Clinical manifestations

Presentation and course

Age of occurrence. Breath-holding spells occur in infants and children typically between the ages of 6 months and 4 years, with a significant percentage sometimes continuing in otherwise normal children until 7 or 8 years of age. Median age of onset is typically 6 months to 12 months and median age of peak occurrence is 12 months to 18 months. A young case was reported at 3 days of postnatal life, with up to 15% of children having onset of spells at less than 6 months of age (19; 26; 36; 10).

Seventy-five percent of children with breath-holding spells begin to have spells by 12 months after birth, and 90% by 18 months. The incidence of breath-holding episodes typically peaks in the second year after birth, with spells occurring more than once a day in some cases.

Types of breath-holding spells. A distinction is made in the medical literature between cyanotic breath-holding spells and pallid breath-holding spells. Cyanotic spells are most common, pallid spells least common, and a mixture of the 2 types intermediate in frequency. Cyanotic and mixed events are precipitated by a brief period of crying resulting typically from anger, fear, pain, or frustration. Occasionally, the inciting event is not evident to an observer. The period of crying is often less than 15 seconds. Generally, the longer the children cry, the less likely are they to have a spell (33). Pallid spells typically involve little crying and often no breath holding.

At the end of expiration, the child stops breathing with his mouth open in a "silent scream" position. Deepening cyanosis or progressive pallor develops with an appearance of being dazed. After approximately 10 seconds to 30 seconds of apnea, the child may become unconscious and develop opisthotonus, jaw clenching, hand fisting, tonic posturing with fine jerking movements, or urinary incontinence, followed by limpness. Tonic or tonic-clonic events are common, occurring in 84% of cases in one study (07). Although these events typically last under 10 seconds, episodes of longer duration may be associated with postictal stupor lasting several minutes (56; 53). Physiological changes consist of tachycardia, hypertension, and decline in oxygen saturation. High-amplitude slowing (33) or electrodecremental changes consistent with transient hypoxia characterize the EEG changes.

A majority of children have 1 spell to 6 spells per week, whereas 25% of the children have 1 spell or less per month (21). Fifteen percent of children with breath-holding spells have more than 1 spell per day.

Prognosis and complications

For the overwhelming majority of children, breath-holding spells, whether cyanotic or pallid, are benign, with normal developmental and intellectual outcome (42). Children with breath-holding spells do not develop epilepsy, but are more likely to have syncopal episodes in adulthood (56; 13), suggesting that autonomic aberrations are common to both breath-holding spells and syncope (49; 41). In rare cases, breath-holding spells have been followed by status epilepticus (29; 66; 65).

Death following a breath-holding episode is exceedingly rare (69; 79; 78; 71), but may be related to intrapulmonary shunting during expiratory apnea. However, most children who have died during "breath-holding" have serious conditions such as brainstem gliomas, tracheoesophageal fistula, or tracheostomy (81). Children with associated airway abnormalities may be at greater risk.

Despite their usually benign outcome for the child, education and a support approach with parents or other caregivers is the essential management strategy due to the traumatic experience of witnessing a child having a breath-holding spell.

Clinical vignette

A 13-month-old male, born full-term, vaccines up-to-date, with no significant medical history was sitting on the floor, playing with a toy that his older sibling took away from him. He immediately began to cry, but after 10 seconds he stopped crying, stopped breathing, and started to turn blue. He then became limp and collapsed onto the carpeted floor. His mother ran over and picked him up. He remained limp, but after an additional 5 seconds, he regained consciousness, and his color returned. No jerking or twitching of his body was noted. He appeared tired but did not fall asleep. He returned to playing a few minutes later. He did not have a fever, cough, or congestion. This was the second episode noted in 2 months.

Biological basis

Etiology and pathogenesis

Family history for breath-holding spells suggests the presence of genetic factors in some cases (77; 26). Analysis of the family members of 57 children with breath-holding spells has suggested a 50:50 risk of inheritance from an affected parent, and that it may represent an autosomal dominant condition with incomplete penetrance or expression (24). A close relationship between cyanotic breath-holding spells and pallid breath-holding spells is further suggested by the presence of both in the same family. Some case reports show other genetic etiologies causing early onset, recurrent, severe breath-holding spells, including 16p11.2 microdeletion syndrome and Riley-Day syndrome (59; 06; 55).

