The effects of beta blockade and clonidine on persistent injury-associated anemia

Abstract

Background

Non-selective beta blockade (BB) and clonidine may abrogate catecholamine-mediated persistent injury-associated anemia. We hypothesized that critically ill trauma patients who received BB or clonidine would have favorable hemoglobin (Hb) trends when adjusting for operative blood loss (OBL), phlebotomy blood loss (PBL), and red blood cell (RBC) transfusion volume, and that the effect would be greatest among the elderly, who have higher catecholamine levels.

Methods

We performed a 4-year retrospective cohort analysis of 280 consecutive trauma patients with ICU stay ≥48 hours and moderate/severe anemia. Patients who received BB or clonidine for ≥25% of their hospital stay were grouped as the BB/clonidine cohort (n=84); all other patients served as controls (n=196). Admission and discharge Hb were used to calculate ΔHb. OBL, PBL, and RBC volume were used to calculate adjusted ΔHb assuming 300 mL RBC = 1 g/dL Hb.

Results

BB/clonidine and control patients had similar age, injury severity, comorbid illness, and admission Hb. BB/clonidine patients received fewer RBCs despite greater OBL, though neither association was statistically significant. BB/clonidine patients had higher discharge Hb (9.9 vs. 9.5, p=0.029) and adjusted ΔHb (+1.0 vs. −0.8, p=0.003). Hb curves separated after hospital day 10. The difference in adjusted ΔHb between groups increased with advanced age (all patients: 1.7, ≥50y: 1.8, ≥60y: 2.4, ≥70y: 3.7).

Conclusions

Critically ill trauma patients receiving BB or clonidine had favorable Hb trends when accounting for OBL, PBL, and RBC transfusions. These findings support the hypothesis that BB and clonidine alleviate persistent injury-associated anemia, with strongest effects among the elderly.

Introduction

Anemia affects about 95% of patients who remain critically ill for three or more days ¹. Fifty-five percent of critically ill trauma patients receive a red blood cell (RBC) transfusion during their intensive care unit (ICU) stay, and about half of all transfusions are given after four days in the ICU ². Anemia of critical illness involves reduced iron bioavailability, suppression of erythropoiesis by inflammatory cytokines, hemodilution, and blood loss ^{1, 3–5}. Persistent injury-associated anemia is a unique subtype of anemia of critical illness that is observed in critically ill trauma patients, characterized by a prolonged hypercatecholamine state mediating bone marrow dysfunction that manifests as decreased red blood cell production with low reticulocyte counts despite adequate iron stores and erythropoietin levels in the peripheral blood ^6–8. Following traumatic injury, high levels of inflammatory mediators and circulating catecholamines compromise iron metabolism and inhibit erythropoiesis ^{1, 4–6, 9}. Persistent injury-associated anemia may disproportionately affect the elderly, based on evidence that plasma norepinephrine levels rise with increasing age in humans and that older mice are unable to replace lost erythrocytes as quickly as younger controls following hemorrhage ^10–12. These factors compound baseline anemia in the elderly, which affects approximately 10% of subjects age ≥65 years and approximately 20% of subjects age ≥85 years ^{13, 14}.

Preclinical studies in rodents indicate that interruption of hypercatecholamine pathways and the neuroendocrine stress response via non-selective beta-adrenergic receptor blockade (BB) or by sympathetic nervous system outflow inhibition with the alpha-2 agonist clonidine effectively restores bone marrow production of erythroid progenitors and alleviates persistent injury-associated anemia ^{7, 8, 15}. However, clinical translation has been difficult to establish, partly due to heterogeneity in practice patterns ⁹. The purpose of this study was to examine the effects of BB and clonidine on persistent injury-asociated anemia among critically ill trauma patients while accounting for confounding impact of operative blood loss (OBL), phlebotomy blood loss (PBL), and RBC transfusions on serial hemoglobin (Hb) levels. We hypothesized that critically ill trauma patients who received BB and clonidine would have improved Hb trends compared to control patients with similar injury severity.

