Hospitalization of children with acute immune thrombocytopenic purpura – is it necessary?

Abstract

OBJECTIVE:

To identify a target group of children with acute immune thrombocytopenic purpura (ITP) that may not require hospitalization for management.

METHODS:

A retrospective chart review was conducted of all children admitted over a two-year period to a tertiary care paediatric hospital with the diagnosis of acute ITP. Patients were classified according to typical and atypical presentations. Typical patients were defined as those aged between one and 10 years, with no hepatomegaly or significant splenomegaly and who had typical laboratory features for ITP. Patients who did not meet these criteria were categorized as atypical. Outcome measures included length of stay (LOS) in hospital; frequency of bone marrow aspiration (BMA); type of treatment; incidence of intracranial hemorrhage (ICH) or severe bleeding; and admission and discharge platelet counts.

RESULTS:

There were 74 patients hospitalized for a mean of 3.6 days. No patients suffered an ICH or bleeding requiring transfusion. Patients with typical presentations (42) were compared with patients with atypical presentations (32) and were not significantly different for clinically important outcomes such as admission and discharge platelet counts, serious complications or type of therapy. Typical patients had significantly fewer BMAs than did atypical patients – 22 of 42 (52%) versus 25 of 32 (78%) (P=0.02), and a shorter LOS – 3.1 (±0.9) days versus 4.2 (±1.8) days (P=0.01).

CONCLUSIONS:

Children presenting with ITP have a low incidence of bleeding complications and many of these patients can be managed as outpatients. A multicentre study is needed to properly delineate a low risk group suited for outpatient medical management.

Immune thrombocytopenic purpura (ITP) is the most common cause of isolated thrombocytopenia in children, with an estimated incidence of four/100, 000 children (1). The typical presentation of this disorder is one in which a previously healthy child between the ages of one and 10 years develops petechiae, purpura and, occasionally, mucous membrane bleeding a few weeks following an upper respiratory tract infection (URTI) or other infection. Physical examination is normal other than petechiae and bruising. Organomegaly and significant adenopathy are usually absent. A peripheral blood smear is normal other than demonstrating thrombocytopenia (2–5).

There is controversy about the optimal way to manage children with typical acute ITP because the disorder is usually self-limited and generally resolves spontaneously, irrespective of the type of therapy (2,4,6–8). Concern regarding the risk of intracranial hemorrhage (ICH) has led to the hospitalization of children with ITP for observation. However, the occurrence of ICH (incidence estimate of less than 1%) in acute ITP is rare (1,9–11). Furthermore, no previous studies have been able to demonstrate that treatment reduces the risk of ICH or that ICH can be better prevented in hospitalized children (1,12–14). Hospitalization has been suggested for children with ITP who have severe life-threatening bleeding or platelet counts of less than 20×10⁹/L and mucous membrane bleeding; but the value of hospitalization for patients presenting with platelet counts of less than 20×10⁹/L, petechiae and purpura, who are otherwise asymptomatic, is uncertain (5,10). Children meeting the definition of typical acute ITP often satisfy the characteristics of this latter group of patients.

Traditionally, reasons given for hospitalizing children with acute ITP, other than enhancing safety, include facilitating a bone marrow aspiration (BMA) and administering therapy. The purpose of the present study was to review the clinical course of all children with acute ITP admitted over a two-year period to The Hospital for Sick Children, Toronto, Ontario, with respect to presentation, investigations, therapy, incidence of complications, and length of stay (LOS) to identify a target group of patients that may not require hospitalization for management.

METHODS

This study involved a retrospective chart review of all children hospitalized between January 1995 and January 1997 in a tertiary health care centre with the first episode of acute ITP. Patients were identified through health records in which patients with admission or discharge diagnoses coded with the words ITP, or immune thrombocytopenic purpura, were eligible; furthermore, all children with bone marrow samples taken for ITP, rule out ITP, or isolated thrombocytopenia during the study period were also eligible. Patients who were initially diagnosed with ITP, but were later found to have conditions other than ITP, were excluded from the study. The health records were reviewed and data were abstracted and recorded on standardized collection forms. Data included age, sex, history of a recent URTI, initial physical examination, admission complete blood count, performance and results of BMA, method of treatment, incidence of ICH or bleeding requiring transfusion, platelet count on discharge, and LOS in hospital. For consistency, duration of admission was based on the dates of admission and discharge, regardless of the time of admission or discharge.

