Epidemiology

Acute inflammatory demyelinating polyneuropathy (AIDP), one of several discrete varieties of Guillain-Barré syndrome, accounts for about 90 percent of the disease in North America and Western Europe. It is the most common paralytic illness affecting children in countries with established immunization programs. The incidence of Guillain-Barré syndrome has been estimated in population-based studies to be between 0.5 and 1.5 cases per 100,000 children younger than 16 years [Rantala et al., 1994]. Based on these estimates, about 400–600 children per year would be diagnosed with Guillain-Barré syndrome in the United States [Prevots and Sutter, 1997]. Similar data have been reported from Canada [McLean et al., 1994], UK, Europe, and the Scandinavian countries [Farkkila et al., 1991; Korinthenberg and Monting, 1996; Kollar et al., 2009; Korinthenberg et al., 2007; Ramirez-Zamora et al., 2009]; Latin America [Asbury and Cornblath, 1990; Hart et al., 1994; Korinthenberg and Monting, 1996; Molinero et al., 2003; Olive et al., 1997]; and the Far East [Hung et al., 1994; Barzegar et al., 2007, Korinthenberg et al., 2007; Ramirez-Zamora et al., 2009]. Both sexes are affected, although most pediatric series have noted a slight male preponderance.

Diagnostic Criteria

As part of a nationwide program to evaluate the possible relationship of swine-flu immunization to Guillain-Barré syndrome, diagnostic criteria were developed under the sponsorship of the National Institute of Neurological Disorders and Stroke and were later modified after analysis of data from several multicenter treatment trials [Asbury and Cornblath, 1990]. Boxes 90-1 and 90-2 provide summaries of the consensus statement describing the clinical characteristics necessary to diagnose Guillain-Barré syndrome, along with those features, including electrophysiologic testing and laboratory measurements, that mitigate for or against the diagnosis.

Box 90-1 Clinical and Laboratory Features in the Diagnosis of Guillain–Barré Syndrome

I. Required for Diagnosis

Progressive motor weakness of more than one limb

Areflexia – loss of ankle-jerk reflex and diminished knee and biceps reflexes suffice, if other features are consistent with the diagnosis

II. Strongly Supportive of the Diagnosis

Progression – weakness may develop rapidly but cease to progress after 4 weeks; roughly 50% will plateau within 2 weeks, 80% by 3 weeks, and 90% after 4 weeks

Relative symmetry

Mild sensory symptoms or signs

Cranial nerve involvement; facial weakness develops in about half of patients

Autonomic dysfunction

Absence of fever at the onset of neurologic symptoms

Recovery – usually recovery begins 2–4 weeks after progression ceases; it may be delayed for months

Variants

Fever at onset of symptoms

Severe sensory loss with pain

Progressive phase longer than 4 weeks

Lack of recovery or major permanent residual deficit

Sphincter dysfunction – sphincters are usually spared, although transient bladder paralysis may occur

CNS involvement

III. Features Casting Doubt on the Diagnosis

Marked persistent asymmetry in motor function

Persistent bowel or bladder dysfunction

Bowel or bladder dysfunction at onset of symptoms

Discrete sensory level

IV. Features that Exclude the Diagnosis

History of recent Hexa carbon abuse

Evidence of porphyria

Recent diphtheria

Features consistent with lead neuropathy and evidence of lead intoxication

A pure sensory syndrome

Definite diagnosis of an alternate paralytic disorder

(From Asbury AK, Cornblath DR: Assessment of current diagnostic criteria for Guillain-Barré syndrome, Ann Neurol 27:S21, 1990.)

Clinical Features

The clinical presentation of Guillain-Barré syndrome is similar in children and adults [Bradshaw and Jones, 1992; Korinthenberg and Monting, 1996; Rantala et al., 1991; Sakakihara and Kamoshita, 1991; Sarada et al., 1994; Sladky, 2004; Asbury, 2000; Barzegar et al., 2007; Gupta et al., 2008; Kollar et al., 2009; Korinthenberg et al., 2007; Vucic et al., 2009; Wu et al., 1999]. Although there is considerable variability in the nature of the initial symptoms, the overall pattern of the evolution of the clinical syndrome and the ultimate severity of disability usually conform to a triphasic model:

The initial phase is generally relentless and rapid. Between 50 and 75 percent of patients develop maximal weakness within 2 weeks, and 90–98 percent do so by 4 weeks [Asbury and Cornblath, 1990; Korinthenberg and Monting, 1996; Italian Guillain-Barré Study Group, 1996]. In several series describing the natural history in adults and children, the mean duration of the progressive phase was in the range of 10–12 days. A small number of patients continue to progress for longer than 4 weeks. This latter group overlaps to some degree with patients diagnosed with chronic inflammatory demyelinating polyneuropathy. The duration of the plateau phase is similar to that of the progressive phase, averaging 10–12 days but ranging from several days to 4 weeks [Italian Guillain-Barré Study Group, 1996]. Patients who develop severe axonal degeneration have a prolonged plateau phase and tend to have more severe residual weakness [Cornblath, 1990]. In a series primarily of adult patients, about half recovered within 6 months and more than 80 percent by 24 months [Italian Guillain-Barré Study Group, 1996].

A multicenter study of 175 children with Guillain-Barré syndrome has reported similar findings [Korinthenberg and Monting, 1996]. At the height of the disease, 26 percent of patients remained able to walk, but 16 percent required artificial ventilation. The median time from onset of symptoms to first recovery was 17 days; to walk unaided, 37 days; and to be free of symptoms, 66 days. The population in this study was heterogeneous, with many children receiving some form of treatment, including plasmapheresis, intravenous immune globulin, and corticosteroids. A large number of children have a relatively benign course, and a smaller group a more severe and protracted course. At 6 months’ follow-up evaluation, 98 of 106 children (92 percent) were free of symptoms and the others were able to walk unaided. Subsequent reports similarly have emphasized a favorable prognosis [Barzegar et al., 2007; Korinthenberg et al., 2007; Lee et al., 2008; Nagasawa et al., 2006; Ramachandran and Kuruvilla, 2004; Shafqat et al., 2006]. As in series of adult patients, the maximum degree of disability at clinical nadir appears to be the most powerful predictor of incomplete recovery.