Methods

We performed a retrospective cohort analysis of 280 trauma patients presenting to our level one trauma center during a four year period ending September 1^st, 2015. Patients were identified by searching our trauma database for all adults (age ≥ 18) who were admitted to the ICU for ≥48 hours and had moderate (< 11 g/dL for adults) or severe (< 8 g/dL for adults) anemia per World Health Organization criteria ¹⁶. This study population was chosen to simulate the rodent model of lung contusion, hemorrhage, and chronic stress that established the pathophysiology of persistent injury-associated anemia and identified BB and clonidine as potential treatment options. We excluded burn patients, outside hospital transfers, and patients with unmeasured blood loss that was unrelated to their initial injury (e.g. gastrointestinal bleed, postoperative hemorrhage). Institutional review board approval was obtained prior to study initiation and chart review.

Clinicopathologic variables were collected from our prospective institutional trauma database and retrospective review of the electronic medical record, including age, sex, Charlson comorbidity index, outpatient prescriptions for BB or clonidine, blunt versus penetrating trauma, injury severity score, head abbreviated injury scale, Glasgow coma scale score, intubation on arrival to the emergency department, vital signs, laboratory values, operations during admission, BB and clonidine administration, operative blood loss, red blood cell transfusions, and length of stay in the hospital and in the ICU.

For each patient, hospital length of stay was divided by the number of days on non-selective BB (e.g. propranolol, labetolol, carvedilol) or clonidine. BB and clonidine were administered at the discretion of the treating physicians. Patients who received BB or clonidine for ≥25% of their hospital stay were allocated to the BB/clonidine group (n = 84); all other patients were allocated to the control group (n = 196). BB and clonidine were considered in a single group because separation into BB, clonidine, and BB + clonidine groups would have created prohibitively small cohorts and reduced statistical power. Of the 84 patients in the BB/clonidine group, 68 received BB alone (81% of the BB/clonidine group, 24% of all patients), 6 patients received clonidine alone (7% of the BB/clonidine group, 2% of all patients), and 10 patients received BB and clonidine (12% of the BB/clonidine group, 4% of all patients). The average interval from admission to initial administration of BB or clonidine was 4.4 (3.7–5.1) days, range 0–18 days. The number of days on which subjects in the BB/clonidine group received BB ranged from 0–152 days. The number of days on which subjects in the BB/clonidine group received clonidine ranged from 0–34 days. The 25% cutoff was chosen based on observations that a 5–6 day course of BB or clonidine is sufficient to improve persistent injury-associated anemia in rodents, and the average hospital length of stay in this study was 19 days. The raw number of days on BB/clonidine may not have accurately represented medication effects, i.e. a patient who receives 5 days of BB/clonidine during a one week admission is different than a patient who receives 5 days of BB/clonidine over the course of one month in the ICU and two weeks of rehabilitation on a hospital ward. Although a higher cutoff value may have magnified the effects of BB/clonidine on post-injury erythropoiesis, the size of the BB/clonidine group would have decreased at higher cutoff values, eroding statistical power.

Admission and discharge Hb levels were used to calculate the change in hemoglobin (ΔHb) for each patient. To account for the effects of OBL, PBL, and RBC transfusions on Hb levels, we recorded these variables for each patient. PBL was estimated by a three step process. First, the amount of blood drawn for laboratory tests was determined. At our institution during the study period, basic metabolic panels and complete blood counts required 5–7 ml of blood, arterial blood gases required 2 ml of blood, and blood cultures required 8–10 ml of blood. Next, the number of laboratory tests performed daily for ICU and floor patients was ascertained from previous reports ^17–20. Based on these parameters, PBL was calculated by multiplying the first 30 ICU days by 55 mL/day, multiplying all additional ICU days by 13 mL/day, multiplying the number of ICU-free days by 9 mL/day, and adding all three values ^17–20.

