A mother with her 1-year-old son, who currently uses a gastric feeding tube, in Vermont.Credit...Elinor Carucci for The New York Times

Feature

When Your Baby Won’t Eat

Our daughter started life on a feeding tube. Then we tried to wean her off it — and began to understand the complexity of how children relate to food.

Violet looked perfectly normal, but she had been slowly suffocating for at least a week, and we had no idea. Violet seemed like an ordinary newborn: From birth, she cried when she was hungry, slept when she was full. When awake, she stared at us intensely; when she slept, she slept a lot — one night, for nine hours. When the time she spent breast-feeding went from 45 minutes to 20, then 10, then five, I thought the two of us were getting better at it. We weren’t. And the purplish tinge to her lips, hands and feet didn’t mean she was chilly. Violet’s heart was failing.

By the time we began to grasp this, the oxygen level in Violet’s blood was only 75 percent of what it should have been. She was admitted to Maria Fareri Children’s Hospital in Valhalla, N.Y., and put on a ventilator, the breathing tube snaked down her throat before she was fully sedated. Even with a machine breathing for her, Violet’s oxygen plummeted. When she hit 20 percent, a cardiologist threaded a catheter into her heart, where he inflated a tiny balloon and tugged, punching a hole through her interatrial septum to release a gush of pent-up oxygenated blood. That was the first time we broke Violet to save her. The next day, we began to learn how several congenital defects had made Violet’s heart ‘‘incompatible with life,’’ and how a cardiothoracic surgeon could cut apart veins and arteries and sew them back together in a life-sustaining pattern over the course of three open-heart operations. When done successfully, this Fontan circulation, as the process is known, enables a child to reach a healthy, if heavily monitored, adulthood.

But this isn’t a story about heart defects. It’s about side effects. When it comes to repairing a heart no bigger than a walnut, the list of things that can fall apart while you’re solving the most obvious problem is virtually endless. In our case, the collateral damage was swift and dangerous: Violet stopped eating.

[The topics parents are talking about. Evidence-based guidance. Personal stories that matter. Sign up now to get NYT Parenting in your inbox every week.]

At first, she didn’t have the energy. This is what had been happening during those initial short feeds — Violet was too oxygen-deprived to make an effort. But our doctors were determined that she regain the half pound lost while she was dying, so she was given a feeding tube almost as soon as she came off the ventilator. A nurse inserted a nasogastric tube into her nose, then pushed it down her esophagus and into her stomach. That tube was connected to a feeding pump beside her Isolette. The nurses ran a cocktail of formula and breast milk through the pump every three hours, pushing as many calories as possible into Violet, whether she was awake or asleep.

Violet’s first open-heart procedure took place a week later. We all assumed that breast-feeding would resume immediately. But Violet continued to tire out so quickly that the doctors figured she was burning more calories trying to eat than she could possibly take in. So a second nasogastric tube went in, this time after very little deliberation.

‘‘It’s a temporary measure,’’ they assured us. ‘‘Just till she gets her strength back.’’ We thought she would be eating normally within two weeks.

Instead we went home with the nasogastric feeding tube still in place, and every three hours, we circled through the same dance: First I tried to get her to nurse as she kicked frantically, turning her head away in fear or fury. Next my husband, Dan, offered a bottle while she cried, gagged, fought. Every three hours, we tried bottle or breast, all the while taking detailed notes to record how long she latched on or how many milliliters she swallowed. (Never more than a teaspoonful.) Every three hours, we tried this most fundamental act of parenthood and failed. Then we connected her nasogastric tube to the blue-and-white plastic feeding pump and listened to the machine whir and beep as it fed our baby.

Because babies begin nursing in the first hours of life, because the cry of hunger is one of our first communications with the world, it’s easy to assume that eating is our most primitive instinct. Yet it’s an instinct that must be reinforced constantly. A baby cries, a breast or bottle is offered; the baby sucks until she feels better. Most newborns do little else in their first few months, until their ability to eat is finely honed and the feeding relationship between parent and child is thoroughly established. In this way, the instinct to eat isn’t just a need for physical nourishment — it also ensures that babies form secure attachments. It’s how they fall in love.

