The Temporal Pattern of Fibrinogen and Albumin Synthesis Rates Perioperatively in Major Abdominal Surgery


 Background: Plasma fibrinogen and albumin concentrations decrease initially after abdominal surgery. On postoperative days 3-5 fibrinogen concentration returns to preoperative level or even higher, while albumin stays low. Fibrinogen and albumin synthesis may be affected differently following major abdominal surgery related to injury and/or loss of liver tissue.The objective of the study was to determine fibrinogen and albumin quantitative synthesis rates in patients undergoing hepatectomies or pancreatectomies with intact liver size.Methods: Patients undergoing liver or pancreatic resection (n=9+6) were studied preoperatively, on postoperative days 1 and 3-5. De novo syntheses of fibrinogen and albumin were determined using the flooding dose technique with 2H5-phenylalanine as a tracer. In addition, several biomarkers indicative of fibrinogen utilization were monitored.Results: After hemi-hepatectomy, fibrinogen synthesis was 2-3-fold higher on postoperative day 1 than preoperatively. On postoperative days 3-5 the synthesis level was still higher than preoperatively. Following major liver resections albumin synthesis was not altered postoperatively compared to preoperative values. After pancreatic resection, on postoperative day 1 fibrinogen synthesis was 5-6-fold higher than preoperatively and albumin synthesis 1.5-fold higher. On postoperative days 3-5 the synthesis levels decreased towards the preoperative levels.Conclusions: De novo synthesis of fibrinogen was markedly stimulated on postoperative day 1 after both hepatectomies and pancreatectomies, while albumin synthesis remained grossly unchanged. The less pronounced changes seen following hepatectomies were possibly related to the loss of liver tissue.


Background
Major liver resections have become increasingly safer due to the improvement of patient selection and surgical techniques. Still, liver surgery is associated with a risk for postoperative liver insu ciency, which is proportional to the resection size and quality of the remnant liver (1,2). The degree of postoperative liver failure is associated with mortality and is the main cause of fatalities after liver surgery (2)(3)(4).
Computer tomography volumetric analysis or magnetic resonance imaging alone or combined with hepatobiliary scintigraphy are used to estimate the functional reserve of the liver and indirectly to assess the risk for postoperative liver failure (5)(6)(7)(8)(9). However, the post-resection functional status of the liver, in particular the synthetic function for export proteins of the remnant liver has not been explored.
Fibrinogen is a positive acute-phase reactant which is synthesized mainly in the liver, and after trauma and surgery increases in plasma concentration can be noticed (10,11).
In a previous exploratory study of patients undergoing hepatic surgery, focusing on coagulation, we found that restitution of the reduction of plasma brinogen concentration seen on the rst postoperative day, was related to the size of resection (12). Patients undergoing hemihepatectomy exhibited an overshoot in plasma concentration on the 4th postoperative day, while patients undergoing extended hemihepatectomy did not. This postoperative pattern is also in agreement with other reports (13).
Plasma concertation of albumin on the other hand falls rapidly, within hours after the start of major abdominal surgery (14), and remains low in the early postoperative days although the absolute synthesis rate is not altered perioperatively (15,16).
In principle changes in plasma protein concentration are attributable to changes in synthesis, degradation, and redistribution (17,18). A possible explanation of our ndings could be that the smaller liver remnant after an extended resection was not capable to reproduce plasma brinogen levels typical of major abdominal surgery because of a compromised synthesis capacity of export proteins.
To explore the mechanism behind the difference in postoperative plasma concentration of brinogen related to hepatic resection, we designed a study to assess the de novo synthesis rates of two liver export proteins brinogen and albumin in patients undergoing liver resections of variable sizes. Patients undergoing pancreatic resections with intact liver size served as controls. We took in parallel biochemical tests to explore the brinogen catabolic pathways and coagulation tests. Thromboelastometry was used to explore coagulability which is affected by brinogen (12,19,20) and, possibly, by albumin plasma concentrations (21,22).

Subjects
Patients with an indication for hepatectomy or pancreatic resection at Karolinska Huddinge University Hospital, Stockholm, Sweden, were recruited between September 2017 and May 2019. Smokers and patients with coagulopathies' or under treatment with anticoagulants or platelets aggregation inhibitors were excluded. The study was registered at ANZCTR, ID ACTRN12617000749303.

