Postoperative Dynamic of Leptin and Fibroblast Growth Factor 21 in 123 Patients Recovering from Cardiac Surgery

Background

Severe stress results in acute abnormalities of metabolic regulation of carbohydrates and lipids, which are detrimental to mortality and morbidity [1–3]. Patients undergoing coronary artery bypass grafting (CABG) seem particularly vulnerable. Prolonged abnormalities in carbohydrate metabolism can sustain ongoing sympathetic over-activation, inflammation, and changes in metabolome [1,2]. These conditions can accelerate arteriosclerosis and sustain sterile neuroinflammation, leading to premature neurodegeneration, denying patients the full long-term benefit of surgery [3,4].

Leptin is a critical hormone intersecting carbohydrate and lipid metabolism, with pleiotropic immunological activities [5]. Abnormal regulation of this hormone leads to the myocardium’s susceptibility to ischemia, inflammation persistence, and hyperglycemia [2,5–9]. Furthermore, leptin regulates satiety, which can be critical for maintaining an optimal body mass index (BMI) after surgery. Prior work demonstrated that leptin serum levels are diminished at 30 days after orthopedic surgery or cardiopulmonary bypass surgery [7]. Another study demonstrated an initial increase that was followed by diminishing levels, but the observation period was less than 24 h, and the study focused on the adult population [5]. The perisurgical increase in leptin levels is limited to serum and adipose tissue [10]. Changes in leptin levels can impact response to shock and infection susceptibility [8,11,12]. Consequently, these data are well aligned with the observation that preoperative abnormal leptin levels are correlated with a composite risk of the adverse cardiovascular level after cardiac surgery [13]. Fibroblast growth factor 21 (FGF-21) has a similar function, driving behavior aimed at increased carbohydrate intake. Combined, the abnormal regulation of these hormones can lead to hyperglycemia, persistent sterile inflammation in the kidney and brain, and susceptibility of the myocardium to ischemia, culminating in unfavorable long-term neurocognitive sequela [1,2]. All these effects mitigate the beneficial effects of heart surgery. It has been suggested that markers may be involved in unfavorable outcomes after cardiac surgery, including cognitive decline and unfavorable cardiac remodeling [14,15]. However, the longitudinal dynamics of the markers after major surgery are unknown. Although the hormones are released during acute stress, it is unclear if their stress-induced secretion outlasts the acute recovery from surgical stress, thereby contributing to long-term recovery from surgery.

This study aimed to investigate the longitudinal dynamics of leptin and FGF-21 in patients recovering from cardiac surgery. We hypothesized that leptin and FGF-21 abnormalities would be limited to the acute inflammatory response and follow serum interleukin 6 (IL-6) and C-reactive protein (CRP). To assess the clinical impact, we assessed serum levels of amyloids β1–40 and amyloids β1–42. In addition, considering hormone function, we investigated the effects of leptin in patients with diabetes or a higher BMI.

Material and Methods

PATIENT ENROLLMENT:

Our study protocol was approved by the Institutional Review Board (IRB) of the University of Pennsylvania and was performed according to the guidelines of the 2003 Helsinki Declaration (#815686; approved March 02, 2020).

All patients scheduled for elective heart surgery were approached for consent. We excluded patients with pre-existing immunological aberrancies who were on immunosuppressant medications in the last 6 months (oral or intravenous prednisone more than 5mg daily; αTNFα, αIL-6, αIL-3, and αCD20 antibody therapy; immunoglobulin, plasmapheresis, methotrexate, and chemotherapy). The study did not include patients with inherited and known dyslipidemias and those who had undergone transplantation.

The demographic characteristics of patients is presented in Table 1.

SAMPLE PROCESSING:

Upon patient consent, serum was isolated and stored at −80°C. Blood was collected at 4 times as follows: The baseline sample (t_baseline) was collected before or shortly after arterial or central line placement. The second sample was collected 24 h after the first sample (t_24h), during the patient’s stay in the Intensive Care Unit (ICU). The third sample was obtained at the patient’s discharge from the hospital or 7 days after the t_baseline sample (t_7d). The last sample was collected no sooner than 3 months after surgery (t_3m) but no later than 4 months after. The 4 samples represents baseline, acute stress response, convalescence, and medium-term recovery, respectively.

