Between July 1st, 2022 and June 30th, 2023, 316 patients were assessed for eligibility. In total, 268 eligible patients were randomly assigned to receive no pharmacological intervention (Control group - CG) or TXA (Test group - TG) (Figure 1).
Additionally, due to missing crucial lab test results, 30 patients were excluded from the trial (CG=12; TG=18). Therefore, a total number of 238 patients (CG=123; TG=115) were included in the analysis. The obligatory 1-month follow-up visit was completed for 97.9% (n=233) of the analyzed patients. Between July 4th, 2022 and June 28th, 2023, all the qualified patients underwent surgery. 22 patients originally assigned to the control group (CG) required an intraoperative dose of TXA. However, according to the intention-to-treat approach, they were still analyzed as the control group. All patients originally assigned to the test group (TG) received the allocated intervention. After 12 months of patient recruitment the trial was concluded. The total data-gathering period ended on August 4th, 2023. The baseline characteristics were comparable between both study groups (Table 1).
Table 1: Baseline demographics (Intention-to-treat and per-protocol analysis).
|
|
Intention-to-treat
|
|
|
Per protocol
|
|
|
CG (n=123)
|
TG (n=115)
|
p-value for group comparison
|
CG (n=102)
|
TG (n=136)
|
p-value for group comparison
|
Age (years)
|
41 [39.5 - 43.1]
|
41.4 [39.7 - 43.1]
|
0.926 ¹
|
41.2 [39.4 - 43.1]
|
41.4 [39.7 - 43.2]
|
0.852 ¹
|
Female gender
|
79 (64%)
|
77 (67%)
|
0.658 ³
|
63 (62%)
|
93 (68%)
|
0.288 ³
|
Height (m)
|
1.72 [1.71 - 1.74]
|
1.70 [1.58 - 1.90]*
|
0.753 ²
|
1.73 [1.71 - 1.74]
|
1.70 [1.58 - 1.90]*
|
0.112 ²
|
Preoperative weight (kg)
|
119 [90 - 172]
|
111 [86 - 175]*
|
0.097 ²
|
121 [91 - 172]*
|
111 [86 - 175]*
|
0.075 ²
|
Preoperative Body Mass Index (kg/m²)
|
40 [33 - 53.1]*
|
38.9 [30.8 - 49.5]*
|
0.261 ²
|
39.9 [33.3 - 53.1]*
|
38.6 [31.4 - 49.8]*
|
0.167 ²
|
Maximal weight (kg)
|
135 [102 - 196]*
|
127 [94 - 190]*
|
0.046 ²
|
135 [104 - 196]*
|
127 [94 - 190]*
|
0.047 ²
|
Maximal Body Mass Index (kg/m²)
|
45 [36.7 - 59.9]*
|
44.3 [35.9 - 56.8]*
|
0.089 ²
|
45.5 [37.8 - 59.9]*
|
44.1 [35.4 - 58.3]*
|
0.163 ²
|
Diabetes melitus
|
21 (17%)
|
20 (17%)
|
0.948 ³
|
18 (18%)
|
23 (17%)
|
0.882 ³
|
Hypertension
|
54 (44%)
|
51 (44%)
|
0.945 ³
|
46 (45%)
|
59 (43%)
|
0.792 ³
|
Dyslipidemia
|
87 (71%)
|
77 (67%)
|
0.530 ³
|
76 (74.51%)
|
88 (65%)
|
0.106 ³
|
Previous abdominal surgeries
|
40 (33%)
|
43 (37.39%)
|
0.431 ³
|
31 (30%)
|
52 (38%)
|
0.209 ³
|
Preoperative lab tests
|
|
|
|
|
|
|
Haemoglobin concentration (g/dl)
|
14 [13.9 - 14.4]
|
14 [13.8 - 14.3]
|
0.495 ¹
|
14.5 [11.5 - 16.1]*
|
14 [13.8 - 14.2]
|
0.790 ²
|
Hematocrit (%)
|
43 [41.9 - 43.4]
|
42.3 [41.7 - 42.9]
|
0.481 ¹
|
42.8 [42 - 43.6]
|
42.3 [41.7 - 42.9]
|
0.277 ¹
|
Platelets count (10³/μl)
|
251 [174 - 394]*
|
263 [252.5 - 272.5]
|
0.746 ²
|
255 [174 - 388]*
|
263 [176 - 355]*
|
0.982 ²
|
White blood cells count (10³/μl)
|
8 [5 - 12]*
|
7.8 [7.4 - 8.1]
|
0.206 ²
|
7.8 [5 - 11.3]*
|
7.6 [4.9 - 11.4]*
|
0.396 ²
|
For normally distrubuted continous variables data are presented as mean with 95% confidence interval otherwise as median with 5th-95th percentile range (*). Nominal variables are presented as N (percentage of the relevant group). Between groups differences were tested using: Anova ¹, Mann-Whitney U ² and chi-square ³ tests, respectively.
