The Platelet Function Analyzer (PFA-100®): a novel in-vitro system for evaluation of primary haemostasis in children


Rand Division of Haematology/Oncology, The Hospital for Sick Children, 555 University Avenue, Toronto, ON M5G 1X8, Canada.


The PFA-100® system provides an in-vitro method of assessing primary platelet-related haemostasis by measuring the time (the closure time, or CT) taken for a platelet plug to occlude a microscopic aperture cut into a membrane coated with collagen and either epinephrine or ADP. We used the system to establish normal ranges for CTs in healthy children, adults and neonates. Mean CTs of healthy children were independent of the needle gauge used (21G or 23G) for blood sampling; they were very similar to the mean CTs of healthy adults, but longer than mean CTs of healthy neonates. Although children with haemophilia had normal CTs, the PFA-100 system was found to be potentially useful in screening for von Willebrand disease in children.

The evaluation of primary, platelet-related, haemostasis in children is difficult. The skin bleeding time (BT) is the most frequently used measure of global platelet function ( Lind, 1991). However, the test is highly operator-dependent, poorly reproducible, and a poor predictor of bleeding risk ( Rodgers & Levin, 1990; Lind, 1991; De Caterina et al, 1994 ). In von Willebrand disease (VWD), the BT has been shown to perform poorly as both a diagnostic and prognostic test ( Lind, 1991; Murray & Lillicrap, 1996). Furthermore, repeated BTs following therapeutic interventions are not possible in the paediatric setting.

The Platelet Function Analyzer (PFA-100®; Dade International Inc., Miami, Fla., U.S.A.) has been developed as a quantitative, simple, rapid in-vitro method of assessing primary haemostasis ( Kundu et al, 1994 , 1995, 1996b; Mammen et al, 1995 ). Whole blood samples anticoagulated with citrate are aspirated at a shear rate of 5000–6000/s through a 150 μm aperture cut into a collagen-coated membrane impregnated with either epinephrine (Col/Epi) or ADP (Col/ADP) ( Kundu et al, 1995, 1996b). Platelets in the sample undergo adhesion and aggregation in response to shear stresses and agonists in the membrane, forming a platelet plug that eventually occludes the aperture ( Kundu et al, 1996b ). The time taken to occlude the aperture is the closure time (CT), a measure of overall platelet-related haemostasis.

The platelet defect arising from ingestion of aspirin (i.e. inhibition of thromboxane A2 formation) is associated with prolonged CTs measured with the Col/Epi cartridge, but not with the Col/ADP cartridge. A prolonged CT with both cartridge types is considered to indicate a disorder with greater severity ( Kundu et al, 1995 ).

The PFA-100 system has been used to evaluate platelet-related haemostasis in adults ( Kundu et al, 1994 , 1996a; Mammen et al, 1995 ); investigations in children have not previously been performed. In our study, normal reference ranges of CTs for healthy children were established and compared with CTs for healthy adults and neonates, and for patients with haemophilia or VWD.


Controls and patients. The healthy children (n = 57; selected by screening children referred for blood testing to the Outpatient Phlebotomy Clinic at The Hospital for Sick Children, Toronto, Canada) and adults (n = 31; volunteers) had no history of systemic illness and had not taken medications known to affect platelet function for at least 10 d before blood sampling. Their platelet counts were > 150 × 109/l. Haemoglobin levels in the children were at least 11.5 g/dl, and in adults >12.0 g/dl.

Healthy neonates (n = 17) were full term (gestational age >37 weeks), had Apgar scores of ≥7 at 1 min and ≥9 at 5 min, and were without evidence of congenital or acquired illnesses, e.g. sepsis.

Patients with haemophilia or VWD, followed in The Hospital for Sick Children Bleeding Disorders Clinic, were also evaluated. 11 male haemophilic patients comprised a broad cross-section of cases of haemophilia A (five mild (6–30% factor VIII), one moderate (1–5%) and two severe (<1%)) and haemophilia B (two moderate (1–5% factor IX) and one severe (<1%)). None of the patients had received factor concentrate in the 6 months prior to blood sampling. Of the VWD patients, five children and two affected parents had type 1 VWD, defined as levels of von Willebrand factor (VWF) and/or ristocetin cofactor (RCoF) < 0.4 U/ml and a positive family or personal bleeding history. An additional affected parent had type 3 VWD, with levels of VWF, RCoF and factor VIIIc of < 0.05 U/ml.

Informed consent was obtained for all subjects and the study was approved by the Hospital Ethics Review Committee.

Methods. Blood was collected by syringe from an antecubital vein with either a 23 gauge (G) needle (30 healthy children), or a 21G needle (all other children and adults), and was anticoagulated with 10.5 m M (final concentration) buffered sodium citrate (Becton Dickinson, Rutherford, N.J., U.S.A.). The same anticoagulant was used for the neonatal blood samples, with the blood being collected from the umbilical vein within 10 min of delivery. CTs were determined on duplicate samples (0.8 ml) within 3 h of collection, using cartridges containing Col/Epi or Col/ADP membranes. The mean coefficient of variation for duplicate samples was 5% ± 3 (n = 124) for both cartridge types. For the VWD patients, diagnostic tests were performed on blood samples taken at the same time as the samples for PFA testing. In a subset of VWD patients, BTs (SurgicuttTM or Surgicutt-Jr.TM, International Technidyne Company, Edison, N.J., U.S.A.) were obtained ( Andrew et al, 1992 ).