CLINICAL DATA:

The electronic medical records were used to collect the demographic and clinical data for all enrolled patients. The patients self-determined their race and ethnicity. Several variables regarding the duration of surgery and anesthesia were collected from the medical records retrospectively. Preoperative hemoglobin A1c (HbA1c) and lipid profiles were collected from routine preoperative laboratory tests, when available. The Acute Physiology and Chronic Health Evaluation II (APACHE II) scores were calculated within 1 h and t24h after admission to the ICU. The burden of chronic disease was calculated using the Charlson Comorbidity Index [16,17]. The severity of illness during the ICU stay was determined using the Marshalls Organ Dysfunction Score [18]. Survival was determined at 28 days and t3m from admission.

ASSESSMENT OF BIOMARKERS:

Leptin and FGF-21 were anlayzed using enzyme-linked immunosorbent assay, according to the manufacturer instructions (BioLegend, San Diego, CA, USA). In addition, inflammatory markers (IL-6, CRP) and neurodegeneration markers (amyloid β1–40, amyloid β1–42) were analyzed using multiplex technology (Theromofisher, Waltham, MA, USA) on a MagPix machine (Luminex, Austin, TX, USA) [19].

STATISTICAL ANALYSIS:

The normality of distribution of studied variables was determined using the Kolmogorov test and descriptive variables. Data were are presented as mean±standard deviation or median and interquartile ranges. The data were compared using the t test (t[n] or Mann-Whitney U test (U[n]). Analysis of variance (F[df;n]) or the Kruskal-Wallis test (H[df;n] was used for variables with multiple subgroups. Longitudinal and pairwise analysis was done in most of the statistical contrasts. Correlations were calculated with the Pearson correlation coefficient (r). Regression analysis was done using the stepwise method. A P value less than 0.05 was considered statistically significant for all tests. SPSS version 26 (IBM Corp, Armonk, NY, USA) was used for analysis.

Results

PATIENT CHARACTERISTICS AND BASELINE LEVELS OF LEPTIN AND FGF-21:

Sex or age over 60 years old did not affect serum leptin and FGF-21 levels at baseline. Asian patients had a significantly higher level of leptin (H[3;102]=9.23; P=0.029) at baseline, but only 5 patients of this race were enrolled in the study (Figure 1).

Patients with pre-existing diabetes had elevated serum leptin at t_baseline (leptin_NoDM=Me[114.72;696] vs leptin_DM=M_e[320.04;802.4]; U[101]=2.205; P=0.027). Charlson Comorbidity Index score was not correlated with leptin or FGF-21 at t_baseline. There was no correlation between HbA1c levels before surgery vs either serum baseline leptin or FGF-21. There were no significant differences between baseline leptin and FGF-21 vs BMI class.

CHANGES IN LEPTIN AND FGF-21 AFTER CARDIAC SURGERY:

Sex, age, race, BMI, and collected history of pre-existing illnesses did not affect the leptin level at t_24h, t_7d, and t_3m. Multiple regression analysis (including age, BMI, Charlson Comorbidity Index, procedure type, stay in ICU, and hospital duration of anesthesia and surgery, as well as APACHE at 1 and t_24h after admission and baseline leptin demonstrated that the preoperative level of leptin accounted for 31% variation in the postoperative level at t_24h (F[12;74]=5.44; P<0.0001), 25% at t_7d (F[12;63]=7.61; P<0.0001), and 11% at t_3m (F[12;44]=1.96; P<0.05).

We found that leptin levels increased significantly at t24h, then declined below baseline t7d after surgery, to normalize at t3m (U[60;3]=7.81; P=0.049) (Figure 2A). FGF-21 remained elevated at t24h and t7d (U[45;3]=21.38; P=0.0009) (Figure 2B). Patients undergoing mitral valve-related surgeries only had a transient increase in the acute postoperative period (P=0.033) (Figure 2C). Interestingly, patients with CABG had a somewhat steady level of leptin throughout all observation periods. Leptin and FGF-21 did not correlate significantly at t24h, t7d, and t3m.

Serum IL-6 levels (U[3;57]=62.16; P<0.00001) normalized at t3m after increases at t24h and t7d (Figure 2D). Serum CRP remained elevated even at t3m ([U3;26]=39.09; P<0.000001) (Figure 2E).