In the past, there has been considerable discussion as to whether the natural history of Guillain-Barré syndrome in children is more benign or comparable to adults. The argument may ultimately be unresolvable because of the absence of large prospective controlled treatment trials in children compared with adults [Klegweg et al., 1989]. There are few case series that include descriptions of the natural course of this syndrome in untreated children with similar end-points. Two adult benchmark studies are the North American Treatment Trial and the French Plasmapheresis Treatment Trial.

Table 90-3 summarizes some of the presenting complaints seen in a series of 49 children hospitalized with Guillain-Barré syndrome. This cohort has proved to be typical of other reported pediatric series. Weakness was the most common initial complaint. Pain, particularly in the back and lower extremities, was also prominent and was near-universal in the most severely affected children. A history of difficulty with balance early in the course of the illness was elicited in almost half the children. The classic triad, characteristic of the Miller Fisher variant of Guillain-Barré syndrome (ataxia, ophthalmoparesis, and areflexia), was present in only two children. All children had evidence of progression. Maximal disability was quantitated according to the 1985 Guillain-Barré Syndrome Study Group grading scale (Table 90-4). At the time of maximal neurologic deficit, 25 percent of children were able to ambulate 5 meters without assistance (grade 2). Another 11 children (22 percent) reached a nadir of grade 3 (unable to walk 5 meters without assistance). The largest group of children (39 percent) were bed- or wheelchair-confined (grade 4), and 7 children (14 percent) required mechanical ventilatory assistance (grade 5). There were no fatalities. Autonomic dysfunction occurred in 14 children (28 percent), and included labile hypertension in 9 children and urinary or bowel incontinence in 6 children.

Table 90-3 Clinical Features in 49 Children Younger than 18 Years with Guillain-Barré Syndrome*

* Two patients had findings consistent with Miller Fisher syndrome.)

(Data from unpublished observations of John T Sladky, MD.

Table 90-4 Clinical Grading Scale from the U.S. Guillain-Barré Syndrome Plasmapheresis Trial

(Adapted from the Guillain-Barré Syndrome Study Group: Plasmapheresis and acute Guillain-Barré syndrome, Neurology 35:1096, 1985.)

The duration of symptoms in children with Guillain-Barré syndrome requiring hospitalization varies in the reported studies. The incidence of admission to pediatric intensive care units for treatment of respiratory failure or other serious medical problems (e.g., infection) or for plasmapheresis ranges between 17 and 68 percent, with the average length of stay about 11 days [Jansen et al., 1993; Ramirez-Zamora et al., 2009; Yata et al., 2003]. The average duration for mechanical ventilation ranges from 17 to 22 days. The mean duration of hospitalization is also quite variable, ranging from 12 to 84 days [Jansen et al., 1993; Jones, 1996]. Time to independent walking reported in a multicenter study in Western Europe [Korinthenberg and Monting, 1996] was somewhat shorter than in other pediatric series, which reported durations of 43 days [Lamont et al., 1991], 52 days [Epstein and Sladky, 1990], and 113 days [Shahar and Leiderman, 2003]. The European study was a questionnaire-based survey that included children from community hospitals along with tertiary referral centers, which the authors thought may represent a less severely affected population. In addition, many of these children received therapies, including corticosteroids, intravenous immune globulin, and plasmapheresis.

Autonomic dysfunction has been reported in 12.5–25 percent of children with Guillain-Barré syndrome [Bradshaw and Jones, 1992; Hung et al., 1994; Lee et al., 2008; Shafqat et al., 2006] and is similar to that in adults [Ropper, 1994; Zochodne, 1994; Korinthenberg et al., 2007; Ryan, 2005; Shafqat et al., 2006]. Symptoms are usually intermittent and include postural hypotension, supraventricular tachycardia, bradycardia, and fluctuating blood pressure. Gastrointestinal complaints are numerous and related to impaired swallowing, gastroesophageal dysmotility, pseudo-obstruction, and constipation. Urinary retention also occurs, and the potential need for catheterization and bladder decompression should be considered in children with more severe symptoms during the acute phase. On rare occasions, cardiac arrest secondary to autonomic dysfunction has been observed in children [Bos et al., 1987; Briscoe et al., 1987].

Although AIDP is often regarded as an entity in which motor symptoms predominate, pain is a common symptom, occurring in up to 79 percent of children [Bradshaw and Jones, 1992; Nguyen et al., 1999; Sladky, 2004] and 72 percent of adult patients [Pentland and Donald, 1994]. The types of pain vary and include paresthesias, dysesthesias, axial and radicular pain, meningismus, myalgia, joint pain, and visceral discomfort. Pain intensity on admission appears to correlate poorly with the degree of initial neurologic disability and is not predictive of prognosis [Moulin et al., 1997]. Patients may also present with or develop pain in a variety of clinical settings, such as the intensive care unit or pediatric ward, or during the more chronic phase of the disease when in rehabilitation or at home. Back and leg pain usually resolves over the first 8 weeks, but dysesthetic extremity pain may persist longer in 5–10 percent of patients, despite motor recovery. Moderate to severe pain requires aggressive treatment. In a recent study of adult patients with Guillain-Barré syndrome, 75 percent required oral opioid analgesics, and 29 percent received parenteral analgesia to provide adequate pain relief [Hahn, 1996]. Despite aggressive therapy, most adults with grade 4 or 5 disease severity reported that their pain was inadequately treated. Few data regarding pain symptoms in children with this syndrome have been published. The principles of pain management for Guillain-Barré syndrome patients are well described [Moulin et al., 1997; Pentland and Donald, 1994].

Clinical Variants of Guillain-Barré Syndrome

Several clinical forms of Guillain-Barré syndrome have been characterized (Table 90-5). Some of the more clearly defined entities are discussed in this section.