Intention to treat analysis:
Primary endpoint:
A statistically significant difference in the mean haemoglobin concentration in the drainage samples was found between the study groups in both the intention-to-treat (CG=1.4g/dL, TG=0.7g/dL; p=0.011) and per protocol analysis (CG=1.7g/dL, TG=0.7g/dL; p=0.001) (Table 2).
Secondary endpoints:
Table 2 presents an analysis of the data of patients. The haemoglobin mass evaluated from the drainage sample differed between the study groups (CG=0.4g, TG=0.2g; p=0.005). There was no difference in mean drainage volume (CG=40ml, TG=40ml; p=0.187). What is more, postoperative blood loss, measured as the difference between pre- and postoperative peripheral blood mean haemoglobin concentration, did not differ significantly between the groups (CG=0.7g/dL, TG=0.5g/dL; p=0.103). No significant difference was found in the mean procedure time (CG=55min, TG=54min; p=0.880). An extended postoperative stay, defined as a deviation from discharge on the first day after the procedure, was noted in 7 patients in the control group; 3 patients in the test group required a prolongation of hospital stay. There was no significant difference between the study groups regarding blood transfusions. In the control group, 2 patients required a transfusion of a total of 5 red blood cell packs and 2 freshly frozen plasma packs; in the study group, 1 patient required a transfusion of 3 red blood cell packs. No VTE or peripheral arterial thrombosis events were noted at the 1-month follow-up visit in both groups.
Table 2: Postoperative outcomes (Intention-to-treat and per-protocol analysis).
|
|
Intention-to-treat
|
|
|
Per protocol
|
|
|
CG (n=123)
|
TG (n=115)
|
p-value for group comparison
|
CG (n=102)
|
TG (n=136)
|
p-value for group comparison
|
Procedure time (min)
|
55 [40 - 90]*
|
54 [39 - 91]*
|
0.880 ²
|
55 [40 - 88]*
|
55 [40 - 95]*
|
0.408 ²
|
Postoperative lab tests from peripheral blood sample:
|
|
|
|
|
|
|
Haemoglobin concentration (g/dl)
|
13.5 [13.2 - 13.7]
|
13.5 [13.3 - 13.8]
|
0.751 ¹
|
13.8 [10.9 - 15.4]*
|
13.5 [13.3 - 13.7]
|
0.407 ²
|
Hematocrit (%)
|
40.6 [39.9 - 41.4]
|
40.9 [40.3 - 41.5]
|
0.613 ¹
|
40.7 [39.9 - 41.5]
|
40.8 [40.2 - 41.4]
|
0.873 ¹
|
Platelets count (10³/μl)
|
255 [171 - 383]*
|
258 [247.4 - 268.6]
|
0.888 ²
|
257.5 [171 - 371]*
|
257 [162 - 346]*
|
0.913 ²
|
WBC count (10³/μl)
|
10.9 [10.4 - 11.4]
|
11.1 [10.6 - 11.5]
|
0.703 ¹
|
10.9 [10.4 - 11.4]
|
11.1 [10.7 - 11.6]
|
0.437 ¹
|
Difference in haemoglobin concentration (g/dl)
|
0.7 [0.5 - 0.9]
|
0.5 [0.4 - 0.7]
|
0.103 ¹
|
0.7 [0.5 - 0.9]
|
0.5 [0.4 - 0.7]
|
0.089 ¹
|
Difference in hematocrite (%)
|
2 [1.5 - 2.5]
|
1.4 [1 - 1.9]
|
0.072 ¹
|
2.1 [1.6 - 2.6]
|
1.5 [1.1 - 1.9]
|
0.065 ¹
|
Difference in platelets count (10³/μl)
|
1.5 [-4.1 - 7.1]
|
7 [-51 - 47]*
|
0.340 ²
|
3 [-2.9 - 8.9]
|
