Protamine Dosing for Heparin Reversal after Cardiopulmonary Bypass: A Double-blinded Prospective Randomized Control Trial Comparing Two Strategies

After approval from the University of Miami (Miami, Florida) Institutional Review Board (approval No. 20220234, April 11, 2022), this prospective randomized double-blinded study was conducted at the University of Miami Hospital from June 2022 to June 2023. The study was registered at www.clinicaltrials.gov (registration No. NCT05426031, Principal Investigator: Michael Fabbro, registration date: June 15, 2022). Informed written consent was obtained from all subjects.

The study included patients at least 18 yr old undergoing elective open heart cardiac surgery under CPB. Only patients with a calculated pre-CPB heparin dose of greater than 27,500 U (using 350 U heparin per kg total body weight as the initial dose) were included, because this would ensure a minimum 10% difference in subsequent initial protamine dosing between the groups (see “Protamine Administration”). Patients undergoing emergency surgery or for whom deep hypothermic circulatory arrest was anticipated were excluded. Other exclusion criteria included pregnancy, end-stage renal disease, known coagulation disorders, inadequate cessation of warfarin (preoperative international normalized ratio on the day of surgery greater than 1.5), inadequate interruption of direct oral anticoagulants or nonaspirin antiplatelet drugs, and ongoing preoperative intravenous administration of unfractionated heparin. Patients with known heparin resistance or contraindications to heparin administration, including heparin-induced thrombocytopenia and heparin allergy, were also excluded.

The patients were assigned to either a fixed protamine dose (group A) or a heparin ratio-based protamine dose (group B) using block randomization with sizes of 2, 4, and 6. The patient, surgeon, attending anesthesiologist, and anesthesia practitioner were blinded to the group allocation and the dose of protamine during the perioperative period. The fixed dose and the titrated dose of protamine were prepared by anesthesia personnel not directly involved in the care of the patient in 50 ml of saline, with the syringe labeled only as “protamine” to maintain blinding. The protocol followed for all patients corresponded to their assigned group.

All study patients were managed in a standardized fashion. After preinduction arterial catheter placement, anesthesia was induced with anesthetic and neuromuscular blocking agents selected by the attending anesthesiologist. Standard American Society of Anesthesiologists (ASA; Schaumburg, Illinois) monitors were applied to all patients. Central venous access and a pulmonary artery catheter were placed either pre- or postinduction, as deemed clinically appropriate. After anesthetic induction and tracheal intubation, anesthesia was maintained with either sevoflurane or isoflurane at the discretion of the attending anesthesiologist. A baseline arterial blood gas was drawn and analyzed on an i-STAT point-of-care analyzer (Abbott, USA), and a baseline point-of-care activated clotting time was run on the Hemochron Signature Elite analyzer (Werfen USA LLC, USA) using activated clotting time + cartridges before heparin administration. All subsequent activated clotting time measurements were performed using the same physical device as used for the baseline measurement.

Before arterial (aortic or femoral) cannulation by the surgeon, 350 U heparin per kilogram of total body weight were administered via the central venous line as the initial pre-CPB heparin dose. An arterial sample was drawn 3 min later to assess the postheparin activated clotting time. Additional heparin, in aliquots of 5,000 U, was administered if the activated clotting time value did not reach 450 s. The sum of this additional heparin and initial pre-CPB heparin constituted the total pre-CPB heparin dose. Per institutional protocol, a further 10,000 U heparin was added to the CPB prime by the perfusionist, and additional heparin was given as needed during CPB to maintain activated clotting time above 450 s. The final heparin dose was calculated as the total pre-CPB heparin plus the heparin administered by the perfusionist. A peripheral cannulation strategy using the femoral vessels was typically employed for minimally invasive (anterior or anterolateral thoracotomy) surgical cases, whereas central cannulation was carried out for surgery performed via sternotomy. CPB was then instituted with a machine utilizing a centrifugal pump (LivaNova, United Kingdom) and membrane oxygenator (Affinity Fusion, Medtronic, USA).

After separation from CPB, protamine was administered from the blinded syringe in accordance with the group allocation. Group A patients received a fixed, single-vial 250-mg protamine dose. Group B patients received 1 mg protamine for every 100 U pre-CPB heparin (1:1 heparin ratio-based dosing). These doses were produced using a single 250-mg vial and additional 50-mg vials as needed, to minimize wastage from opening a second 250-mg vial. Based on the minimum calculated pre-CPB heparin dose of 27,500 U, patients in group B would be expected to receive at minimum 275 mg protamine, 10% more than group A. Protamine was administered over 5 to 6 min via a central venous line. If hypotension or pulmonary hypertension attributed to protamine occurred, administration was halted, and hemodynamic support with intravenous pressors or inotropic agents was carried out as deemed appropriate. Then, the remainder of the dose was administered slowly. A postprotamine arterial activated clotting time was measured 3 min after completion of the full initial dose. Additional protamine was administered if the activated clotting time exceeded the baseline activated clotting time by 20% or per anesthesiologist or surgical discretion if the activated clotting time was higher than baseline but less than 20% over baseline.

