A 17-Year-Old Boy With High-Functioning Autism, Gastrointestinal Illness, and Seizures

A previously healthy 17-year-old boy with a high-functioning pervasive developmental disorder was seen by his primary care physician for 1 day of vomiting, abdominal pain, loose stools, and poor tolerance of oral intake. At the office, he was alert but tired, afebrile, with mild tachycardia, dry mucous membranes, and slightly delayed capillary refill. Preliminary diagnoses of acute viral gastroenteritis and dehydration were noted, and his primary care physician recommended treatment with intravenous (IV) fluids and antiemetics at a local emergency department (ED). Because the patient and his mother preferred to trial oral rehydration at home, a prescription was provided for ondansetron to reduce nausea. During the car ride home, the patient had a 4-minute episode of eye rolling, shaking, and posturing that spontaneously resolved without incontinence or tongue biting. Emergency medical services was called and noted on arrival that the patient appeared sleepy and minimally able to communicate. Ambulance transport to the ED was uneventful.

Dr Klig, can you comment on the approach to parents who decline their providers’ advice, as occurred with the recommendation to go directly to the ED?

The start of this case presents a challenging situation, especially in the office setting. There are 3 ethical pieces to the scenario in which a parent declines a provider’s advice: physician fiduciary responsibility, the harm principle, and the best interest principle. If the provider determines that the best course of treatment is IV rehydration in the ED, then he or she assumes the duty to act in good faith to propose and facilitate this plan. The parent can decline the plan if safety is not an issue; otherwise the harm principle, “to identify a harm threshold below which parental decisions cannot be tolerated,”¹ can apply, which includes the provider's role as a mandated reporter to the state child protection authorities. In most cases, the provider and parent together can find reasonable alternatives in which the best interests of the patient are incorporated. The alternative in this case was to trial oral fluids at home with ondansetron to control nausea, which seems to be a reasonable compromise given that the parent agreed to bring her son to an ED if the plan failed. Indeed, single-dose ondansetron has been demonstrated to control nausea and avoid IV rehydration in the ED.^2,3 One could not have predicted the subsequent seizure episode in the car, and his mother responded appropriately by calling 911.

The patient arrived to the ED alert and sitting up, yet he was pale, fatigued, and responding slowly with few words. His mother reported that he was usually talkative and an academically strong high school senior. She added that the patient had attempted to stay hydrated despite his vomiting and diarrhea by drinking water but had become unable to hold down liquids, prompting the office visit. A review of systems, provided by the patient with his mother’s help, was notable for headache and recent pain on urination. The patient denied fever, abdominal pain, or trauma.

At baseline, he was tall with a thin and muscular physique. Vital signs revealed an elevated blood pressure of 147 over 70 mmHg, a heart rate of 88 beats per minute, and an oral temperature of 36.8°C. He had equal, round, and reactive pupils and nonicteric sclera. Mucous membranes were dry. The lungs were clear, and cardiac examination results were normal. Ongoing dry heaves made the abdomen difficult to assess for tenderness or focality. There were no visible rashes found on a brief skin examination. The hands and feet were cool to touch, and capillary refill was 3 seconds. There was good strength in all extremities, and initial neurologic examination was nonfocal.

Dr Klig, what would you consider to be the immediate priorities for this patient? What are the limitations to his evaluation on ED arrival?

A top priority for this patient is to address his hypovolemia, or acute dehydration from gastrointestinal losses. The heart rate is seemingly normal yet elevated from baseline because the patient is an athlete. His slow response to questions may be from severe nausea yet also may reflect an ongoing central nervous system process, such as a postictal state or nonconvulsive seizure (absence or other). Metabolic abnormalities, toxins, early encephalopathy, and other intracranial etiologies are also possible causes. Given the extent of hypovolemia, immediate treatment with 20 mL/kg IV boluses of 0.9% saline is initiated as laboratory tests (complete metabolic panel, complete blood count with differential, lactate, venous blood gas, and urinalysis) are sent. Continuous monitoring of vital signs is important, and serial assessments of the airway, breathing, and circulation are used to further guide our early management.