Protocol
De novo synthesis rates of brinogen and albumin were studied in a longitudinal protocol at three times: preoperatively and on postoperative day 1 and postoperative day 3-5. The patients who underwent portal vein embolization with two-stage hemi-hepatectomy were studied preoperatively and after the second stage of the operation (postoperative day 1 and 3-5). At the same time, blood samples for coagulation and biochemical tests were collected.
All procedures in conjunction with anesthesia, surgery, and postoperative care were in accord with the local routines at Karolinska University Hospital. The general anesthesia was performed with propofolfentanyl induction and maintenance with sevo urane-fentanyl. Epidural analgesia was administrated intra-and postoperatively for 5-7 days. Intraoperatively uids were administrated intravenously in the form of glucose 2.5% 1 ml/kg/hour and Ringer acetate 4 ml/kg/hour. All patients received daily thrombosis prophylaxis with 5000 IU dalteparin (Fragmin®) at 8:00 p.m. under the entire hospitalization starting the day before the operation.
Preoperatively the patients were investigated in a post-absorptive state (more than 6 hours after a warm meal, and more than 4 hours after coffee or tea. Tap water was allowed without restriction). Postoperatively enteral nutrition was initiated as soon as possible and was stopped 6 hours before measurement. Intravenous glucose solution, if given, was paused not less than 2 hours before measurement. The intravenous crystalloid solution Ringer acetate 1-2 ml/kg/hour was administrated postoperatively.
Analytic methods for de novo protein synthesis We used the ooding dose technique with 2 H 5 -phenylalanine for the determination of synthesis rates of brinogen and albumin which was described previously (17,23).
Gas chromatography-mass-spectrometry (Agilent 5975C, Agilent, Kista, Sweden) was used to assess i) the enrichment of 2 H 5 -phenylalanine in the free phenylalanine pool and ii) the enrichment of 2 H 5phenylalanine in brinogen and albumin.

Calculation of fractional and absolute synthesis rates
Fractional synthesis rates (FSR), representing the ratio (in percentage) of the intravascular protein mass synthesized per 24 hours, were calculated by dividing the increase of enrichment of the 2 H 5 -phenylalanine incorporated in the protein ( brinogen or albumin) to the area under the curve for 2 H 5 -phenylalanine in its free pool (adjusted for the secretion time of the protein) and multiplied by 100 (17,23).
The absolute synthesis rates (ASR), representing the total amount of protein synthesized per day (mg/kg/day), were calculated by multiplying corresponding FSR with the total plasma protein ( brinogen or albumin) mass, divided by the bodyweight (23). Plasma volume was estimated using a dedicated formula based on anthropometric data (24).

Statistics
Normality was tested with Shapiro-Wilk test. The repeated measures analysis of variances within groups was made by one-way ANOVA, or mixed-effects model-analysis (in case of missing values), and between groups with mixed-effects model analysis. Tukey's method was used for multiple comparisons. Pearsons's correlation test was used for normally distributed variables and Spearman's correlation test for non-normally distributed variables. The level for acceptable statistical signi cance (p) was < 0.05.
For the statistical analysis was used GraphPad Prism 8 (GraphPad Software Inc, La Jolla, CA).

Results
In total 15 patients were enrolled in two groups: 1) large liver resections (n = 9) and 2) pancreas resections (n = 6). Patients' characteristics, diagnosis, and surgical interventions are given in Table 1.  Mixed-effects model analyses between the two groups are given in Fig. 3 (panel A). The differences were not statistically signi cant for trends of brinogen plasma concentrations. Differences in albumin plasma concentration trends between the two groups were attributed to the lower preoperative values in patients undergoing liver surgery.

