All statistical analyses were performed using SPSS (IBM, Version 31). Additional verification of results and figure generation were conducted in Python (version 3.11). Normality of all continuous variables was assessed using the Shapiro–Wilk test. Data following a normal distribution were expressed as mean ± SD, and non-normally distributed data as median [IQR, 25th-75th percentile]. Within-subject comparisons of continuous variables across time points were conducted using repeated-measures ANOVA for normally distributed variables or Friedman test for non-normally distributed data. Post-hoc pairwise comparisons were performed using paired T-tests or Wilcoxon signed-rank tests, respectively. Between-group comparisons were analyzed using Mann–Whitney U tests for both absolute values and relative changes from baseline (baseline = pre, Δ values). To increase group sizes and reduce the number of necessary comparisons, the end-tidal concentrations of desflurane and isoflurane, which were used in only a few cases, were converted into equipotent sevoflurane concentrations [26]. Associations between continuous variables were assessed using Spearman rank correlation for both absolute and Δ values. To evaluate potential additive or interaction effects of volatile anesthetics and opioids on HR and MAP, a two-way analysis of variance was performed with both agents as factors and a nonparametric aligned-rank transform ANOVA was applied. Both absolute values and Δ values were analyzed at each time point. To account for potential confounding by catecholamine administration, time-matched partial Spearman correlations were computed, evaluating the relationship of catecholamine infusion rate and hemodynamic parameters while controlling for opioid dose and volatile anesthetics concentration, and vice versa. All variables were rank-transformed before partial correlation, and residuals after regression on covariates were correlated to yield partial Spearman’s ρ. All tests were two-tailed. To control for multiple testing, Bonferroni correction was applied, and a corrected p ≤ 0.05 was considered statistically significant. The detailed description and results of all subsequent statistical analyses are provided in the Supplementary Material.

Since no evidence-based recommendations are available, management was mainly determined by the attending senior anesthesiologist. Volatile anesthetics were administered in 41% DBD donors during organ retrieval, with sevoflurane used in all but three cases. No intravenous hypnotic agents were used in any of the cases. Sufentanil was the only opioid used and was administered in 80% of the cases prior to incision, while NMBA were used prior to incision in 92% (Table 2). At least one catecholamine was used in 88% of the cases (Table 3).

Infusion rates of catecholamines were analyzed across all time points. Group sizes for epinephrine, dobutamine, and dopamine were small, and therefore the Friedman test was used as a more robust approach although distributions appeared to be normal. Only vasopressin infusion rates showed a significant difference over time (p = 0.016, partial η² = 0.171). However, post hoc analysis showed no significant differences between time points (p ≥ 0.206 for all, Cohen´s d_z = 0.243–0.512) (Table 4; Supplementary Table S1).

HR remained stable throughout the period from before incision (pre) until 15 min after incision (Figure 1A; Supplementary Table S2). There were no significant differences between time points (p = 0.308, Kendall´s W = 0.014). A significant difference in MAP was found across all time points (p = 0.007, Kendall´s W = 0.042), with post hoc analysis revealing only a modest decrease at 15 min compared with 5 min after incision (p = 0.034, r = −0.317) (Figure 1B; Supplementary Table S2). Six (7%) DBD donors exhibited a MAP <65 mmHg at one time point, seven (8%) at two consecutive time points, and two (2%) at four time points.

Sufentanil administration showed no significant influence on either HR or MAP at any time point. Across all comparisons, median absolute HR values were comparable between donors with and without sufentanil (p = 1.000 for all, r = −0.092 to 0.024) (Figure 2A; Supplementary Table S3). For absolute MAP values, no time point reached significance, though a weak tendency toward higher values at 15 min in the group with sufentanil was observed (p = 0.345, r = −0.287) (Figure 2B; Supplementary Table S3). Changes from baseline HR (Δ HR, p = 1.000 for all, r = −0.116 to 0.064) and MAP (Δ MAP, p ≥ 0.600 for all, r = −0.228 to −0.156) were likewise similar between groups (Figures 2C,D; Supplementary Table S3).