Statistical analysis. Values reported are means ±1 SD. Data were analysed using two-sample t-tests or analysis of variance (ANOVA) and Tukey's pairwise comparisons. P < 0.05 was considered to indicate statistically significant differences.


PFA-100 closure times for healthy children, adults and neonates, and children with haemophilia, are detailed in Table I. For healthy children, mean CTs obtained for samples taken with 21G v 23G needles were not significantly different for either cartridge type. Thus, the data were combined ( I. PFA-1Table I, Fig 11), establishing normal reference ranges (5th to 95th percentile). Mean CTs for healthy children (combined data) were slightly longer than for adults ( Table I), but the differences were not statistically significant for either cartridge type. Children had a slightly lower mean haemoglobin level (P < 0.01) than adults ( I. PFA-1 Table I); this difference was statistically, but not clinically, significant.

Table I. PFA-1.  00® closure times in healthy children and adults, and in children with haemophilia. Needle gauges are those used for blood sampling. Data are means ± 1 SD.* Ranges in parentheses (minimum and maximum values).† Normal ranges in parentheses (5th to 95th percentile).‡ Significantly different compared with children (combined data), P < 0.01. Thumbnail image of
Figure 1.

Fig 1. Scattergraph of closure times obtained with cartridges containing membranes coated with (A) collagen/epinephrine or (B) collagen/ADP, in healthy children, adults and neonates, and in patients with haemophilia or von Willebrand disease. Shaded areas indicate normal ranges (5th to 95th percentile) for healthy children, n = 57. NC, non-closure (e.g. closure time > 300 s).

Mean CTs for healthy neonates were significantly shorter than those for healthy children (P < 0.01) ( I. PFA-1Table I, Fig 1) with the majority of CTs falling below the normal range established for children. Neonates had a higher mean haemoglobin level (P < 0.01) than children, but even after adjusting for differences in haemoglobin levels, mean CTs remained significantly different (P < 0.01, analysis of covariance).

For children with haemophilia, mean CTs, haemoglobin levels, and platelet counts were not significantly different compared with healthy children ( Table I). Closure times for all but one patient were within the normal range established for children; CTs for this patient were shorter (Fig 1).

Closure times for the patients with VWD (n = 8) are shown in Fig 1. Six of the seven patients with type 1 VWD had prolonged CTs or showed non-closure (e.g. CT >300 s) with both cartridge types; BTs were measured in two of these six patients and were normal (4 and 7 min). The seventh patient with type 1 VWD had normal CTs with both cartridge types (BT not done). The patient with type 3 VWD showed non-closure with both cartridge types (BT not done).


The PFA-100 has been developed as a simple, inexpensive and reproducible system for the assessment of primary haemostasis. In establishing normal ranges of CTs in healthy children, we found that mean CTs in healthy children and adults were similar. Blood samples from children are commonly drawn via smaller-gauge needles than from adults; we found that needle gauge (23G v 21G) did not affect CTs. Therefore blood can be collected with either needle gauge provided that flow through the needle is uninterrupted and that the sample is collected with ease.

Healthy neonates had significantly shorter CTs than healthy children. Although CTs vary inversely with haemoglobin levels ( Kundu et al, 1996b ), the observed differences were still apparent after adjustment for the increased haemoglobin levels in the cord blood samples. Since VWF has been shown to be important in the formation of the platelet plug at the membrane apertures of the PFA-100 cartridges ( Kundu et al, 1996b ), the decreased CTs observed in neonates may be due to their elevated levels of plasma VWF ( Blanchette & Rand, 1997).

Most haemophiliacs do not have prolonged BTs ( Bithell, 1993); CTs of the haemophiliacs in this study were not prolonged, indicating that factors VIII and IX are not involved in platelet plug formation in the PFA-100 system.

In contrast with haemophilic patients, all but one patient with type 1 VWD had prolonged CTs, as did the patient with type 3 VWD, probably due to abnormalities/deficiencies in VWF.

Therefore our study indicates that the PFA-100 may have potential as a screening tool for VWD in children and adults. Further studies are warranted to determine its usefulness in the monitoring of therapeutic interventions in paediatric VWD patients as well as in the evaluation of primary haemostasis in children with acquired or congenital platelet disorders.


We are grateful to Dade International Inc. (Miami, Fla., U.S.A.) for providing the PFA-100 instrument and test cartridges, to International Technidyne Company for providing the Surgicutt and Surgicutt-Jr. bleeding time devices, and to Ms L. McWhirter, Division of Perinatology, Mount Sinai Hospital, for help in obtaining cord blood samples. This paper was prepared with the assistance of Editorial Services, The Hospital for Sick Children, Toronto, Ontario, Canada.