As a baby grows, the act of eating becomes increasingly intricate. Learning to chew solid foods, use a spoon and drink from an open cup all require fine motor skills and constant practice. Yet for most babies, the process happens so seamlessly that this learning looks instinctive: The infant happily gums toys, then graduates to slurping down spoonfuls of applesauce, quickly associating such foods with the same satiety experienced through breast-feeding. ‘‘Babies come into the world predisposed to learn all kinds of different things,’’ says Leann Birch, a psychologist who studies infant feeding and childhood obesity at the University of Georgia. ‘‘There is a developmental timetable for when a baby can swallow food or move her tongue and jaws in certain ways. But without the right learning experiences, it won’t come together.’’

It turns out that the instinct to eat is surprisingly fragile. Only around 100,000 children in the United States have problems severe enough to require the use of a feeding tube, according to estimates by the Feeding Tube Awareness Foundation. But 25 to 45 percent of all children develop the kind of habits that pediatricians and therapists see as the hallmarks of a ‘‘problem feeder.’’ They may refuse to eat certain flavors, textures or even entire food groups; others eat too much. Colic, reflux or a poor latch can cause an otherwise healthy infant to go on a temporary hunger strike. An increased suspicion toward new foods is expected with toddlers. But while some parents and pediatricians may panic over these normal developmental stages, others may dismiss a sensory processing problem or weak oral motor skills as just picky eating.

Either reaction can result in a kind of daily instinct override for the finicky toddler whose parents turn every meal into a battle over ‘‘just one more bite,’’ and, conversely, for the stocky 5-year-old whose worried parents ban second helpings. Over time, a child can be conditioned by parental instructions to ignore her own instincts, though not always with the desired result. Studies have found that when children are rewarded for eating healthy foods, they tend to like those foods less and crave sweet treats more. ‘‘There’s a tension here, because we need children to become socialized to eat at mealtimes,’’ Birch acknowledges. ‘‘And yet parents often think they need to take more control of this than they should.’’

In Violet’s case, the eating instinct was destroyed almost as soon as it emerged, by what’s known medically as an ‘‘oral aversion.’’ Also referred to as ‘‘oral defensiveness,’’ and more unnervingly as ‘‘infantile anorexia,’’ this condition results when a child refuses to eat as a way of protecting herself from perceived trauma. Somehow, as a result of those early nursing struggles, the emergency intubation in the hospital or perhaps our own ceaseless efforts to get her to eat, Violet forged a connection between eating and pain, just as dogs learned to salivate at the sound of a bell in Ivan Pavlov’s classic experiment on conditioning. A baby with an oral aversion can lose those digestive reflexes and instead feel nauseated at the sight of a breast or bottle. She might not ever feel hungry, especially with a feeding tube supplying all her nutrition.

Whatever causes the initial interruption, the results seem to be the same: a child who no longer connects to her own internal sense of hunger and satiety, but instead relies on external cues to decide whether and how much to eat. In this way, successful eating requires both our most primal instinct and the right set of learned behaviors. When eating goes wrong, whether it’s a life-threatening aversion like Violet’s or a common case of pickiness, parents and medical professionals find themselves at a version of the same crossroads: Do you try to correct the behavior — training a child to eat well, Pavlov-style — or do you try to rediscover that primal urge and trust her to take it from there? It’s a divisive question among the doctors and therapists who work with children like Violet, as well as a debate unfolding, consciously or not, around most kitchen tables in the country.

By the time she was 2 months old, Violet was entirely dependent on the feeding tube and no longer displayed the slightest interest in eating. Our pediatrician connected us with Lynne Westgate, a speech-language pathologist at MidHudson Regional Hospital in Poughkeepsie, N.Y. Speech-language pathologists often work as feeding therapists because good oral motor skills are needed both to speak and to eat; some physical therapists, occupational therapists, pediatricians and psychologists also work in this area. In most cases, a family’s choice is dictated by geographic availability; we had no idea what kind of therapy Westgate would pursue.

