Sage Journals: Discover world-class research

Abstract

Clinically evident and subclinical peri-procedural bleeding following interventional therapies are associated with adverse cardiovascular outcomes. The risk factors for clinically evident bleeding have been well described. Despite the well-documented association of adverse outcomes for patients with a subclinical peri-procedural hemoglobin drop, the clinical predictors have not yet been defined. We identified 1176 consecutive patients with a subclinical drop in hemoglobin (fall of ≥ 1 g/dl in patients without clinical bleeding) following percutaneous coronary interventions (PCI) and peripheral vascular interventions (PVI). Multivariate logistic regression analysis was performed. A subclinical peri-procedural hemoglobin drop ≥ 1 g/dl was identified in 41% (400/972) of PCI and in 49% (213/435) of PVI. More than one access site predicted a higher risk of a subclinical drop in hemoglobin in both groups. A body mass index ≥ 30 predicted a lower risk of a subclinical drop in hemoglobin in both groups. For PCI, creatinine clearance < 60 ml/min was associated with a higher risk of a subclinical drop in hemoglobin. In conclusion, clinically silent peri-procedural hemoglobin fall ≥ 1 g/dl is common in patients undergoing both coronary and peripheral percutaneous intervention. Predictors identified in our study will need prospective validation.

Keywords

complication percutaneous interventions peri-procedural hemoglobin drop predictors

Introduction

Peri-procedural bleeding has emerged as a common complication of percutaneous coronary interventions (PCI) that negatively impacts a patient’s prognosis.^1,2 In addition to an increased mortality risk, other potential deleterious consequences of bleeding, including a hypercoagulable state, interruption of anti-platelet therapy, exacerbation of coronary ischemia and the harmful effect of a blood transfusion.^3–6

Predictors of PCI-related peri-procedural bleeding include increasing age,⁷ female sex,⁸ lower body weight,⁹ chronic renal impairment,¹⁰ history of hypertension,¹¹ baseline anemia,⁷ glycoprotein 2b3a inhibitor,⁷ low molecular weight heparin,⁷ activated clotting time > 250 seconds,^12–14 femoral as compared to radial access,¹⁵ intra-aortic balloon pump,⁷ larger sheath diameter,¹⁶ delayed sheath removal,^15,17 and procedure time greater than 60 minutes.¹⁷ These reports include patients with both clinically evident and clinically silent (subclinical) peri-procedural bleeding events.^{1,7,8,11,12,18–21} Recent studies have shown that clinically silent peri-procedural blood loss is strongly associated with increased mortality and contrast-induced nephropathy.^22,23

The predictors of a clinically silent or subclinical peri-procedural drop in hemoglobin have not been defined. We performed a retrospective analysis of a large cohort of patients undergoing both coronary and peripheral vascular percutaneous interventions to identify predictors of a subclinical drop in hemoglobin.

Methods

We reviewed the records of patients undergoing PCI and peripheral vascular interventions (PVI) from January 2007 to September 2009. The Institutional Review Board approved the study protocol. Consecutive patients, with a documented baseline and post-procedural hemoglobin were identified from the John Ochsner Heart and Vascular Institute’s catheterization laboratory database. We excluded patients with clinically evident bleeding noted in the chart, such as a hematoma, retroperitoneal hemorrhage, gastrointestinal bleeding, hemoptysis, hematuria, those with transfusions and patients who underwent next-day surgery.

Baseline clinical and demographic information as well as details of the percutaneous interventions were extracted from the electronic medical record system. PVI included cerebral, renal, mesenteric, and upper and lower extremity interventions. A history of chronic kidney disease was defined as an estimated glomerular filtration rate of < 60 ml/min [Modification of Diet in Renal Disease study formulae was used: glomerular filtration rate in ml/min per 1.73 m² = 175 × serum creatinine (exp[–1.154]) × age (exp[–0.203]) × (0.742 if female) × (1.21 if black)].24 All patients received aspirin and Plavix post-procedure.

Our primary outcome was a post-procedural hemoglobin fall ≥ 1 g/dl without a clinically evident reason. We divided the study population into a PCI group (n = 972) and a PVI group (n = 435). Patients who underwent combined coronary and peripheral interventions at the same sitting (n = 52) were excluded for the purpose of this analysis. The entire study population was followed via an electronic medical record and the social security death index was used for the 6-month adverse cardiovascular outcomes (all-cause mortality, myocardial infarction and stroke).

