnostic classification systems were searched from inception to January 2021; 145 studies and two systems, ICD-10 and SNOMED, were retained. Definition criteria were extracted in duplicate and grouped into eight categories: location, age at onset, pain profile, trajectory, pain types and risk factors, relation to activity, severity and impact, physical examination and further investigations¹.

The result is a finding of failure, and the authors put it this way: consensus between sources is extremely poor. The only more or less stable element is the location in the lower limbs, present in 50% of sources. Then come evening or night pain (48%), an episodic or recurrent course (42%), a normal physical examination (35%) and bilateral involvement (31%). No other criterion reaches 30%. More than 80% of sources give no age at onset. And 93% make no reference to growth, even though that is the word that gives the entity its name¹.

A diagnosis whose published definitions fail, in 93% of cases, to mention the very thing that gives it its name is not a diagnosis: it is a way of saying that nothing was found.

And what about growth in all this?

The question has been asked directly, by two teams, on two continents, with two different methods.

In Denmark, a study nested in a cohort of preschool children followed 777 children aged 3 to 6 at inclusion, between 2016 and 2019. Parents received a message every two weeks asking whether the child had been in pain; when the answer was yes, a telephone interview detailed the characteristics of the pain. The prevalence of growing pains comes out between 24 and 43% depending on the definition used, which in itself restates the imprecision of the framework. The pain occurred most often one to three times a week, was mainly in the legs, could be unilateral as well as bilateral, and most often had no consequences. On the point that concerns us, the authors are explicit: they find no relation with periods of rapid growth and suggest that the term is a misnomer².

In Taiwan, a retrospective study carried out at the national university children's hospital reviewed 268 patients under 12 diagnosed between 2006 and 2019, with a mean age of 4.7 years, and compared their weight and height z-scores at diagnosis and then up to two years later. Here again, no significant association between growing pains and rapid growth in the two years following diagnosis³.

A clinical review published in Sports Health reaches the same conclusion by another route, observing that the peak incidence of these pains does not correspond to a period of rapid growth¹⁵. This is therefore not a semantic detail: the mechanism the name suggests has never been demonstrated, and it has been looked for.

What the term really covers, and why that changes management

Two pictures that have nothing in common but the patient's age are filed under the same label.

On one side, so-called growing pains in the strict sense: most often bilateral, in the evening or at night, diffuse, with no precise painful point, with a strictly normal examination, no limp, and no relation to the day's exertion. The Sports Health review insists on this last point: the pain is not provoked by activity and does not cause a limp¹⁵. It is a diagnosis of exclusion, whose prevalence in 4- to 6-year-olds has been estimated at 36.9% with a 95% confidence interval of 32.7 to 41.1 from 1,445 valid parental responses in South Australia¹⁶.

On the other, overload apophysitis: mechanical pain, triggered and paced by exertion, precisely localised at a tendon insertion at a secondary ossification centre, often unilateral, with a reproducible exquisite tenderness on palpation. At the heel, this is Sever's disease. At the tibial tubercle, Osgood-Schlatter disease. At the patellar apex, Sinding-Larsen-Johansson disease.

The difference in management is immediate. Faced with the first picture, we reassure, we do not investigate, we do not treat, and we have nothing to change about sport. Faced with the second, we have a mechanical lesion at an insertion, a causal factor which is load, a therapeutic lever which is the modulation of that load, and a prognosis counted in weeks. Saying "it's growth" to a child with calcaneal apophysitis is sending him away without changing anything about what produces his pain.

Is Sever's disease really an apophysitis?

The question looks academic. It is not: the name given to the lesion drives the logic of treatment, and the one that has been used since 1912 may point to the wrong structure.

The classic description fits into one sentence of pathophysiology. The posterior calcaneal apophysis is a secondary ossification centre, separated from the body of the calcaneus by a growth plate until it fuses with it; it receives the pull of the calcaneal tendon behind and that of the plantar fascia below. Repeated impacts and traction are said to produce inflammation there: an apophysitis.

The timing of this ossification is usually given in precise years in textbooks, without the primary source being cited, and we do not reproduce it. What is measured, and enough for practice, is simpler: ultrasound of 336 young baseball players classified this apophysis into five maturation stages, and heel pain existed at every stage except the last, that of completed fusion¹⁷. The treatment trials, for their part, recruit between 8 and 15 years¹¹, with mean ages of 10.3 to 13.3 years¹⁰.

What MRI showed, and what contradicts the name

Ogden and colleagues performed MRI on children with a presumptive diagnosis of Sever's disease whose pain persisted despite conservative treatment. What they saw was not where the name predicted. The signal abnormality lay in the trabecular bone of the metaphyseal region, adjacent to the apophysis, with limited involvement of the secondary ossification centre itself. After immobilisation, the metaphyseal signal regressed or disappeared, in parallel with clinical improvement, whereas the apophyseal signal did not change⁵.