We often see patients with acute nausea, vomiting, or dry heaves that limit both the response to questions and tolerance of physical examination (especially the abdomen). Early use of an antiemetic offers therapeutic and diagnostic value in many cases. Our patient is alert, sitting up, and able to slowly answer questions, all of which are reassuring signs of tissue perfusion and brain function. If severe nausea is causing his pallor, fatigue, dry heaves, and slow responses, these should improve with antiemetic medication. Physical examination results should likewise improve because the patient may better tolerate and localize symptoms, particularly for his abdomen.

The patient's dry heaves resolved and he became slightly more communicative after initial IV fluids and ondansetron as initial laboratory tests were pending. Approximately 45 minutes after ED arrival, he developed an acute right-sided forehead-sparing facial droop, right-sided ptosis, and expressive aphasia. His pupils appeared to be reactive, but extraocular movements were difficult to assess. The stroke team was activated, and the patient was urgently brought to radiology for a CT scan of the head to rule out intracranial hemorrhage or large acute infarction; CT angiogram scan of the head and neck to rule out large vessel occlusion, stenosis, and aneurysm; and brain MRI to evaluate for intracranial enhancement or other structural abnormalities. This acute change resolved within 20 minutes during imaging. The neurology team reevaluated the patient promptly and found him to be alert and oriented to name, place, and year but not to month. Ptosis was resolved. Extraocular movements were intact without nystagmus. The patient’s cranial nerves were intact without any dysarthria. His motor, sensation, reflex, and cerebellar examinations were without deficits; his gait was not assessed. The mother felt that the patient had returned to his baseline.

Dr Krishnamoorthy, what would be your differential diagnosis for acute-onset facial droop and aphasia in this clinical presentation? Does his history of autism or diagnosis of acute dehydration broaden the differential?

The differential diagnosis of acute-onset facial droop and aphasia in a 17-year-old boy with autism spectrum disorder (ASD) is broad, and it includes stroke and seizure among other etiologies. A case review by Mackay et al⁴ of 287 children presenting to the ED with acute neurologic symptoms suspicious for stroke revealed the most common diagnoses to be migraine (28%), seizures (15%), Bell palsy (10%), stroke (7%), and conversion disorders (6%). Most cases of strokelike presentation in this study turned out not to be classic vascular strokes.

The incidence of epilepsy in autism is generally 5% to 21% (in contrast to 0.5%–1% in the general population).^5,6 Advancing age appears to affect the incidence of epilepsy in ASD,^5,–7 with seizures often presenting in teenagers. There is no known preponderance of strokes in patients with ASD.

The occurrence of forehead-sparing unilateral facial droop and aphasia raises a red flag for a stroke in the contralateral hemisphere in the distribution of the middle cerebral artery. The difference between an upper versus lower motor neuron lesion in the localization of facial nerve weakness is based on the supranuclear innervation of the seventh cranial nerve nucleus, where the forehead and eye closure have a dual innervation. Therefore, in a typical hemisphere lesion causing a stroke, there will be detectable weakness in the opposite lower face while forehead movements remain intact. In a lower motor neuron lesion of the facial nerve, all the muscles innervated by the seventh nerve will be affected, causing a complete loss of facial movements on the affected side.

Strokes in teenagers are often due to thromboembolic arterial occlusion from trauma, vascular dissection, or thrombophila.⁵ Rarely, severe dehydration can lead to cortical venous sinus thrombosis and a strokelike presentation. Head CT scan results are often normal in cerebral venous thrombosis, whereas brain MRI in combination with magnetic resonance venography is the most informative technique.⁸ Normal brain CT scan, MRI, and magnetic resonance venography results would exonerate the possibility of an ischemic, hemorrhagic, or venous stroke in this patient.

An unlikely cause of acute facial droop is seventh cranial nerve palsy, which has many possible causes that include idiopathic Bell palsy, otitis media, Lyme disease, and postviral (eg, herpes simplex virus types 2 and 6 and Varicella) causes.⁹ Given the rapid resolution of our patient’s facial droop, we would strongly argue against Bell palsy.