The most common approach, used by almost all the nearly 30 feeding programs found in children’s hospitals and private clinics around the country, is a ritualistic method known as one-to-one reinforcement. Think of it as the Pavlovian approach: It’s a form of ‘‘behavior modification,’’ a psychological tactic in which food refusal is classified as negative behaviors to be systematically replaced with positive ones. ‘‘Babies can grasp cause and effect very early,’’ says Amy Kathryn Drayton, a leading behaviorist who directs the feeding program at the University of Michigan’s C.S. Mott Children’s Hospital. ‘‘I’ve seen children who learned to vomit at the sight of the bottle, and an 8-month-old who could fake-cough because he knew that would make the feeding stop.’’

In late 2014, I sat in an observation room at the Children’s Hospital of Philadelphia’s Pediatric Feeding and Swallowing Center with Colleen Lukens, one of the program’s behavioral psychologists. Lukens watched through a two-way mirror as Ivy, a 2-year-old recovering from a stroke, was spoon-fed her lunch by a clinical feeding specialist named Julie Quenzer. A bin of toys sat at their feet; whenever Ivy swallowed some puréed broccoli, Quenzer pulled out a fire truck and thrust it at her: ‘‘Good job, Ivy! Way to swallow your bite!’’ Then she took the toy back and offered more food. When Ivy spit out a bite, Quenzer scooped it up and replaced it. ‘‘You’ve got to keep it in, Ivy!’’ When Ivy spit it out once more, the process repeated itself. Ivy swallowed on the fifth try, and the toy reappeared. ‘‘Good job, Ivy!’’

Image
Clockwise: Violet, 5 months old, with a nasogastric feeding tube inserted. Violet, 6 months old, just after her permanent gastric feeding tube was placed. Violet, at home, trying to use ‘‘my button’’ all by herself. The last meal Violet ‘‘ate’’ through her gastric feeding tube.Credit...Top left photograph by Stephen Upham. All other photographs by Virginia Sole-Smith.

The behavioral model presumes that children who don’t eat need external motivations. Drayton and Lukens don’t deny the existence of internal cues about hunger and fullness, but they say that many of these children are no longer responding to them. ‘‘The early tube-feeding experience often disrupts all of that,’’ Drayton says. ‘‘Every inch of their being says stop eating, stop eating, stop eating. If we let these children make all of their own choices, they would make bad choices. That’s why we don’t let 2-year-olds get their own apartments.’’

When Ivy left the hospital after four weeks, she was eating 3.5 ounces of puréed food and drinking four ounces of milk three times daily. Nearly all the children who come through the program achieve their feeding goals by the month’s end, but the hospital does not track long-term outcomes. Families often struggle to maintain the rigid feeding routine at home; Lukens admitted that their protocol is ‘‘very tedious.’’ Yet behavioral intervention remains the only treatment with ‘‘well-documented empirical support’’ for pediatric feeding disorders, according to an evidence review of 48 single-case research studies spanning 40 years, published in 2010 in Clinical Child and Family Psychology Review.

Westgate draws on the behavioral model, but whether by training or by personality, she is much more laid back. She showed us how to tap on Violet’s cheeks with our fingers or a small teething toy and then make our way over to her lips, encouraging her to gum on the toy as long as she could tolerate it. Sometimes the game lasted only seconds before Violet began to cry and gag. Whenever she did, we stopped; the idea of ‘‘replacing’’ bites, even when the bite was a toy, didn’t sit right with any of us. ‘‘Our goal is to give Violet positive associations with her mouth,’’ Westgate told us.

But real progress was impossible because Violet’s nasogastric feeding tube was worsening her aversion. Every other Friday, I pinned her down and sang ‘‘You Are My Sunshine’’ while Dan threaded a new tube down to her stomach. When Violet screamed so hard that her throat closed, we would wait until she breathed again. When she choked and sputtered until the tube came out of her mouth, we would start all over, hoping that Dan didn’t twist it into her lung by mistake. Even trained hospital nurses misplace feeding tubes as many as 8,000 times per year, the American Society for Parental and Enteral Nutrition estimates, causing serious complications or even death. Nobody tracks how frequently parents make this mistake. The only way Dan could steel himself to keep going was to pretend in those moments that Violet wasn’t his daughter at all.