Table 90-5 Variants of Immune-Mediated Acute and Chronic Demyelinating and Axonal Disorders Seen in Children

* References listed are for pediatric case series, if available.

CMAP, compound motor unit action potentials; CSF, cerebrospinal fluid; EMG, electromyography.

Differential Diagnosis

A variety of other peripheral and occasionally CNS disorders can be confused with Guillain-Barré syndrome. These are outlined in Box 90-3, and the reader is referred to the specific chapters that discuss these conditions.

Box 90-3 Differential Diagnosis of Guillain-Barré Syndrome

Cerebral

Bilateral strokes

Hysteria

Cerebellar

Acute cerebellar ataxia syndrome (multiple etiologies)

Posterior fossa structural lesion

Spinal

Compressive myelopathy

Transverse myelitis

Anterior spinal artery syndrome

Peripheral Nerve

Toxic neuropathy

Drugs: amitriptyline, dapsone, glutethimide, hydralazine, isoniazid, nitrofurantoin, nitrous oxide, incrusting

Toxins: acrylamide, glue sniffing, fish toxins, heavy metals (lead, arsenic, mercury, thallium), insecticides, N-hexane and other solvents, organophosphates

Critical illness neuropathy

Diphtheria

Tick paralysis

Porphyria

Neuromuscular Junction

Botulism

Myasthenia gravis

Neuromuscular-blocking agents

Muscle Diseases

Acute viral myositis

Acute inflammatory myopathies (polymyositis, dermatomyositis)

Metabolic myopathies (multiple types)

Periodic paralysis

(Modified from Evans OB: Guillain-Barré syndrome in children, Pediatr Rev 8:69, 1986; Jones HR: Childhood Guillain-Barré syndrome: Clinical presentation, diagnosis, and therapy, J Child Neurol 11:4, 1996.)

Clinical Laboratory Evaluation

The diagnosis of Guillain-Barré syndrome is based on clinical features. Confirmatory laboratory studies include cerebrospinal fluid examination, electrophysiologic testing, and occasionally MRI to look for enhancement in spinal roots after contrast administration.

Pathogenesis

The pathologic findings of Guillain-Barré syndrome are confined predominantly to spinal nerve roots and peripheral nerves [Asbury et al., 1969]. The pathogenesis of the disorder is immune-orchestrated, with macrophage-mediated injury to sensory and motor axons within the peripheral nervous system [Hafer-Macko et al., 1996b; Hartung et al., 1996; Hughes et al., 1999]. In contradistinction to the traditional notion that Schwann cell/myelin-related constituents are the primary targets of the autoimmune process and that associated axonal damage is a nonspecific “bystander effect” of the inflammatory reaction, there is evidence that epitopes expressed exclusively on axolemma may also be singled out for immune attack and result in concomitant axonal degeneration [Hafer-Macko et al., 1996b; Hartung and Kieseier, 1999; Hughes et al., 1999; Asbury, 2000; Hughes and Cornblath, 2005; Jones, 2000; van Doorn, 2009; van Doorn et al., 2008]. It is likely that myelin/Schwann cell and axonal surface epitopes are immunologically targeted to variable degrees during the active phase of Guillain-Barré syndrome, resulting in the spectrum of clinical manifestations that are features of the syndrome. The typical pathogenic mechanism in Guillain-Barré syndrome involves macrophages insinuating themselves under the myelin sheath at the node of Ranvier or between myelin lamellae, stripping myelin from the underlying axon [Hafer-Macko et al., 1996b; Hartung et al., 1996, 2002; Hughes et al., 1992, 1999; Kiefer et al., 2001; Steck et al., 1998; Vucic et al., 2009; Willison et al., 2002, 2008; Willison, 2005b]. Documenting the early events in the immune-mediated attack on the peripheral nervous system is, for obvious reasons, difficult in humans. The early pathologic changes have been characterized in a laboratory model, experimental allergic neuritis (EAN), a disorder nearly homologous to acute inflammatory demyelinating polyneuropathy. Experimental allergic neuritis can be induced in laboratory rodents immunized with constituents of peripheral nerve myelin, or with transfer of activated T cells from animals with the disease (Figure 90-1) [Rostami, 1997]. A different pattern of injury has been described (Figure 90-2) in acute motor axonal neuropathy related to Campylobacter infection in children from northern China. In these patients, macrophages have been found to interrupt axoglial tight junctions at nodes of Ranvier and attack and destroy the motor axon, sparing the circumferential myelin, which then secondarily breaks down [Griffin et al., 1995, 1996a, b; Hafer-Macko et al., 1996a; McKhann et al., 1993]. This process results in a relatively pure motor syndrome without evidence of primary demyelination on histopathologic examination or electrophysiologic testing [Griffin et al., 1995, 1996a, b; Hafer-Macko et al., 1996a; McKhann et al., 1993; Hughes and Cornblath, 2005; Kuwabara et al., 2001; Lu et al., 2000; Vucic et al., 2009; Willison et al., 2002, 2008; Willison, 2005b].

Fig. 90-1 Cross-section through a sciatic nerve from a Lewis rat during the acute phase of experimental allergic neuritis.

There is a large collection of mononuclear cells within the endoneurium during this acute phase of the illness (bounded by arrows) (×250).

(Courtesy of John Sladky, MD.)

Fig. 90-2 Cross-section through a sciatic nerve from a Lewis rat during the acute phase of experimental allergic neuritis.

Note that there are several demyelinated axons (some indicated with asterisks) and thinly myelinated axons in the early stages of remyelination. Large-caliber myelinated axons demonstrating splitting and fragmentation of myelin lamellae in the early stage of macrophage-mediated demyelination are present (arrows) (×650).

(Courtesy of John Sladky, MD.)