Patients were classified as typical and atypical presentations. Patients with typical presentations included those aged between one and 10 years who had no hepatomegaly or significant splenomegaly as recorded in the note of the most responsible physician, and typical laboratory features for ITP based on previous studies done at the authors’ institution and the laboratory reference ranges at The Hospital for Sick Children (15). Typical features include a platelet count of less than 50×10⁹/L and a hemoglobin level of greater than 100 g/L (six to 12 months of age) or greater than 110 g/L (older than one year of age), a total white blood cell count of greater than 5×10⁹/L (six months to six years of age) or greater than 4×10⁹/L (older than six years of age) and neutrophil count of greater than 1.5×10⁹/L (six months to six years of age) or greater than 2×10⁹/L (older than six years of age). Children whose initial presentation did not meet the above criteria were categorized as atypical ITP. Mucous membrane bleeding was not considered to be a factor that distinguished typical ITP from atypical ITP.

Differences between typical and atypical groups were tested using t tests for continuous variables (eg, age), χ²test for categorical variables (eg, presence or absence of BMA), and Mann-Whitney two-sample statistic for count data (eg, LOS). The STATA statistical software (STATA Corporation, USA) package was used for all calculations.

RESULTS

There were 74 children admitted with a primary diagnosis of acute ITP. Patient characteristics, clinical presentation, admission hemoglobin, admission platelets and discharge platelets are listed in Table 1. No patient had ICH or severe bleeding requiring transfusion.

TABLE 1.

Patient characteristics of children with a primary diagnosis of acute immune thrombocytopenic purpura

Total sample (n=74)
Number of male patients	38 (51%)
Prevalence of mucocutaneous bleeding	23 (31%)
Prevalence of upper respiratory tract infection	42 (57%)
Hemoglobin level at admission	119.8 (SD±15.28)
Platelet level at admission	7.3 (SD±6.35)
Platelet level at discharge	51.7 (SD±36.29)

Prednisone (4 mg/kg/day for four days) was the most common form of treatment administered in 37 (50%) patients. Intravenous immune globulin (IVIG) (1 g/kg/dose for one or two doses separated by 24 h) was used in 23 (31%) patients. Combined therapy was used in five (7%) patients. In one typical case, treatment was initially started with prednisone, but changed to IVIG because of poor response. In a second case, treatment with prednisone was started but changed to oral dexamethasone because of poor tolerance to the bitter taste of oral prednisone. Two atypical cases did not respond to prednisone and were changed to IVIG and one patient was switched from IVIG to prednisone. Nine cases (12%) received no therapy other than observation in hospital, and all those patients did well.

There were 42 patients with typical presentations and 32 with atypical presentations. The only differences found between typical and atypical patients were for age, frequency of BMA and LOS. Typical patients were significantly younger (4.3±2.3 years) than atypical patients (7.4±6.3 years) (P=0.004). Bone marrow aspiration was performed in 22 of 42 (52%) patients with typical ITP compared with 25 of 32 (78%) patients with atypical ITP (P=0.02) and confirmed the diagnosis of ITP in all patients. The LOS was significantly shorter in the group presenting with typical ITP (Figure 1) with a mean of 3.1±0.9 (median 3.0, range one to five) days, compared with the group presenting with atypical ITP with a mean of 4.2±1.8 (median three, range two to nine) days (P=0.01).

Figure 1).

Length of hospital stay in typical and atypical presentations of acute immune thrombocytopenic purpura (ITP). Boxplot comparing the median (bold line in each box) and the range of length of hospital stay in typical and atypical presentations of acute ITP. The upper and lower boundaries of each box represent the 25th and 75th percentiles, respectively. Length of stay was significantly shorter in the typical group (P=0.01)

DISCUSSION

In the present study, we looked at typical versus atypical presentations of ITP as a possible group (typical) to target for outpatient management, but were unable to find a significant difference in bleeding between the two groups due to the low incidence of bleeding complications.