I gave up on breast-feeding before Violet was 3 months old, soon after she began vomiting up every meal, another common side effect of nasogas­tric feeding. Two months later, Westgate suggested gently that we put the bottle away as well. Then, before Violet was half a year old, a surgeon cut a hole in the side of her abdomen and implanted a permanent gastric feeding tube directly into her stomach wall. This would be easier to live with: nothing taped to her face, no more torturous tube replacements, just a small plastic button next to her belly button. Westgate knew it was our best shot at healing the aversion, but I couldn’t see it as anything other than failure. We could now plug Violet in for food as if we were charging an iPhone. It was devastating. And also a relief.

Around this time, I discovered a group of thera­pists who offer not only a more moderate take on the behavioral model but also programs that reject it outright. These therapists believe that all children have some internal motivation to eat, as well as an innate ability to effectively self-regulate their intake. ‘‘This is a scientific conversation, but it’s also a deeply philosophical one,’’ said Suzanne Evans Morris, a speech-language pathologist and founder of the New Visions education and therapy program in Faber, Va., when I called her to discuss Violet’s case. ‘‘Does an aversion or a complicated medical history erase a child’s internal motivation to eat? Or can we help them rediscover it? I believe there is a tremendous amount of wisdom in these little kids and that they will transition to eating in their own time if we give them the right support.’’

Morris is a kind of guru in the speech-pathology world; along with the pediatric occupational therapist Marsha Dunn Klein, she wrote ‘‘Pre-Feeding Skills: A Comprehensive Resource for Mealtime Development,’’ the 798-page bible of speech-language pathologists. She also helped write the profession’s standard skills checklist, which Westgate used to first evaluate Violet’s oral aversion, and pioneered a ‘‘child-centered’’ approach to feeding therapy, training parents to read their child’s cues and offer food only when clearly invited to do so. To encourage children to issue such invitations, Morris turns food into play, racing crackers balanced on top of a child’s toy cars.

Disciples of this child-centered approach teach the ‘‘division of responsibility,’’ a concept developed in the 1980s by Ellyn Satter, a registered dietitian and family therapist. Although medical professionals have long viewed over- and undereating as separate issues, the former rooted in a lack of willpower and the latter in a self-destructive need for control, Satter sees them as related. ‘‘In both cases, these kids are reacting to distortions in their feeding relationship with their caregivers,’’ she told me. To restore balance, Satter says, a parent need only take responsibility for deciding what kinds of food to offer and, as children reach the end of their first year, when and where meals take place. She leaves children in charge of how much and even whether they eat. Satter believes the preservation of that ability to self-regulate is at the crux of solving both childhood obesity and pediatric feeding disorders.

Satter and Morris have published numerous books and case studies but are still developing a way to test their methods in controlled clinical trials. Their lack of published empirical studies raises questions with behaviorists. ‘‘I think most of the children we work with have already tried a child-led approach, and it’s failed,’’ Lukens says. But Morris listed many clients who were dissatisfied with the behavioral approach as well. And I met several formally trained behavioral therapists who had crossed over to the child-centered model.

Image
Violet, almost 17 months old, eating chocolate-avocado pudding shortly after her tube meals were stopped.Credit...Photograph from Virginia Sole-Smith

‘‘I realized that teaching a child to eat when their body is telling them not to is not only counterproductive, it’s dangerous,’’ says Jennifer Berry, an occupational therapist and founder of the Spectrum Pediatrics Tube Weaning Program, based in Alexandria, Va. She resists the term ‘‘aversion’’ altogether because it implies a dysfunctional behavior. ‘‘It’s not a problem,’’ she told me, ‘‘it’s an adaptive skill to know when eating isn’t safe.’’

Studies suggest that children may eat less healthfully when parents exert too much control over the process. For example, in a 2006 trial published in the journal Appetite, children instructed to ‘‘finish their soup’’ complied begrudgingly but still ate less than children who were unpressured. Research on eating-disorder patients also suggests that highly pressured mealtimes early in life might play a key role in the development of those conditions.