It is clear that, early in the immunological cascade that ultimately results in neurological disease, an exquisitely precise targeting mechanism is activated, which directs subsequent downstream events, resulting in focal tissue or cellular damage. Previous hypotheses that these disorders are directed principally at myelin-related antigens and that the extensive axonal injury that frequently occurs is a nonspecific “bystander effect” have been questioned by empirical observations, especially among variants within the Guillain-Barré syndrome family of diseases. These are among the best-studied disorders in which firm evidence exists that antibodies elicited against antigenic components found on the membrane surfaces of an infectious organism play a significant role in the pathogenesis of human disease.

In the most directly relevant examples, there are recognizable homologies between the infectious agent and constituents of the nervous system that reinforce the notion of cross-reactivity underpinning immune-mediated neurological injury, a process that has been termed “molecular mimicry.” This case can best be made with C. jejuni, where epidemiological studies confirm a significant relation between antecedent Campylobacter enteritis and subsequent peripheral neuropathy [Hadden and Gregson, 2001; Kuwabara et al., 2004; Liu et al., 2003; Nachamkin et al., 2007; Takahashi et al., 2005; Wu et al., 1999]. Investigators also have established that infection with a relatively small group of specific C. jejuni serotypes accounts for the majority of cases of Guillain-Barré syndrome-like illness worldwide. The precise relation among serotypes, specific antibodies, and clinical phenotypes requires further study. There are compositional identities between oligosaccharides expressed on the lipo-oligosaccharide membrane of the bacterium and neural membrane gangliosides from which oligosaccharide moieties protrude, where they are recognizable by immune surveillance mechanisms. Antibodies directed toward these surface epitopes can be demonstrated in many patients and almost certainly play a role in disease pathogenesis [Hughes and Cornblath, 2005; van Doorn et al., 2008; Willison, 2005b; Willison and Plomp, 2008]. There are even stronger correlations between some phenotypical subtypes in which discrete populations of peripheral axons and Schwann cells are attacked by antibodies directed against specific gangliosides, as has been persuasively demonstrated with anti-GQ1b antibody expression in patients with the Miller Fisher syndrome variant [Akinci et al., 2007; Koga et al., 2001; Nagashima et al., 2004; Nagashima, 2007; Nishimoto et al., 2004; Takahashi et al., 2005]. Similarly, in the acute motor axonal variant of Guillain-Barré syndrome, acute motor axonal neuropathy, anti-GM1 or GD1a, ganglioside antibodies specifically target motor nerve axolemmal membranes.

The sequence of events that engenders autoimmune neuromuscular disorders such as Guillain-Barré syndrome is incompletely understood. Dalakas [1995] emphasized that several sequential steps are required to orchestrate autoimmune-mediated tissue injury:

The presence of proinflammatory and anti-inflammatory cytokines acts to modulate the intensity of the autoimmune attack on endoneurial contents through multiple pathways. Nerve fibers in peripheral nerves are relatively protected from systemic immune responses and inflammatory reactions by the specialized endoneurial, endothelial, and perineural barriers [Dalakas, 1995; Hartung et al., 2001, 2002; Hughes et al., 1999]. Compared with peripheral nerve trunks, there are more macrophage-like cells and more permeable blood vessels in the dorsal root ganglia and spinal roots that are preferred sites for the inflammatory response in experimental allergic neuritis. As noted earlier, experimental allergic neuritis can be induced by immunization with myelin proteins or fragments of the proteins. P0 and P2 proteins have been most consistently used to induce this model; a similar but less inflammatory, more chronic demyelinating neuropathy can be induced in rabbits with the glycolipid, galactocerebroside [Harvey et al., 1987; Ang et al., 2001b, 2002a, b; Willison, 2005b; Willison and O’Hanlon, 1999; Willison and Plomp, 2008]. In myelin protein-induced experimental allergic neuritis, CD4+ helper T-cell responses are sufficient to transfer full-blown disease, but antibodies of undefined specificity also may contribute to the disease process. In rabbit galactocerebroside neuritis, antibodies are probably more important and are directed predominantly against a haptenic group, including galactose. A small proportion of patients with Guillain-Barré syndrome and chronic immune demyelinating polyneuropathy show T-cell responses to P0 or P2, which may mark important pathogenetic reactions, or may simply represent an epiphenomenon. In addition, a fraction of patients with Guillain-Barré syndrome and chronic inflammatory demyelinating polyneuropathy have antibodies to gangliosides – in particular, the most abundant peripheral nerve ganglioside, GM1. There is an association between patients with acute severe Guillain-Barré syndrome and marked axonal degeneration and immunoglobin G1 antibodies to GM1 and GD1b; similarly, a strong correlation exists, linking Miller Fisher syndrome with antibodies to GQ1b and GD1a [Ang et al., 2001a, b, 2002a, b; Kusonoki et al., 2008; Nagashima et al., 2004; Nagashima, 2007; Nishimoto et al., 2004, 2008b; Willison, 2005b; O’Hanlon et al., 2001; Willison and Plomp, 2008]. In acute inflammatory demyelinating polyneuropathy, macrophage-mediated demyelination is probably orchestrated by autoantibody- and T-cell-mediated autoimmune responses. The nature of the autoantigen and the relative importance of T- versus B-cell responses are likely to vary, depending on the nature of the initial immunologic stimulus.