The authors draw a head-on conclusion: the disorder commonly called Sever's apophysitis would be a metaphyseal trabecular stress fracture, comparable to the calcaneal stress fracture of the toddler, and should not be called an apophysitis. They see it as an overuse lesion producing microdamage in an "equivalent" metaphyseal trabecular bone that has not yet adapted to the biomechanical demands of the growing, active child⁵.

Why the word matters for management

If the lesion is an apophyseal inflammation, the logic of treatment leans towards anti-inflammatories and rest. If it is a bone stress lesion from insufficient adaptation, the logic flips: what matters is no longer calming an inflammation but bringing load below the threshold the bone tolerates, then raising it again slowly enough for it to adapt. That is exactly the reasoning applied to adult stress fractures, and it leads to instructions different from "rest until it stops hurting".

This hypothesis finds indirect support in the fact that the pain is paced by exertion, that it reappears on resumption if that is too quick, and that 13.6% of cases in an elite football cohort are recurrences¹². It also explains why doing nothing works about as well as everything else: what heals, in the three arms of the pragmatic Dutch trial, is probably the reduction in load that each of them imposes by a different route¹¹.

We do not treat an inflammation we have not seen. We treat a load the bone no longer tolerates, and the lever is not rest: it is dosage.

The window closes when the apophysis fuses

A Japanese cross-sectional study ultrasounded the calcaneal apophysis of 336 young baseball players and classified its maturation into five stages, from stage 1, apophysis not visible, to stage 5, complete fusion. Heel pain was defined as pain felt on the squeeze test. Forty-nine players had pain, a prevalence of 14.6%, of whom 18 unilaterally (5.4%) and 31 bilaterally (9.2%)¹⁷.

The main result is negative: there was no difference in maturation stage between players with pain and the others. The stage the apophysis has reached therefore does not predict pain. But a secondary result is very useful in practice: no player had pain at stage 5, that of completed fusion¹⁷. In other words, maturation does not say who will have pain, but it does say when it can no longer happen. That is what underpins what we tell the family: this pain has an expiry date, and it is the closure of the calcaneal growth plate.

Who has heel pain, at what age, and for how long?

The figures vary by a factor of forty depending on the population surveyed. This is not a contradiction: it is a reminder that the question "how many cases?" only makes sense alongside "among whom?".

In the general population

Two studies have measured incidence outside the world of sport, and they agree.

In the Netherlands, the records of patients aged 6 to 17 from 34 general practices were analysed over three years, 2008 to 2010, using computerised registration networks. Out of 16,383 records examined, 61 children carried a diagnosis of calcaneal apophysitis, an incidence of 3.7 per 1,000 registered patients. The authors point out that the true incidence is probably higher, their inclusion criteria having been strict⁴.

In Istanbul, a retrospective audit of the records of all paediatric patients aged 6 to 17 seen between January 2014 and December 2017 finds 74 cases out of 20,967 children, a four-year incidence of 0.35 %. Boys were more often affected, bilateral forms were more common than unilateral ones, and consultations increased in spring, which the authors link to the resumption of physical activity¹⁸.

In athletes, and by sport

The figures change scale as soon as a sporting population is surveyed, and they change again according to how a case is defined.

In a German football academy, all injuries were recorded over ten years, from 2009 to 2018: 4,326 cases in 612 players, with a weekly on-site clinic. Twenty-two cases of calcaneal apophysitis were identified, an incidence of 0.36 cases per 100 athletes per year. The mean age at diagnosis was 11.8 ± 2.1 years, with unilateral involvement in 20 cases and bilateral in 2¹². This low figure is explained by the definition used (a case diagnosed in clinic, with time off) and not by any rarity of heel pain in this population.

That is what the Japanese study already cited shows, which did not wait for players to consult: by actively screening 336 young baseball players with a squeeze test, it finds 14.6% of children in pain¹⁷. Between 0.36% per year and 14.6% at a given moment there is no contradiction; there is the difference between the children who consult and those who have pain.

An Italian study assessed 430 young athletes aged 6 to 14 (48.1% basketball players, 29.5% footballers, 22.3% volleyball players), with a physical examination, the foot posture index and questionnaires. Prevalence was higher in the youngest, with no difference by sex, foot posture index, body mass index, surface type or sport⁹.

Who, exactly

A systematic review with thematic analysis reviewed 28 studies and 1,362 cases published up to June 2021. Boys accounted for 973 cases, or 71.4%. The presentation was bilateral in 589 cases (43.2%) and unilateral in 433 (31.8%), laterality not being specified for the rest. Radiographic imaging had been used for diagnosis in 358 cases, or 26.3%. All reported treatments were conservative, and 733 cases (53.7%) reported improvement, against 32 cases (2.3%) without improvement, outcome not being reported for 44% of cases¹⁹.

How long it lasts, and what should be announced

This is the question families ask first, and the honest answer comes in two parts: the order of magnitude is two months, and the spread is such that this figure predicts nothing for a given child.