PERIOPERATIVE COURSE AND DYNAMICS OF LEPTIN AND FGF-21:

The increase in leptin and FGF-21 at t_24h and t_7d did not correlate with the duration of anesthesia, surgery, or time of the bypass. Estimated blood loss was weakly correlated with leptin level at t_24h (r=0.2; P=0.047), t_7d (r=0.399; P=0.0001), and t_3m (r=0.395; P=0.001), but not with volume of crystalloids given during the procedure. Intake of aspirin, ketorolac, and acetaminophen showed no effect on leptin or FGF-21 at any time point. No patients were given steroids in our cohort. FGF-21 at t_24h (r=0.59; P<0.0001) correlated strongly with the intake of benzodiazepines but not with that of opiates.

LEPTIN AND FGF-21 AND SHORT- AND LONG-TERM OUTCOME:

Leptin levels at t24h correlated weakly yet significantly with amyloid β1–42 but not with amyloid β1–40 at t3m (Figure 3). Central nervous system failure resulted in aberrations in FGF-21 at t24h (U[85;7]=2.42; P=0.01). Patients experiencing a stroke during surgery had lower baseline levels of FGF-21 (U[91;7]=2.24; P=0.023).

Leptin and FGF-21 serum levels did not correlate with APACHE scores at admission, t24h, and 48 h. Length of stay in the hospital or ICU did not correlate with leptin or FGF-21 levels at any time. Disposition at the discharge from the hospital indicated that FGF-21 at t24h (U[76;15]=1.97; P=0.049) and t7d (U[71;8]=2.06; P=0.038) was lower if the patient was transferred to the long-term facility versus to home discharge (Figure 4). Only 1 patient died in that group, rendering statistical analysis not feasible.

Discussion

This study encompassed a longitudinal sampling of patients who underwent heart surgery, with a follow-up at t3m. Our primary goal was to establish the relationship between serum leptin and FGF-21 longitudinal dynamics and postsurgical inflammatory and stress response, considering the function of these cytokines [9, 20–22]. We were particularly interested in the relationship of leptin and FGF-21 to presurgery BMI and pre-existing diabetes because of the role of leptin in long-term carbohydrate metabolism [2,4,5,9,21,23,24]. Pre-existing data suggested a link of leptin levels to susceptibility to pneumonia and sepsis as well as to long-term cognitive decline and cancer propagation [8,11,25–28]. FGF-21 modulates and augments several leptin functions, but no longitudinal studies of the serum hormone levels after cardiac surgery have been conducted to date [21,22].

Our study demonstrated that leptin levels peaked after around t7d and recovered to pre-baseline levels after t3m. This observation was similar to that of a prior study which detected normalization of the level by day 30 in patients undergoing cardiopulmonary bypass [5,7,13]. Our results showed that at t7d, leptin levels were lower, in contrast to another study that showed normalization of levels by day 4 [5,7,13]. Normalization by day 4 is likely part of a downward trend. We demonstrated that the type of surgery is one of the critical factors in determining the postsurgical dynamics of leptin. Patients undergoing arch repair had the highest baseline level but also had the steepest decline in leptin levels after surgery. Patients undergoing valve surgery exhibited a much more modest trend, while patients undergoing CABG had minimal leptin dynamics. The reason for these differences is unclear, but they have been related to the different etiologies leading to a particular surgery. The most common indication for arch surgery is aneurysmal changes, and leptin can accelerate the progression of this disease in animal models [29,30]. What is particularly interesting is that surgery of the arch dramatically reduces the level of leptin, suggesting removal of the source. This somewhat puzzling finding warrants further study to establish the causative effect. We expected that CABG would have the most pronounced level of leptin changes because of the role of leptin in sugar metabolism and progression of atherosclerosis via smoldering inflammation [5,8,9]. However, in the present study, CRP and IL-6 did not correlate with leptin levels, suggesting that leptin was an independent factor in postoperative inflammation. Also, the level of surgical insult measured via several perioperative factors demonstrated low correlations with leptin dynamics, suggesting that heart surgery by itself is a strong enough trigger for leptin release. This nonspecific role of leptin in modulating postsurgical response was noted in a study describing the pericardial fat release of leptin [10]. Interestingly, YKL-40 and leptin levels correlated, suggesting that ongoing tissue repair and leptin release may be interconnected, and that the serum leptin level may be potentially driven by epicardial fat [31,32]. Also, the pre-existing leptin levels were the most crucial variable in determining the postsurgical leptin response. A similar observation was made by Gu et al, in which preoperative leptin level was a risk factor for adverse effects after cardiac surgery [13]. That study was done in older Asian adults, suggesting that race may be a determining factor as it was described before [13]. This may suggest that there is a more complex interaction between long-term outcomes of surgery and race via leptin-driven mechanisms.