OBL, PBL, and RBC transfusion volume were recorded in mL. Each 300 mL change in blood volume was expressed as a 1 g/dL change in Hb (e.g., 600 mL blood loss = Hb decrease by 2 g/dL, 900 mL RBC = Hb increase by 3 g/dL, etc.) based on previous studies ^21–24. Thus, ΔHb was adjusted to account for the effects of OBL, PBL, and RBC transfusion volume. For example, a patient with admission hemoglobin 12.0 g/dL and discharge hemoglobin 10.0 g/dL who spent seven days in the ICU and ten days on a hospital ward, receiving 600 mL PRBC and suffering 500 mL operative blood loss during admission would have raw ΔHb = −2.0 g/dL during admission; PBL effect: (55 mL/day × 7 days) + (9mL/day × 10 days) = 475 mL/(300 mL/1 g/dL) = +1.6 g/dL; OBL effect: 500 mL/(300 mL/1 g/dL) = +1.7 g/dL; RBC transfusion effect: 600 mL/(300 mL/1 g/dL) = −2.0 g/dL; adjusted ΔHb: −2.0 g/dL + 1.6 g/dL + 1.7 g/dL – 2.0 g/dL = −0.7 g/dL.

Statistical analysis was performed using SPSS (version 23, IBM, Armonk, NY). BB/clonidine and control groups were compared by one-way analysis of variance for continuous variables (reported as mean (95% confidence interval) and Fisher’s Exact test for discrete variables (reported as n (%)) with significance set at α= 0.05. pH values were missing at random (MAR) for 31 patients (11%); base deficit values were MAR for 33 patients (12%). Available pH and base deficit data were analyzed without imputing missing values.

Results

BB/clonidine and control patients had similar age, injury severity, comorbid illness, and admission Hb levels (Table 1). Approximately two out of three patients were male, 11% had penetrating injuries, and overall mean injury severity score was 24 (23–27). Head abbreviated injury scale scores were similar between patients in the BB/clonidine group (1.8) compared with controls (1.9, p = 0.751). In the entire study population, 32/280 patients (11%) had an outpatient prescription for BB or clonidine, without significant differences between groups (10% of the BB/clonidine group had an outpatient prescription for BB or clonidine vs. 12% of the control group had an outpatient prescription for BB or clonidine, p = 0.682). Initial blood pressures were higher in the BB/clonidine group, though the difference was not statistically significant (134 vs. 127 mmHg, p = 0.058), and indicators of cellular hypoxia (pH and base deficit) were similar between BB/clonidine and control groups on admission (pH: 7.28 vs. 7.30 respectively, p = 0.133; base deficit: 4.1 vs. 3.8 mEq/L respectively, p = 0.609).

Table 1.

Summary of patient characteristics on admission for patients who received non-selective beta blockade or clonidine on ≥25% of hospital days (BB/clonidine group) compared with controls.

Patient characteristics	BB/clonidine (n = 84)	Control (n = 196)	p
Age (years)	48 (46–52)	44 (45–50)	0.626
Male	58 (69%)	127 (65%)	0.582
Charlson comorbidity index	0.9 (0.6–1.3)	0.6 (0.4–0.8)	0.107
Outpatient BB/clonidine prescription	8 (10%)	24 (12%)	0.682
Penetrating trauma	9 (11%)	23 (12%)	0.842
Injury severity score	25 (23–26)	24 (21–26)	0.537
Head abbreviated injury scale	1.8 (1.5–2.2)	1.9 (1.7–2.1)	0.751
Glasgow Coma Scale score	10.9 (9.7–12.0)	10.6 (9.8–11.3)	0.709
Intubated on arrival	24 (29%)	49 (25%)	0.554
Heart rate	96 (91–101)	97 (94–99)	0.864
Systolic blood pressure (mmHg)	134 (128–140)	127 (123–131)	0.058
Mean arterial pressure (mmHg)	91 (87–95)	87 (84–90)	0.071
pH	7.28 (7.26–7.30)	7.30 (7.29–7.31)	0.133
Base deficit (mEq/L)	4.1 (3.1–5.1)	3.8 (3.1–4.5)	0.609
Hemoglobin (g/dL)	11.7 (11.3–12.2)	11.4 (11.1–11.7)	0.276

Data are presented as mean (95% confidence interval) or n (%).