Antecedent events associated with the subsequent development of Guillain-Barré syndrome have included infection, immunization, drugs [Awong et al., 1996], trauma, or parturition. It is likely that many of the reported associations are purely coincidental. The relations between several common infections and immunizations and Guillain-Barré syndrome that seem relatively solid are summarized in Table 90-6. Some of the better-established infections include cytomegalovirus [Jacobs et al., 1997c], Epstein–Barr virus [Grose and Feorino, 1972; Jacobs et al., 1998] and other herpesviruses [Dowling and Cook, 1981], hepatitis [Tabor, 1987], and Mycoplasma spp. infections [Susuki et al., 2004; Thomas et al., 1993]. Of the acute viral infections implicated as possible antecedent events, cytomegalovirus and Epstein–Barr virus have been most consistently reported, using serologic markers, to be statistically related to subsequent Guillain-Barré syndrome illnesses [Jacobs et al., 1998; Winer et al., 1988; Hughes and Cornblath, 2005]. The subgroups of Guillain-Barré syndrome – acute inflammatory demyelinating polyneuropathy, acute motor and sensory axonal neuropathy, and acute motor axonal neuropathy – have all been associated with antecedent or coincident systemic infection with Campylobacter spp. [Hadden and Gregson, 2001; Hughes, 2004; Jacobs et al., 1997b; McCarthy and Giesecke, 2001; Melendez-Vasquez et al., 1997; Rees et al., 1995a; Toyka, 1999; Ang et al., 2000; Nachamkin et al., 2007; Ogawara et al., 2000; Schessl et al., 2006, 2007]. Controversy exists concerning the oral polio vaccine and its relation to Guillain-Barré syndrome. In 1993, the Institute of Medicine found that the evidence favored acceptance of a causal relationship between oral polio vaccine and Guillain-Barré syndrome [Ismail et al., 1998; Stratton et al., 1994]. Numerous subsequent epidemiologic studies have found no such correlation [Rantala et al., 1994; Hughes et al., 2006c; Kuitwaard et al., 2009]. This issue is discussed further in Chapter 104.

Table 90-6 Infectious Agents Associated with Guillain-Barré Syndrome

Treatment

Immunomodulating strategies for the treatment of Guillain-Barré syndrome can include agents to interrupt the autoimmune process at several stages from T-cell activation to macrophage-mediated tissue injury [Hahn, 1997; Hughes et al., 1999; Saida, 1996]. There is reasonable evidence that it is possible to blunt the severity of tissue injury by interfering with the autoimmune process at multiple levels, including interventions that:

The mechanisms by which plasmapheresis and intravenous immune globulin administration act in Guillain-Barré syndrome or chronic inflammatory demyelinating polyneuropathy are not entirely understood. It is likely that the primary effect of plasmapheresis is to remove circulating antibody directed toward peripheral nerve antigens. The predominant effect of intravenous immune globulin is speculative. High-dose intravenous immune globulin administration may act by binding anti-idiotypic antibodies, absorbing complement, or downregulating B-cell-mediated antibody production. Other actions that may also play a role include blockade of activated receptors, enhancement of suppressor-cell activity, and interference with lymphocyte proliferation [Thornton and Griggs, 1994]. Both of these therapies also downregulate production of proinflammatory cytokines, which play an important part in enhancing the severity of the immune-mediated injury [Aarli, 2003; Creange et al., 1996; Jander and Stoll, 2001; Lisak et al., 1997; Sharief et al., 1999; Stoll, 2002].

There are no definitive, prospectively derived data to indicate that treatment of children with Guillain-Barré syndrome using plasmapheresis or intravenous immune globulin is efficacious. Nevertheless, several limited studies indicate that both treatments may be effective for this disorder [Abd-Allah et al., 1997; Baskin et al., 1996; Epstein and Sladky, 1990; Gurses et al., 1995; Jansen et al., 1993; Kanra et al., 1997; Khatri et al., 1990; Lamont et al., 1991; McKhann, 1990; Shahar and Leiderman, 2003; Singhi et al., 1999; Sladky, 2004; Vajsar et al., 1994; Hughes et al., 2003b, 2004, 2006b, d, 2009; Korinthenberg et al., 2005]. Evidence in large prospective trials of plasmapheresis and intravenous immune globulin involving predominantly adults has confirmed that both treatment modalities reduce morbidity associated with Guillain-Barré syndrome and that they are comparable in efficacy [French Cooperative Group on Plasma Exchange in Guillain-Barré Syndrome, 1987; Plasma Exchange/Sandoglobulin Guillain-Barré Syndrome Trial Group, 1997; Hughes and Cornblath, 2005; Hughes et al., 2003b]. Both modalities appear to reduce the duration of the disease and improve the neurologic outcome [Plasma Exchange/Sandoglobulin Guillain-Barré Syndrome Trial Group, 1997]. In addition, the Guillain-Barré Syndrome Study Group documented that treatment with high-dose intravenous corticosteroids not only was ineffective, but also was likely to be harmful in adults due to the increase of adverse events in the steroid-treated cohort [Guillain-Barré Syndrome Steroid Trial Group, 1993; Hughes et al., 2010]. In a subsequent study in adults by these same investigators, plasmapheresis and intravenous immune globulin were comparable in efficacy, and these two treatments in combination were not superior to either treatment alone [Plasma Exchange/ Sandoglobulin Guillain-Barré Syndrome Trial Group, 1997]. Complementary studies have examined the benefit of performing two, four, or six plasmaphereses in patients with varying degrees of severity of Guillain-Barré syndrome [French Cooperative Group on Plasma Exchange in Guillain-Barré Syndrome, 1987]. The conclusions of this study recommended that patients with mild Guillain-Barré syndrome receive two cycles of plasmapheresis on admission, whereas patients with moderate or severe forms would benefit from two additional exchanges. More than four exchanges did not enhance the pace or completeness of recovery.

Prognosis

Most case series indicate that the prognosis of Guillain-Barré syndrome in children is generally excellent and likely better than that for older age groups [Briscoe et al., 1987; Cole and Matthew, 1987; Kleyweg et al., 1989; Korinthenberg and Monting, 1996; Ortiz Corredor and Mieth Alviar, 2003; Rantala et al., 1991; Sladky, 2004; Vajsar et al., 2003; Korinthenberg et al., 2005, 2007; Koul and Alfutaisi, 2008; Kuitwaard et al., 2009; Ramachandran and Kuruvilla, 2004; Ramirez-Zamora et al., 2009; Ryan, 2005]. Opinion in the literature is not unanimous, and dissenting viewpoints argue that the clinical course and prognosis are similar [Kleyweg et al., 1989]. In the group of 49 children listed in Table 90-3, 80 percent achieved complete functional recovery within 6 months of the onset of illness. About 20 percent had mild residual neurologic deficits, including mild weakness in the facial or lower-extremity muscles and decreased deep tendon reflexes [Barbara Martin, MD, personal communication]. Similar findings, as reviewed earlier in this chapter, were reported from a retrospective international multicenter study of 175 children with Guillain-Barré syndrome [Korinthenberg and Monting, 1996].

