From neck to hip, he is encased in plaster of Paris. It's hard, unrelenting and no substitute for the tiny body hidden behind it for the past 12 months. Worse than that, it was me who chose to have him put inside this dreadful second skin.
Edward suffers from infantile scoliosis, a condition causing his spine to curve and twist to one side. Unchecked, some cases get better. But Edward is one of the minority who needs help. Without his plaster jacket, his spine would continue to curve and he would become terribly deformed with a characteristic and unsightly rib hump on his back.
Opting for the plaster was easy. No mother wants her child to be physically deformed, and I believed a cast offered better prospects than a brace. Results are hard for the lay person to find, but the few I found in a 10-year-old medical publication convinced me that I was making the right decision.
Getting the treatment, however, was fraught with problems. Although plaster was pioneered almost 20 years ago, there are still only a handful of orthopaedic consultants practising the technique.
Others favour the traditional method of braces and are sceptical about plaster, partly because of the lack of published results and partly because they feel there are risks attached.
The three-month delay in treating Edward, while I tried to find out who would put him in plaster, has cost my son dearly. His curve, initially what I now consider a modest 30 degrees, deteriorated to more than 50 degrees. He was, according to the orthopaedic surgeon now treating him, completely off balance.
Brad Williamson is based at one of the North-west's specialist scoliosis units at the Royal Manchester Children's Hospital. We were referred to him after I refused to wait until Edward was two for bracing treatment to begin, which his first consultant recommended.
Mr Williamson has been using plaster casts on children with infantile scoliosis for the past three years and says none of them have regressed.
He is openly sceptical that bracing could achieve the same results. "All scoliosis textbooks tell you that bracing works," he said, "but I've never seen any scientific evidence to back that up."
Mr Williamson's chief aim is to straighten the spine completely. "As long as you can get the spine straight and keep it that way for a period, it will grow straight," he says. This has been achieved using plaster jackets by the orthopaedic surgeon who pioneered the technique - Min Mehta from the Royal National Orthopaedic Hospital Trust in Stanmore, Middlesex. Some of her earliest patients are now adults whose spines remained straight after treatment.
Because there is still a great deal of ignorance about infantile scoliosis, even among professionals, children are often referred too late for Mr Williamson to achieve his goal. "I have children with curves of 90 degrees and they are a surgical nightmare," he said.
His best alternative for children such as these, and for Edward, is to reduce or "hold" the curve with plaster jackets to delay or minimise surgery.
Eleven of his patients, with ages ranging from 18 months to nine years, are currently in casts. It is, he says, a benign and generally well-tolerated treatment.
It is not known exactly how many children suffer from infantile scoliosis. Some studies estimate up to four per cent of school-aged children have curved spines. Of these, some will have developed the scoliosis in infancy - before the age of three.
Most cases of scoliosis do not develop until adolescence, when management and treatment of the curves is less of a problem than in young children still growing vigorously.
It is partly because of the adolescent form of the condition that youngsters with infantile scoliosis are suffering through late referral.
While adolescent scoliosis tends to need surgery when the curve reaches about 40 degrees, the same criterion cannot be applied to scoliosis in a baby.
"By the time infantile scoliosis has reached 40 degrees, it's quite a bad curve," says Mr Williamson. "If we could get the children young with smaller curves, we'd be able to straighten them out rather than try to tide them over until they're teenagers. The casts are most effective while the child is still growing. Early referral is the key."
To assess scoliosis, X-rays are taken and calculations made based on the angles between a specific vertebra and ribs. Generally, if the outcome is less than 20 degrees, the curve is likely to resolve; anything above 20 degrees tends to get worse. The method is not foolproof, however, and there are other features that need consideration, as Mr Williamson points out.
"Sometimes, the children have asymmetric heads or pelvises," he said. "They may have congenital heart defects or suffer from a syndrome. On the whole, their curves tend to progress. It's the big, healthy, normal babies who only have scoliosis that do better."
Edward's treatment is not pleasant. The cast is applied while he is under a general anaesthetic, with gentle traction helping to straighten his spine.
It is held in this new position by the cast, which is carefully moulded by the consultant to push the spine as straight as possible and to flatten any rib prominence.
Casts remain in place from six to 12 weeks, depending on the child's rate of growth, with only one or two hours of freedom in between for bathing and cuddling.
For those in nappies, however, accidents occasionally force early removal and older children have been known to immerse themselves in paddling pools - with obvious results.
It is true that children adapt quickly to their plaster "skin". Edward was back on his feet within 24 hours of having his first cast fitted. For parents such as me, however, the treatment can be emotionally devastating.
Exchanging the warmth and softness of my son's body for a rigid and unfeeling plaster is something that, almost 12 months on, I still have not grown accustomed to. At worst, he could be in and out of plaster until he is seven or eight.
Nobody knows what triggers scoliosis, although research is under way into its possible causes. One theory is that it is due to a fault in the way messages from the spine are transmitted to or processed by the brain.
The reason for the high incidence of the condition in Britain is also a mystery. Although infantile scoliosis is seen throughout Europe, it is virtually unknown in the United States.
The knowledge that Edward's casts could make the difference between a straight spine and a badly deformed one, or reduce the scale of any surgery from major to minor - or even none at all - is often the only bright light at the end of a very dark tunnel.
SAUK, the Scoliosis Association (UK), 2 Ivebury Court, 323-327 Latimer Road. London, W10 6AR.
Arise, The Scoliosis Research Trust, Royal National Orthopaedic Hospital, Stanmore, Middlesex HA7 4LP.