Management parameters were also similar between groups (Table 2). On average, BB/clonidine patients had more OBL than control patients, slightly less PBL, and fewer RBC transfusions, though none of these differences were statistically significant. Discharge Hb levels were higher in the BB/clonidine group (9.9 (9.6–10.2) vs. 9.5 (9.3–9.7) g/dL, p = 0.029). Hb trends over time are illustrated in Figure 1, which demonstrates that separation between groups occurred after hospital day 10. Although raw ΔHb was not significantly different between groups, the BB/clonidine group had a smaller decrease in Hb during admission despite having more OBL and fewer transfusions, such that BB/clonidine patients had favorable adjusted ΔHb, representing change in Hb from admission to discharge, adjusted for OBL, PBL, and RBC transfusion volume as described in the Methods section (+1.0 (0.1 to +1.9) vs. −0.8 (−1.2 to −0.3), p = 0.003) (Figure 2). There were 38 patients in the control group who received BB or clonidine for <25% of their hospital stay, composing 19% of the control group. When these 38 patients were excluded, adjusted ΔHb was unchanged, albeit with a wider 95% confidence interval (−0.8 (−1.3 to −0.2), suggesting that the primary analysis was not significantly affected by allocation of patients who received BB or clonidine for <25% of their hospital stay to the control group.

Table 2.

Summary of management and outcome parameters for patients who received non-selective beta blockade or clonidine on ≥25% of hospital days (BB/clonidine group) compared with controls.

Management and outcomes	BB/clonidine (n = 84)	Control (n = 196)	p
During admission
Operations	3.2 (2.6–3.8)	2.9 (2.6–3.3)	0.451
Operative blood loss (mL)	853 (613–1,093)	721 (595–847)	0.335
Phlebotomy blood loss (mL)	609 (503–714)	587 (517–657)	0.738
RBC transfusion within 24 hours	38 (45%)	91 (46%)	0.896
Number of transfusions	2.0 (1.5–2.6)	2.2 (1.8–2.6)	0.679
RBC transfusion after 24 hours	59 (70%)	128 (65%)	0.489
Number of transfusions	2.5 (1.9–3.1)	2.8 (2.4–3.2)	0.410
Discharge hemoglobin (g/dL)	9.9 (9.6–10.2)	9.5 (9.3–9.7)	0.029
ICU length of stay (days)	11 (9–13)	12 (11–14)	0.236
Hospital length of stay (days)	19 (16–22)	19 (17–21)	0.843

RBC: red blood cell, ICU: intensive care unit. Data are presented as mean (95% confidence interval) or n (%).

Figure 1.

Patients who received non-selective beta blockade or clonidine on ≥25% of hospital days (BB/clonidine group) had a trend toward increased late hemoglobin levels

Figure 2.

Change in hemoglobin from admission to discharge was favorable among patients who received non-selective beta blockade or clonidine on ≥25% of hospital days (BB/clonidine group) when controlling for operating room (OR) blood loss, phlebotomy blood loss, and red blood cell transfusions during admission to calculate adjusted Δhemoglobin (*p = 0.003).

To assess the impact of outpatient BB or clonidine prescriptions on initial Hb levels and early changes in Hb levels, a subgroup analysis was performed in which subjects in the control group with an outpatient prescription for BB or clonidine (n=24) were compared with subjects in the control group who did not have an outpatient prescription for BB or clonidine (n=172). Interestingly, admission Hb was lower among control group patients who had an outpatient prescription for BB or clonidine (10.6 (9.7–11.5) vs. 11.5 (11.2–11.9) g/dL, p = 0.059). This finding may have been influenced by significantly older age among patients in the prescription subgroup (63 (57–69) vs. 45 (42–48) years, p < 0.001) accompanied by greater chronic disease burden (Charlson comorbidity index 1.8 (0.9–2.5) vs. 0.5 (0.3–0.6), p = 0.003). Within the control group, subjects with an outpatient prescription for BB or clonidine had a similar incidence of PRBC transfusion within 24 hours of admission compared with patients in the no prescription subgroup (46% vs. 47%, p > 0.999). Hemoglobin levels were also similar between prescription and no-prescription subgroups on hospital days one (9.8 (9.0–10.7) vs. 10.1 (9.8–10.4) g/dL, respectively, p = 0.558) and three (8.3 (8.0–8.8) vs. 8.8 (8.5–9.0) g/dL, respectively, p = 0.107).