To date, the present study is the first long-term observation of the FGF-21 dynamic after cardiac surgery, except for one abstract that was published 10 years ago [33]. In our study, FGF-21 levels increased at t24h and fell below the pre-baseline levels at t3m. This profile was similar to that of IL-6 and CRP dynamics. Elevated FGF-21 levels correlated with central nervous system failure and the intake of benzodiazepines. Although central nervous system failure was defined alongside the clinical score, this observation was correlated with another study that showed an interaction between FGF-21 and the emergence of delirium or neurodegeneration [15,18,34]. The correlation between FGF-21 levels and the use of benzodiazepines may contribute to the difference in observed central nervous system failure. However, long-term decline was a predictor of unfavorable discharge, a finding well aligned with the critical role of FGF-21 in the metabolome, tissue healing, and neurodegeneration [34–37].

This study had several strengths. It had a large patient population, and we were able to study the presurgical state. Our sample was large and allowed us to make comparisons between baseline and later values. Because of the significant variability in baseline levels, the use of a longitudinal design was one of the most important advantages, specifically because hyperleptinemia is linked to persistent inflammation [22]. Also, we studied diverse adult patients in regards to sex, age, pre-existing medical conditions, and surgery type. We collected clinical data on the severity of surgery, perioperative management, and serum levels of inflammation markers (CRP, IL-6). Several pre-existing conditions important for leptin and FGF-21 levels were included in the data collection and analysis [15,22,24]. Finally, the methodology (enzyme-linked immunosorbent assay, multiplex) used to assess the serum level of the markers is very well established and robust, especially for measurements of IL-6, CRP, and FGF-21.

This was an exploratory study. Consequently, we could not conduct power analysis for FGF-21, in particular. In addition, our patient cohort was biased toward Whites and men; however, leptin levels are elevated in patients of Asian descent [13]. The significant limitation of the techniques used is relatively low sensitivity; however, only amyloid serum levels were close to the detection limits, potentially biasing the results. Furthermore, leptin levels are affected by glucocorticosteroids, circadian rhythm, and level of inflammation [9,38,39]. Our study did not account for these variables in an optimal way. Finally, FGF-21 levels may be affected by genetic variance [40].

Leptin and FGF are highly regulated by muscle and the liver [24]. Changes in muscle mass were not accounted for in this study. Liver function remained essentially nominal judging from liver injury markers; however, they can be a poor predictor of impaired synthetic function. There is a considerable amount of data linking leptin to immune-compromised immunity; however, in our sample, no patient experienced septic shock, and postoperative infectious complications were low [5,8,9]. Also, leptin biological activity is modulated by smoking and alcohol consumption, and we did not control for these factors. The correlations between leptin and FGF-21 and neuro-injury (amyloid) need to be quantified in more detail; however, FGF-21 seems to be generally protective in neuroinjury [19,34,35]. The abnormalities in FGF-21 and leptin may be a part of broader hyperglycemia and inflammation instead of being directly involved in the postsurgical neuroinflammation process. Next, studies should analyze the changes in leptin in conjunction with glucocorticosteroid metabolism, as leptin is dependent on steroids released during stress, in particular [9,24]. This is particularly important as exogenous steroids are used during some cardiac cases. There is a growing need to establish the role of leptin as the independent risk factor for postoperative complications in the context of pre-existing levels, race, and diabetes. Our pilot data demonstrated specific clinical correlates with neurodegeneration and cardiac remodeling markers; however, these findings must be validated with appropriate functional testing.

Conclusions

Leptin and FGF-21 evolve independently from the inflammatory response in the aftermath of cardiac surgery and correlate with cardiac remodeling and markers of neurodegeneration. Functional correlates of these correlations need to be established in the next study.