De novo synthesis of brinogen and albumin
Results for brinogen synthesis are given in Fig. 1B. The fractional synthesis rates were markedly increased on postoperative day 1 for both groups of patients, followed by a relative lowering of values on postoperative days 3-5. The interindividual scatter was larger, so the values on postoperative days 3-5 remained more marginally elevated for hepatectomies but were not different from preoperative ones for pancreatectomy. When values for absolute synthesis rate were calculated, they followed the same pattern.
Albumin synthesis rates are given in Fig. 2B. For the hepatectomy group the synthesis rates were unaltered on postoperative day 1, but the fractional synthesis rate increased on postoperative days 3-5.
The pancreatectomy group showed an increase in both fractional and absolute synthesis rate of albumin on postoperative day 1, which remained higher for the fractional synthesis rate, but not for the absolute synthesis rate on postoperative days 3-5.   . The temporal pattern of brinogen and albumin plasma concentrations (panel A) and synthesis rates (panel B) in patients undergoing major liver surgery (n = 9) (depicted in red) and pancreas resection (n = 6) (depicted in blue). Values are provided as the mean + 0.95 con dence intervals and the level of statistical signi cance for mixed-effects model analysis between groups are given in the upper right corner.
Mixed-effects model analyses between the two groups are given in Fig. 3 (panel B). There were no signi cant differences between pancreas and liver surgery regarding the synthesis of brinogen, but for albumin the synthesis rate trends were different.

Biochemical and coagulation analyses
Laboratory analyses are given in Table 2. The bilirubin plasma concentrations were increased following liver surgery during the entire observed postoperative period, while in pancreas surgery they were unchanged. For PT-INR there were increased values both for hepatectomy and pancreatectomy patients especially on postoperative day 1. APTT and platelet count were not altered postoperatively regardless of the type of operation. Levels of statistical signi cance (p 0 ) for the ANOVA (or mixed-effects analysis within groups) and (p) for mixed-effects model analysis between groups are given. (L = liver surgery, P = pancreas surgery). For multiple comparisons *, **, ***, **** denotes the statistical signi cance p < 0.05, p < 0.01, p < 0.001 and < 0.0001 respectively of the difference from preoperative values, while # have the same signi cance from postoperative day 1.
Following both liver and pancreas surgery thrombin-antithrombin complexes (TAT) levels were higher on postoperative day 1 than preoperatively, decreasing towards normalization on postoperative days 3-5. Soluble brin concentrations were increased on postoperative day 1 only in liver surgery, while the brinopeptide B (fpB) levels were unaltered in both groups regardless of the postoperative day. D-dimer concentrations increased postoperatively, more accentuated following liver surgery. There were no signi cant differences between liver and pancreas surgery regarding the coagulation markers except for soluble brin. A graphical representation of data for TAT, soluble brin, fpB and D-dimer is given is given in an additional le (see Additional le 1).

The results from the ROTEM® analyses
The thromboelastometric results are given in an additional le (see Additional le 2). For both groups, the mean values of most ROTEM® parameters remained within the normal range during the entire study period. The exception was MCF-FIBTEM values which tended to exceed the hypercoagulability on POD 3-5. There were no signi cant differences between the two groups regarding the temporal pattern.
The small temporal changes regarding CFT and MCF were parallel with the brinogen plasma concentrations and not with brinogen synthesis rates which were moving in opposite directions (see Additional le 2).

Correlations
At any of the three determinations in liver or pancreas surgery, there were no correlations between brinogen plasma concentrations (or changes in plasma concentrations) and FSR or ASR of brinogen, or between plasma albumin concentrations and FSR or ASR of albumin.
No correlations were found between the ROTEM parameters and brinogen or albumin synthesis rates in any group at any time.
There were no correlations between bilirubin plasma concentrations or PT-INR and brinogen or albumin synthesis rates regardless of the type of operation.

Hypercoagulability signs
On POD 3-5 most patients who underwent pancreatectomy had plasma brinogen concentrations, as well MCF FIBTEM, over the normal range (see Additional le 2).
Two patients operated with distal pancreatectomy had on POD 4 and 5 respectively the ROTEM parameter MCF increased over the normal in all three curves INTEM, EXTEM and FIBTEM, and brinogen plasma level was over 6 g/L. In one of these patients, a routine CT scan on postoperative day 8 disclosed a fresh portal vein thrombosis.