The only series to have measured it in a structured sporting setting reports a mean return to play of 60.7 days with a standard deviation of 64.9 days. Athletes with a recurrence had longer recovery times and return to play than those diagnosed for the first time. Neither age nor body mass index at diagnosis had any effect on the time to return¹².

A standard deviation almost equal to the mean means that the distribution is very skewed: many children resume within a few weeks, a few take six months. That is the information to convey, and not the mean alone, on pain of creating a programmed disappointment in the eighth week.

How is the diagnosis made, and is an X-ray needed?

The diagnosis rests on three manoeuvres that take a minute. The X-ray, for its part, raises a question that two good studies answer in opposite ways, and one has to know which applies to the child in front of you.

Three tests, and their measured values

A Swedish team assessed 30 consecutive children with a Sever's injury, all in considerable pain and very active in sport, and 15 matched control children without pain. Three clinical tests were compared, and each was set against the true status⁶ :

• Single-leg toe standing. The child stands on one foot with the heel lifted. Sensitivity 100 %, specificity 100%.
• The medio-lateral squeeze test of the calcaneus, pressing the two sides of the heel between thumb and index finger. Sensitivity 97 %, specificity 100%.
• Direct palpation of the calcaneal insertion. Sensitivity 80 %, specificity 100%.

The authors propose basing the diagnosis on the history and the results of two specific clinical tests, and conclude that the diagnosis of Sever's injury is clinical and not radiological⁶.

What the X-ray shows in children who have no pain

This is the most useful result of that same study, and it should close an old misunderstanding. Increased density and fragmentation of the calcaneal apophysis have long been presented as the radiological signs of the disease, supposedly reflecting the inflammation linked to the pull of the calcaneal tendon. The authors therefore X-rayed the controls too.

The result: all the children, cases as well as controls, had increased density of the apophysis. And half the controls with no pain at all showed fragmentation, against nearly 90% of the children with pain⁶. In other words, what was shown to parents as proof of the disease is, in the young athlete, a picture of normal growth.

And yet there is a serious argument for X-raying

The preceding reasoning logically concludes that a child's heel pain should not be X-rayed. Another study, just as honest, reaches the opposite conclusion, and both must be read.

A team in Memphis reviewed all the records of children aged 4 to 17 who presented with heel pain, excluding acute trauma and calcaneal tendinopathies. Ninety-eight patients and 134 feet were retained, with a mean age of 10.8 years, all carrying a clinical diagnosis of calcaneal apophysitis. The radiographs were then re-read by three orthopaedic surgeons blinded to the clinical diagnosis⁷.

Five patients had an abnormality, all visible on the lateral view: three solitary bone cysts of the calcaneus, one non-ossifying fibroma of the distal tibia and two calcaneal stress fractures, one patient having both a cyst and a fracture in the same foot. The rate of abnormalities was therefore 5.1% of patients. These findings changed management, towards closer radiographic monitoring or immobilisation⁷.

The point that should hold the physiotherapist's attention is not the figure, it is the sentence that goes with it: no common feature of the history or the examination made it possible to identify the patients likely to have an abnormal X-ray⁷. One cannot therefore take comfort in the idea that atypical cases will be spotted on examination: in this series, they were not.

In practice, in the French setting, the physiotherapist does not prescribe this imaging: he has it requested. The wording matters, and "rule out a bone lesion in a child with heel pain, lateral view" obtains a more useful examination than "suspected Sever's disease".

What should take a case out of the benign frame?

The serious diagnoses in this chapter are rare. Yet they share with Sever's disease the age, the site and sometimes the mode of onset, and the only reliable filter is the behaviour of the pain, not its intensity.

One preliminary remark, because it changes how what follows should be read. A ten-year-old with heel pain has, in the vast majority of cases, calcaneal apophysitis. None of the pictures described here should lead to systematic investigation. What should change is the way the complaint is listened to: the question is not "is this serious?" but "does this pain behave like a mechanical pain?". A mechanical pain increases with load, decreases with rest, allows sleep, and comes with nothing else. As soon as one of those four points is missing, we are outside the frame.

Infection, which presents at the same age

A systematic review gathered 42 studies and 128 cases of paediatric calcaneal osteomyelitis published between 2000 and 2021. The sex ratio was two boys to one girl, the median age 8 years, the median duration of symptoms 0.6 months. Trauma was the main cause in 41 cases (54%), and limitation of activity the most frequent symptom, in 68 cases. Cultures were positive in 75.4% of samples, Staphylococcus aureus being the most frequent organism (57.1%). Surgery was performed in 51% of patients, and the rate of infectious recurrence was 6.8%¹³.

What to retain for the clinic: "limitation of activity" as the dominant symptom is exactly what a child with Sever's disease reports. What separates the two is fever, the constant character of the pain, and the absence of a mechanical rhythm. A child with heel pain at night, at rest, sitting down, and with a fever, is not a child to rehabilitate.