Another possible diagnosis is Todd paralysis, which presents as a transient focal paralytic disturbance after a seizure. The mechanism of focal weakness from seizures is speculative and might be related to neuronal exhaustion from hypoxia, substrate depletion, or insufficient focal metabolism.¹⁰ It typically occurs immediately after a prolonged focal seizure but may also occur after generalized seizures.¹¹ Our patient’s facial droop, ptosis, and aphasia were noted ∼45 minutes after the observed seizure, making Todd paralysis less likely.

Given the history of present illness, a likely cause of the neurologic complications in this patient could be hyponatremia from excessive fluid loss (and antidiuretic hormone activation) during acute gastroenteritis. Hyponatremia alone may cause a range of neurologic symptoms, including focal neurologic deficits. Plasma sodium <120 mEq/L can provoke seizures and other neurologic disturbances, particularly in the setting of a rapid-drop serum sodium concentration.⁵ In children, signs of hyponatremic encephalopathy may vary from nonspecific symptoms (such as headache, weakness, lethargy, and confusion) to more severe manifestations (such as seizures and coma).^12,13 In a study described by Bokemeyer et al¹⁴ of 261 adult patients with hyponatremia, 140 (54%) had neurologic symptoms, with unspecified weakness and confusion being the most common presenting symptoms (seen in 59% of patients). Focal neurologic signs, such as cranial nerve palsy, hemiparesis, gaze palsy, and hemiataxia, were present in 31% of patients.¹⁴ Thus, it is possible that our patient’s seizures and focal neurologic deficits were caused by severe hyponatremia in the setting of dehydration.

The results of the CT scan and brain MRI were without abnormality. A serum sodium of 119 mmol/L was reported. Other electrolyte results were as follows: potassium 4.6 mmol/L, chloride 82 mmol/L, bicarbonate 17 mmol/L, glucose 155 mg/dL, serum urea nitrogen 12 mg/dL, creatinine 0.87 mg/dL, anion gap 20 mEq/L, and lactate 6.3 mmol/L. The pH was 7.39 on blood gas analysis. The complete blood count was notable for a leukocytosis of 15.18 k/μL with an 85% neutrophil predominance and no band forms. Liver function test results were within normal limits. Urinalysis obtained after IV fluids had a specific gravity of 1.040, a negative result for ketones, and did not reveal infection. Our patient’s seizure episodes and later focal neurologic deficits were thus ultimately attributed to significant hyponatremia from acute dehydration from gastrointestinal losses with limited water intake only.

Dr Murphy, can you help us with key points on the evaluation and treatment of a patient with hyponatremia? What is the initial approach to correcting hyponatremia as in this case?

Our patient presented with seizures as a symptom of his hyponatremia. Hyponatremia, defined as a serum sodium concentration <135 mmol/L, is the most common electrolyte disorder encountered in clinical medicine.¹⁵ It is associated with increased morbidity and mortality, and hyponatremic encephalopathy is an ominous complication. This results from water movement out of the hypotonic intravascular compartment, causing brain cell swelling, tissue edema, and increased intracranial pressure. Hyponatremic encephalopathy is a life-threatening emergency, and clinical detection of neurologic impairment is essential: signs such as vomiting, somnolence or coma (Glasgow Coma Scale score <9), seizures, or cardiorespiratory compromise should prompt immediate concern for symptomatic brain edema. In these cases, the serum sodium should be rapidly increased by 4 to 5 mmol/L until clinical symptoms are reversed.

A guideline for correction is to expect that serum sodium will rise by ∼1 to 2 mmol/L for each 1 mL/kg of 3% saline infused.^16,17 Clinical responses to hypertonic saline can vary and may be exaggerated in patients with more severe hyponatremia (<120 mmol/L).¹⁸ An approach for patients who are symptomatic with hyponatremia that is supported by published expert guidelines¹⁹ is to administer a 2 mL/kg bolus of 3% saline (513 mEq/L) over 20 minutes; if symptoms continue, send a blood specimen for urgent sodium level measurement while administering a second 3% saline infusion of 1 to 2 mL/kg and watch for clinical effects.

