Younger children usually do better; studies put the age where it starts to matter at 6 to 8Strong evidence
In a US multicentre study of 345 hips in children aged 6 to 12 at onset, treatment made no difference in children aged 8 or under, and hips with the most severe X-ray grade did poorly whatever was done. In a Norwegian study of 368 children, treatment made no difference under 6, and having more than half of the ball affected was the strongest predictor of the outcome. In the UK study, onset over 6 and being a girl each predicted a worse hip shape at 2 years.
So the direction is clear: younger children usually do better. The exact age where it starts to matter differs between studies, at 6 to 8, and an NHS leaflet gives 6 or younger as the best outlook and 8 or older as a sign of a poorer one.
Sources 12, 13, 6, 2
Surgery or not: the specialist's decision, with real uncertaintyMixed evidence
In the UK study, two-thirds of hips were treated without surgery over 2 years, most by watching and waiting alone, and about a third had an operation. Surgeons' decisions turned mainly on stiffness and an age over 8, and surgery made no detectable difference to the hip's shape at 2 years.
Other studies point the other way for some children: in the US study, children over 8 with a middle X-ray grade did better with surgery, and in Norway, children over 6 with more than half of the ball affected did better with an operation on the thigh bone (an osteotomy). Because of this uncertainty, a UK national trial, Op NON-STOP, is randomly allocating children aged 5 to 12 to surgery or the best non-surgical care.
Sources 6, 12, 13, 14, 15
Physiotherapy keeps the hip moving; there's no good evidence it changes the final shapeEmerging evidence
A 2020 review found 15 studies comparing non-surgical treatments in 1,745 children, none of them randomised trials, and said no recommendation could be made for any of them. In the one controlled physiotherapy study, 8 children who did stretching, strengthening and balance work gained hip movement and strength, while 9 who were only observed lost movement, but the X-rays didn't change. In the US multicentre study, hips treated with range-of-movement exercises ended up no different from those with no treatment or a brace.
Physiotherapy is still standard, and sensible: it was recommended for 61% of children in the UK study, and in a UK case review, 26 of 32 children were referred and seen a median of about 10 times in 2 years. In an international survey of 160 surgeons, 71% rated keeping the hip moving as very important, most supported stretching in the early stages, and they disagreed about strengthening then. Surgeons and physiotherapists in Sweden agreed on stretching at every stage and disagreed on strengthening in the first two. So the exercises in this guide are for movement, strength and comfort, not to reshape the ball.
Sources 16, 17, 12, 6, 18, 19, 20
Activity: high impact paused in the early stages, swimming and cycling keptEmerging evidence
Surgeons in the international survey strongly agreed on discouraging high-impact activity, such as trampolining, gymnastics and ball sports, including school sport, in the early stages, easing these limits somewhat as the hip heals, and they allowed or recommended swimming and cycling in almost every case. Many used pain, as well as the stage, to guide activity, and there was no agreement on when full activity can return.
UK leaflets say the same: avoid running, jumping (including trampolines) and contact sport; swimming is a good activity because the hip moves through its full range without the pull of gravity; cycling, with supervision; and in PE, balance, ball skills and climbing without jumping off the apparatus. Australian state health advice is the same. None of this has been tested in a trial.
Sources 19, 20, 2, 21, 22, 23
Walking and weight-bearing: the team decides, and restrictions have costsMixed evidence
UK and Scandinavian surgeons mostly let children walk as comfort allows, while surgeons in Eastern Europe recommended no weight on the leg; none of the Swedish surgeons or physiotherapists surveyed recommended keeping all weight off it. In a US study, at least 6 months with little or no weight on the leg, started at the earliest stage, went with less flattening (good X-ray results in 50% against 0% with treatment of the symptoms only), and a review cites another study favouring longer weight relief.
Restrictions have costs too. In 100 US children aged 8 to 14, restrictions went with worse mobility scores, and in 92 children aged 5 to 7, moderate restrictions went with more anxiety and low mood. They didn't lead to extra weight gain in 130 children. These are observational studies of selected children, so this is a decision for your child's team, and if your child is restricted, ask about support for mood, play and school.
Sources 19, 20, 24, 4, 25, 26, 27
Braces and castsEmerging evidence
In the Norwegian study, a brace that held the legs apart did no better than physiotherapy, and the authors advised giving it up. The 2020 review found similar rates of poor outcomes with braces and with no treatment. Casts that hold the legs apart are still used in some countries, usually by specialist decision.
Sources 13, 16, 3
Hip replacement in adult life: two long-term cohortsMixed evidence
In Denmark, 13% of 167 patients treated in childhood with a splint had had a hip replacement at a mean of 47 years later, against none of a matched group without Perthes disease. In Norway, 19% of 244 hips treated without surgery had been replaced at a mean of 48 years, at an average age of 46.
Other figures vary with the era, treatment and length of follow-up: 6.8% after non-surgical treatment across 27 long-term studies, rising to 23% in studies that followed people for over 40 years, and 22% in a web survey of 1,182 adults, who chose to take part and may include more people with problems. In a small Dutch group with moderate disease, over half showed signs of arthritis on X-ray at a median of 41 years. So many adults who had Perthes have good hips for decades, and some need a replacement in middle age.
Some leaflets give much gloomier figures: one NHS leaflet estimates that about half will need a hip replacement by 50 to 60, and an older Australian fact sheet that a third have arthritis needing treatment by 30. The long-term studies above don't support numbers that high, so if you've read them, ask the team what they expect for your child.
Sources 28, 29, 30, 31, 32, 2, 33
Hip shape and the risk later: one cohortEmerging evidence
In the Norwegian cohort, 3% of hips that healed round needed a replacement, against 25% of oval ones and 46% of those that healed flat. Onset at 6 or older went with more replacements (28% against 10%).
That is one cohort's finding: the meta-analysis of 27 studies didn't find a direct link between the X-ray shape grade and hip replacement, though it did find more replacements with older age at onset.
Sources 29, 30