A limitation in our ability to draw definitive conclusions regarding the prognosis of Guillain-Barré syndrome in children is the absence of prospectively defined, systematically collected clinical data. The number of children included in individual published studies is small and contributes to the variable fashion in which these limited observations are interpreted. The consensus in the literature is that children will do at least as well as, if not better than, older patients with Guillain-Barré syndrome [Jones, 2000; Korinthenberg and Monting, 1996; Sladky, 2004]. This opinion is reinforced by several nonprospective studies looking at treatment, in which children recover faster after plasmapheresis or intravenous immune globulin administration than published recovery rates in adults [Epstein and Sladky, 1990; Jansen et al., 1993; Jones, 2000; Korinthenberg and Monting, 1996; Lamont et al., 1991; Reisin et al., 1996; Sladky, 2004]. These studies are not directly comparable and need to be interpreted with caution; however, there is emerging consensus that the prognosis in childhood Guillain-Barré syndrome is good and may be enhanced as new translational initiatives bring novel treatment modalities to the bedside.

Epidemiology

The true incidence of this disorder in children is unknown and is probably unlikely to be determined. Although it represents the most common acquired, potentially treatable chronic neuropathy in childhood, it none the less is rare [Iijima et al., 2008]. Ascertaining the incidence of this disease in children is further complicated by its propensity for mimicking genetically determined neuropathies, except that the often-anticipated history of dominant transmission is lacking. As many children affected with this disorder respond well to immunosuppressive therapy, it is critical to consider this disease in the differential diagnosis of any child with chronic neuropathy, particularly when there are features of demyelination and when evidence for a hereditary pattern cannot be elicited [Connolly, 2001; Korinthenberg, 1999; McLeod et al., 1999; Nevo, 1998; Rodriguez-Casero et al., 2003; Ryan et al., 2000; Sladky, 1996; Hughes et al., 2006; Markowitz et al., 2008; Vallat et al., 2010].

Clinical Features

The clinical features of CIDP are protean and diagnosis can be difficult. It can present as an indolently progressive, predominantly motor neuropathy and closely resemble hereditary motor and sensory neuropathy type I, including several genotypic variants [Carter et al., 2004]. Active demyelination may be confined to proximal nerve segments, and electrophysiologic testing may be most compatible with an axonal neuropathy with the disease masquerading as hereditary motor and sensory neuropathy type II. It has been reported to present in infancy with extremely slow nerve conduction velocity measurements resembling type III hereditary motor and sensory neuropathy [Sladky, 1996]. It may also present with features of sensory ataxia and can mimic Friedreich’s ataxia [Quan and Kleinschmidt-DeMasters, 2005]. Alternatively, the clinical manifestations may include any combination of these presentations. In some children, the disease presentation is more typical, with the distinct onset of neuropathic symptoms in a previously healthy child, a negative family history with normal examination of both parents, and confirmatory laboratory studies. This presentation is usually the exception and the likelihood of diagnosing children with more subtle presentations of CIDP is directly related to the persistence and expertise of the neurologist evaluating such patients. In summary, the disease may present in any age group, the rate of progression may be subacute to indolent, and it may affect selective populations of peripheral nerve axons and interrupt the task of any organ system that relies on peripheral nerve function for feedback or regulatory input.

There have been several series reporting the clinical, electrophysiologic, histopathological, and prognostic features in children with CIDP [Connolly, 2001; Hattori et al., 1998; Korinthenberg, 1999; Nevo et al., 1996; Simmons et al., 1997a, 1997b; Sladky, 1996; Markowitz et al., 2008; Rabie and Nevo, 2009; Rossignol et al., 2007]. As one might predict, there are divergent observations among individual reports, presumably due to the relatively small number of patients in each series. However, several common themes emerge. First, as in adults with CIDP, the clinical course and response to treatment are variable. Children may present with a subacute course or a more subtle indolent progression of symptoms. Lower-extremity weakness is the most common presenting complaint, whereas sensory symptoms are relatively uncommon. Areflexia is the rule. Second, the pace of progression of the initial symptoms may range from fulminant to subacute over 2–3 months [Rodriguez-Casero et al., 2005; Ruts et al., 2010] or may exhibit a more protracted course. In some individuals, the initial course may be so indolent as to mimic genetically determined neuropathies. Those individuals with a more rapid onset of symptoms may have a better likelihood of responding to treatment and a more benign outcome. Third, electrophysiologic testing is a useful diagnostic tool and will, in most cases, confirm the acquired nature of the neuropathy. These tests, however, have only limited predictive value regarding prognosis. Fourth, CIDP is an eminently treatable disease in children. Therapeutic outcomes fall roughly into three categories. Most children initially improve in response to treatment with corticosteroids or intravenous immune globulin. About half the children with CIDP have a monophasic course and are able to be weaned from their initial treatment, although, in many cases, over a protracted time period. The remaining patients become refractory to steroid treatment and require alternative immunosuppressant treatment or suffer one or more relapses. Most patients in this group will respond to other immunosuppressants and many achieve remission. A small fraction of children with CIDP may become refractory to aggressive immunotherapy and become dependent on chronic immunosuppression or sustain severe, widespread peripheral nerve injury with substantial residual neurologic disability. As in Guillain-Barré syndrome, the prognosis for children with CIDP is more favorable than in adults [Korinthenberg, 1999; Koul et al., 2002; Kuitwaard and van Doorn, 2009; Markowitz et al., 2008; Ryan et al., 2000].

