Medicine requires lifelong learning. So we read journals, attend lectures, and talk about cases with colleagues. However, the most powerful teacher is experience. As we practice, we watch outcomes. When something we do goes well, we remember it. When something goes wrong, we cannot forget it. Experience here is defined the way "student" uses this term in Pediatrics (September 1997, p. A60), "experience is what you get from not having it when you needed it most." A disproportionate share of my learning by experience has come from neurosurgical conditions, the subjects of this issue. A few examples follow.
A 12-year-old girl with a history of myelomeningococcal at birth, hydrocephalus requiring shunt placement during infancy, and mild mental retardation came to the emergency center because of headaches and irritability. It was about 10 PM. There were several children waiting to be seen, so I went in to take a quick look to be sure she could wait. She was sitting up on the examination table, able to answer a couple of simple questions and afebrile with stable vital signs. It was clear that shunt malfunction would have to be ruled out before she could be sent home. However, it looked as if she could wait while I caught up with patients. So she was asked to rest for a short while. Approximately 30 minutes later, the nurse called: "Doctor, you better come here and look at this." This phrase rarely communicates good news. The nurse had happened by to check vital signs (I had failed to put the patient on a cardiorespiratory monitor) and noted she was unconscious with depressed respirations and bradycardia. The patient was revived by sitting her up and stimulation, and she did well after an emergency shunt revision. A more detailed history revealed the girl had been alternating between lucid but irritable states and depression for a few days. Apparently, the supine position increased intracranial pressure enough to compromise cerebral perfusion or brain stem function. This happened approximately 25 years ago, but ever since, I have had more respect for patients who could have increased intracranial pressure and have used monitoring more liberally whenever control of a patient's situation was in question.
A 2-month-old boy was referred because of a flattened occiput. The child had been routinely placed in the supine position to sleep, according to recommendations of the Academy of Pediatrics, and the parents may have left him in that position more than usual when awake. The examination revealed posterior plagiocephaly, the absence of a ridge over any of the sutures, and normal development with no other dysmorphology. The sutures looked open on plain skull films, but to be sure, I asked the pediatric radiologist to read them. He agreed that the lambdoid and other sutures were not fused. So the patient was sent back to the referring physician for further observation with parental instructions to move the child's head frequently so he was not always lying directly on the back of his head. Approximately 5 months later, the referring physician called back because the head shape was worse and a ridge was now palpable along the lambdoid sutures. This time the plain films revealed clear fusion here, and the patient required cranial reconstruction. The reason this diagnosis was missed was a mystery until I read the article by Dr. Keating in this issue (pp. 600-612): during early stages of craniosynostosis, the sutures may not be fused on skull radiographs. A computed tomography (CT) scan is apparently more sensitive. The lesson has been earlier referral to neurosurgery, more aggressive use of CT scans with bone windows, and more of the pediatrician's best friend-close follow up-for the patient with possible craniosynostosis.
There were others. A 9-year-old girl from a dysfunctional family went from being an "A" student to failing grades over a 4-month period. She had also become depressed and withdrawn. When she developed morning vomiting with severe headaches, a CT scan revealed a brain tumor. A 5-month-old girl with an abnormal CT scan was referred because her legs were rigid and muscle spasm had to be overcome to flex her knees and hips. When there was no change after 6 weeks of gentle exercise, I mentioned the possibility of cerebral palsy to the parents. Eleven months later, and after 11 months of parental anxiety, the neurological examination became and remained completely normal. Development was always otherwise normal. Recent reading revealed that signs of cerebral palsy, such as differences in limb muscle tone or strength and prolonged primitive reflexes, often disappear by the age of 2 years. So many pediatricians do not make this diagnosis until these signs persist beyond this, especially when the child is otherwise doing well.
Somehow, the children who had head injuries from abuse were less of a diagnostic challenge. Child abuse was common where I worked and thus at the top of the list when unexplained lethargy, coma, or injury presented to us. There were problems-like the father who said he would "blow my head off" if another social worker from the state came near his house. (He was a substantial threat and I took him at his word.) Also, there was the night an abusive father called my wife while 1 was out of town and told her he was on his way to pay her a visit (he never arrived). However, the diagnosis and management of these abused children was fairly clear and, fortunately, there were less "opportunities" to learn from experience (or mistakes). Some keys to recognizing non-accidental brain injury follow. Many hold for all forms of child abuse whereas others are more specific for head injuries.
* The infant with unexplained depression of consciousness (no satisfactory accident history or other physical evidence of trauma; normal serum electrolytes, glucose, metabolic screening tests; negative laboratory data and history for toxins and no clear explanation on physical examination) should be considered to have an overwhelming infection or to have been shaken until proven otherwise.
* A history of accidental trauma that does not fully explain a brain injury is suspicious. Falling 18 to 24 inches from a bed should not cause a severe injury, and a skull fracture rarely if ever occurs in this circumstance.
* Do a retinal examination early and confirm your findings with an ophthalmological consult whenever there is a chance of shaking or head trauma. Retinal hemorrhages are typical of the shaken baby syndrome.
* As for abuse in general, a prolonged period between when the injury occurred and when medical attention was first sought is suspicious. Ordinarily, parental anxiety is greatest immediately after an accidental injury, and this is when parents usually seek care. However, several hours may be required for anger to abate and fear about the seriousness of non-accidental injuries to grow before a parent will risk detection by seeking care.
* Use long bone, chest, and skull films or a bone scan when abuse is suspected, even when no evidence of trauma is evident at these sites on examination. The bone changes of abuse are often asymptomatic because they are due to old healed fractures, or from calcified subperiosteal hematomas without fracture or small metaphyseal fractures from pulling or twisting an extremity. A skull fracture, identified on a plain radiograph or CT scan indicates that significant head trauma has occurred. Documenting this is one of the few remaining indications for skull films.
Enjoy this issue. I hope it opens doors for you as it has for me. In thinking about why experience (ie, mistakes) has been such an important learning tool for neurosurgical conditions, the reason is probably that my training and reading have left more gaps for surgical conditions than with the medical aspects of pediatrics. This may or may not be true for you, but it is clear that we will often be the first to see children who have neurosurgical conditions.