But the behavioral method gets tube-dependent children to eat; when the stakes are that high, worrying about emotional consequences may feel like an afterthought. ‘‘The question of when children start to enjoy food for its own sake remains a mystery to all of us in the feeding world,’’ Lukens says. ‘‘Never once do I say to a parent, ‘By the end of the four weeks, I think she’ll love food.’ ’’

But I yearned for Violet to love food. Even though we had switched to a permanent feeding tube, I wanted to believe we could unlock her internal motivation to eat. Still, whenever Dan and I discussed the idea, we got stuck. How can you trust a child to eat enough after she has shown you that she is willing to choose starvation — even if that was what Morris and Berry would call ‘‘a good choice at the time’’? For that matter, given how badly I misread her cues in those first two months, how could I trust myself to understand the choices she was making now?

In Palatine, Ill., Joey giggled as he painted chocolate pudding onto a glass sliding door with Heidi Liefer Moreland, a speech-language pathologist with Spectrum Pediatrics. We were in a Chicago-area rental apartment that the team was using for Joey’s feeding-tube ‘‘wean,’’ as the transition to normal eating is known. ‘‘Kids shouldn’t learn to eat in a clinic,’’ Berry told me. ‘‘They should learn to eat at the family table, so that’s where we work.’’ Or in the case of Joey, who was 2 and had depended entirely on his feeding tube ever since his premature birth, that’s near where they worked. There was nothing that resembled force-feeding; I never saw either therapist hold a spoon. Mostly, they were there to hang out with Joey and his family while everybody waited for him to realize he hadn’t been tube-fed more than a few ounces in almost a week — and that eating would be the only way to erase his nagging hunger.

This is where Berry’s approach splits off from the traditional child-led model and becomes radical. Calories from the feeding tube are cut significantly over a five-day period, so a tube-fed infant or child begins the wean on around 50 percent of his normal daily caloric intake and 80 percent of his optimal fluid needs. Over a 10-day period, Berry and Moreland are on call around the clock, giving support and coaching as the child, they say, rediscovers the drive to eat. After the initial wean, therapy continues as needed for six months; by then, 95 percent of Berry’s patients, she says, are eating all of their daily calories by mouth, although her findings have yet to be replicated or published in an American scientific journal.

Many behavioral programs also incorporate a modified version of this ‘‘appetite manipulation’’; Ivy’s tube feeds were gradually reduced by 60 percent over the course of her month at the Children’s Hospital of Philadelphia, and Lukens credited her hunger for the success. But only a few mainstream programs (including one at Seattle Children’s Hospital) have adopted a child-centered, hunger-based approach to tube weaning. Most consider it too risky because of the potential for weight loss. Berry’s program requires a doctor to set safety parameters for each child and to sign off on the amount of weight loss that can be tolerated.

Later that night, Joey sat listlessly on the couch with his mother, Angela Reid. A bag of Terra Chips and an Elmo sippy cup lay on the coffee table in front of them, and every so often, Joey would pick up one or the other and moan.

‘‘Is this normal?’’ Reid asked Berry. ‘‘He’s never like this. What’s going on?’’

Berry sat against the far wall, assessing the situation. ‘‘I think he’s very, very hungry,’’ she said. ‘‘We have to decide whether he’s going to find this motivating, or does he need a little support?’’ After a brief discussion, Reid fed Joey 30 milliliters of Pedialyte through his gastric tube. Within minutes, he bounced up, all smiles: ‘‘Elmo!’’ By the time we left, he had nibbled three chips.

‘‘Usually that’s just enough to take the edge off that horrible new feeling of hunger,’’ Berry explained to me later. I asked her how Joey’s palpable frustration was any different from Ivy’s slack-jawed compliance with her feeder. ‘‘Because we’re following his cues,’’ she said. ‘‘We didn’t let him suffer. But we didn’t force him to eat before he was ready either.’’ The next day, Joey ate more than 20 chips. A few days later, he discovered a love of fried chicken. The family went home a day early; over the next month, they used his feeding tube for occasional supplementation, but Joey continued to eat well and gained four ounces.

We offered Violet her first bite of banana when she was 5 months old, about a week after her permanent feeding tube went in. She sat on Dan’s lap at breakfast and seemed fascinated as she watched us eat. I held my breath as I offered the spoonful: Why would this time be any different? But Violet took a taste. Solid foods intrigued her. They didn’t trigger her aversion in quite the same way.