Discussions
The focus of our present study was to explore how plasma concentrations and de novo synthesis of two major hepatic export proteins were affected perioperatively by major abdominal surgery with and without a reduction in liver mass.
Following both major liver resections and pancreatic resections there was a dramatic increase in the de novo synthesis rate of brinogen on postoperative day 1, most pronounced in the pancreatectomy group with an intact liver mass. As far as we are aware, this is the rst time this physiological response has been described. Postoperative levels (or change in levels) of plasma brinogen concentrations were not indicators for levels (or change in levels) of synthesis. On the contrary, plasma concentrations were marginally decreased (in all subjects undergoing hepatectomy) or unaltered (pancreatectomy) on postoperative day 1, which was in agreement with previous studies (12,13,25). At this time-point the plasma brinogen concentrations likely were on an upward trend, as their nadir, according to previous studies, is at the end of the operation (13,26). On postoperative days 3-5 this increase in de novo synthesis was levelled off, but plasma concentration showed an overshoot compared to preoperative values in the pancreatectomy patients with the intact liver mass. The indirect biomarkers of brinogen utilization showed an activation on postoperative day 1 (TAT and soluble brin), and a more prolonged activation (D-dimer), where the difference between hepatectomy and pancreatectomy groups was equivocal. For coagulation competence, the global measures via tromboelastometry showed balanced coagulation for both groups of patients.
The concentrations of plasma brinogen on postoperative day 1 can only be explained through the enhanced use of important amounts of de novo synthesized brinogen, which was on average 2 g/day for hepatectomies, and 6 g/day for pancreatectomies. Taking into consideration the brinogen half-time, which in normal conditions is around 4 days (27), this additional output of brinogen should have been noticed on plasma concentrations unless a fast utilization occurred. The brinogen exit from the plasma pool is commonly associated with brin formation. The utilization of brinogen in the coagulation process is increased intraoperatively compared to normal conditions (28). In our study TAT levels on postoperative day 1 indicated a certain coagulation activation and implicitly an increased brinogen utilization, to the same extent in liver and pancreas surgery. However, the levels of soluble brin monomers showed that the ongoing cleavage of brinogen by thrombin, which is exacerbated intraoperatively (29,30), decreased hastily after pancreatectomies, while following major liver surgery the process was slower. Although the magnitude of the active brin formation is hard to be estimated, as indicated by fpB levels, this seems not to be high on postoperative day 1 and onwards. There is evidence that even under trauma or surgical operations brin formation might not be the main utilization pathway for brinogen (28). Hence, alternative utilization may contribute to the plasma brinogen levels. Unfortunately, although intuited, these pathways are not described until now (31,32). An increased transcapillary escape might also be involved.
Plasma brinogen is essential for balanced coagulation. ROTEM mostly showed balanced coagulation following both types of surgery, which for liver surgery is in agreement with previous studies (12,26).
However, following pancreatectomies, increased plasma brinogen concentrations on postoperative days 3-5 affected the ROTEM rmness parameter MCF towards the hypercoagulability area. The portal vein thrombosis observed in one of the patients with signs of hypercoagulability in three ROTEM parameters might be a consequence of a pro-thrombotic state generated by high plasma brinogen concentration.
For albumin there is a well-characterized redistribution starting already during the surgical procedure, resulting in a sharp decrease in plasma concentrations (14,15). There are no major losses of albumin by degradation or loss out of the body to explain the decrease in concentration, so the initial step is redistribution to the extravascular space. De novo synthesis rates are unaffected or increased. In the hepatectomy group, the synthesis rate for albumin was unchanged under the entire explored period. By all means, taking into consideration the massive loss of liver tissue following hemihepatectomies, there was indeed an increase in albumin synthesis per tissue unit in the remnant liver. In this study on postoperative day 1 there was an increase of albumin synthesis rates in the pancreatectomy group, compared to earlier studies when they are found unaltered on postoperative day 2 (15). The increase of the novo synthesis of albumin, following pancreatectomies on postoperative day 1 over the preoperative baseline, was in absolute value around 4 g/day; this was insu cient to in uence plasma albumin concentrations.
The albumin synthesis is suggested to be regulated primarily by the colloid osmotic pressure (18).
In ammation in general has a stimulatory effect on albumin synthesis as well (33). In our study albumin synthesis was stimulated on postoperative day 1 but not on postoperative days 3-5, which, under the condition of constant low albumin plasma concentrations, favors the in ammation as a stimulatory factor.