Statistical analysis

Demographic characteristics of the coronary and peripheral groups were analyzed using the Student’s t-test for independent continuous data and the chi-square test or Fisher’s exact test for categorical data. Some patients had more than one record in the database so generalized estimating equations (GEE) regression methods were used to derive the logistic regression models, corrected for multiple records. Separate models were generated for the PCI and PVI groups following the same strategy. We used non-directional stepwise modeling methods to find the best prediction models. Univariate analyses were used to determine whether the baseline clinical, demographic, and procedural variables listed in Tables 1 and 2 were associated with a hemoglobin drop ≥ 1 g/dl, using a liberal criterion of p < 0.20. Those variables associated with the outcome were then used as possible predictors in a non-directional stepwise procedure. All analyses were performed with SAS statistical software (version 9.1; SAS Institute Inc., Cary, NC, USA).

Table 1.

Baseline clinical and demographic characteristics

Variable (%)	Total population (n = 1176)	Coronary intervention (n = 849)	Peripheral intervention (n = 327)	p-value
Age ± SD	66 ± 13	65 ± 13	68 ± 13	0.0005
Male	761 (65)	562 (66)	199 (61)	0.0860
Race				0.9342
Caucasian	809 (69)	586 (69)	223 (68)
Black	247 (21)	178 (21)	69 (21)
Other	120 (10)	85 (10)	35 (11)
Body mass index ± SD^a	29 ± 6.5	30 ± 6.6	29 ± 6.0	0.0007
Obese (BMI > 30)^a	410 (41)	307 (43)	103 (36)	0.0268
Hypertension	997 (85)	723 (85)	274 (84)	0.5588
Diabetes mellitus	200 (17)	149 (18)	51 (16)	0.4243
Hyperlipidemia	969 (82)	712 (84)	257 (79)	0.0335
Current smoker	225 (19)	152 (18)	73 (22)	0.0842
History of chronic kidney disease	462 (33)	272 (28)	190 (44)	< 0.0001
LVEF ≤ 40%^b	205 (21)	157 (21)	48 (21)	0.8336
Hemoglobin ± SD	13.1 ± 1.8	13.3 ± 1.8	12.7 ± 1.8	< 0.0001
Platelet count	229 ± 73	228 ± 72	231 ± 77	0.6346
INR	1.1 ± 0.3	1.1 ± 0.2	1.1 ± 0.3	0.4998
Aspirin	1162 (99)	839 (99)	323 (99)	0.9487
Clopidogrel	69 (5.8)	59 (6.3)	10 (3.0)	0.0003

n = 999; 710 with coronary intervention and 289 with peripheral intervention.

n = 972; 739 with coronary intervention and 233 with peripheral intervention.

LVEF, left ventricular ejection fraction; INR, international normalized ratio.

Table 2.

Baseline clinical and demographic characteristics between patients with a hemoglobin drop more than or equal to 1 g/dl versus less than 1 g/dl

Variable (%)	Total population (n = 1176)	Hemoglobin drop ≥ 1 g/dl (n = 550)	Hemoglobin drop < 1 g/dl (n = 626)	p-value
Age ± SD	66 ± 13	67 ± 13	65 ± 13	0.0087
Male	761 (65)	334 (61)	427 (68)	0.0074
Race				0.0281
Caucasian	809 (69)	393 (71)	416 (66)
Black	247 (21)	97 (18)	150 (24)
Other	120 (10)	60 (11)	60 (10)
Body mass index ± SD	29 ± 6.5	29 ± 6	30 ± 7	0.0001
Obese (BMI > 30)^a	410 (41)	150 (33)	260 (47)	< 0.0001
Hypertension	997 (85)	462 (84)	535 (85)	0.4858
Diabetes mellitus	200 (17)	74 (13)	126 (20)	0.0024
Hyperlipidemia	969 (82)	438 (80)	531 (85)	0.0198
Current smoker	225 (19)	102 (19)	123 (20)	0.6313
History of chronic kidney disease	397 (34)	206 (38)	191 (31)	0.0079
LVEF ≤ 40%^b	205 (21)	119 (27)	148 (28)	0.8185
Hemoglobin ± SD	13.1 ± 1.8	13.4 ± 1.8	12.9 ± 1.9	< 0.0001
Platelet count	229 ± 73	235 ± 76	224 ± 70	0.0184
INR	1.1 ± 0.3	1.1 ± 0.3	1.1 ± 0.2	0.3369
Aspirin	1162 (99)	543 (99)	619 (99)	0.8074
Clopidogrel	61 (5)	32 (5)	37 (5)	0.5149

n = 999; 710 with coronary intervention and 289 with peripheral intervention.

n = 972; 739 with coronary intervention and 233 with peripheral intervention.

BMI, body mass index; LVEF, left ventricular ejection fraction; INR, international normalized ratio.