The rest of the intraoperative anesthetic management was per institutional protocol. Transfusion of packed red blood cells was reserved for patients with hemoglobin values of less than 7 g/dl or evidence of ongoing bleeding. Platelets were transfused in patients who had evidence of clinical bleeding with preoperative thrombocytopenia or ongoing transfusion requirements. Fresh frozen plasma was reserved for patients with evidence of ongoing bleeding and continued red cell transfusion requirements.

Patient data were recorded on individual study sheets by the investigators. Baseline patient information including age, sex, weight, ASA classification, surgical approach (sternotomy vs. minimally invasive thoracotomy), and surgery type were noted. The surgery type was categorized as: (1) isolated coronary artery bypass grafting (CABG); (2) valve surgery (mitral, tricuspid or aortic valve surgery, or a combination of these procedures, with or without ancillary procedures including patent foramen ovale closure, aortic root replacement, or ascending aortic surgery not utilizing deep hypothermic circulatory arrest); (3) CABG plus valve surgery; or (4) other. The baseline activated clotting time and postheparin activated clotting time were documented. Also noted were the initial pre-CPB, total pre-CPB, and final heparin doses, as well as the CPB time, cross-clamp time, postprotamine activated clotting time, and dose of additional protamine, if administered. Other information recorded included the number of units of packed cells, platelets, fresh frozen plasma, and cryoprecipitate transfused during surgery, and the chest tube output during the first 24 h after surgery.

The primary outcome was the postprotamine activated clotting time obtained after the initial protamine dose. Secondary outcomes included the administration of additional protamine, the total protamine dose, the number of units of blood components transfused during surgery, and the total chest tube output in the 24-h period after surgery. An analysis of net savings of protamine in units of 50-mg vials was performed. The odds of having an activated clotting time after the initial protamine dose more than 10% higher than baseline and the odds of requiring additional protamine based on the group assignment were assessed.

Because studies of the activated clotting time have shown that these values are normally distributed,^9,10  we computed the power requirements for the current study based on a two-group, two-sided t test using the activated clotting time values after the initial doses of protamine in the two groups from our previous study⁷  and on studies demonstrating that a 10% difference in the activated clotting time is relevant with respect to clinical decision-making.¹¹  Using Stata version 17.0 (StataCorp, USA) with a power of 80% and type I error of 5%, a minimum of 100 patients (50 per group) would be needed to be enrolled to detect a two-sided difference of 10 s in post-CPB activated clotting time. Fifteen additional patients in each group were enrolled to compensate for potential patient exclusions due to protocol violations, inability to separate from CPB, or withdrawal of consent after randomization.

Categorical binary variables were compared between groups using Stata’s two-sample test of proportions or by Fisher’s exact test if the variables were nonbinary. Continuous variables were compared between groups using t tests incorporating Satterthwaite’s approximation and unequal variances. Mean differences and the 95% CI are reported along with the median and interquartile range as summary statistics for each group. Multivariable logistic regression was performed in Stata to evaluate the group assignment on the binary occurrences of an activated clotting time greater than 110% of baseline and for the need for additional protamine after the initial, protocol-based dose.

There were 130 enrolled patients, equally allocated to the two groups. Three patients in group A and two patients in group B were dropped for an unanticipated need for circulatory arrest. These patients were given a standard 1:1 protocol and not included in analysis. Thus, there were 62 patients in group A and 63 in group B who were analyzed (fig. 1).

There were no significant differences with respect to sex, ASA Physical Status, a minimally invasive surgical approach, or type of surgery (table 1). There was imbalance with respect to the occurrence of a previous cardiac surgery, with more patients in group B (P = 0.017).

As expected, the initial protamine dose was lower in group A (mean, −85.6 mg; 95% CI, −97.8 to −73.5 mg; P < 0.0001; table 2). There were no significant differences between groups with respect to age, weight, heparin doses (initial, total pre-CPB, total including the pump), the activated clotting time (baseline, postheparin), CPB time, or cross-clamp time (table 2).

