Premature babies have fewer complications if a lower platelet count is accepted

Fewer premature babies die or have major bleeding if platelet transfusions are withheld until platelet numbers drop to a lower level. At 28 days, death or new major bleeding occurred in 19% of neonates transfused when they had less than 25,000/mm3 platelets compared to 26% of neonates transfused when they had less than 50,000/mm3 platelets.

This trial included 660 premature babies with low platelet counts.

The results suggest that in the absence of actual bleeding, platelet transfusions may be safer to give when blood platelet concentrations are lower than currently recommended trigger thresholds. It remains unclear why there was an increased risk of the combined outcome of either death or major bleeding, when neither outcome was increased when looked at alone. However, some caution needs to be applied to the findings as 39% of the neonates had already had a platelet transfusion before entering the trial.


Why was this study needed?

Platelets circulate in the blood and are essential for blood clotting. Low numbers of platelets may be due to inadequate production or to higher consumption, both common in preterm babies. Platelet transfusions can prevent and treat bleeding, but carry risks such as transfusion reactions.

Decisions on when to give platelet transfusions in these infants are based on clinical opinion as to the individual baby’s risks. These decisions can include a particular threshold of platelet count in addition to the clinical picture. Although there are suggested thresholds for when to give transfusions based on consensus-based guidelines, there is wide variation across the UK.

This multi-centre trial aimed to assess the effects of transfusing at two different platelet trigger thresholds.


What did this study do?

This trial included 660 premature infants born at less than 34 weeks’ gestation with severely reduced blood platelet count from three countries including the UK. Average gestational age was 26.6 weeks and median birth weight was 740g. A third were on antibiotics for sepsis and 19% had major bleeding.

The included infants were randomised to receive a platelet transfusion of 15ml/kg if their platelet count dropped to either less than 25,000/mm3 or less than 50,000/mm3.

There were instances were transfusions were either indicated but not administered or administered but not indicated. Additionally, 39% of the infants had already had a transfusion before entering the trial (126 in the 50,000/mm3 group and 121 in the 25,000/mm3 group).

Although clinicians were aware of allocation, outcome assessors were blinded which allows us to have more confidence in the findings.


What did it find?

  • At 28 days, death or new major bleeding occurred in 26% (85/325) of infants transfused at the 50,000/mm3 threshold compared with 19% (61/329) of infants transfused at the 25,000/mm3 threshold (odds ratio [OR] 1.57, 95% confidence interval [CI] 1.06 to 2.32).
  • A total of 90% (296/328) of infants in the 50,000/mm3 threshold group received at least one platelet transfusion compared with 53% (177/331) of infants in the 25,000/mm3 threshold group (hazard ratio [HR] 2.75, 95% CI 2.36 to 3.21).
  • There was no significant difference in the number of infants who died by 28 days: 15% (48/326) in the 50,000/mm3 threshold group compared with 10% (33/330) in the 25,000mm3 threshold group (OR 1.56, 95% CI 0.95 to 2.55).
  • A similar proportion of infants had at least one major bleeding episode, occurring in 14% (45/328) of infants in the 50,000/mm3 threshold group compared with 11% (35/330) in the 25,000/mm3 threshold group (HR 1.32, 95% CI 1.00 to 1.74). Minor or worse bleeding occurred in two-thirds of each group and moderate bleeding in a third of each group.
  • Serious adverse events were common in both groups, occurring in 25% (81/324) of the 50,000/mm3 threshold group compared to 22% (74/336) of the 25,000/mm3 threshold group (OR 1.14, 95% CI 0.78 to 1.67).


What does current guidance say on this issue?

The UK’s 2014 transfusion guideline for neonates states that there is no clear correlation between the severity of low platelet count and major bleeding and suggests that other clinical factors are important. The guidance acknowledges this trial as being underway and suggests platelet transfusions where platelets are:

  • below 25,000 or 30,000/mm3 in the absence of bleeding
  • below 50,000/mm3 where there is bleeding, current coagulopathy (bleeding disorder) or exchange transfusion (replacing abnormal blood components with donor blood)
  • below 100,000/mm3 where there is major bleeding or major surgery.


What are the implications?

This study suggests that preventative platelet transfusions for premature babies with low platelets may be safer to give at lower levels than currently recommended.

As the rate of pre-randomisation transfusion was similar in both groups, this is unlikely to account for the between-group difference seen in the composite primary outcome of death or major bleeding. However, it does make it more difficult to provide clear guidance.

Even so, these findings are likely to impact practice and aid clinicians with decisions on when to offer platelet transfusions for premature babies.


Citation and Funding

Curley A, Stanworth SJ, Willoughby K et al.; PlaNeT2 MATISSE Collaborators. Randomized trial of platelet-transfusion thresholds in neonates. N Engl J Med. 2019;380(3):242-51.

This study was funded by the National Health Service Blood and Transplant Research and Development Committee; Sanquin Research, Amsterdam; Addenbrooke’s Charitable Trust; the Neonatal Breath of Life Fund.



NICE. Blood transfusion. NG24. London. National Institute for Health and Care Excellence; 2015.

Norfolk D (editor); Joint United Kingdom Blood Transfusion and Tissue Transplantation Services Professional Advisory Committee. Handbook of Transfusion Medicine 5th Edition. Norwich: TSO; 2014.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre



Neonates are defined as infants up to 28 days after birth.


Expert commentary

Over 250,000 adult doses of platelets are transfused each year in the UK. Outside haemato-oncology practice there is limited evidence with usage based on presumed benefit and historical practice. Curley et al. looked specifically at use in neonates. The results are striking and unexpected. Platelets were associated with harm not benefit. A recent prospective study in adults on anti-platelet agents and with intra-cranial bleeding, the PATCH study, had a similarly unexpected negative finding.

These studies ring alarm bells. Instead of an assumed benefit of uncertain magnitude we are looking at possible harm from these biological hand grenades. We should look carefully at our practice in neonates following this report, but also consider other areas, such as cardiac surgery where platelet transfusion is essentially an evidence-free zone.

Dr Jonathan Wallis, Consultant Haematologist, Chair of the National Blood Transfusion Committee

The commentator is collaborating with one of the authors on a separate study not involving platelets