Correlation analyses revealed no clear dose-dependent effects of sufentanil on HR or MAP when considering absolute values (p ≥ 0.754 for all, ρ = −0.003–0.138). When changes relative to baseline were analyzed, no time point reached statistical significance for Δ HR (p ≥ 0.919 for all, ρ = −0.073–0.132). For Δ MAP, a weak positive association between sufentanil dose and Δ MAP was observed at 0 min (p = 0.079, ρ = 0.252) and 10 min (p = 0.094, ρ = 0.245) (Supplementary Table S3).

In contrast, sevoflurane administration was associated with consistently lower HR and MAP.

Absolute HR values were significantly lower in the sevoflurane group at all time points (p ≤ 0.001 for all, r = 0.470–0.571). Absolute MAP values were likewise reduced, but only significantly at 0 min (p = 0.020, r = 0.386), and with trends at pre (p = 0.093, r = 0.302), 5 min (p = 0.114, r = 0.292), 10 min (p = 0.075, r = 0.312), and 15 min (p = 0.057, r = 0.325), respectively (Figures 3A,B; Supplementary Table S4). No significant changes were observed for Δ HR (p = 1.000 for all, r = −0.114 to 0.048) or Δ MAP (p ≥ 0.193 for all, r = 0.030–0.253) at any time point (Figures 3C,D; Supplementary Table S4).

Dose–response analyses confirmed a negative correlation between sevoflurane concentration and both HR and MAP. For absolute HR values, correlations were strongly negative across all time points (p < 0.001 for all, ρ = −0.499 to −0.415). For absolute MAP values, significant negative correlations occurred at pre, 0 min, 10 min, and 15 min (p ≤ 0.032 for all, ρ = −0.322 to −0.294), and a trend was seen at 5 min (p = 0.059, ρ = −0.272). There were no significant correlations for Δ HR (p = 1.000 for all, ρ = −0.026–0.123) or Δ MAP (p ≥ 0.270 for all, ρ = −0.199 to −0.040) (Supplementary Table S4).

Across all time points, no significant interaction effects between sufentanil and sevoflurane were observed for absolute values of HR (p ≥ 0.466 for all, partial η² = 0.007–0.035) and MAP (p = 1.000 for all, partial η² = 0.001–0.005) (Figures 4A,B; Supplementary Table S5). Analyses of relative changes yielded similar results, with no interaction between sufentanil and sevoflurane on the magnitude or direction of changes in Δ HR (p ≥ 0.887 for all, partial η² = 0.000–0.018) or Δ MAP (p ≥ 0.453 for all, partial η² = 0.003–0.031) (Supplementary Table S5).

Since norepinephrine was the predominant catecholamine, and dosing of all catecholamines did not change significantly over time, subsequent analyses focused solely on the effects of norepinephrine. Partial correlation analyses revealed time-specific relationships between norepinephrine infusion rate and hemodynamics. After controlling for sufentanil and sevoflurane, norepinephrine positively correlated with HR in the early phase from pre to 5 min (p ≤ 0.038 for all, ρ = 0.287–0.318) and negatively with MAP during the later phase from 10 min to 15 min (p ≤ 0.032 for all, ρ = −0.303 to −0.294) (Supplementary Figures S1A, S1B; Supplementary Table S6). Thus, higher norepinephrine rates were associated with increased HR and decreased MAP at corresponding times. No significant partial correlations were found between sufentanil dose and HR or MAP after norepinephrine adjustment (p ≥ 0.547 for all, ρ = −0.048–0.175) (Supplementary Figures S1C, S1D; Supplementary Table S6), whereas the inverse relationship between sevoflurane concentration and HR remained significant throughout all time points (p < 0.001 for all, ρ = −0.502 to −0.414), and significant for MAP for 0 min, 10 min, and 15 min (p ≤ 0.033 for all, ρ = −0.316 to −0.292) (Supplementary Figures S1E, S1F; Supplementary Table S6).