Laboratory Evaluation

Although there are persuasive arguments to the contrary, it is sometimes useful to look at Guillain-Barré syndrome and CIDP as different ends of the temporal spectrum of the same disease. In that context, the similarities between the disorders are more striking than the differences. Like the children with the more fulminant Guillain-Barré syndrome, children with CIDP may show evidence of pathognomonic features of multifocal demyelination on electrophysiologic testing. These features have been reviewed extensively [Albers and Kelly, 1989; Connolly, 2001; Hattori et al., 1998; McCombe et al., 1987; Simmons et al., 1997b; Markowitz et al., 2008; Rabie and Nevo, 2009; Ruts et al., 2010].

Sensory and motor nerve conduction studies are the most useful and practical method to assess the physiologic properties of saltatory conduction in large-caliber myelinated fibers. Such data may be useful in distinguishing inherited from acquired disorders. However, none of the electrophysiologic measures is, in itself, an absolute discriminator. Apart from those rare kinships with steroid-responsive inherited demyelinating neuropathy [Bird and Sladky, 1991; Dyck et al., 1982c; Ginsberg et al., 2004], some children with typical hereditary motor and sensory neuropathy type I may demonstrate dispersion of proximal evoked compound motor action potentials and conduction block. Conversely, some children with CIDP may have electrophysiologic characteristics that suggest a genetically determined disease. When the clinical history is suggestive of an acquired disorder and electrophysiologic testing is compatible with an axonal neuropathy, cerebrospinal fluid examination may be helpful in demonstrating an elevated protein concentration typical of demyelinating neuropathies. Measurement of circulating antibodies to myelin protein constituents is rarely useful in children. Contrast-enhanced MRI may indicate blood–nerve barrier disruption caused by spinal nerve root inflammation [Bertorini et al., 1995; Crino et al., 1993; Kuwabara et al., 1997; Midroni et al., 1999]. Nerve biopsy may be necessary to assist in discriminating acquired from sporadic or recessively transmitted genetic disorders [Gabreels-Festen et al., 1993; Ginsberg et al., 2004; Hu et al., 2007; Schmidt et al., 1996; Vallat et al., 2003b]. Because de novo mutations can result in the appearance of a sporadic disorder, testing for an abnormality in the more common gene mutations associated with hereditary motor and sensory neuropathies phenotype may establish a diagnosis and define implications for prognosis and genetic counseling. Finally, when the clinical and laboratory data remain inconclusive, including clinical and electrophysiologic examination of family members, a therapeutic trial of immunosuppressive therapy may be warranted.

Pathologic Findings

The most characteristic histopathologic features on nerve biopsy of patients with CIDP are loss of myelinated axons, particularly larger-caliber axons; evidence of segmental demyelination and remyelination; and regional variability in the severity of fiber loss and inflammation [Baba et al., 1993; Gabreels-Festen et al., 1993; Matsumuro et al., 1994; Sladky et al., 1986]. The presence of inflammatory cells and selective fascicular variability in the expression of pathologic abnormalities are the only characteristics helpful in distinguishing an acquired from a genetically determined disorder [Gabreels-Festen et al., 1993]. The presence of thinly remyelinated internodes, segmental demyelination and remyelination, and onion bulb formations (indicative of chronicity) may be seen in any chronic demyelinating neuropathy. Specific features of inflammation (i.e., mononuclear cellular infiltrates) are present in a minority of biopsies from children and infants with CIDP. Most children with CIDP demonstrate evidence of subperineurial and endoneurial edema and an increased number of inflammatory cells dispersed within the endoneurium [Sladky et al., 1986]. Discrete perivascular collections of inflammatory cells around epineurial or endoneurial vessels are relatively rare. Teased fiber preparations demonstrating isolated demyelinated internodes or random variability in internodal length and myelin thickness are also suggestive of a multifocal inflammatory demyelinating process (Figure 90-3). Electron microscopy may confirm these features and demonstrate fragmentation of myelin sheaths, with splitting of myelin lamellae and intralamellar insinuation of macrophages ingesting internodal myelin. Immunocytochemical testing may be helpful to confirm activated lymphocytic subpopulations within the endoneurium and to demonstrate expression of class I and II major histocompatibility complex markers on Schwann cell surfaces in areas of active demyelination. Similar findings have been reported in Guillain-Barré syndrome [Asbury et al., 1969; Hughes et al., 1992; Midroni et al., 1999; van Doorn, 2009; van Doorn et al., 2008; Vucic et al., 2009; Willison, 2005b; Willison et al., 2008]. More recently, circulating antibodies to myelin-related gangliosides have been reported and associated to some degree with specific Guillain-Barré syndrome subtypes. However, antibodies to any of these candidate myelin autoantigens were not significantly more common in patients with CIDP than in controls or other groups studied. In addition, serologic evidence of C. jejuni or cytomegalovirus infection was not detected.

Fig. 90-3 Teased fiber preparation of osmicated axons from a patient with chronic immune demyelinating polyneuropathy.

Sequential internodes along the course of an individual myelinated axon are shown. (Individual nodes of Ranvier are indicated with arrowheads.) The random variability in internodal length and myelin thickness is pathognomonic for a multifocal demyelinating process and could be seen in chronic immune demyelinating polyneuropathy or during the recovery stage in Guillain-Barré syndrome (×68).

(Courtesy of John Sladky, MD.)

The ability to make a conclusive diagnosis from a biopsy specimen of a multifocal process such as CIDP is limited by the problem of sampling error. Since only a finite amount of tissue is removed, pathognomonic histopathologic features may not be present in the sample obtained for pathologic studies. This situation is of particular concern in CIDP, in which inflammation with secondary axonal injury may be confined to proximal nerve segments. A biopsy specimen from a sural nerve site well “downstream” from the area of active inflammation may indicate only evidence of Wallerian degeneration of distal axons below the level of primary injury.