Tumour, which is paid for in delay

The Greek series of 12 children operated on for an osteoid osteoma of the foot between 1995 and 2010 is instructive less for what it describes than for what it measures. Mean age 12 years, from 8 to 16. The lesion was in the talus in 8 cases, in the calcaneus in 3, in the fourth metatarsal once. And above all: symptoms lasted on average two years, from 14 months to 4 years, before diagnosis. The authors note atypical presentations (pain referred to the hip, extensor tenosynovitis, atrophy of the calcaneal tendon) and conclude that the diversity of symptoms and the delay in diagnosis remain problematic²⁰.

The case of calcaneal osteosarcoma published in 2023 tells the same story in a graver key: a 19-year-old woman whose heel pain had resisted analgesics for four months. The authors stress that these tumours are frequently taken clinically for a traumatic or inflammatory process, and call for osteosarcoma to be kept in the differential diagnosis of chronic heel pain in the adolescent, despite its rarity²¹.

The signal is not the intensity of the pain, it is its refusal to follow load. Heel pain that does not improve when load is reduced is no longer an apophysitis: it is an open question.

The decision tree

The differential table: age, site, condition

The table crosses the two variables that sort out a child's pain attributed to growth, and adds for each condition the step that must not be missed. It runs from the heel down to the knee, and refers to the dedicated page where one exists: the knee conditions are not covered again here.

One line deserves to be read twice, that of the cysts and the fibroma. These are lesions that gave no distinctive sign, in children whose clinical diagnosis was calcaneal apophysitis, and that nevertheless changed management⁷. They do not justify irradiating every child's heel, but they explain why heel pain that does not follow load should be imaged rather than rehabilitated harder.

Which risk factors really hold up?

The usual story fits in one sentence: a heavy child, who trains too much, with stiff calves. Three studies have tested those three elements. None comes out intact.

What the systematic review says

A systematic review published in BMJ Open searched for the risk factors, associated factors and consequences of calcaneal apophysitis in six databases from inception to April 2021. Of 736 studies identified, 11 observational studies met the criteria, totalling 1,265 participants with a mean age of 10.72 years. Four studied extrinsic factors, ten intrinsic factors, three both²⁴.

The factors listed are limited ankle dorsiflexion, foot alignment, midfoot stiffness and mobility, plantar pressures and ground reaction forces, body mass index, age, sex, the presence of other osteochondroses and sports participation. The conclusion is cautious and worth quoting in its reservation: limited dorsiflexion is the most frequently studied intrinsic factor, followed by peak plantar pressures and foot malalignment, but disagreements exist between the investigators of the included studies, and there is no unanimity on what counts as a risk factor, an associated factor or a consequence²⁴.

This last distinction is decisive and rarely made. A child who has had heel pain for three months walks differently, loads the heel less, and loses dorsiflexion. Measuring his dorsiflexion at the moment he consults does not say whether the stiffness caused the pain or the reverse.

Weight and training volume: the story that does not hold

An Australian prospective study compared children with Sever's disease with a non-symptomatic population of the same age, testing the causal models most cited in the literature: biomechanical foot malalignment measured by the Root method and by the foot posture index, ankle dorsiflexion measured with modified equipment, body mass index, and the volume and type of sporting activity⁸.

The result is clear-cut and runs against intuition: this study finds no evidence that weight and activity level are risk factors. The odds ratio for reduced ankle dorsiflexion, statistically significant on the left side, was 0.93, which the authors themselves describe as clinically negligible. Only the forefoot to rearfoot ratio emerged with a clinically stronger and bilateral odds ratio, which leads the authors to point to biomechanical malalignment as a lead to explore, and no more⁸.

The Italian study of 430 young athletes goes the same way and adds a frankly counter-intuitive result. Body mass index, sex, surface type, sport and the foot posture index should not be considered risk factors. On the other hand, a significantly higher risk was found in the youngest subjects, in those who had fewer training sessions per week, and in those whose sessions were shorter⁹.

What is measured in these children all the same

The cross-sectional study nested in the Australian trial measured height, weight, waist circumference, body mass index, foot posture and ankle range in 124 children aged 8 to 14 with a clinical diagnosis of calcaneal apophysitis, 72 of them boys. Compared with normative values, these children had a higher body mass index, a higher weight and a greater height, all with p below 0.001, as well as differences in foot posture and ankle range²⁵.

There is no contradiction with the previous study, but a difference of question: comparing affected children with population norms is not the same as comparing cases with controls matched from the same sporting environment. The first method also captures what distinguishes a young athlete from a child in general.

The most directly useful result of this work lies elsewhere, in the multivariate analysis: the older participants (p = 0.046) and those whose pain had lasted longer (p = 0.043) reported more intense pain. The authors conclude that early management may reduce the intensity and duration of pain²⁵. In other words, the only clearly identified modifiable factor in this literature is the delay before management, which leads straight back to the problem of the term "growing pain", which defers the consultation.