Iron deficiency, with or without anemia, also appears to be a factor to contributing breath-holding spells (16; 89; 18; 64; 91; 08; 39; 52; 45; 38). Investigators in Turkey used serum transferrin receptor analysis to assist in detection of iron-deficiency anemia (39). A study has also suggested a possible relationship between maternal iron deficiency anemia and children with breath-holding spells (45).

Because breath-holding spells can be precipitated by a fall, injury, or other noxious stimulus, there has been a long-standing misconception that breath-holding spells are a symptom of an underlying psychosocial problem. Child psychiatrists once maintained that children with breath-holding spells were "neuropathic children of neuropathic parents" (01).

However, studies both in the United States and internationally dispel the misconception that children who have breath-holding spells are "difficult" children. The frequency of temper tantrums may be no greater in children with breath-holding spells than in children without breath-holding spells (22). Similarly, there also seem to be no apparent differences in parental attitudes or in children’s habits related to feeding, toilet training, or sleeping (07).

A correlation between breath-holding spells and later susceptibility to syncope has been suspected based on clinical experience. Pallid breath-holding spells and neurogenic syncope have been proposed to be the same entities based on vagal stimulation response in an adolescent with neurogenic syncope and a history of breath-holding spells (76).

Finally, some case reports suggest other unusual associations of breath-holding spells. Fujisawa and colleagues reported occurrence of severe cyanotic breath-holding spells as a post-operative complication of cervicomedullary ganglioglioma resection in a 2-year-old child (31). Furthermore, hyperventilation-provoked breath-holding spells were noted in 2 young adults during inpatient hospitalizations for somatoform disorder (46).

Traditional concepts of breath-holding spells invoked clearly distinct mechanisms for cyanotic and for pallid spells. Cyanotic breath-holding spells were thought to result from oxygen desaturation during a period of apnea. Pallid spells were thought to be mediated by vagally mediated bradycardia or asystole. Loss of consciousness in either type was attributed to transient cerebral ischemia as a result of the Valsalva maneuver and hypocapnia (29).

However, there appears to be evidence of parasympathetic reflex cardio-respiratory inhibition in both types, as well as subtle generalized autonomic dysregulation (21; 05; 52). Furthermore, mixed and cyanotic breath-holding spells have a complex pathogenesis involving interactions among hyperventilation, end-expiratory apnea, increased intrathoracic pressure, and intrinsic pulmonary mechanisms (09). This “autonomic dysregulation” may be detected in some patients by a high resting heart rate and diastolic blood pressure, positive orthostatic signs, as well as hypersensitivity of pupils to pilocarpine instillation (05).

“Excessive vagal tone” or “vagal overactivity” related to autonomic dysregulation (57; 52) is currently believed to be the primary biological basis of breath-holding spells. Numerous investigators have proposed that there is dysfunction of the autonomic nervous system in infants and children who have breath-holding spells. Studies in a small group of children with pallid breath-holding spells showed a significant decrease in mean arterial blood pressure and increase in pulse rate during the change from lying to standing (23; 21). Findings further suggested greater sympathetic activity in children with cyanotic breath-holding spells.

DiMario and colleagues demonstrated a marked difference in respiratory sinus arrhythmia in children with pallid breath-holding spells when compared to controls or subjects with cyanotic breath-holding spells. This suggested that autonomic dysregulation in pallid breath-holding spells is “caused by a primary central parasympathetic disturbance distinct from the dysregulation found in cyanotic breath-holding spells” (20).

Some investigators maintain that children with breath-holding spells are more likely to have a positive ocular compression test (80; 57; 40). Ocular compression is a maneuver during EEG that demonstrates increased vagal responses and reproduces clinical symptoms of syncope or breath-holding. Ocular compression (triggering the oculovagal or oculocardiac reflex) can provoke periods of asystole and clinical episodes in as many as one-half of these patients, a finding that suggests hyperreactivity of the vagal nucleus (23). An abnormal oculocardiac reflex, manifested by significant slowing of the heart rate and lengthening of the beat-to-beat interval, is present in most children with pallid breath-holding spells and in many with cyanotic spells (20).