Note that an overly rapid correction of hyponatremia can lead to osmotic demyelination, which appears to occur in cases in which severe hyponatremia is present for >48 hours. Nonetheless, a patient with moderate or severe symptomatic hyponatremia is at a greater risk of progressing to brain stem herniation or death if untreated than of developing osmotic demyelination if treated. A quick yet modest increase in serum sodium (5 mmol/L) is sufficient to prevent herniation.^20,–22 Once severe symptoms are mitigated by sodium repletion, and the herniation risk is reduced, a slower correction rate is advised.

The patient received 2.5 L (∼42 mL/kg) of IV normal saline during initial resuscitation, before the results of his initial basic metabolic panel were known. Serum sodium was reported as 131 mmol/L on repeat laboratory draw several hours later. IV fluids were then held.

Dr Murphy, was this correction too rapid in a patient with acute hyponatremia?

In this case, isotonic fluid was appropriately given for initial volume resuscitation before the electrolytes were sent or resulted. Hyponatremia was an unexpected finding given the relatively brief course of his illness. Although the change in a patient’s serum sodium can be estimated by using the Adrogué and Madias²³ calculation, the actual rate of sodium correction is clinically variable. It is thus important to closely monitor sodium levels to gauge the individual response to treatment.

For our patient, the serum sodium began to correct quickly even though he was only treated with IV isotonic fluid. An overly rapid correction of sodium is a well-recognized problem in the treatment of hyponatremia, particularly if the hyponatremia is driven by volume depletion (as in this case), cortisol deficiency, or desmopressin or thiazide diuretics. In these situations, discontinuation of the medication, steroid replacement, or volume repletion remove the factor that is driving the hyponatremia or vasopressin secretion. With this eliminated, patients then begin to appropriately produce maximally diluted urine, which can increase serum sodium by >2 mmol/L per hour and drive a rapid correction.

If there is severe hyponatremia <120 mmol/L and/or a concern that hyponatremia has been present for ≥2 days, overcorrection should be slowed or reversed to limit sodium rise to 10 to 12 mmol/L in the first 24 hours and 18 mmol/L in 48 hours. A quick rise in sodium should prompt all ongoing measures to increase sodium to be halted. In this case, all sodium containing IV fluids were held when an acute increase of 11 mEq/L was detected. The use of 5% dextrose in sterile water should then be given intravenously to replace urine losses only, or urine losses can be replaced orally with water for the remainder of the 24-hour period.

After the patient returned from head imaging, a repeat physical examination revealed abdominal tenderness in the bilateral lower quadrants and suprapubic region with notable rebound tenderness over the right-lower quadrant and a positive obturator sign.

Dr Klig, what is your differential diagnosis given these new physical examination findings?

The patient now has abdominal right-lower quadrant rebound tenderness, which is a significant change from ED arrival. Acute gastroenteritis is still a possible culprit, yet our primary differential diagnosis is expanded to include appendicitis, focal enteritis, pyelonephritis or nephrolithiasis, or referred pain from the scrotum (testicular torsion, appendix testis torsion, or epididymitis) or hip. Urinary sources are unlikely given a negative urine dipstick and microscopy result. Acute testicular problems or hip problems can also be eliminated if there are no findings on examination. Focal enteritis is possible given his symptoms and mild elevation of C-reactive protein to 11.2 mg/L. Appendicitis is possible given the acute onset of symptoms, with the presence of rebound tenderness suggesting peritoneal irritation. The obturator sign (right-lower quadrant pain on flexion and internal rotation at the hip from irritation of the obturator internus muscle) is present in those cases in which the inflamed appendix is located in the pelvis (8% sensitivity and 94% specificity).²⁴ Our patient has a higher probability of appendicitis, with 6 of the following 7 criteria of the modified Alvarado²⁵ score seen: migratory right-lower quadrant pain, anorexia, nausea or vomiting, tenderness in the right-lower quadrant, leukocytosis (white blood cell count >10 k/μL), and fever >37.5°C. Further evaluation for appendicitis is thus clearly warranted. There is also literature to suggest that hyponatremia can be associated with complicated appendicitis, so early consultation with a pediatric surgeon is important.²⁶ 

The high value of serial reassessments must be emphasized here. A quick repeat examination can make the difference in diagnostic, treatment, and disposition decisions, especially in an outpatient office, clinic, or ED setting. Reassessments are particularly helpful if some form of treatment (eg, antipyretic or pain medication) is given. Current evidence supports the use of pain medication even when appendicitis is a concern.²⁷ In our patient, a quick recheck of the abdomen was critical, and the use of antiemetic and pain medication did more to help than hinder the detection of focal findings on the clinical examination.

