Evaluation of the safety and efficacy of a new hemostatic powder using a quantitative surface bleeding severity scale

The safety and efficacy of a hemostatic powder (HP) versus a control agent, absorbable gelatin sponge and thrombin (G + T), were assessed, using a validated, quantitative bleeding severity scale.

G + T for achieving hemostasis within 6 min; and superiority of HP relative to G + T for success for achieving hemostasis within 3 min.
Results: A total of 388 subjects were included in the primary efficacy analysis. At 6 min, hemostasis was achieved in 93.0% (238/256) of the HP group compared to 77.3% (102/132) of the G + T group (non-inferiority P < 0.0001, superiority P < 0.0001). All secondary endpoints were met. Complications were comparable between treatment groups.
Conclusions: HP had superior rates of hemostasis, shorter preparation time, and a similar safety profile compared to G + T in this prospective, randomized trial using quantitative bleeding severity criteria.

K E Y W O R D S
bleeding scale, collagen, hemostat, hemostatic agent, hemostatic powder, thrombin

| INTRODUCTION
Level I evidence of the benefit of local hemostats in surgical operations is limited, yet research in this area is continuing to grow.  Although the ideal surgical hemostat-one combining safety, efficacy, usability, cost, and approvability-has yet to be created, continued development work is resulting in new agents. [40][41][42] The objective of this clinical trial was to evaluate the safety and efficacy of a novel hemostatic powder (HP) containing collagen, chondroitin sulfate, and thrombin compared to an established absorbable gelatin sponge and thrombin hemostatic agent. 43 This multicenter trial compared efficacy based on levels of bleeding using a unique, validated, quantitative surface bleeding severity scale (SBSS), the SPOT GRADE. 44 This SBSS is a bleeding severity scale based on quantitative measures that has been specifically validated in a clinical setting. 45 Other bleeding severity scales are based on qualitative assessments or have not been validated clinically. 39,46 2 | MATERIALS AND METHODS  Eligibility was also assessed intraoperatively using the SBSS to confirm identification of a target bleeding site (TBS) with minimal, mild, or moderate bleeding for which conventional means of hemostasis were ineffective or impractical. 43,44 Subjects were evaluated preoperatively, intraoperatively, postoperatively, and at 6 ± 2 weeks ( Figure 1).

| Study groups
The investigational device (HEMOBLAST Bellows, Biom'up, Lyon, France) was supplied in a bellows applicator preloaded with 1.65 g of ARDEHALI ET AL. These control hemostats are frequently used as comparators in clinical trials of effectiveness. 4,5,15,26,34 Subjects in the trial were randomized to receive HP or G + T in a 2:1 ratio. A 2:1 ratio was chosen to collect additional safety information on HP. Randomization was stratified by surgery type and a randomized block design was implemented. Concealed allocation was achieved by requiring that all personnel present in the operating room were blinded to the treatment group assignment until after FIGURE 1 Clinical investigation flow. A, One subject with withdrawn intraoperatively; 46 subjects were withdrawn prior to surgery due to early termination of the study for efficacy at the interim analysis. B, Unreported or missing data not presented in subsequent tables. C, Gelatin sponge and thrombin identification of the TBS and confirmation of intraoperative eligibility.
Subjects were blinded to treatment assignment.
The first subject for each investigator without previous experience or hands-on training with HP was treated as a lead-in subject. All leadin subjects received HP and were followed for safety, but not for efficacy. The full cohort study enrollment was stopped early on HP or G + T was applied to the source of minimal, mild, or moderate bleeding and maintained at the TBS until 3, 6, and 10 min after initial application (Figures 3a-3d). 43,44 If hemostasis was not achieved at 3 and/or 6 min, repeat application of the original randomized hemostat was performed. If hemostasis was not achieved at 10 min, a rescue method of hemostasis treatment was used. Rescue treatment using another hemostatic agent containing thrombin was not allowed.
Bleeding severity and hemostasis were assessed using the SBSS at baseline and at each assessment time point. Hemostasis was defined as an SBSS score of 0 with bleeding scored from 1 (mild) to 5 (extreme) ( Figure 4). 44 Clinical investigators underwent training and testing on the SBSS prior to the enrollment of any subjects. In cases where hemostasis was initially achieved, but bleeding recurred prior to subject closure, the re-bleeding was documented as an adverse event.
Adverse events (AEs), serious adverse events (SAEs), rescue treatment, and reoperation due to bleeding were collected. Antibody testing for porcine collagen was performed preoperatively and at the 6 ± 2 weeks follow-up visit.

| Study endpoints
The primary efficacy endpoint of this clinical investigation was noninferiority of HP relative to G + T for success at achieving hemostasis within 6 min. The secondary efficacy endpoints were, in rank order: 1.
Superiority of HP relative to G + T in mean preparation time from the opening of the package to the product being ready-to-use; 2. Noninferiority of HP relative to G + T for success at achieving hemostasis within 3 minutes; 3. Superiority of HP relative to G + T for success at achieving hemostasis within 6 min; and 4. Superiority of HP relative to G + T for success at achieving hemostasis within 3 min.