A Japanese prospective cohort study provides the only element capable of guiding an exercise. Thirty-seven young gymnasts, that is 74 limbs, were assessed on flexibility, foot alignment, trunk function and functional balance tests, then followed for six months. The gymnasts who developed Sever's disease had a lower arch height ratio and lower values on the star excursion balance test in the anterior direction. The authors suggest that support of the medial longitudinal arch and dynamic postural balance work could prevent the condition²⁶ : a prospective suggestion, in 37 subjects from a single sport, which has never been tested in a trial.

What is each treatment worth, and at what level of evidence?

The short answer is disappointing and must be given as it stands: ten randomised trials in all, 654 children, and a certainty of evidence rated low to very low for almost everything. The long answer is more useful, because what these trials show in outline is worth more than what they show directly.

The state of the evidence, as the Cochrane review describes it

The Cochrane review devoted to non-surgical treatment of lower limb apophyseal injuries searched six databases with no language restriction up to 4 January 2025. It includes 10 randomised trials, seven of them on calcaneal apophysitis and three on traction apophysitis of the tibial tubercle, totalling 654 children with a mean age of 10.3 to 13.3 years, 73% of them boys. Risk of bias was assessed with the RoB 2 tool, and the certainty of the evidence with the GRADE method¹⁰.

The comparisons that produced a usable result are few¹⁰ :

• Foot orthoses versus heel lifts, calcaneal apophysitis: probably little or no difference in overall pain (mean difference 0.00, 95% confidence interval −0.44 to 0.44; 1 study, 123 participants) or in physical function (−1.30, −7.58 to 4.98; 1 study, 124 participants), in the short term. Moderate certainty.
• Kinesio tape versus placebo, calcaneal apophysitis: the data are very uncertain, for pain (0.10, −1.25 to 1.45) as for function (6.10, −0.08 to 12.28), from a single study of 22 participants. Very low certainty.
• Heel pad versus heel brace : very uncertain for physical function (−2.00, −12.48 to 8.48; 1 study, 43 participants). Very low certainty.
• No adverse effect reported for orthoses and heel lifts, across 11 studies and 101 participants, moderate certainty.

The authors' conclusion is explicit: the evidence is limited, rated for the most part as low to very low certainty, the outcomes were heterogeneous, and the trials did not specifically target children with persistent symptoms and functional limitation. No trial measured quality of life, whereas cohort studies have shown that apophyseal injuries can affect it in the long term, and none examined the economic impact¹⁰.

The trial that should guide the consultation

It is a pragmatic Dutch single-blind trial that compared three commonly used approaches in children aged 8 to 15 with symptoms for at least four weeks and a revised faces pain scale of at least 3 points: a protocol of wait and see, or a heel lift, or an eccentric exercise programme supervised by a physiotherapist. Treatment lasted ten weeks, the primary outcome was the faces pain scale at three months, with intention-to-treat analysis. One hundred and one children were included, three lost to follow-up¹¹.

At six weeks, the children with a heel lift were more satisfied than those in the other two groups, and improved significantly on the Oxford Ankle and Foot Questionnaire, child version, compared with wait and see; the physiotherapy group improved significantly on the parent version. But at the end of follow-up, all three approaches had produced a clinically relevant and statistically significant reduction in pain, with no clinically relevant difference between them. The authors conclude that clinicians should deliberate with patients and parents about the preferred treatment¹¹.

This result is richer than it looks. It does not say that nothing works: it says that everything works, including doing nothing in particular, which shifts the question away from the choice of technique towards support, load dosage and the explanation given to the family.

The twelve-month follow-up, which erases the early differences

The Australian 2 × 2 factorial trial recruited 124 children aged 8 to 14 with a clinical diagnosis of calcaneal apophysitis. Two factors were crossed: the type of in-shoe insert, heel lift or prefabricated orthosis, and whether or not the shoe was replaced. The primary outcome was functional disability, the secondary outcomes pain and dorsiflexion range, with follow-up at 1, 2, 6 and 12 months²⁷.

At 1 and 2 months there was a main effect of the heel lift, but only in the physical domain of the Oxford questionnaire (p = 0.04). At 6 and 12 months, no main effect and no interaction remained, for any outcome. The authors conclude that beyond two months the choice of treatment may be a matter of clinical judgement, cost minimisation or patient preference²⁷.

The table of approaches

How can load be managed rather than stopping sport?

This is the heart of the work, and it is also the part least supported by trials. We therefore have to distinguish what is demonstrated, what is deduced from a mechanism, and what rests on expert agreement, and say so to the family in those terms.

Why a complete stop is a poor tool

Three arguments converge, none decisive on its own.

The first is mechanistic. If the lesion is a stress lesion of an immature bone that has not yet adapted to the biomechanical demands of an active child⁵, then removing load entirely does not advance adaptation: it defers the problem until the resumption, where the stress met again will be relatively greater on a bone that has lost what it had gained.