Treatment

Several therapeutic modalities have been found to be effective in the management of CIDP, including corticosteroids, plasmapheresis, intravenous immune globulin, and other immunosuppressive agents [Baba et al., 1993, 1996; Barnett et al., 1998; Berger et al., 2003; Brannagan, 2002; Brannagan et al., 2002; Choudhary et al., 2005b; Comi et al., 2003; Connolly, 2001; Dalakas, 2002, 2004a, b; Dyck et al., 1982a, b, 1985b, 1986, 1994; Dyck and Kurtzke, 1985a; Good et al., 1998; Gorson et al., 1998, 2004; Hahn et al., 1996a, b; Hodgkinson et al., 1990; Hughes et al., 2003a; Kissel, 2003; Mahattanakul et al., 1996; Mendell et al., 2001; Meriggioli and Rowin, 2000; Molenaar et al., 1997; Mowzoon et al., 2001; Vallat et al., 2003a; van Doorn and Ruts, 2004; Eftimov et al., 2009; Hughes, 2008, 2009; Hughes et al., 2009; Kieseier et al., 2008; Magy and Vallat, 2009; Midroni et al., 1999; van Doorn, 2009; van Doorn et al., 2008; Vucic et al., 2009; Willison, 2005b; Willison et al., 2008]. However, there are few data comparing the relative efficacy of these treatments [Choudhary and Hughes, 1995a; Kieseier et al., 2008; Kuitwaard and van Doorn, 2009; Lopate et al., 2005; Lunn and Willison, 2009]. It remains unclear which treatment should be used first. The subject is one of on-going and enthusiastic debate [Sladky, 2008; Teasley, 2008]. General consensus is for beginning with either corticosteroids or intravenous immune globulin. If initial favorable response to either treatment among naive patients is probably greater than 50 percent, and once efficacy can be confirmed for either regimen and the patient firmly established on a trajectory of improvement, undesirable aspects of either treatment can be addressed.

In patients with suspected or definite CIDP, initial treatment with oral methlyprednisolone at a dose of 2 mg/kg/day for 6–8 weeks has been successful. Most previously reported untreated children with the disease demonstrate substantial functional improvement within that timeframe [Baba et al., 1996; Connolly, 2001; Korinthenberg, 1999; Nevo, 1998; Sladky, 1996; Markowitz et al., 2008; Rossignol et al., 2007]. Lack of significant clinical improvement in this setting suggests that:

Patients nonresponsive to corticosteroids can be tapered off medication over several weeks. In steroid-responsive individuals, it is desirable to reduce the dosage schedule gradually over 6 months to an alternate-day dose of about 1 mg/kg/day, or a dose of 30 mg on alternate days in children weighing more than 30 kg. A cumulative daily dose of 0.5 mg/kg is considered to pose little significant risk for opportunistic infection, osteopenia, growth retardation, or cataract formation. This regimen can be slowly reduced over the next 4–6 months. About half the children will remain in remission after treatment is discontinued. There are scientific and empiric rationales for pulse corticosteroid regimens in CIDP. These have included oral and parenteral protocols using methylprednisolone or dexamethasone given at weekly or monthly intervals [Molenaar et al., 1997; Lopate et al., 2005; Eftimov and van Schaik, 2008; van Schaik et al., 2010].

Although most children with CIDP initially respond to corticosteroids, some become refractory or develop unacceptable side effects. Human intravenous immune globulin or plasmapheresis should then be considered [Barnett et al., 1998; Gay and Katz, 1986; Hahn, 2000; Hahn et al., 1996a, b; Kiprov and Hofmann, 2003; Tanaka, 1998; Vedanarayanan et al., 1991; Walk et al., 2004; Brannagan, 2009; Hughes et al., 2006d, 2009; Kieseier et al., 2008; Kuitwaard and van Doorn, 2009]. Intravenous immune globulin has the advantage of permitting treatment at home using visiting nursing teams, if the child has exhibited no untoward reactions to an initial course of treatment in the hospital or clinic. In older children who are large enough for adequate peripheral venous access for plasmapheresis to be established, the procedure can be initiated on an outpatient basis. The choice between therapies depends more on logistical and technical factors than on the inherent superiority of one or the other modality. Some adolescents strongly resent the forfeiture of autonomy that frequent visits to the clinic or sessions with a home health-care provider symbolize. Under these circumstances, the option of self-administered subcutaneous immunoglobulin can be explored.

A small number of children become refractory to these treatments and require more potent immunosuppressive agents to maintain remission. There are no data to guide the clinician as to which immunosuppressant will be best for an individual patient. It is probably wisest to begin with an agent with which the treating physician has the most experience [Brannagan, 2009; Hughes, 2008, 2009; Hughes et al., 2006d; Kuitwaard and van Doorn, 2009; Lunn and Willison, 2009; Magy and Vallat, 2009]. Clinicians have been inclined to avoid cyclophosphamide and azathioprine in the past because of the potential risk of developing future malignancies. However, there are probably only small differences in the risk for this late complication among different chemotherapeutic agents. Methotrexate is not commonly helpful in CIDP, but it is worth consideration in refractory cases because it can be given orally on a weekly basis and is associated with few complications when serum levels are carefully monitored. Cyclosporine can be useful in treating selected refractory cases [Barnett et al., 1998; Jongen et al., 1988; Mahattanakul et al., 1996; Ryan et al., 2000]. This agent is given orally at an initial dose of 3–5 mg/kg/day. It is generally a safe drug, if renal function and serum levels are carefully followed.

Prognosis

The prognosis in children is generally favorable [Ryan et al., 2000; Simmons et al., 1997b; Sladky, 1996; Sladky et al., 1986]. About 50 percent respond well to corticosteroids and remain in remission after treatment is discontinued. Another 25–30 percent respond to additional treatment modalities and require a prolonged course of treatment. The long-term prognosis in this group of patients remains uncertain. A small number of children respond poorly to all treatment strategies or, more typically, become refractory to treatment or develop significant side effects to protracted immunosuppression and suffer substantial neurologic disability [Connolly, 2001; Nevo, 1998; Nevo et al., 1996; Pollard, 1994, 2002; Ryan et al., 2000; Simmons et al., 1997a, b; Markowitz et al., 2008; Rabie and Nevo, 2009; Rossignol et al., 2007].