We began to put Violet in her highchair every time we sat down to a meal, trying to catch on quickly to the slight shake of a head that meant ‘‘no’’ before it escalated to gagging and crying. We didn’t always succeed. At times, we offered books or toys to reward Violet, or at least maintain her interest. She didn’t eat much of anything, and the food that did pass her lips was gagged on or spit out. Still, it felt close to normal, like how our friends sat at the table with their babies.

Then Violet underwent her second open-heart operation. Over the next three months, we spent 50 days in the hospital as she fought off various complications. Eating was forgotten. But one day, Violet took a previously rejected sippy cup and drank an ounce of water. She was hooked up to oxygen to keep her breathing stable, and the nurses told us that breathing on a nasal cannula feels not unlike driving down a highway with your head out the window. So Violet was thirsty for the very first time — and somehow she knew that she could use her mouth to make herself feel better. We now had a bit of proof that her internal drive was still in there, somewhere.

After Violet recovered, we went home and spent the summer watching for any sign that she was ready to try eating for real. We began working with a pediatric dietitian named Margaret Ruzzi, who instructed us to offer table foods before every tube feed. By her first birthday, Violet was taking a few bites of food at every meal. She loved flavor — chicken tikka masala, pad thai, blueberries. One afternoon, sitting on my lap, she happily gummed an apple core, and I knew: Food once again meant comfort. The aversion was gone.

But Violet still received all her calories through her feeding tube. Because she never got hungry, eating was only recreational. Sometimes it felt as if she were baiting us; she would pack spoonfuls into her cheeks and then — just as we thought here was true eating, at last! — spit it all out. When I described the constant spitting, Morris suggested I reframe it as a critical part of Violet’s learning curve. ‘‘Spitting helps Violet know she can get the food out,’’ Morris explained. ‘‘That makes it safe to experiment with taking another bite.’’ So we let Violet spit. And started to think about what might happen if we starved her a little.

It felt counterintuitive and maybe even self-indulgent. I wondered if we were pushing too hard because the feeding tube — with its formula to blend, syringes to clean, equipment sure to malfunction at 1 a.m. — was driving us insane. But Violet wasn’t eating because she didn’t know she needed to eat. And so with Ruzzi and Westgate’s support, we decided to drop her tube calories by 20 percent for two weeks.

Our plan was to start by cutting out a ‘‘morning snack’’ tube meal, which Violet normally received at 9 o’clock; instead, she would have six hours between breakfast and lunch to experience hunger for the first time in over a year. In the first two weeks, not much happened. Violet cried more, particularly around 11 a.m., when presumably her stomach felt empty, but she had no idea what to do about it. We held her a lot. She didn’t eat more. At our next checkup with Ruzzi, she had lost 10 ounces but also grown an inch. ‘‘Let’s keep going,’’ Ruzzi said.

So we cut back her tube feedings by 40 percent, and over the next two weeks, Violet began to eat. A tiny wheel of cheese. A pouch of applesauce. At our next check-in, she had gained five ounces. We cut back more on the tube calories. By November, we estimated that Violet was eating between 150 and 200 calories by mouth per day. In December, she ate her final tube meal.

Not long before that, we went to a diner for brunch. It was the first time in over a year that we went anywhere without packing the pump, syringes and tube. We ordered egg sandwiches and, from the kids’ menu, a grilled cheese, which Dan carefully cut up into postage-stamp-size bites. All around us, other families were tucking into their Sunday pancakes, chatting and clinking forks. Violet scribbled with crayons on her place mat, threw my French fries on the floor, giggled at her dad’s funny faces. And then she ate everything but the crusts.

Virginia Sole-Smith is a writer whose work has appeared in Real Simple, Harper’s Magazine and Newsweek. She last wrote for the magazine about a mermaid show in Florida.

Sign up for our newsletter to get the best of The New York Times Magazine delivered to your inbox every week.

A version of this article appears in print on  , Page 47 of the Sunday Magazine with the headline: Hand to Mouth. Order Reprints | Today’s Paper | Subscribe

Advertisement

SKIP ADVERTISEMENT