The second is empirical and comes from the pragmatic Dutch trial. The "wait" arm was not a "stop sport" arm: it was a wait-and-see protocol. And it produced a clinically relevant reduction in pain, like the other two¹¹. Nothing in this literature shows that immobilising does better.

The third concerns consequences. The American Academy of Pediatrics clinical report on overuse injuries, overtraining and burnout in young athletes recalls that taking part in sport brings the child considerable physical and psychological benefits, and that burnout, defined as a state of physical or mental exhaustion with a reduced sense of accomplishment leading to devaluation of the sport, is one of the leading causes of dropout in youth sport, and a direct threat to the aim of lifelong physical activity³¹. A child kept away from his club for three months for a condition whose prognosis is spontaneously favourable pays a price that appears in no trial.

Pain as a compass, not as a prohibition

The practical principle rests on a distinction families grasp immediately: between pain one works through and pain that settles in.

In concrete terms, the child can continue an activity that wakes a tolerable discomfort during exertion, that subsides within the following two hours, and that has not increased by the next morning. He must cut back as soon as one of those three points is crossed: pain that forces him to change his running or to limp, pain that persists in the evening, pain on the first steps the next day. This framework has not been validated by a trial in calcaneal apophysitis, none has tested it, but it follows directly from the load reasoning and it gives the child a criterion he can apply on his own, which "stop if it hurts" does not.

A child told to stop as soon as it hurts learns that he can no longer do anything. A child given three criteria learns to dose, and that is what we leave him with for the future.

The concrete levers, in the order of what they cost least

Load modulation is not binary, and the most frequent error is to jump straight to cutting volume when three less costly levers exist upstream.

1. The surface and the shoe. This is the socially least costly lever. The heel lift comes in here: it showed an early advantage on satisfaction and on the physical domain at six weeks and two months^11,27, and the fact that this advantage disappears at twelve months takes nothing away from its value during the painful phase.
2. The content of the session rather than its duration. It is repeated impacts and traction that load the apophysis: sprints, jumps, changes of direction, running on the toes. Removing the twenty minutes of footwork from a football session costs the child far less than removing him from the session.
3. The frequency of peaks rather than total volume. Two training sessions and a match in the same weekend concentrate the stress; spreading them out changes the exposure without removing anything.
4. Volume, last of all. And never down to zero, unless the pain is present in everyday walking.

What is left for the physiotherapist to do

A quick reading of the literature could lead to discouragement: if supported wait and see does as well as a supervised programme, what is rehabilitation for? The answer is in the very structure of the trial that raises the question. All three arms included a consultation, an explanation, a framework and ten weeks of follow-up. What was compared was three techniques, not the presence against the absence of care.

And one of the few modifiable factors identified in this field points in exactly that direction: in 124 children, pain was more intense in those in whom it had lasted longer, which led the authors to recommend early management²⁵. The physiotherapist who recognises Sever's disease in a single consultation, rules out the red flags, explains the mechanism and puts the child back on a dosed load is doing precisely what that finding suggests.

What should be said to parents and to the coach?

This is the real intervention, and it is also the only one for which a measure exists: parents and children do not perceive this condition in the same way, and the gap narrows over time.

What parents see and the child does not say

A longitudinal study nested in the Australian trial had the Oxford Ankle Foot Questionnaire, child version, completed by the child and by the parent, at inclusion and then at 1, 2, 6 and 12 months. One hundred and thirty-three children were recruited, 124 took part, 101 completed the questionnaire at all five time points³⁴.

Agreement between child and parent ranged from poor (0.06) to good (0.77) for the physical domain, and from poor (0.09) to good (0.66) for the footwear domain. The school and emotional domains held up better, from moderate (0.46) to good (0.77). Above all: parents initially reported a greater impact than their child did, with the two viewpoints converging over the follow-up. The authors conclude that it is essential to understand the impact from the child's point of view as well as the parent's³⁴.

The consequence in consultation is direct. Questioning only the parent overestimates the impact at the start, and leads to heavier restrictions than the child experiences. Questioning only the child, who wants to play, underestimates it. Both are needed, separately, and it helps to name the gap when it appears: "you see him suffering more than he feels it, that is usual, and the two will come closer together."

The script for parents

The script for the coach

This is the most often forgotten party, and the one who in practice decides the real load. Three messages are enough, and they must be phrased in terms of the session, not of the diagnosis.

• What he can do: anything that does not cause a limp, all ball work with the foot flat, technical work, strength work, cycling, swimming.
• What we take out for a few weeks: sets of sprints, repeated jumps, prolonged work on the toes, hard surfaces if there is a choice.
• How we will know it is coming back up: no limp at the end of the session, no pain on the first steps the next morning. On those two conditions, we add one block of impact work per week.

The systematic review of 1,362 cases concludes, moreover, after noting the absence of any robust treatment data, that educating parents and coaches about the symptoms, aetiology and treatment of this condition is essential in order to diagnose it earlier and obtain better outcomes¹⁹. When the literature of a field cannot separate the techniques, it ends up pointing to what is left.

