Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Partial plasma exchange transfusion in polycythaemic neonates

Three Part Question

In [an asymptomatic neonate with an incidental finding of raised venous haematocrit] does [treatment with a partial plasma exchange transfusion] [reduce adverse neurological outcome]?

Clinical Scenario

You are a neonatal junior doctor looking after the special care nursery. You process a capillary blood sample taken on the morning blood round by someone who has now gone home, and find the haematocrit to be high at 69%. You go back to the baby and find the child to be term, of low birthweight, and admitted to the special care nursery because of low Apgars the previous day. He has not been feeding too well, but the neonatal nurses are not otherwise concerned. He is normal on examination, and his venous haematocrit comes back at 68%. Does this baby need a partial plasma exchange transfusion (PPET)?

Search Strategy

Cochrane Controlled Trials Register and Medline.
CCTR: (polycythaemia)
Medline: ({polycythemia/} AND {infant, newborn/} AND RCT filter

Search Outcome

6 papers identified from Medline, of which one irrelevant and one of insufficient quality.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Malan AF et al,
South Africa
49 neonates selected by: appearance of polycythaemia, central haematocrit >=65%, and no symptoms, or mildly symptomatic Randomised to PPET or symptomatic careRCT (1b)Developmental scoreNo difference8 months follow up 86% follow-up, one death (NEC) in treatment group
Neurological examinationNo difference (normal in all)
Goldberg K et al,
20 neonates, selected by: screening heel prick on all babies, result >=68%, and hyperviscous on venous blood, excludes neurologically symptomatic Randomised to PPET or symptomatic careRCT (1b)Bayley psychomotor development index (control score 112)8 months 80%
Bayley mental development index Mean difference 2.6 (95% CI -16 to 21) (control score 112)
Abnormal postural reflexes1/6 of treatment group, 5/10 observed group, NNT = 3(95% CI -10 to 1). 6/10 in treatment group, 0/6 in observed group, NNT = 2 (95% CI 1 to 3)
Abnormal neurological findingsMean difference -0.8 (95% CI -13 to 15) (control score 108)
Black VD et al,
93 neonates selected by: admission to nursery at 4-6 hours of age, screened by heel prick testing, and found to have both venous polycythaemia (HCT >=65%) and hyperviscosity, includes symptomatic and asymptomatic. Randomised to PPET or symptomatic care.RCT (1b)Neurological diagnosis rateReduced by PPET (25% vs 55%) p<0.052 years, follow up 65% follow up rate includes 2 deaths in observed group, one of head trauma and one of hepatitis
Motor delay rateReduced by PPET (19% vs 35%),
Mental delay rateReduced by PPET, (18% vs 13%)
Bada HS et al,
28 neonates screening cord blood haematocrit, and found to have both arterial haematocrit >=63% with hyperviscosity, and asymptomatic. Randomised to PPET or symptomatic care.RCT (1b)Mental development index/IQMean score in treated group 85 vs observed 88; difference -3 (95% CI -19 to +13) Mean 27.5 months follow up, assessors blind to neonatal course 71% follow up rate
Mental retardation rateHigher borderline retardation rate in treated group (66% vs 45%; ARR = 0.21 (95% CI -0.21 to 0.63)


The research papers all use hyperviscosity rather than polycythaemia as the basis for their studies. The test for this is not routinely available, and so in the clinical setting, polycythaemia is used as a marker for hyperviscosity. Hyperviscosity is the increased internal friction of blood. It differs according to the flow rate, and it is in the capillaries, where the flow rate is low, that the hyperviscosity is thought to contribute to reduced tissue perfusion. Red cell concentration (i.e. haematocrit) contributes significantly to viscosity; Bada et al in the study above found a moderate correlation between the two (r=0.5, p<0.001). Neonatal polycythaemia is associated with adverse neurological outcome. All of the above papers included a control group of non-polycythaemic or non hyperviscous infants. The group of infants in the Black et al study showed a significant overall risk of mild developmental delay, and this risk is greater than the modest treatment effect, compared with infants controlled for gestational age, birth weight, but not for perinatal risk factors. Whereas, Bada et al matched the control group for gestational age, birth weight, and perinatal risk factors of fetal distress or asphyxia, pre-eclampsia, maternal diabetes. They performed regression analysis to determine the effects of these risk factors on outcome. They found that the risk of adverse neurological outcome was more strongly predicted by perinatal risk factors than the presence of hyperviscosity. It is notable that there was no abnormal neurology in the infants studied at 8 months follow up by van der Elst, which may be due to the type of patients included, the small numbers of patients, or the relatively short follow up period. The subjects of the Black study included some of who were neurologically normal at one year, but had developed signs by 2 years. However, the relatively low follow up rate (65% for those randomised, 51% for control infants) makes the findings tentative at best. Similarly, although the Goldberg et al study has 80% follow up of the 20 patients; only 6/10 of the observation group were studied, in contrast to 10/10 of the treatment group. It is important to assess whether these patients are comparable to the clinical scenario. Only the van der Elst patients were selected on clinical grounds ("appearing polycythaemic"), the other patients being selected by a screening process. The other factor that makes the van der Elst group of patients comparable in particular to this scenario is that the infants included those who had "minor signs", defined as lethargy, irritability, peripheral cyanosis, vomiting and poor feeding.

Clinical Bottom Line

Polycythaemia is a risk factor for later neurological abnormality. Partial plasma exchange transfusion has no clear long-term benefit in asymptomatic polycythaemia or hyperviscosity, though evidence either way is weak and inconsistent.


  1. Malan A, de V Heese H. The management of polycythaemia in the newborn infant. Early Hum Dev 1980:4(4);393-403.
  2. Goldberg K, Wirth FH, Hathaway WE et al. Neonatal hyperviscosity II Effect of partial plasma exchange transfusion. Pediatrics 1982:69(4);419-25.
  3. Black VD, Lubchenco LO, Koops BL, et al. Neonatal hyperviscosity: randomized study of effect of partial plasma exchange transfusion on long-term outcome. Pediatrics 1985:75(6);1048-53.
  4. Bada HS, Korones SB, Pourcyrons M et al. Asymptomatic syndrome of polycythaemic hyperviscosity: effect of partial plasma exchange tranfusion J Pediatr 1992:120(4 Pt 1);579-85.