Ocular compression also induced longer asystoles in infants with pallid breath-holding spells, compared to those with cyanotic breath-holding spells (48). Vagal mediation of pallid breath-holding spells is further suggested by the observation that atropine can sometimes mitigate this abnormal response to ocular compression (60; 56). It is unlikely, however, that autonomic dysregulation exerts its primary clinical effect on the heart. DiMario and colleagues has postulated that significant arteriovenous shunting at peripheral sites may precipitate breath-holding spells (20).

Stephenson reviewed the use of ocular compression during EEG (82). This procedure has been reported as useful in distinguishing breath-holding spells and syncope from epileptic seizures. However, Stephenson argues that use of this procedure is not an appropriate way to obtain a diagnosis of episodic loss of consciousness in children. Accurate and thorough clinical history remains the mainstays of diagnosis.

Disturbances of respiration during sleep have also been noted in infants with breath-holding spells. A polysomnographic study of infants with breath-holding spells, compared with age-matched controls, disclosed that the breath-holding infants (1) were more likely to be covered with sweat during sleep or wakefulness, (2) had significantly less NREM sleep, more indeterminate sleep, more arousals, and more sleep-stage changes, and (3) had more brief airway obstructions and snoring during sleep (47).

Guilleminault and colleagues demonstrated in 14 children with cyanotic breath-holding spells, that sleep-disordered breathing and abnormal respiratory disturbance index may be present during polysomnography, related to the presence of a narrow upper-airway. Surgical intervention with adenotonsillectomy eliminated sleep disordered breathing in this group of children (35).

Another study showed decreased average numbers of both REMs and bursts of REMs in REM sleep in children with breath-holding spells. The authors hypothesized that because REMs in REM sleep are presumed to be generated in the brain stem, a functional brainstem disturbance may be involved in the occurrence of breath-holding spells (51).

Some children with breath-holding spells appear to be able to voluntarily hold their breath beyond the point at which most individuals feel an overwhelming urge to breathe. However, ventilatory chemosensitivity (increase in breathing in response to hypercapnia or hypoxia) is similar in individuals with and without a history of cyanotic breath-holding in infancy (04).

The presence of lower hemoglobin values in children with breath-holding spells (43; 07) and improvement in or amelioration of spells with oral iron therapy (56; 16; 18; 84; 91), suggests that anemia may be a contributing factor. The study by Daoud and colleagues further suggests that iron deficiency, with or without anemia, may adversely affect autonomic function (18).

Epidemiology

Breath-holding spells are relatively common and have been reported worldwide. In Lombroso and Lerman's study, 4.6% of children had cyanotic breath-holding (56). Some studies have shown a predilection for boys (male:female ratio of 3:1) (07), with males demonstrating an earlier peak (13 to 18 months) than females (18 to 24 months) (19). Having a near relative with a history of breath-holding spells occur in 20% to 35% of cases (54). Early-onset, recurrent, severe breath-holding spells have been associated with 16p11.2 microdeletion syndrome and Riley-Day syndrome (59; 06).

Prevention

No medication has been shown to effectively prevent the development of breath-holding spells, although a variety of strategies have been tried to decrease the frequency of severity of spells. Parents of children with breath-holding spells often mistakenly attempt to shelter the children from any troublesome or unpleasant event. Rather than attempt what is inevitable, there should be education of the parents regarding the pathophysiology of the disorder and “management” of individual episodes.

Mattie-Luksic and colleagues showed that “mothers of children with breath-holding spells or seizure disorders experience more overall stress and disruption in their attachment or understanding of their child, compared with control mothers. These groups of mothers (breath-holding spells and seizure disorders) also perceived their child as more distractible or hyperactive, less adaptive, and more demanding than did control mothers. However, mothers of the breath-holding spells group alone showed significant disruption in their sense of competence as a parent, maintaining self-identity, and receiving positive reinforcement from their child. Mothers of children with seizure disorders showed a similar trend, but it was not significant. These findings were not related to maternal health or feelings of depression, isolation, insufficient spousal support, child's mood, or other life stresses.” They concluded that “parenting a child with breath-holding spells or seizure disorders impacts a greater degree of life stress on mothers of these children. . .although the stress is greater for the breath-holding spells group than for the seizure disorders group” (61).