On the basis of these physical examination findings, an urgent abdominal ultrasound (AUS) was obtained that was highly suggestive, but not definitive for, acute appendicitis (Figs 1 and 2). Pediatric surgery was consulted and recommended delaying operative plans given the risks associated with the patient’s initial metabolic abnormalities. The patient was treated with piperacillin and tazobactam and admitted to the PICU.

Dr Masiakos, what are the benefits of delaying appendectomy in this case?

For decades, emergency appendectomy has been the standard of care for the treatment of acute appendicitis. A delay in operation was thought to increase postoperative morbidity or progression to complicated appendicitis, such as perforated appendicitis or periappendiceal abscess. However, the concept of emergency appendectomy has been recently challenged by studies in which researchers suggest that acute appendicitis could be treated medically or delaying surgery did not result in any increasing morbidity.^28,–31 Many studies have revealed no relationship between the postponement of surgery and risk of postoperative complications in children.³² In fact, many groups have shown that a procedural delay of ≥12 hours causes no further harm to the patient and that early intervention in complicated appendicitis may be ill advised. Therefore, the decision to operate on a patient with examination and radiographic findings that are concerning for appendicitis is superseded by the overall clinical status of the patient.

In this case, although the patient had an examination result that was concerning for appendicitis, I was not confident that this was the exact surgical pathology. The exact cause of the patient’s abdominal pain was confounded by several factors that have been described earlier. Therefore, we could not say for sure that we were not dealing with a chronic problem (abscess, phlegmon, or inflammatory bowel disease). Collectively, we wanted to correct the electrolyte abnormalities, address the volume loss, try to elucidate the cause of the seizure, and perform a CT scan of the abdomen to ascertain a more certain diagnosis before operating on the patient.

A CT scan of the abdomen performed the next morning revealed acute appendicitis (Fig 3) with moderate-volume free fluid and subtle peritoneal enhancement along the right pelvic sidewall, suggesting perforated appendicitis. Dilated loops of small bowel were present, without a transition point, likely reflecting ileus.

Dr Masiakos, what clinical information does the CT scan result of the abdomen add for this patient given the positive ultrasound findings? In what cases is an ultrasound reliable?

AUS has supplanted abdominal CT scanning as a first diagnostic tool that is used in concert with physical examination in patients who are suspected of having appendicitis. It is also used when the physical examination result is equivocal or unreliable (in young children). AUS examinations have been shown to be user dependent and have a sensitivity, specificity, and accuracy of 55.6%, 85%, and 71%, respectively.³³ Together with the physical examinations and history, providers in the United States can improve diagnostic clarity when clinical suspicion is present.

Although the ultrasound findings in this patient were concerning for an inflammatory process in the right-lower abdomen, the diagnosis was not clearly defined by this modality. Also, because the patient’s history was unreliable, and the presenting symptoms were confounding, we collectively decided to admit the patient and, after he was rehydrated, better elucidate the abdominal pathology with a CT scan.

The patient was taken to the operating room for laparoscopic appendectomy and washout and was confirmed to have perforated appendicitis.

Dr Masiakos, if the diagnosis of perforated appendicitis was known at initial presentation, would this have changed the decision to go to the operating room sooner?

The moral of the surgical portion of this story is that appendicitis, whether it is acute or advanced, is rarely a surgical emergency. In this complicated patient, appendicitis was just 1 possible cause of his abdominal pain and leukocytosis (maybe even his hyponatremia and seizure). That said, if a patient presents with a history or physical findings that are inconsistent for appendicitis or symptoms that are confounding, such as long-standing diarrhea, weight loss, seizure, and electrolyte abnormality, it is prudent to correct the metabolic and fluid derangements and clearly identify the cause before embarking on a definitive operation (if the patient remains stable).