And faced with a painful adolescent knee, where are we heading?

"His knee hurts, it's growth" is the second most frequent phrasing after the heel one. This chapter does not go over any of the knee conditions again: it starts from that sentence and says which condition we are heading towards, with the link to the page that covers it.

First, why this chapter is short. Each of the conditions mentioned here has its own page on this site, written in full. Going over them again would create two texts competing for the same query, and the reader would lose the depth. What was missing was not one more description: it was the triage, that is, the translation of the complaint as it arrives into a direction.

What "growing pain of the knee" covers in consultation

The phrase is used for at least five distinct situations, only one of which is benign by definition.

1. An apophysitis of the knee. Mechanical pain, exquisite tenderness at an insertion. At the tibial tubercle this is Osgood-Schlatter disease, at the patellar apex Sinding-Larsen-Johansson disease.
2. Patellofemoral pain. Peri-patellar pain, on stairs, on squatting, on prolonged sitting. It is not a growth condition, but it is the most frequent one: the meta-analysis of 23 studies gives an annual prevalence of 28.9% in adolescents, a pooled point prevalence of 7.2% (95% CI 6.3 to 8.3) and of 22.7% in adolescent female athletes (95% CI 17.4 to 28.0)²³.
3. Pain referred from the hip. This is the costly trap, and it is dealt with in detail in the dedicated article: the knee can be the only site of the complaint while the lesion is a slipped capital femoral epiphysis.
4. Osteochondritis dissecans. Persistent exertional pain, poorly localised, with effusion, locking or giving way when the lesion is unstable.
5. A true growing pain, in the strict sense of the previous chapter: bilateral, in the evening or at night, diffuse, without limp, with a normal examination and no relation to exertion.

Triage in four questions

These four questions take less than two minutes and are enough to give a direction.

1. Is the pain bilateral, nocturnal and unrelated to exertion, in a child whose examination is strictly normal and who does not limp? If yes, and only if all four conditions are met, this is the framework of so-called growing pains. Nothing to investigate, nothing to treat, and it should be said clearly rather than leaving a doubt.
2. Can the pain be pointed to with one finger on an insertion? The tibial tubercle two to three centimetres below the patella, or the apex of the patella. If yes, this is an apophysitis, and the mechanism is the same as that of this heel: traction stress on an immature ossification centre, identical therapeutic lever.
3. Is hip internal rotation symmetrical? This is the manoeuvre that costs thirty seconds and avoids the most expensive delay in all of paediatric practice. An adolescent with knee pain, whose knee is normal on examination, and one of whose hips has lost its internal rotation, has a hip to image before he has a knee to rehabilitate.
4. Is there an effusion, locking or giving way? These three signs fall outside the framework of apophysitis and of patellofemoral pain syndrome, and call for imaging.

What do published clinical cases teach us?

Four genuinely published cases or series, with their identifier. None is reconstructed or composite: cases that are too clean are exactly what makes a review article untrustworthy.

Two monozygotic twins, the same heels

Two 10-year-old Indian monozygotic twin sisters, from a disadvantaged rural background, developed bilateral Sever's disease at the same time. Both reported bilateral heel pain of gradual onset for seven months, worsened by walking and physical activity. On examination, average build, weights of 26 and 27 kilograms, bilateral hallux varus in both, and a positive calcaneal squeeze test on both sides, which was enough to make the diagnosis. Anteroposterior and lateral radiographs of both feet confirmed the apophysitis. Faced with an identical clinical and radiographic presentation in two twins, the authors raise the possibility of a genetic predisposition or shared biomechanical risk factors³⁵.

What this case teaches: seven months of symptoms before diagnosis, in two children whose examination was characteristic in every respect and for whom a thirty-second test was enough. It is the exact illustration of the problem the word "growth" creates: these are not difficult cases that drag on, they are easy cases that no one brings in. And the finding that goes with it is James's: the longer the pain lasts, the more intense it is²⁵.

A heel investigated with ultrasound

A 10-year-old girl was managed for persistent left heel pain. Clinical examination and diagnostic ultrasound confirmed Sever's disease, the ultrasound also making it possible to document associated involvement of the calcaneal tendon. Treatment combined medication, immobilisation, rehabilitation modalities and exercises, with improvement after ten weeks³⁶.

What this case teaches: the order of magnitude of the timescale, consistent with the ten weeks of treatment in the pragmatic Dutch trial¹¹, and the value of ultrasound when the question is not to confirm Sever's disease but to find out whether something else is added to it. It is also a single case, with no comparison group: it demonstrates the effectiveness of none of the modalities used, and the authors do not claim otherwise.