Differential diagnosis

Confusing conditions

Without a careful clinical history, breath-holding spells may easily be confused with seizures or epilepsy because of their paroxysmal character and clinical manifestations, thus, the essential importance of comprehensive history and exam being obtained in every case (30). They occur in children at ages common to benign febrile seizures, but are associated with neither febrile seizures nor epilepsy. The issue may be confused, however, because on occasion breath-holding spells may precipitate seizures or, rarely, status epilepticus (65; 53; 44). Stephenson and colleagues reviewed true reflex-anoxic seizures, including video recordings (83; 44). Although no specific test can diagnose breath-holding spells, the spells should be differentiated from epilepsy and other types of paroxysmal events such as infantile syncope, vasovagal syncope, prolonged QT syndrome (11), and cardiac dysrhythmia.

There appears to be some overlap in the clinical presentation and mechanism of pallid breath-holding spells and reflex apnea associated with familial rectal pain (75; 37). Careful consideration should also be given to the possibility of anatomic or functional airway abnormalities such as lingual thyroglossal duct cyst (87), gastroesophageal reflux, and perhaps hypoglycemia, hypocalcemia, and benign paroxysmal vertigo. Apparent life-threatening events (ALTE) occur in young infants and are characterized by apnea and cyanosis, hypotonia, and gagging. They differ from breath-holding spells because they occur without apparent provocation (although it may sometimes be difficult to identify a trigger for patients with pallid breath-holding spells).

Anoxic seizures should be differentiated from breath-holding episodes. Seizures due to anoxia may be induced by smothering, ventricular tachyarrhythmias, or sick sinus syndrome (81). A misdiagnosis of anoxic seizures may be made because brief jerks resembling tonic-clonic movements may be produced by both breath-holding and syncopal spells. Particularly with younger children, careful consideration should be given to possible Münchhausen syndrome-by-proxy, in which events may be fictitiously reported or provoked by intentional suffocation.

The literature confirms the importance of including increased QT dispersion, long QT syndrome, or other paroxysmal cardiac rhythm abnormality, in the differential diagnosis of children present with breath-holding spells (02; 27; 03).

Other diagnoses to consider include sudden breath-holding in sleep, apneic episodes associated with sepsis, hyperekplexia, shuddering, congenital laryngeal stridor, laryngospasm, and whooping cough. These respiratory diagnoses are usually associated with characteristic inspiratory sounds that are typically absent in breath-holding spells (55).

Diagnostic workup

• An accurate, detailed history is critical to the correct diagnosis of breath-holding spells and the determination of their cyanotic, pallid, or mixed nature.

It is important to verify that spells are provoked by anger or frustration, discomfort or pain, startle or fear, and are associated with crying. The clinician must also establish that there is a characteristic progression of events (crying, apnea, cyanosis, or pallor preceding the loss of consciousness, and convulsions).

Evaluation should include an ECG to evaluate possible prolonged QT syndrome or other paroxysmal cardiac dysrhythmia and a complete blood count to ascertain the presence of anemia or microcytosis. At times, pediatric cardiology consultation may be indicated, should more than history and ECG with rhythm strip be required (73). EEG and neuroimaging are typically not indicated. However, if an EEG is performed, it may disclose transient high-voltage slowing following brief asystole if recorded during an event (56).

Ocular compression with ECG and EEG monitoring has been proposed as a method of diagnosis for pallid syncope, though carrying the risk of asystole arguing against its use in standard clinical practice. In 2 series, approximately 50% of children had a positive response to ocular compression (33; 56) or an electrodecremental response. Video EEG monitoring may also play an important role in a limited number of cases when there may be confusion over the exact etiology of events. This is especially true if Münchhausen syndrome-by-proxy is suspected.

If there are difficulties in distinguishing breath-holding spells from seizures, prolactin and cortisol levels have been measured in various clinical settings, as both have been reported to increase following generalized convulsive seizures, but not after syncope or breath-holding spells (90). However, normal levels neither confirm a diagnosis of breath-holding spells nor exclude a diagnosis of a convulsive seizure, and guidelines from the American Academy of Neurology suggest prolactin levels do not distinguish epileptic seizures from syncope (15).

If anemia is suspected, a complete blood count and serum ferritin should be performed (55).

Management

• The principal goal in the management of children with breath-holding spells is to educate the parents regarding the association between the precipitating behavior and subsequent breath-holding spell, and to explain that breath-holding spells are reflexic in nature.