The mean initial postprotamine activated clotting time (the primary outcome) was not different between groups (−2.0 s; 95% CI, −7.2 to 3.3 s; P = 0.47; table 3). The odds of having an activated clotting time that was more than 10% higher than the baseline value was not influenced by the group assignment (odds ratio, 0.73; 95% CI, 0.31 to 1.71; P = 0.47; supplemental digital content, https://links.lww.com/ALN/D719), after controlling for age, weight, ASA Physical Status, previous cardiac surgery, sternotomy, the baseline international normalized ratio or activated clotting time, heparin pre-CPB, heparin including the pump, during of CPB, and the cross-clamp time.

Less total protamine was administered in group A, with a mean difference of −2.1 50-mg vials per case (95% CI, −2.4 to −1.8; P < 0.0001; table 3). There was no difference in the number of patients in groups A and B requiring additional protamine, 18 (29.0%) and 20 (31.7%), respectively, nor was there a difference among those patients in the additional dose administered (mean difference, −5.1 mg; 95% CI, −17.8 to 7.6 mg; P = 0.42). The odds of requiring additional protamine was not affected by the group assignment (odds ratio, 0.81; 95% CI, 0.33 to 1.94; P = 0.63; supplemental digital content, https://links.lww.com/ALN/D719), after controlling for the same variables listed above for the postprotamine activated clotting time. In the approximately 30% of patients requiring additional protamine, the mean additional protamine doses were 46.4 and 51. 5 mg (P = 0.89) in group A and group B, respectively. Among patients receiving additional protamine, 7 (39%) in group A and 8 (40%) in group B had an initial postprotamine activated clotting time exceeding 20% of baseline. The remaining patients received protamine at the discretion of the physicians.

The protamine-to-heparin ratio was lower in group A versus group B (difference = −0.26; 95% CI, −0.30 to −0.22; P < 0.0001; table 3). The mean effective protamine-to-heparin ratio in group A was 0.79. There was no difference in the cumulative chest tube output measured at 24 h postsurgery (difference, −77 ml; 95% CI, −220 to 65 ml; table 3). Blood product administration rate was low, overall, precluding meaningful analysis. Only four patients received an intraoperative red cell transfusion (two in each group). Two patients in group A received platelets and plasma, whereas five patients in group B received intraoperative platelets, four of whom also received plasma.

In this randomized, double-blinded, prospective trial comparing a fixed dose of protamine versus a 1:1 ratio-based dose to the total amount of heparin given pre-CPB, there was no difference in the initial postprotamine activated clotting time. This is clinically important because the activated clotting time is the most common measure used to determine the adequacy of heparin reversal. There were also no significant differences between the groups in terms of demographics, heparin dosing, CPB, or cross-clamp times or secondary outcomes like chest tube output. The only difference was that the fixed-dose group received considerably less protamine than the ratio-based group. Because 1:1 ratio-based dosing was without clinical benefit, this strategy reflects poor pharmaceutical stewardship of a resource that, at times, is in short supply. Thus, a reasonable approach to the management of heparin neutralization after CPB for most adult elective surgery patients may be to start with a fixed, 250-mg dose irrespective of the actual pre-CPB heparin dose and then titrate additional doses per clinical discretion or as indicated by the activated clotting time. These results are in line with the findings of our earlier, retrospective, nonrandomized observational study.⁷ 

In this study, the median heparin doses were 32,000 and 33,000 U in groups A and B, respectively, resulting in a median protamine-to-heparin ratio of 0.79 in the fixed-dose group. As noted, the conventionally used, higher ratio-based doses of protamine carries a risk of overdosing given there is no consideration for time since heparin administration or its clearance.¹²  Higher protamine doses are known to impair platelet aggregation,¹³  reduce thrombin generation,¹⁴  and, overall, result in a higher activated clotting time value with increased bleeding risk.¹⁵  Randomized controlled trials comparing protamine-to-heparin ratios of 0.8 versus 1.3 and 0.8 versus 2.0 demonstrated worse coagulopathy and bleeding in the higher protamine dose groups, suggesting that the lower ratio of 0.8 is likely optimal.^16,17  In a separate study comparing a protamine-to-heparin ratio of 0.8 with 1.0 found no difference in indices of postoperative bleeding.¹⁸  The protamine-to-heparin ratio of 0.79 found in the fixed-dose group therefore corresponds to a well documented, effective reversal strategy. However, using a protamine-to-heparin ratio of 0.8 for all would still result in the use of more than one 250-mg vial of protamine in patients requiring greater than 32,000 U heparin (anyone greater than 91 kg) before CPB. Although using a protamine-to-heparin ratio of 0.8 may be considered, centers could further limit the initial protamine dose to within a single vial of 250 mg in these patients when shortages arise.

