Special diagnostic considerations in older children
Children over the age of about 10 years may be more likely to have a chronic course (Walker & Walker, 1984; Reid, 1995). Other autoimmune diseases associated with thrombocytopenia should be considered, particularly systemic lupus erythematosus (SLE) and antiphospholipid syndrome. Older children should have additional investigations performed (see below). Thrombocytopenia has not commonly been reported as the initial presentation of HIV infection in children, but can occur during the course of the disease.
Special diagnostic considerations in the Accident & Emergency (A & E) Department
The possibility of non-accidental injury (NAI) and meningococcal disease must be considered by A & E staff when dealing with a young child presenting with bruising and purpura for the first time. Children with infection usually have other features and non-accidental injury does not present with generalized purpura.
Investigations
Full blood count and careful film examination by an experienced morphologist. If a low platelet count does not fit the clinical picture the count should be repeated to exclude a spurious result.
Coagulation screening. Only necessary if there is a possibility of meningococcal infection, features to suggest an inherited bleeding disorder in addition, or a suspicion of NAI.
Antiplatelet antibodies. Measurement of antiplatelet antibodies does not assist in the diagnosis (Taub et al, 1995; George et al, 1998).
H. pylori infection. This is discussed above. There is no information about this infection as a precipitating cause of ITP in children, although H. pylori infection does occur in childhood. There is therefore no indication to screen children with ITP unless there are other clinical indicators to suggest infection (Sherman & Macarthur, 2001).
Bone marrow aspiration. Bone marrow examination excludes some causes of thrombocytopenia but can only confirm that the picture is consistent with peripheral destruction. Although it is argued that if a child has a sinister underlying disorder there will be other clues in the blood picture or the clinical examination (Halperin & Doyle, 1988; Calpin et al, 1998) this is not always reliable (Reid, 1992).
Recommendation for investigation of suspected childhood ITP:
In a child with typical clinical and laboratory features who needs no treatment, a bone marrow examination is not required (Grade B recommendation). If therapy is considered, it must be recognised that occasionally the diagnosis of ITP will be incorrect. Bone marrow examination will usually reveal this. Marrow examination is necessary in the presence of atypical clinical features or if there is no response to treatment. It is also recommended that the bone marrow be examined before steroid therapy is given (Evidence Level IV). Marrow aspiration should ideally be accompanied by a trephine biopsy because this gives a better estimate of megakaryopoiesis (Beardsley & Nathan, 1998). Marrow examination is painful and is more acceptable when performed under general anaesthesia (Evidence levels: IIb, III).
Management of ITP in childhood: general measures
It is essential that children are classified clinically and not by platelet count, because children with severe thrombocytopenia (less than 10×109/ l) have usually "mild" clinical symptoms. Pronounced skin purpura and bruising, however, extensive, do not indicate a serious bleeding risk on their own. To date there has been a universal tendency to treat the count alone rather than the child’s symptoms. Two UK national surveys of children with ITP have demonstrated that only 4% of children with ITP have serious symptoms such as severe epistaxis or GI bleeding (Bolton-Maggs & Moon, 1997; Bolton-Maggs & Moon, 2001). Similar data are available from Germany (Sutor et al, 2001) and Norway (Zeller et al, 2000). Several studies have con?rmed that the incidence of intracranial haemorrhage (ICH) is much less than the 1–3% widely quoted, and is closer to 0.1–0.5% (Lilleyman, 1994). In the two UK national surveys there were two ICH in 703 cases (0.3%), with complete recovery in both cases, similar to the incidence derived from a retrospective national survey (Lilleyman, 1994, 1997). Similarly, a recent Japanese study found an incidence of 4 in 772 (0.52%) children, with no deaths (Iyori et al, 2000). It is impossible to predict which children will develop an ICH, and it may be that there are other predisposing factors that contribute to this for example an underlying vascular anomaly.
ICH has occurred in children who have been treated and in one study this was the case in a third of children reviewed (Lee & Kim, 1998). It is worth noting that in the ?rst randomized study of IVIg the only child to have ICH did so early, was on IVIg and died 6 d after presentation (Imbach et al, 1985). As discussed earlier, the need for treatment in children with acute ITP should not be driven by the platelet count alone, but mainly by symptoms. Children who continue to be severely thrombocytopenic with signi?cant bleeding symptoms are very rare indeed, and should be referred to a specialist centre for management (Lilleyman, 1999).
Clinical classifications
Two clinical scoring systems have been used, the first in the two UK national surveys for analysis of 703 newly diagnosed children (Bolton-Maggs & Moon, 1997, 2001), and the second was reported from a single centre in Texas where it has been analysed on 109 occasions in 54 patients, both at diagnosis and in established ITP (Buchanan & Adix, 2001). Both of these confirm that the majority of children do not have serious bleeding problems despite very low platelet counts. The severity of bleeding at any given time, especially at presentation does not predict the risk of subsequent episodes of serious bleeding. Children should not be treated on the basis of cutaneous signs alone, however, dramatic and widespread the purpura.
Recommendation:
Clinical severity, in addition to platelet count, should be used to de?ne the severity of acute ITP in children. Treatment should be considered on the basis of other clinical symptoms in addition to cutaneous signs, and not the platelet count alone (Level IIb evidence, Grade B recommendation).
