From the Journals

Can high hematocrit predict early shunt thrombosis?


Key clinical point: Higher hematocrit levels have been associated with early shunt occlusion in newborns having systemic to pulmonary artery shunt placement.

Major finding: For every 5 additional percentage points of hematocrit, an infant’s odds of early shunt occlusion more than doubled (odds ratio, 2.70; P = .009).

Data source: Retrospective study of all newborns who underwent primary systemic to pulmonary artery shunt placement from January 2010 to July 2015 at a single center.

Disclosure: Dr. Anderson receives salary support from the National Center for Advancing Translational Sciences. Coauthor Jennifer M. Duchon, MDCM, MPH, receives salary support from the National Institute of Allergy and Infectious Disease. The remaining coauthors had no financial relationships to disclose.

View on the News

Is this a slow-moving train?

In her invited commentary, Nancy S. Ghanayem, MD, of the Medical College of Wisconsin, gives credit to Dr. Anderson and coauthors for introducing the notion that high initial postoperative hematocrit may increase the risk of early shunt thrombosis in newborns, but with a caveat: “we remain somewhat hesitant regarding wholesale acceptance of the validity of the conclusions for several reasons” (J Thorac Cardiovasc Surg. 2017;153:956).

Those reasons include the low number of reported events, failure to list the actual hematocrits of the five patients who experienced early shunt occlusion and the lack of hemodynamic data – the latter of which she called “a significant limitation.”

A patient with low cardiac output, especially one who is cyanotic or has a single ventricle, is at higher risk for thrombosis and more likely to be transfused, “which in this case would potentially be masked by the method of comparing transfused volumes,” Dr. Ghanayem said. The inotrope score, while useful, is not a surrogate for actual cardiac output. She asks, “Accordingly, is early shunt occlusion due predominantly to passenger (corpuscular) overload or to a slow-moving train?”

Dr. Ghanayem had no financial relationships to disclose.



Shunt occlusion has been a well-documented cause of shunt failure in newborns who have had systemic to pulmonary shunt placement, and little has been known about why shunts occlude. However, researchers have reported in a small retrospective study that higher postoperative hematocrit levels immediately after surgery may be predictors of shunt occlusion.

Reporting in the April 2017 issue of The Journal of Thoracic and Cardiovascular Surgery, Brett R. Anderson, MD, of New York-Presbyterian/Morgan Stanley Children’s Hospital, Columbia University Medical Center in New York, and coauthors, found that every 5-point increase in postoperative hematocrit more than doubled an infant’s odds of having early shunt occlusion (J Thorac Cardiovasc Surg. 2017;153:947-55).

“Beginning in the latter half of 2014, we noticed an increase in the incidence of early shunt occlusions in our neonatal cardiac intensive care unit,” Dr. Anderson and coauthors said. So they conducted a retrospective chart review of 80 infants who had undergone systemic to pulmonary shunt placement from January 2010 to July 2015, hypothesizing that increased hematocrit in the early postoperative period might have caused early shunt occlusion. They investigated the association between the first postoperative hematocrit and early shunt occlusion and in-hospital mortality in these patients.

Five patients (6.3%) experienced early shunt occlusion – that is, within 24 hours of placement (actually, within 10 hours of placement). Overall, 12 infants (15%) experienced shunt occlusion. The physicians at New York-Presbyterian do not administer anticoagulation in these patients during the first 12 hours after shunt placement.

The median initial postoperative hematocrit was 41.7%, with a range of 31.7%-55.8%. The survival analysis the researchers performed found that for every 5 additional percentage points, the hazard ratio for early shunt occlusion was 2.7 (P = .007) and 1.74 for any shunt occlusion (P less than .001). Incidentally, four cases of early shunt occlusion occurred in the later study period after 2014, during which the average first postoperative hematocrit was significantly higher than in the pre-2014 study period, 45.3% vs. 41.5% (P = .21), and the odds of early shunt occlusion were 16 times higher (P less than .001). Dr. Anderson and coauthors said the possible explanation for this variation was a switch to a new point-of-care analyzer in 2013.

With regard to mortality, six infants overall (7.4%) died before discharge, and four (5%) within 30 days of shunt placement. No infants with early shunt occlusion died, although two with late shunt occlusion died. Increased inotrope score and first postoperative arterial oxygen tension were the only factors associated with increased mortality. “No significant association was identified between hematocrit and 30-day mortality,” Dr. Anderson and coauthors said.

In the first 24 hours after surgery, 11 infants (13.8%) received packed red blood cell (PRBC) transfusions, seven (8.8%) received platelets, and four (5%) received fresh frozen plasma/cryoprecipitate. Higher postoperative PRBC transfusion volumes were associated with increased odds of mortality (P = .001), but none of these factors were significantly associated with early shunt occlusion.

Dr. Anderson and coauthors acknowledged that shunt occlusion is a “vexing problem” in infants with cyanotic heart disease. While other researchers studied postoperative hematocrit levels and possible associations with outcomes, including shunt occlusion and mortality, the New York-Presbyterian investigators said this is the first study of the first postoperative hematocrit.

Dr. Anderson and coauthors said their findings raise the question about the ideal perioperative prophylactic antithrombotic therapy in these patients. These researchers initiate aspirin therapy 12 hours after surgery if hemostasis is established.

As a result of this study, Dr. Anderson and coauthors instituted a number of practice changes at their center. They include:

• Cardiac anesthesiologists have been asked not to transfuse shunted neonates with hematocrit level of greater than or equal to 35%, and hematocrits are then immediately repeated when a patient returns to the cardiac ICU.

• Patients with hematocrits greater than or equal to 55% get partial exchange transfusions.

• An individualized approach for patients with lower hematocrits who are more cyanotic than expected. This includes a diagnostic echocardiogram, nitric oxide, oxygen and heparin, escalated inotropic support if necessary and sometimes a cautious approach to transfusions if symptoms do not resolve and an acute shunt occlusion if not likely.

Dr. Anderson and coauthors acknowledged limits to their study, most notably its retrospective nature and a small population at a single center, and that large investigations are needed to validate their findings.

Dr. Anderson disclosed receiving salary support from the National Center for Advancing Translational Sciences. Coauthor Jennifer M. Duchon, MDCM, MPH, receives salary support from the National Institute of Allergy and Infectious Disease. The remaining coauthors had no financial relationships to disclose.

Next Article:

   Comments ()