Data given as n (%).

Results

There were 849 patients who underwent 972 PCIs and 327 patients with 435 PVIs. The average time to obtain the post-procedural hemoglobin was 21.6 ± 9.7 hours. A post-procedural hemoglobin fall of ≥ 1 g/dl occurred in 44% (613/1407) of all interventions: 41% (400/972) among the PCI group and 49% (213/435) in the PVI group. Baseline clinical and demographic characteristics of the entire study population, and the PCI and PVI subgroups are shown in Table 1. Compared to PCI patients, the PVI patients were older (68 ± 13 versus 65 ± 13, p = 0.0005), had a lower pre-procedural hemoglobin (12.7 ± 1.8 versus 13.3 ± 1.8, p < 0.0001), a higher prevalence of chronic kidney disease (44% versus 28%, p < 0.0001), and a lower body mass index (29 ± 6.0 versus 30 ± 6.6, p = 0.0007).

Among PCI patients, 35% (340/972) underwent interventions for acute coronary syndromes (ACS) and 65% (632/972) were for other indications. Among PVIs, 23% were cervicocerebral (99/435), 14% (60/435) were visceral and 63% (276/435) were extremity arterial interventions. A comparison between patients with a hemoglobin fall of more than or equal to 1 g/dl versus less than 1 g/dl is shown in Table 2. Compared to PCI, those undergoing PVI more frequently had more than one access site, more catheter exchanges, larger sheath sizes, longer procedural times and more frequently received manual compression to achieve access site hemostasis. Other procedural differences are shown in Table 3. The majority of patients had a hemoglobin fall between 1 and 2 g/dl (84% PCI, 86% PVI). Significant predictors of a subclinical drop in hemoglobin for PCI and PVI procedures by logistic regression analysis are shown in Table 4.

Table 3.

Procedural characteristics

Variable (%)	Total (n = 1407)	PCI (n = 972)	PVI (n = 435)
Emergent	49 (3)	33 (3)	16 (4)
Day of procedure
Weekend	56 (4)	41 (4)	15 (3)
Weekday	1351 (96)	931 (96)	420 (97)
Time of procedure
Day	1320 (94)	899 (92)	421 (97)
Night	87 (6)	73 (8)	14 (3)
Access site
Femoral	1303 (93)	903 (93)	400 (92)
Radial	79 (6)	62 (6)	17 (4)
Brachial	25 (2)	7 (1)	18 (4)
> 1 access site	245 (17)	143 (15)	102 (24)
Catheter exchanges
≤ 3	1174 (83)	800 (82)	374 (86)
4–6	207 (15)	161 (17)	46 (11)
> 6	26 (2)	11 (1)	15 (3)
Sheath size
4–6	1099 (78)	807 (83)	292 (67)
> 6	308 (22)	165 (17)	143 (33)
Hemostasis
Closure device	918 (65)	686 (71)	232 (53)
Manual compression	489 (35)	286 (29)	203 (47)
Time of post-procedure Hgb draw	21.6 ± 9.7	20.9 ± 9.4	23.2 ± 10.3
Procedure duration (min)	58 ± 32	53 ± 27	68 ± 39
Heparin	1402 (99)	967 (99)	435 (100)
Glycoprotein 2b3a	167 (12)	161 (17)	6 (1)
Thrombin inhibitor	94 (7)	88 (9)	6 (1)

PCI, percutaneous coronary interventions; PVI, peripheral vascular interventions.

Table 4.

Logistic regression analysis for predictors of a hemoglobin fall ≥ 1 g/dl

Variable	Odds ratio	95% Confidence interval	p-value
Coronary intervention
> 1 access site	1.91	1.23, 2.98	0.0041
BMI > 30	0.64	0.47, 0.86	0.0032
Chronic kidney disease^a	1.58	1.14, 2.19	0.0065
Peripheral intervention
> 1 access site	2.35	1.34, 4.12	0.0029
BMI > 30	0.46	0.29, 0.85	0.0014

Chronic kidney disease = creatinine clearance. Estimated glomerular filtration rate (eGFR) of < 60 ml/min.

BMI, body mass index.

Variables included in multivariate analysis included age, body mass index, sex, ejection fraction, hypertension, weekday versus weekend procedure, night-time versus day-time procedure, number of catheter exchanges, number of sheath exchanges, vascular sheath size, more than one access site, radial versus other access site, manual compression versus closure device, procedure duration, antiplatelet agent, antithrombin inhibitor, glycoprotein 2b3a inhibitor use, international normalized ratio (INR) and creatinine clearance.