Only few studies have systematically examined hemodynamic responses to surgical stimuli in DBD donors. The earliest available investigation reported an immediate increase in HR and MAP after incision in 10 donors but provided detailed information on anesthetic agent or catecholamine administration for only one case. In this donor, HR and MAP decreased 11 minutes after incision following administration of enflurane, while no opioids or catecholamines were given. In all other cases, only an unspecified reduction in catecholamine dosing was described, with HR and MAP normalizing within 25 min in all donors [13]. A subsequent case report documented similar increases in HR and MAP after incision in two donors but did not specify anesthetic or catecholamine use. Notably, only these two analyzed cases of a total of 30 reviewed DBD donors exhibited a hemodynamic response [11]. Another study involving 14 DBD donors reported elevations in HR and MAP 6 minutes after incision, without administration of anesthetic agents to any donor, but with dopamine in nine cases at constant doses [12]. Fitzgerald and colleagues reported on 11 DBD donors without anesthetic agents, seven of whom received dopamine. MAP increased immediately after incision and then decreased, while HR remained stable [9]. In a more recent study, donors with anesthetic agents exhibited higher HR and more frequent episodes of hypotension, although maximal MAP did not differ between groups. However, the study only distinguished between donors with and without anesthetic agent administration. Among those receiving agents (62% in total), the majority were given opioids alone (72%), while volatile anesthetics were administered in only 16% of cases. Additionally, hemodynamic parameters were not compared with baseline values prior to incision. Consequently, the observed hemodynamic effects cannot be attributed to any specific drug class or a possible reaction to surgical stimuli [4]. Taken together, these studies highlight the lack of standardized methodology for assessing hemodynamic parameters during organ retrieval, which are likely to contribute to inconsistent findings.

The observed hemodynamic responses in DBD donors have traditionally been interpreted as sympathetic reflexes to noxious stimuli mediated via spinal cord pathways in the absence of supraspinal control [20, 27]. Moreover, increases in systemic vascular resistance and endogen catecholamine levels after incision have been demonstrated [9, 12]. These findings have led to the hypothesis that anesthetic agents may attenuate such hemodynamic responses and thereby exert organ-protective effects [6].

Our findings add new evidence to this conflicting body of literature. The total cohort in our study exhibited no increase in HR or MAP over time or after incision compared with baseline. Sufentanil did not alter hemodynamic values. There was even a non-significant trend to higher MAP values in donors receiving sufentanil, which aligns with results from previous studies [18, 19]. Thus, our results do not support the use of opioids for hemodynamic control in DBD donors, contrary to prior recommendations [20].

Opioid receptors are widely distributed within the central nervous system (CNS), including the substantia gelatinosa of the dorsal horn and to a lesser extent in peripheral tissues. Within the CNS, activation of opioid receptors in the midbrain is thought to be a major mechanism of opioid-induced analgesia. They stimulate descending inhibitory pathways, which results in a reduction of nociceptive transmission from the periphery to the thalamus [28]. Peripheral opioid effects are variable and are primarily observed under conditions of tissue injury, such as inflammation or neuropathy [29]. Although opioids can exert direct inhibitory effects when administered epidural, spinal or peripheral, clinically relevant analgesia following systemic administration seems to be predominantly mediated via central mechanisms [28]. In DBD donors, the absence of supraspinal integration and functioning descending inhibitory pathways likely limits the ability of systemically administered opioids to attenuate nociceptive transmission and autonomic responses. Moreover, the absence of effective cerebral perfusion in the brain-dead state prevents systemically administered opioids from reaching the brain, rendering potential central antinociceptive and reflex-modulating effects unlikely [6]. Hence, the isolated use of opioids during organ retrieval appears ineffective for controlling possible neurovegetative reflexes or conferring tissue protection [4]. In contrast, sevoflurane significantly decreased HR and MAP, but these characteristics were already seen at baseline and were independent of surgical stimuli. This pattern is consistent with the general pharmacodynamic properties of volatile anesthetics, which induce vasodilation predominantly through actions at the spinal cord level, independent of any effects on the CNS [30], and may also involve direct myocardial effects or reductions in systemic vascular resistance [31]. Importantly, this effect does not reflect modulation of responses to surgical stimuli.