These results have significant implications in an era of drug shortages and increased emphasis on medication stewardship. Conservation of critical medications like protamine and dosing strategies aimed at conserving vials of medication can extend access to services within a given institution during times when the drug supply is constrained. More importantly, these findings question the purported benefits of conventional protamine-to-heparin ratio of 1.0 dosing in preventing underdosing or heparin rebound. In turn, our results suggest that this approach is wasteful and exposes the patient to potential coagulopathy.

Alternatives to fixed or ratio-based approaches to protamine administration include point-of-care titration-based heparin neutralization systems. These systems recommend heparin bolus doses based on the heparin dose–response curves. They also determine heparin concentration using quantitative and functional assays and suggest protamine dosing based on the heparin concentration. Although some evidence favors its use, these systems remain expensive¹⁹  and are unavailable in many institutions, including the study hospital.²⁰  Furthermore, the benefit of these systems remains controversial, because the evidence favoring their use is conflicting.²¹  Current recommendations only provide weak support for routine use of heparin concentrations assays.²²  Excess protamine dosing due to overestimation of heparin concentrations or by use of calculations that overestimate patient blood volume,²³  especially in patients with high body mass indices, has been reported. Higher protamine estimates may also be provided by these systems if the circulating heparin concentrations are below detectible range.²⁴  Variability in protamine dosing may still exist depending on whether pump volumes are added to total blood volumes, serving as an additional cause of protamine overdose with use of these systems.²⁵  Finally, residual heparinization after protamine administration may occur despite the use of these devices.²⁶ 

The activated clotting time still remains the most widely available and used “gold standard” test for assessing anticoagulation during CPB.²⁷  In this study, we used the activated clotting time to test for the adequacy of heparin reversal, in line with our institutional practice. Although activated partial thromboplastin time and anti–factor Xa are more sensitive tests for detecting residual heparin effects, they are not conventionally performed at the point of care. Thus, they are therefore neither suitable for real-time clinical decision-making regarding additional protamine administration in the operating room nor for detecting heparin rebound in the intensive care unit. Additionally, the activated partial thromboplastin time has been found to correlate poorly with circulating heparin, and elevated values may be attributed to factor deficiency or protamine itself.²⁸ 

The study was conducted with a single surgeon at a single center where all cases were elective, and the majority were done using a minimally invasive approach. Although it is unlikely that these limitations had any effect on the primary outcome measure of adequate heparin reversal, secondary endpoints may have been affected. Additionally, conclusions drawn from our study may not be universally applicable to all clinical practices. Giving additional protamine at the discretion of the clinician, as opposed to a clinical endpoint, also introduces bias; however, this would not have affected the primary endpoint. Follow-up activated clotting times were not routinely performed when additional protamine was administered, which limits any conclusions about this practice. Although viscoelastic testing provides objective parameters and has utility in evaluating postprotamine residual heparin or heparin rebound, this test was not available to us.^29,30  We also had no way of determining whether perceived bleeding was related to an excess of protamine. The lack of viscoelastic testing-driven transfusion endpoints limits possible conclusions regarding transfusion, but intraoperative transfusion rates in this study were too low for statistical comparison, regardless. It should be noted that cryoprecipitate was not administered to any study subjects. Another limitation is that blood product transfusion in the postoperative period as a secondary endpoint was not assessed; we used chest tube output as a marker for ongoing coagulopathy. Last, the number of patients with previous cardiac surgery was higher in group B, but no other meaningful differences were found in terms of CPB times or chest tube output. This is unlikely to have affected postprotamine activated clotting time, although reoperation procedures are known to have increased bleeding risk. However, the difference would have had no effect on the primary endpoint of the initial postprotamine activated clotting time.

A conservative heparin reversal strategy based on the initial administration of a single 250-mg dose of protamine is feasible and effective. It resulted in activated clotting time values that were not different from a heparin ratio-based strategy and allowed protamine conservation without worsening postoperative bleeding. This approach can extend access to cardiac surgical services requiring CPB to patients during protamine shortages and represents a responsible approach to medication stewardship.

The authors thank Alberto Cruz, M.P.H. (University of Miami Department of Public Health Sciences, Division of Biostatistics, Miami, Florida).

Support was provided solely from institutional and/or departmental sources.

Dr. Lamelas is a paid consultant for Edwards Lifesciences (Irvine, California) and Medtronic (Minneapolis, Minnesota), which have no influence over this article. The other authors declare no competing interests.

Full protocol available at: mxf790@med.miami.edu. Raw data available at: mxf790@med.miami.edu.

Supplemental tables, https://links.lww.com/ALN/D719