Treatment: Expectant 'Watch and Wait' policy
More than 80% of children with acute ITP do not have signi?cant bleeding symptoms and can be managed without specific therapy directed at raising the platelet count (Buchanan et al, 1997; Baronci et al, 1998; Dickerhoff & von Ruecker, 2000; Bolton-Maggs et al, 2001; Bolton-Maggs & Moon, 1997, 2001; Sutor et al, 2001) (Evidence levels III and IV). It is essential that the parents, and child where able, have an explanation that this is usually a self-limiting benign disorder (Dickerhoff & von Ruecker, 2000). Most children can be managed well at home, and do not require hospital admission, which should be reserved for children with clinically important bleeding (severe epistaxis, i.e. lasting more than 30 min with heavy bleeding, GI bleeding, etc.). Parents should be advised to watch for other signs of bleeding and be given a contact name and 24 h telephone number; similar advice as that given to families with a child with severe haemophilia is often the most appropriate, with, as far as possible, avoidance of formal contact sports or activities with high risk of trauma or head injury. Other activities can be continued as normal, and the child should be encouraged to continue schooling on the basis that ITP is a disorder that may last some weeks or months (Evidence levels: III and IV, Grade B and C recommendation).
Most children with mild or moderate symptoms only (the majority) can be safely managed as outpatients with weekly or less frequent visits (level III evidence) (Dickerhoff & von Ruecker, 2000; Bolton-Maggs et al, 2001). A repeat platelet count is not necessary while a child still has purpura as this symptom is an indication that the count is likely to be less than 20×109/ l; however, a repeat count should be performed within the first 7–10 d to check that there is no evidence of a serious marrow disorder emerging, particularly aplasia. Otherwise the count can be performed when clinically indicated by a change of symptoms, or if the family feel more at ease knowing the result. However, too close a focus on numbers is not helpful. When the thrombocytopenia persists but the child remains well, the intervals between visits can be stretched out in order to minimise interference with schooling. With increasing duration of a very low platelet count and limitation of some sporting activities, lifestyle issues become relatively more important and should be sympathetically discussed. These issues may in?uence decisions about treatment. ‘Most parents and patients can live quite comfortably with petechiae and low platelets awaiting spontaneous remission providing their physician can’ (Dickerhoff, 1994).
Recommendation:
Children with acute ITP and mild clinical disease may be managed expectantly with supportive advice and a 24 h contact point, irrespective of platelet count (Grade B recommendation). The full blood count should be repeated within 10 d of diagnosis to ensure there is no evidence of evolution to a serious marrow disorder; thereafter the count need not be monitored until resolution of clinical symptoms suggests the onset of remission or there are other clinical indicators suggesting the need.
Specific treatment to raise the platelet count
Several therapies raise the count faster than no treatment (Ib evidence). However, all have significant side effects and none alters the underlying pathology nor increases the chance of complete remission. These strategies are appropriate for children with severe bleeding symptoms.
Corticosteroids
Prednisolone. Conventional doses of 1–2 mg /kg/ d for a maximum of 14 d may be effective in raising the count but perhaps is little faster than no therapy (Buchanan & Holtkamp, 1984) (Level 1b evidence). Prednisolone should be discontinued after a maximum of 2–3 weeks irrespective of platelet count because of the serious side effects associated with prolonged treatment. There is evidence that smaller doses (0.25 mg / kg / d for 21 d) may produce a rise in count (Bellucci et al, 1988). A high dose regimen (prednisolone 4 mg / kg / d) has been compared in two randomized controlled trials and been shown to be more effective (Level Ib evidence). Fifty three children were randomized between IVIg, high dose (4 mg /kg / d) oral prednisolone with tapering and cessation by 21 d and no therapy in one study (Blanchette et al, 1993). The median time to reach a count of > 50×109/ l was 4 d in the prednisolone group compared with 16 d in the 'no treatment group, and 2 d in those given IVIg. The same regimen was used in a later study of 146 children which included a treatment arm with anti-D (Blanchette et al, 1994). In the latter study 28 / 39 (72%) of children in the high dose steroid arm achieved a count > 50×109/ l within 72 h.
The disadvantage of this steroid regimen is the duration at high dose. A short course may be suf?cient. A pilot study of 25 children given 4 mg / kg for 4 d (Carcao et al, 1998) demonstrated that 22 children achieved a count of > 20×109/ l within a week with minimal side effects (Level III evidence). If a child requires treatment for symptoms, an early effect on count is reassuring. It is essential that steroids are not continued for longer than 2 to 3 weeks or titrated against the platelet count as this may lead to inappropriate doses and duration, with very signi?cant long-term side effects.
Other steroid regimens
High dose methyl prednisolone (HDMP). This has been used as an alternative to IVIg because it is cheaper and effective. A small study of 22 children given an oral 7 d course (30 mg / kg/ d for 3 d followed by 20 mg / kg / d for 4 d) demonstrated that all patients achieved platelet counts > 50×109/ l by d 7. Courses were repeated monthly if the count was less than 20×109/ l on d 30, for up to six courses (Ozer et al, 2000). Other small studies have demonstrated HDMP to be as effective as IVIg in raising the platelet count (Ozsoylu et al, 1989, 1993). Two studies demonstrated the ef?ciency of a combined approach using HDMP with IVIg but these were small uncontrolled studies (Barrios et al, 1993; Gereige & Barrios, 2000).
Pulsed high dose dexamethasone. This treatment appears to be less effective in children than in adults in producing long-term remission, but may be useful as a temporary measure. There is no evidence that this treatment, or any of the treatments that can produce an increase in the platelet count, increases the chance of complete remission. Three small studies of 11 (Kuhne et al, 1997), 17 (Borgna-Pignatti et al, 1997) and seven children (Chen et al, 1997) demonstrated some benefit in some children – 78% children achieved a platelet count > 100×109/ l within 72 h in 41 cycles of treatment given to 11 children with chronic ITP (Kuhne et al, 1997). However, side effects were unacceptable in 3 / 11 children. This treatment cannot be recommended as ?rst line therapy in symptomatic children.