The patient did well after surgery and was discharged on postoperative day 3. He remained on IV piperacillin and tazobactam in the hospital and was discharged from the hospital on amoxicillin clavulanate. There were no additional seizures or neurologic symptoms during admission. A spot EEG performed the day after surgery was unremarkable. Serum electrolytes at the time of discharge revealed a sodium level of 140 mmol/L and a white blood cell count of 5.39 k/μL. The hyponatremia and leukocytosis had resolved.

Our patient ultimately was diagnosed with a common clinical problem that can present to any outpatient setting. The unexpected and evolving changes in the case reveal the value of repeat clinical assessments, notably with thorough serial physical examinations. As clinicians, we strive to align the standard of care with many factors, such as patient preferences, changes in clinical course, unexpected responses to treatment, and/or atypical presentations of common diagnoses. Four unexpected clinical events make this case pertinent to many providers.

The first unexpected event in this case was the patient and his mother’s decision to go home from the office rather than to the ED for IV rehydration. Without question, this is a dilemma that both primary care and ED providers face. What is considered safest for our patients must be balanced with shared decision-making in which we respect patient autonomy and maintain an alliance with the patient and family. Crafting a modified plan in which we integrate patient preferences can foster more reliable follow-up and use of the health care system.³⁴ It also can turn an adverse and risky situation into a safer and more rewarding interaction.

The second unexpected event was neurologic: seizure followed 1 hour later by acute facial droop and aphasia. The facial droop and aphasia were ultimately considered to be a part of significant hyponatremia yet also remind us that acute ischemic stroke is underrecognized in this population³⁵ as an important consideration. Although brain CT and MRI results were normal, the risks of missing acute stroke or bleeding in this young patient were sufficient to justify immediate imaging. The somewhat dramatic acute neurologic events ultimately distracted from the patient’s evolving abdominal pain, which underscores the high value of serial reassessments during clinical care.

A third unforeseen event was the rapid rise in serum sodium during the initial treatment of dehydration with IV 0.9% saline. The patient’s hyponatremia was unknown, and thus stepwise correction did not occur. In hindsight, a more gradual fluid resuscitation guided by vital signs and neurologic status may have been helpful along with serial electrolyte testing. Nonetheless, the rapid rise in serum sodium yielded no harm to the patient and reveals a conundrum in clinical medicine: textbook pathophysiology is at times the exception rather than the norm. Evidence-based treatment is important, with a margin of safety that is ensured by frequent evaluations to account for individual response.

A fourth and final unforeseen event was that right-lower quadrant abdominal tenderness became evident once the patient’s seizure, vomiting, and then brief paralysis resolved. Although the finding was initially unrecognized, given the circumstances, it was discovered in a timely manner and could easily have gone missed for even longer. Reassessment is paramount when the clinical course does not fit the working diagnosis or mitigating factors confound the assessment. Additionally, patients with altered mental status, intoxication, critical illness, developmental delay, or autism may not express discomfort as one may expect,^36,37 increasing the importance of a thorough physical examination with a low threshold for a broader workup.

This case reminds us of the importance of keeping a broad differential diagnosis, considering alternative diagnoses, and being thorough in reassessment even for seemingly common or straightforward presentations, such as acute gastroenteritis. We highlight the challenges of clinical care with unforeseen events that launch on departure from the primary care office, with unexpected clinical progression, and ultimately yield an elusive underlying diagnosis of appendicitis. The use of reassessment as a clinical tool allowed for the relatively early identification of the patient’s multiple diagnoses and produced an excellent outcome from acute perforated appendicitis.

ASD

autism spectrum disorder

AUS

abdominal ultrasound

CT

computed tomography

ED

emergency department

IV

intravenous

We thank the patient and his mother for allowing us to tell his story. We also thank Dr Pallavi Sagar (pediatric radiology) for her assistance in obtaining and reviewing images for this article.