Twelve children, an osteoid osteoma, two years of delay

Twelve children (7 girls, 5 boys) were operated on for an osteoid osteoma of the foot in a Greek department between February 1995 and February 2010, with a mean follow-up of 5 years. Mean age 12 years, from 8 to 16. The lesion was in the talus in 8 cases, in the calcaneus in 3 cases, in the fourth metatarsal once. Symptoms lasted on average two years, from 14 months to 4 years. A crucial point: night pain and relief with anti-inflammatories, held to be characteristic of this tumour, were present in only half the cases. Atypical presentations were observed: pain referred to the hip, tenosynovitis of the foot extensors, atrophy of the calcaneal tendon. En bloc excision achieved complete cure with no recurrence²⁰.

What this series teaches: the trap is not failing to know about osteoid osteoma, it is believing that it announces itself with its classic triad. In half the cases it does not. The only reliable signal is that of the red flags chapter: foot pain that lasts and does not follow load. Two years, in a child of 12, is two sporting seasons.

An osteosarcoma of the calcaneus

A 19-year-old woman consulted for heel pain resistant to analgesics for four months. The diagnosis retained was calcaneal osteosarcoma, an extremely rare entity since it accounts for less than 1% of all osteosarcomas. The authors stress that these tumours typically present with swelling and chronic heel pain, and that they are frequently taken clinically for a traumatic or inflammatory process²¹.

What this case teaches: it lies outside the age window of Sever's disease, and that is precisely why it is useful. At 19, the calcaneal apophysis is fused, and chronic heel pain can no longer be an apophysitis. The child's age is as much a criterion for exclusion as for inclusion.

What these four cases have in common

None is a case of spectacular diagnostic error. Three out of four share the same feature, and it is the one worth retaining: a long duration of symptoms before anyone asked the question. Seven months, two years, four months. In the first case it was an ordinary apophysitis that nobody had brought in to be looked at; in the other two, a serious lesion nobody had thought of.

The common thread is therefore not the difficulty of the diagnosis. It is the delay, and the delay has the same cause in both situations: a child's pain that is not frightening does not prompt a consultation, and "it's growth" is the sentence that prevents it.

How is this applied concretely in practice?

What is done at first contact, in what order, and the mistakes that cost time or credibility.

A five-step sequence

Step one, four questions in the history. Since when? How does the pain behave in relation to exertion: does it increase during, after, the next morning? Does it wake him at night? Has there been fever, weight loss, unusual tiredness? These four questions separate a mechanical pain from one that is not, and that is the only separation that matters at this stage.

Step two, three manoeuvres. Single-leg toe standing, on the painful side then the other. Medio-lateral squeeze of the calcaneus between thumb and index finger. Palpation of the insertion. The measured values of these three tests are in the diagnosis chapter; the first is the most sensitive and is also the one the child understands best⁶. Add a comparative ankle dorsiflexion measurement and a quick examination of the hip and knee, because one apophysitis does not exclude another²⁴.

Step three, decide whether we are leaving the frame. The red flags are in their box. In their absence, we do not request imaging to confirm, and we say so explicitly to the family, who often expect it: the X-ray does not make this diagnosis, and its signs are present in all pain-free young athletes⁶.

Step four, set out the load plan. With the child, not only with the parent. Name the three criteria he will apply on his own: no limp during the session, pain subsided within two hours, no pain on the first steps the next day. Choose the levers in the order of the corresponding chapter, starting with the content of the session and not its volume. Offer a heel lift if it helps: it has a measured early advantage and no reported adverse effect^10,11.

Step five, give the prognosis with its spread, and set the checkpoint. About two months, with wide variability¹². And an explicit rule: if after six weeks of properly dosed load nothing has changed, we do not treat harder, we reopen the diagnosis.

The mistakes that cost

• Using the word "growth" without qualifying it. It sends the child away without a plan, and delays the next consultation. Two of the published cases cited above had seven months and two years of symptoms^35,20.
• Stopping sport altogether. No data support it, and the cost in sporting dropout is documented elsewhere³¹.
• Having an X-ray taken to confirm. It is radiation that can confirm nothing, and it gives the family a picture of fragmentation that will then have to be defused⁶.
• Not having an X-ray taken when the pain does not follow load. This is the mirror-image mistake, and it is the more serious one: 5.1% of serious lesions in a series of 98 children, with no clinical warning sign⁷.
• Promising recovery in three weeks. The mean is two months and the standard deviation almost as large¹².
• Blaming the condition on the child's excess weight in front of him. It is humiliating, and it is not supported: two studies do not find body mass index to be a risk factor^8,9.
• Speaking only to the parent. Perceptions diverge, and the parent's initially overestimates the impact³⁴.
• Forgetting the coach. It is he who sets the real load six days out of seven.

Honesty about the level of evidence

It must be said to the family in terms they understand, because it changes the relationship. The wording that works: "we know this condition well, we know it gets better, and we do not know well which of the things we offer speeds it up; that is why I am suggesting we choose together what suits you best." That is exactly the conclusion of the authors of the pragmatic Dutch trial, who recommend deliberating with patients and parents about the preferred treatment¹¹, and it is also what a systematic review of eight randomised trials says, concluding that conservative treatment is effective in general without being able to single out a best one³⁷.