In current times of fiscal restraint and decreasing numbers of hospital beds, paediatric institutions are becoming increasingly discriminating with regard to which children require admission and which ones can be managed on an outpatient basis. In North America, the diagnosis of acute ITP has traditionally resulted in hospitalization for BMA, therapeutic intervention, and in-hospital observation for potential complications. In this study, we focused on a group of patients presenting with features of typical ITP (5). Though descriptions outlining the characteristics that define typical ITP appear in the literature (2–5) the incidence of this typical presentation has not been established. In the present study, patients with typical features of acute ITP represented more than half of the cohort of children with acute ITP during the two-year study period. This group may serve as an ideal prospective target group for outpatient medical management.

Performance of a BMA is often facilitated through admission to hospital. However, recent retrospective studies at our institution have seriously questioned the need for BMA in acute ITP. Halperin and Doyle (16) reviewed 127 patients presenting with ITP over a 3.5-year period. BMA confirmed ITP in all but three cases (one case was diagnosed as Wiscott-Aldrich syndrome and two cases were diagnosed as amegakaryrocytosis) and no diagnosis of leukemia was made. Calpin et al (15) examined 324 patients over a 12.5-year period who presented with typical ITP. With the exception of one case of aplastic anemia, ITP was confirmed in all cases. Our findings concur with these studies. Forty-seven patients underwent BMA in this study population and no patient was found to have leukemia or other bone marrow disorder.

The most feared complication of ITP is ICH. The incidence of ICH is extremely rare (estimated at less than 1%) (1,9,10). In our study, no patient had an ICH or bleeding episode severe enough to require transfusion. However, our sample size was insufficiently powered to generate a reliable estimate of the incidence of ICH. This would require a cohort of several thousand patients with acute ITP (9).

The treatment of acute ITP in children has been extremely controversial. The natural history of ITP is generally benign with self-resolution regardless of therapy (1,2,4,6–8). However, the fear of ICH occurring in patients with acute ITP has led to a tendency to treat children with low platelet counts. The modalities of treatment in our study reflect the current trends in use by paediatricians. These include prescribing oral corticosteroids (4 mg/kg/day for four days) (17) or IVIG (1 mg/kg) for a platelet count of less than 20×10⁹/L. Overall, nine children received no therapy (strictly observed) and five children received combined therapy. There was no significant difference in the choice of therapies between the typical and atypical groups, with oral prednisone used more commonly than IVIG in both groups. The choice of therapy may have been influenced by the presence of minor mucous membrane bleeding, the comfort level of the attending paediatrician with a particular approach, or the preference of the consulting hematologist. All patients improved regardless of the management strategy used. Platelet counts at admission and discharge were examined and not found to be significantly different between typical and atypical groups.

The focus of this study was the need for hospitalization in acute ITP. Routine admission for BMA or observation for potential ICH appears to be unnecessary and is not recommended by either the British Guidelines for Management of ITP (10) or the American Society of Hematology practice guideline (5). Hospitalization for the provision of treatment may also be unnecessary. The distribution in our study likely reflects the current management practices of practitioners. Oral prednisone is still used most commonly as first line therapy and can be administered on an outpatient basis. IVIG may be used less frequently because of a lack of conclusive evidence that it is clinically superior to steroids; the theoretical risk of viral transmission; side effects such as fever, headache and vomiting; recent supply shortages; and high cost (18). Treatment requires 4 to 8 h infusions, sometimes for two doses. The practicality of administering IVIG on an outpatient basis depends on community resources and availability. Nonetheless, hospitalization may be required for those children deemed to be at high risk of ICH because of a very low platelet count (less than 5×10⁹/L), significant mucosal bleeding, or for families where geographical location and community resources may hinder obtaining urgent medical care.

Our study of hospitalized children with acute ITP demonstrated that children with typical presentations had a significantly shorter LOS than patients with atypical presentations. This trend may be accounted for by the increased number of investigations performed on the atypical group, such as BMA, and their threshold level for discharge platelet count. The short hospital stay and rapid increase in platelet count to greater than 20×10⁹/L in the typical group may serve as an indicator that these children have a benign course and recover particularly quickly regardless of therapy and, thus, may be a good target group for outpatient management.

CONCLUSIONS

Children presenting with acute ITP have a low incidence of bleeding complications and many of these patients could be managed as outpatients. A multicentre study is needed to properly delineate a low risk group suited for outpatient medical management.