Because persistent injury-associated anemia may disproportionately affect the elderly, subgroup analyses were performed to calculate adjusted ΔHb (representing change in Hb from admission to discharge, adjusted for OBL, PBL, and RBC transfusion volume, as described in the Methods section) for BB/clonidine and control groups stratified by age (≥40 years, ≥ 50 years, ≥ 60 years, and ≥ 70 years). The magnitude of the difference in adjusted ΔHb between groups increased with increasing patient age, though the differences became non-significant as sample sizes diminished (Figure 3). There were only 79 study patients age ≥60 and 37 study patients age ≥70, hindering statistical comparisons between BB/clonidine and control groups within these age groups.

Figure 3.

Differences in adjusted Δhemoglobin (representing change in Hb from admission to discharge, adjusted for OBL, PBL, and RBC transfusion volume, as described in the Methods section) between patients who received non-selective beta blockade or clonidine on ≥25% of hospital days (BB/clonidine group) and controls became greater with increasing patient age (*p < 0.033).

Discussion

Our data support animal studies indicating that BB and clonidine attenuate persistent injury-associated anemia ^6–8. BB/clonidine and control patients had similar age, injury severity, comorbid illness, and admission Hb levels, indicating that differences in outcomes were likely attributable to management parameters, including the influence of BB and clonidine. Although the net influences of OBL, PBL, and RBC transfusion favored higher Hb levels in the control group, the BB/clonidine group had significantly higher discharge Hb levels. When ΔHb was adjusted for the effects of OBL, PBL, and RBC transfusion, the BB/clonidine group had a significantly higher adjusted ΔHb. This observation was primarily attributable to late differences in Hb trends, after hospital day 10. In subgroup analyses stratified by patient age, trends in adjusted ΔHb indicated that older patients may have benefited most from BB/clonidine, consistent with our hypothesis that the elderly are especially vulnerable to persistent injury-associated anemia due to higher plasma norepinephrine levels, exaggerated bone marrow suppression, and baseline anemia.

To our knowledge, the only human randomized controlled trial studying the effects of post-injury BB on erythropoiesis was conducted by Bible et al. ⁹. In this pilot study, the authors randomized 45 severely injured trauma patients (mean injury severity score 29) to receive non-selective BB with propranolol versus standard care. BB was administered within 24 hours of admission after lactate was ≤4 mg/dL, and then titrated to decrease heart rate by 10–20%. The BB group had lower peripheral blood levels of hematopoietic progenitor cells and non-significant trends toward higher discharge Hb levels (9.9 vs. 9.1 g/dL) and fewer RBC transfusions (2.0 vs. 2.4). The similar but less disparate Hb levels in our study did reach statistical significance, suggesting that the Bible study was underpowered to demonstrate statistical significance in discharge Hb. OBL and PBL were not reported. This trial was performed before the potential benefits of clonidine were realized, and therefore did not assess the effects of clonidine therapy.

The major limitations of this study are its retrospective design, absence of data regarding iron metabolism, and small sample size (n = 280). It appears that a larger study would be necessary to generate a robust analysis of the interaction between age and persistent injury-associated anemia. In addition, estimating the effects of OBL, PBL, and RBC transfusion on Hb is subject to inaccuracy on several levels. However, we have no reason to believe that such inaccuracies would disproportionately affect either study group. Despite these limitations, this study supports the hypothesis that BB and clonidine therapy may benefit critically ill trauma patients who are at risk for persistent injury-associated anemia Future research should seek to build upon the trial performed by Bible et al. ⁹ by enrolling more patients, quantifying operative and phlebotomy blood losses, and assessing the effects of clonidine as well as BB on anemia. Future studies should also investigate other mechanisms by which BB may improve outcomes among critically ill patients, a phenomenon which has been observed in the presence and absence of traumatic brain injury ^{25, 26}.

Conclusions

Critically ill trauma patients who received BB or clonidine for ≥25% of their hospital stay had favorable Hb trends when accounting for OBL, PBL, and RBC transfusion. These effects were primarily attributable to late changes in Hb and were more pronounced among elderly subjects. These observations support the hypothesis that BB and clonidine abrogate catecholamine-mediated post-injury bone marrow suppression and anemia. Future adequately powered randomized control trials would help to answer this question.