Discussion

Our results corroborate previously published data on subclinical bleeding.²² In our study, 41% of PCI and 49% of PVI had a peri-procedural hemoglobin fall ≥ 1 g/dl in the absence of clinically evident bleeding. More than one access site predicted a higher risk of a subclinical drop in hemoglobin in both groups. Similar to recently published data on clinical bleeding, a body mass index ≥ 30 kg/m² predicted a lower risk of a subclinical drop in hemoglobin in both groups. For patients undergoing PCI, chronic kidney disease increased the risk of a subclinical drop in hemoglobin.

Published data demonstrate a U-shaped relationship between serum hemoglobin levels and mortality in patients undergoing PCI, with a 20% increase in cumulative 2-year mortality per 1 g/dl change in hemoglobin.²⁵ Identification of risk factors for clinically evident peri-procedural blood loss allows preprocedural assessment of the patient’s bleeding risk, which provides an opportunity to modify procedural variables. One example would be choosing a radial artery access in a high-risk patient, rather than a femoral artery access site.²⁶ With recent data demonstrating an increase in the risk of adverse outcomes in patients undergoing PCI with a peri-procedural drop in hemoglobin,^22,23 identification of predictors of a clinically silent procedural drop in hemoglobin will impact daily practice.

The increased prevalence of medical comorbidities in patients undergoing PVI when compared to PCI, highlights that patients presenting for a peripheral arterial intervention are often sicker. PVI, when compared to PCI, had a longer procedural time and required larger sheaths. Both these variables have been previously identified to increase the risk of a clinical bleeding event,^16,17 underscoring the risk of peripheral vascular interventions.

In patients undergoing PCI, the two predictors for a subclinical drop in hemoglobin were similar to those identified in prior studies on clinical bleeding.^7,10,27 Chronic kidney disease was associated with a higher risk of a subclinical drop in hemoglobin^7,10 and an increased body mass index (> 30 kg/m²) was associated with a lower risk of subclinical bleeding.²⁷ For the PVI subgroup, one or more access sites predicted an increased risk and, similar to the PCI subgroup, a body mass index > 30 predicted a decreased risk of a subclinical drop in hemoglobin. Recently published data on clinical bleeding and body mass index discusses the existence of a bleeding–obesity paradox.²⁷ The pathophysiology behind this paradox remains unclear. An alteration in circulation coagulation factors has been suggested, with obesity being associated with higher levels of multiple coagulation factors such as factor VII, VIII, fibrinogen and plasminogen activator inhibitor-1.²⁸

Limitations

Our study is limited by its single center retrospective study design. In addition, we lack data on peri-procedural intravenous fluid administration (hemodilution could possibly affect the results), undocumented access site blood loss and activated clotting time values.

Conclusion

Ours is the first study to address a subclinical peri-procedural hemoglobin drop in patients undergoing PCI and PVI. We found that almost half of the patients undergoing these interventions have a subclinical peri-procedural hemoglobin drop of ≥ 1 g/dl. Predictors identified in our study will need prospective validation.

Footnotes

The research was internally funded by the John Ochsner Heart and Vascular Institute.

None of the authors has any conflict of interest related to this manuscript.

References

Manoukian

Feit

Mehran

. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY Trial. J Am Coll Cardiol 2007; 49: 1362–1368.

Ndrepepa

Berger

Mehilli

. Periprocedural bleeding and 1-year outcome after percutaneous coronary interventions: appropriateness of including bleeding as a component of a quadruple end point. J Am Coll Cardiol 2008; 51: 690–697.

Doyle

Rihal

Gastineau

Holmes

Jr . Bleeding, blood transfusion, and increased mortality after percutaneous coronary intervention: implications for contemporary practice. J Am Coll Cardiol 2009; 53: 2019–2027.

Sparrow

Veale

Healey

Payne

. Red blood cell (RBC) age at collection and storage influences RBC membrane-associated carbohydrates and lectin binding. Transfusion 2007; 47: 966–968.

Sandhu

Doyle

Singh

. Frequency, etiology, treatment, and outcomes of drug-eluting stent thrombosis during one year of follow-up. Am J Cardiol 2007; 99: 465–469.

Metivier

Marchais

Guerin

Pannier

London

. Pathophysiology of anaemia: focus on the heart and blood vessels. Nephrol Dial Transplant 2000; 15(suppl 3): 14–18.

Nikolsky

Mehran

Dangas

. Development and validation of a prognostic risk score for major bleeding in patients undergoing percutaneous coronary intervention via the femoral approach. Eur Heart J 2007; 28: 1936–1945.