It was within our hypothesis that any differences in plasma concentrations of brinogen between the two groups of surgical patients studied may be attributable to a difference in liver mass resulting from the surgical procedure. However, there were also some other differences between the two groups regarding brinogen utilization, so it is not possible to conclude anything to be attributable to the difference in liver mass. Still, we intended to have a comparable surgical trauma in the upper abdomen to see what was related to the surgical procedure per se and the reduction in liver mass.
Our ndings in this descriptive observational study gain their full meaning when related to their clinical contexts. A predictable postoperative rate of increase in plasma brinogen may affect the decision on brinogen supplementation. Studies performed in cardiac surgery or cystectomy with massive bleeding show that during the rst 24 postoperative hours the plasma brinogen concentrations increase at a predictable rate if brinogen is not supplemented (34,35). On postoperative day 1, they reach the same values as preoperative, regardless of the administration or not of brinogen concentrate (34)(35)(36), suggesting that the administration of brinogen in excess is disposed by the organism, or slows the synthesis. We showed that the liver synthesis of brinogen following major abdominal surgery is modulated postoperatively to eventually reach the levels of plasma brinogen to which the organism is accustomed, which are the preoperative ones. This makes the postoperative plasma brinogen concentrations predictable even following hemihepatectomies with no risk for post-hepatectomy liver failure related to low brinogen availability. Concerning albumin, in our patients following hepatectomies the tendency of albumin synthesis to increase towards postoperative days 3-5, when in ammation stimuli were, by all means, negligible, suggested a good synthetic function of the regenerating liver. Studies demonstrate that the regeneration of the liver tissue volume is considerable during the rst week after liver resections (37), but there were no reports until now regarding the in-vivo synthetic functionality of the new liver tissue in humans. The role of brinogen in liver regeneration following hepatectomies is discussed increasingly (38) and our study of brinogen synthesis in the remnant liver holds signi cance even in this area of research.
The strengths of our study are the techniques to assess the synthesis rate of liver export proteins quantitatively and to put these in relation to functional parameters related to plasma concentrations and to liver function in connection with surgical trauma. There are also some limitations to our study, including the relatively small number of subjects available, which meant that describing the synthesis trends for patients following extended hemihepatectomies was unattainable. Another notable limitation was the lack of techniques to quantitatively assess the disappearance rates for brinogen and albumin in vivo.
In this study we were able to report patterns for brinogen and albumin in terms of plasma concentrations and de novo synthesis perioperatively in conjunction with larger upper abdominal surgery with and without reduction of liver mass. For brinogen, there was a dramatic increase in de novo synthesis rate on postoperative day 1, although plasma concentrations were decreased following liver resections or unaltered following pancreatic resections. Also, on postoperative days 3-5 the hepatectomy group had a higher de novo synthesis rate of brinogen compared to the preoperatively values. By the end, the results of these synthesis modi cations were a return to the preoperative plasma brinogen concentrations in the hepatectomy group, and an overshoot in plasma concentrations of brinogen in the pancreatectomy group with an intact liver mass. Albumin de novo synthesis was maintained or marginally increased in contrast to the dramatically decreased concentrations. So, in spite of an over 50% reduction in liver mass after hemihepatectomy the characteristic response in changes in concentrations and de novo synthesis of brinogen and albumin perioperatively remained grossly similar after a major reduction in liver mass as compared to surgery of comparative size. However, the alterations were less pronounced after hepatectomy compared to pancreatectomy, possibly related to the reduced liver mass.

Conclusion
The well-known drops in plasma concentrations of brinogen and albumin immediately following major abdominal surgery were associated with increases in the de novo synthesis rates of both proteins. For brinogen the increase was dramatic resulting in a restitution of brinogen plasma concentration within a few days. The relation between the plasma concentration and upregulated synthesis rate suggested that utilization of brinogen in the postoperative period also to be increased. For albumin the well-known altered distribution volume, made plasma concentration to stay low despite the maintained or even increased de novo synthesis rate. Despite the loss of a substantial portion of liver tissue in the hepatectomy group, the above described increase in brinogen synthesis rate was not different from the pancreatectomy group with intact liver tissue.

Declarations
Ethics approval and consent to participate: The patients gave written consent after being informed about the study protocol orally and in writing. The protocol was approved by the Regional Ethics Committee in Stockholm, Sweden, and was in accordance with the Declaration of Helsinki of 2013.

Consent for publication: Not applicable
Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.