| Analytic methods
Efficacy analyses were conducted on the time to hemostasis (TTH) population, defined as all subjects who were randomized, received study intervention, and had a TTH assessment recorded regardless of whether the measurement was censored. Lead-in subjects were not part of the TTH population. Safety analyses were conducted on the Safety Population, defined as all subjects enrolled in the study, which included lead-in subjects.
The primary analysis compared the probability of TTH within 6 min in subjects receiving HP to those receiving G + T. The estimated difference in the probability of TTH at 6 min between treatment arms was adjusted for surgical indication by weighting the stratum-specific differences in observed proportions using Cochran-Mantel-Haenszel weights. 49 Secondary endpoints were a priori specified and analyses were conducted using a rank-order test sequence to maintain the family-wise Type I error rate at 0.05.
The IDMC was provided with blinded efficacy data at the planned interim analysis. The IDMC was guided by a formal stopping rule based on the primary efficacy endpoint; the clinical trial could be stopped for reasons of futility or efficacy.
Summaries of the number and percent of subjects with at least one AE or SAE were computed for each treatment arm. AEs were further stratified by adverse event type.

| Sample size justification
The sample size for the study was calculated based on a level 0.025 (one-sided) test to exclude a probability of TTH within 6 min that was 10% less among subjects treated with HP compared to those treated with G + T. The 10% non-inferiority margin is based upon an FDA guidance for industry on non-inferiority clinical trials. 50 With an  subjects met all eligibility criteria and were enrolled; 24 were lead-in and 388 were randomized to receive HP or G + T in a 2:1 ratio ( Figure 1). All randomized subjects completed the primary efficacy assessment, which was completed intraoperatively, while a total of 29 enrolled subjects did not complete the study as planned (withdrew from the study, non-compliant, subject death, or lost to follow-up). Demographic and baseline clinical characteristics were recorded for all enrolled subjects (Table 1). Age, race, and ethnicity data were similar between treatment groups and representative of the target intended use population, per FDA Guidance for Industry. 51 The mean subject age was 55.7 years, with 39.8% of subjects being male. The surgical indication, TBS locations, and TBS tissue types -including soft tissue, muscle, bone, and parenchyma-were similar between groups (Table 2).

| Efficacy
HP met all pre-specified primary and secondary efficacy endpoints. HP was non-inferior to G + T in achieving hemostasis at 6 min (93.0% for HP; 77.3% for G + T; P < 0.0001) ( Table 3). The preparation time for HP group was significantly shorter than for the G + T group, with a mean of 0.38 min (23 s) for the HP group and a mean of 2.28 min (2 min, 17 s) for the G + T group (P < 0.0001) ( Table 4). HP also demonstrated non-inferiority and superiority at 3 min as well as superiority at 6 min compared to G + T (Table 3). For subjects with the highest degree of bleeding permitted for enrollment in the trial, a baseline SBSS of 3 (moderate bleeding), 89.6% of the HP group compared to 56.3% of the G + T group achieved hemostasis within 6 min (P < 0.0001 for non-inferiority; P = 0.0003 for superiority) (Table 5)  Kaplan-Meier estimates for the probability of achieving hemostasis at all scheduled assessment times were determined ( Figure 5). There were no significant differences in intraoperative red blood cell, platelet, and fresh frozen plasma administration between the test and control groups (P = 0.6597, P = 0.7784, and P = 0.3297, respectively).

| Safety results
Serum samples were collected preoperatively and at the 6 ± 2week follow-up visit to test for the development of antibodies to porcine collagen. In the HP group, 2.5% of subjects displayed preoperative antibodies against porcine collagen and 12.1% demonstrated anti-collagen antibodies at 6 ± 2 weeks postsurgery (Table 7).
In the G + T group, 4.1% of subjects had preoperative antibodies against porcine collagen and 6.8% had positive titers postoperatively.
Subjects with postoperative titers against porcine collagen showed no       Table 2). The difference between groups was most pronounced at a baseline SBSS of 3, which is defined as moderate bleeding with a maximum flow rate of 117 mL/min. These findings imply that HP may be most beneficial at higher levels of bleeding within the range indicated for the product. As a ready-to-use powder, HP may offer advantages in preparation time when compared to other currently available hemostatic agents. 1,[40][41][42]52,53 The   This investigation utilized a quantitative and validated method for assessing bleeding severity and hemostasis. 44,45 Previous investigations have used subjective methods for assessing bleeding severity and success (hemostasis); other bleeding scales developed to assess hemostat performance have not been validated in a clinical setting. 10,39,46 This study is unique in that it is the first clinical investigation to employ a clinically validated and quantitative bleeding scale, thus providing a higher level of evidence in assessing comparative performance. 39,44,45,46

| Limitations
Investigators were not blinded to study treatment due to the visual differences between HP and G + T. Once TBSs were identified and baseline SBSS scores assigned, subjects were then randomized and the investigator was notified of the treatment assignment. Therefore, investigators were aware which intervention the subject would undergo during assessment of bleeding severity at 3, 6, and 10 min.
This lack of allocation concealment represented a possible source of bias in the study. Additionally, HP was not compared to the latest thrombin agents, sealants, or adhesives. G + T was selected as the control agent for several reasons. G + T is one of the most frequently used hemostatic agents in the United States. 56 The components of the control article are very similar to those of the investigational device (porcine gelatin with recombinant human thrombin vs. porcine collagen with human pooled plasma thrombin, respectively). G + T is approved for use in all three surgical areas evaluated in this trial and is recognized by FDA as an appropriate control that has been used in multiple other hemostatic studies. 26   ELISA, enzyme-linked immunosorbent assay.