Robertson

Kennard

Mehta

. Influence of gender on in-hospital clinical and angiographic outcomes and on one-year follow-up in the New Approaches to Coronary Intervention (NACI) registry. Am J Cardiol 1997; 80: 26K–39K.

Cox

Resnic

Popma

Simon

Eisenhauer

Rogers

. Comparison of the risk of vascular complications associated with femoral and radial access coronary catheterization procedures in obese versus nonobese patients. Am J Cardiol 2004; 94: 1174–1177.

10.

Berger

Best

Topol

. The relation of renal function to ischemic and bleeding outcomes with 2 different glycoprotein IIb/IIIa inhibitors: the do Tirofiban and ReoPro Give Similar Efficacy Outcome (TARGET) trial. Am Heart J 2005; 149: 869–875.

11.

Kinnaird

Stabile

Mintz

. Incidence, predictors, and prognostic implications of bleeding and blood transfusion following percutaneous coronary interventions. Am J Cardiol 2003; 92: 930–935.

12.

Hillegass

Brott

Chapman

. Relationship between activated clotting time during percutaneous intervention and subsequent bleeding complications. Am Heart J 2002; 144: 501–507.

13.

Tolleson

O’Shea

Bittl

. Relationship between heparin anticoagulation and clinical outcomes in coronary stent intervention: observations from the ESPRIT trial. J Am Coll Cardiol 2003; 41: 386–393.

14.

Brener

Moliterno

Lincoff

Steinhubl

Wolski

Topol

. Relationship between activated clotting time and ischemic or hemorrhagic complications: analysis of 4 recent randomized clinical trials of percutaneous coronary intervention. Circulation 2004; 110: 994–998.

15.

Cantor

Mahaffey

Huang

. Bleeding complications in patients with acute coronary syndrome undergoing early invasive management can be reduced with radial access, smaller sheath sizes, and timely sheath removal. Catheter Cardiovasc Interv 2007; 69: 73–83.

16.

Lincoff

Bittl

Kleiman

. Comparison of bivalirudin versus heparin during percutaneous coronary intervention (the Randomized Evaluation of PCI Linking Angiomax to Reduced Clinical Events [REPLACE]-1 trial). Am J Cardiol 2004; 93: 1092–1096.

17.

Feit

Voeltz

Attubato

. Predictors and impact of major hemorrhage on mortality following percutaneous coronary intervention from the REPLACE-2 Trial. Am J Cardiol 2007; 100: 1364–1369.

18.

Byrne

Spence

Fretz

. Body mass index, periprocedural bleeding, and outcome following percutaneous coronary intervention (from the British Columbia Cardiac Registry). Am J Cardiol 2009; 103: 507–511.

19.

Tiroch

Arora

Matheny

Liu

Lee

Resnic

. Risk predictors of retroperitoneal hemorrhage following percutaneous coronary intervention. Am J Cardiol 2008; 102: 1473–1476.

20.

Farouque

Tremmel

Raissi Shabari

. Risk factors for the development of retroperitoneal hematoma after percutaneous coronary intervention in the era of glycoprotein IIb/IIIa inhibitors and vascular closure devices. J Am Coll Cardiol 2005; 45: 363–368.

21.

Manoukian

. Predictors and impact of bleeding complications in percutaneous coronary intervention, acute coronary syndromes, and ST-segment elevation myocardial infarction. Am J Cardiol 2009; 104: 9C–15C.

22.

Lee

Kang

. Periprocedural hemoglobin drop and contrast-induced nephropathy in percutaneous coronary intervention patients. Korean Circ J 2010; 40: 68–73.

23.

Sattur

Harjai

Narula

Devarakonda

Orshaw

Yaeger

. The influence of anemia after percutaneous coronary intervention on clinical outcomes. Clin Cardiol 2009; 32: 373–379.

24.

Levey

Bosch

Lewis

Greene

Rogers

Roth

. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999; 130: 461–470.

25.

Reinecke

Trey

Wellmann

. Haemoglobin-related mortality in patients undergoing percutaneous coronary interventions. Eur Heart J 2003; 24: 2142–2150.

26.

Agostoni

Biondi-Zoccai

de Benedictis

. Radial versus femoral approach for percutaneous coronary diagnostic and interventional procedures; Systematic overview and meta-analysis of randomized trials. J Am Coll Cardiol 2004; 44: 349–356.

27.

Delhaye

Wakabayashi

Maluenda

. Body mass index and bleeding complications after percutaneous coronary intervention: does bivalirudin make a difference? Am Heart J 2010; 159: 1139–1146.

28.

Reiner

Siscovick

Rosendaal

. Hemostatic risk factors and arterial thrombotic disease. Thromb Haemost 2001; 85: 584–595.