6 [-52 - 47]*
|
0.771 ²
|
Difference in WBC count (10³/μl)
|
-2.9 [-3.4 - -2.5]
|
-3.3 [-3.7 - -2.9]
|
0.245 ¹
|
-3 [-3.4 - -2.5]
|
-3.2 [-3.6 - -2.8]
|
0.428 ¹
|
Postoperative lab tests from drainage sample:
|
|
|
|
|
|
|
Drainage volume (ml)
|
40 [10 - 140]*
|
40 [5 - 101]*
|
0.187 ²
|
40 [5 - 160]*
|
40 [5 - 101]*
|
0.520 ²
|
Haemoglobin concentartion (g/dl)
|
1.4 [0.1 - 7.1]*
|
0.7 [0.1 - 6.1]*
|
0.011 ²
|
1.7 [0.1 - 7.7]*
|
0.7 [0.1 - 6.1]*
|
0.001 ²
|
Hematocrit (%)
|
4 [0.1 - 22.9]*
|
1.8 [0.1 - 19.4]*
|
0.029 ²
|
5 [0.1 - 23.4]*
|
1.8 [0.1 - 19.4]*
|
0.003 ²
|
Haemoglobin mass (volume x concentration) (g)
|
0.4 [0 - 5.6]*
|
0.2 [0 - 3.4]*
|
0.005 ²
|
0.6 [0 - 9.6]*
|
0.2 [0 - 3.5]*
|
0.006 ²
|
For normally distrubuted continous variables data are presented as mean with 95% confidence interval otherwise as median with 5th-95th percentile range (*). Nominal variables are presented as N (percentage of the relevant group). Between groups differences were tested using: Anova ¹, Mann-Whitney U ² and chi-square ³ tests, respectively.
Subgroup analyses:
As mentioned in the study protocol, due to clinical relevance, 2 additional subgroup analyses were performed. The first analysis considered the preoperative presence of metabolic syndrome components during the examination of patients on admittance to the hospital (obesity, arterial blood hypertension, type 2 diabetes, hyperlipidemia), known to increase the general complication rate in bariatric surgery. The analysis was carried out in subgroups characterized by the sum of metabolic syndrome components. Patients with a higher number of obesity complications had a significantly smaller drainage volume after TXA administration intraoperatively (Figure 2); however, no significant differences were noted in haemoglobin concentration and mass in drainage samples in those subgroups.
An additional subgroup analysis was created considering preoperative INR values. Patients with INR<1 did not have any significant differences in volume, haemoglobin concentration and mass in the drainage sample. However, patients with the INR value >1 had significantly smaller haemoglobin concentration and mass in the drainage sample (Figure 3).
Unintended effects and harm
The Clavien-Dindo scale was used to evaluate complications during the obligatory 1-month follow-up visit. Events considered as major complications (CD grade ³3) were noted in 6 patients in the control group, and 3 patients in the test group. All complications are listed in Supplementary Table 1. Additionally, there was no need for upper endoscopy for bleeding in both study groups.
Per protocol analysis:
Both primary and secondary outcomes were analysed as stated above for the patients allocated to control group (CG) and test group (TG) according to per protocol principle. Results are presented in Table 2.