The clinician must reassure the parents that virtually every child with breath-holding spells is behaviorally and neurologically normal. It is essential that the clinician encourage the parents not to overprotect the child, but to implement appropriate child-training principles and practices (22).

Because anemia or iron deficiency may be a factor contributing to breath-holding spells, use of an oral iron preparation is appropriate, with the goal of ameliorating or ending the spells (16). A placebo-controlled study with ferrous sulfate solution 5 mg/kg per day for 16 weeks produced a complete response in 51% and a partial response in 36% of children with breath-holding spells, compared to a complete response in none and a partial response in 6% of children without breath-holding spells (18).

Medical management. In Turkey, Mocan and colleagues found that in treating 91 children with breath-holding spells, two thirds of whom had iron deficiency anemia, 84.1% of children showed a “significant difference for correction of cyanotic spells” treated with 6 mg/kg per day (63). Improvement in autonomic dysregulation and cessation of breath-holding spells was demonstrated in 3 children treated with iron supplementation (67).

Historically, treatment with traditional anticonvulsants, sedatives, or stimulants is inappropriate and ineffective (29). Two randomized trials (one unblinded) were shown to reduce the frequency of breath-holding spells using a piracetam, a GABA-like medication given in 2 divided doses totaling 40 to 50 mg/kg per day (28; 74; 17). Further study is indicated before this could be recommended as standard therapy, but piracetam is not currently available in the United States. Levetiracetam has reportedly been effective in ameliorating unconsciousness following breath-holding spells in a toddler (58).

A retrospective study demonstrated successful treatment of pallid breath-holding spells with fluoxetine (86). Six patients with severe pallid breath-holding were identified. These were noted to be complex cases with severe and frequent spells, which were associated with seizures, life-threatening bradycardia, or asystole. Fluoxetine, between 1 and 4 mg daily was titrated to effect. Mean follow-up of 7 years demonstrated either complete resolution of breath-holding or less clinically significant events.

An unusual case of status epilepticus following breath-holding spells was reported by Kuhle and colleagues, and treatment with psychotherapy, as opposed to anticonvulsant medication, was beneficial due to an apparent “severely disturbed mother-daughter relationship that was the trigger of the breath-holding spells” (53).

In rare cases of severe breath-holding, treatment with atropine sulfate, methonitrate (62), or scopolamine (68) has been attempted, although this is not dramatically helpful in most circumstances. Tracheostomy may be considered for the small minority of children with associated anatomic or functional airway abnormalities. Prolonged asystole, which occurs rarely with breath-holding spells, has been selectively managed with implantation of cardiac pacemakers (72; 70).

Regarding cardiac pacemakers, Villain and colleagues chose to implant a pacemaker (ventricular demand) in children with severe breath-holding spells. Eleven children, over 15 years, were paced. The children were studied prior to surgery with ECG and ocular compression, showing normal ECG but bradycardic responses to ocular pressure. The study showed results that were “spectacular with disappearance of spells and restoration of normal activities.” Follow-up Holter monitoring showed normal pacemaker function, having pre-operative studies documenting ECG pauses, sudden bradycardia, and sinus arrhythmia in individual patients. They conclude that “although psychosocial factors play a part in breath-holding spells, pacemaker implantation is effective in suppressing symptoms in severely affected children.”

Kelly and colleagues implanted permanent ventricular demand pacemakers in 10 pediatric patients with breath-holding spells and associated bradycardia and concluded this intervention to be “safe, efficacious, and warranted” in these patients (50). Further studies will be needed to assess which and when children, may benefit from pacemaker implantation (85; 25).

Melatonin has also been proposed as an effective strategy in the treatment of childhood breath-holding spells (12). Lastly, there has been a case report of successful treatment of this condition with glycopyrrolate (88).

Theophylline was described in a case report as treatment for severe and frequent breath-holding spells, but further studies are needed to confirm or refute its usefulness (14; 32).

In a multicentric, retrospective series of 84 children with severe pallid breath-holding spells, belladonna was used as treatment; 44% (20 patients) had complete resolution of spells and no major adverse reactions (34).

Outcomes

For the overwhelming majority of children, breath-holding spells, whether cyanotic or pallid, are benign, with normal developmental and intellectual outcome (42).