Administering anesthetic agents during organ retrieval could nonetheless be considered an anesthetic preconditioning strategy to attenuate IRI. Several studies have suggested that volatile anesthetics exert cardioprotective effects in cardiac surgery [32]. Under brain death conditions, these mechanisms have only been investigated scarcely. Experimental data suggest that pharmacological modulation could possibly mitigate IRI [33], and volatile anesthetics have been proposed to exert similar protective effects through distinct pathways [14, 17]. One single clinical study showed that preconditioning with sevoflurane during organ retrieval improved liver graft function [34], whereas a later study was unable to confirm a protective effect in early or long-term graft survival for kidney, liver, lung, or heart transplantation [16]. Another study likewise found no association between the use of volatile anesthetics and/or opioids and kidney graft function [4]. Overall, both approaches - administering or omitting anesthetic agents - lack robust clinical evaluation and high-quality evidence regarding their impact on graft outcomes [4].

Norepinephrine was the predominant catecholamine used in our cohort, while other catecholamines were administered in only a few donors, consistent with previous findings [4, 8]. Whether their use was intended as part of an organ-protective strategy, for hemodynamic stabilization, or simply continued from preexisting ICU therapy cannot be determined in the retrospective setting of our study. Overall, catecholamine dosing did not change significantly over time, and a MAP ≥65 mmHg was maintained in most donors throughout the observation period.

Very few other studies have investigated hemodynamic management during organ retrieval. One study reported that the target MAP of 65 mmHg was not maintained for a considerable period in 62% of DBD donors [4]. In another study, 24% of donors experienced hypotension lasting from a minimum of 10 min to a maximum of 96 min [8]. Interestingly, our data show that shortly after incision, a decline in MAP was particularly evident in the cohort receiving VA. Notably, higher norepinephrine doses correlated with lower MAP values, likely reflecting the reactive use of vasopressors in response to hypotension rather than a direct hypotensive effect of norepinephrine.

This phenomenon is consistent with earlier studies demonstrating a rapid decline in endogenous catecholamine levels shortly after incision [9, 11]. The combination of reduced catecholamine concentrations and the vasodilatory properties of volatile anesthetics may increase the risk of hypotension and should therefore be carefully considered when volatile anesthetics are administered. Opioids, in contrast, did not contribute to hemodynamic modulation in our cohort.

Continuous waveform monitoring enables real-time detection of hemodynamic changes in critically ill patients and is essential for timely intervention [35]. Its use in DBD donors has not been described in the literature but may represent a promising approach to ensure hemodynamic stability during organ retrieval and to improve understanding of the effects of anesthetic agents and catecholamines in this setting.

Beyond physiological considerations, the choice to administer anesthetic agents may reflect clinicians’ unease with the concept of brain death—a difficulty frequently noted in the literature [24, 25]. Administering anesthesia to a DBD donor for reasons other than a potential protective effect risks undermining the conceptual and ethical clarity of brain death and could have detrimental implications for both public perception and healthcare professionals [36]. Consequently, Turner suggested that the term ‘anesthesia´ itself should perhaps be avoided in this context to minimize misunderstanding of the transplantation process [37].

Potential confounders could not be controlled for, such as details of preceding ICU management strategies or preexisting cardiovascular or other organ specific conditions. Hemodynamic measurements were referenced to the time of the first incision, which may not correspond to the most relevant noxious stimulus. The administration of other drugs (e.g., corticosteroids) or fluids was not considered, although these may have influenced hemodynamic responses to an unknown extent. As a single-center study, generalizability is limited, and management strategies may also have evolved over the long observation period, introducing potential heterogeneity. The decision to administer anesthetic agents was not randomized and may have been influenced by clinical judgement or donor characteristics. Finally, graft outcomes could not be analyzed. Therefore, no conclusions can be drawn regarding the potential effect of anesthetic management on transplantation results.