- Email: [email protected]
Predictors of Inpatient Outcomes in Hospitalized Patients After Left Heart Catheterization
Predictors of Inpatient Outcomes in Hospitalized Patients After Left Heart Catheterization Yun You Li, MD, PhD, Charles A. Bush, MD, Anthony Orsini, RICS, Zhibao Mi, PhD, and Carl V. Leier, MD* Clinical and laboratory factors predicting inpatient outcomes, specifically in-hospital mortality and length of stay (LOS), have not been defined for hospitalized patients specifically referred for left heart catheterization and coronary angiography (LHC). The objective of the study was to determine these outcomes and their predictors in hospitalized patients after LHC. Multivariate logistic regression models were used to identify risk factors for in-hospital mortality and Cox proportional hazards models were used to identify factors determining LOS in 9,420 consecutive patients hospitalized for LHC. Odds ratio for in-hospital mortality and hazard ratio for prolonged LOS were derived. The strongest predictors of mortality were advanced age, left ventricular (LV) end-diastolic pressure (EDP), LV ejection fraction (EF), systemic blood pressure, and renal insufficiency. Predictors of prolonged LOS were LVEDP, LVEF, 3-vessel coronary artery disease, and valvular disease. Clinical and laboratory characteristics of patients with an LVEF >50% were also compared with those of patients with an LVEF <50%. Predictors of mortality and LOS remained the same for patients with an LVEF <50%. For an LVEF >50%, LVEDP also determined LOS and chronic renal insufficiency provided predictive power to mortality and LOS in this subgroup. In conclusion, several readily attainable clinical and laboratory parameters predict inpatient mortality and LOS in hospitalized patients referred for LHC. © 2009 Elsevier Inc. (Am J Cardiol 2009;103:486 – 490)
We hypothesized that certain clinical characteristics and laboratory findings can serve as independent prognostic factors for inpatient outcomes in hospitalized patients referred for left heart catheterization and coronary angiography (LHC). We conducted this retrospective, single-center study to determine the predictive factors for in-hospital mortality and length of stay (LOS) in hospitalized patients undergoing LHC. Methods From February 1, 1999, to August 31, 2005, all patients hospitalized at The Ohio State University Medical Center and referred for LHC were included in this study. Patients undergoing LHC as an outpatient procedure were not included in the data set. Approval for this study was obtained from our institutional review board for human subjects. The Health Insurance Portability and Accountability Act guidelines for patient confidentiality were applied to data collection, analysis, and reporting. All study patients underwent LHC consisting of left ventricular (LV) pressure measurements and LV and coronary artery angiography. During LHC, a fluid-filled 5Fr or 6Fr pigtail catheter was positioned in the mid-LV cavity. Transducers were balanced with the 0 level at the midaxil-
Division of Cardiovascular Medicine, The Ohio State University Medical Center, Columbus, Ohio. Manuscript received July 24, 2008; revised manuscript received and accepted October 7, 2008. *Corresponding author: Tel: 614-293-8963; fax: 614-293-5614. E-mail address: [email protected] (C.V. Leier). 0002-9149/09/$ – see front matter © 2009 Elsevier Inc. doi:10.1016/j.amjcard.2008.10.012
lary line with the patient in a fully supine position. Pressures were measured and recorded digitally before diagnostic LV and coronary angiography. LV angiography was performed in a 30° right anterior-oblique projection. LV end-diastolic and end-systolic endocardial contours were traced in the frames of maximal and minimal dimensions (volumes), from which LV ejection fraction (LVEF) was calculated using Innova (General Electric Marquette, Milwaukee, Wisconsin). Ectopic and immediate postectopic beats were excluded from analysis. Primary inpatient outcomes, specifically death and LOS, after LHC were obtained from the comprehensive cardiovascular database. We collected for analyses the major clinical and laboratory factors felt to be associated with in-hospital mortality or LOS, including gender, race, age, body mass index, co-morbid conditions, standard risk factors for atherosclerosis, LV end-diastolic pressure (LVEDP), LVEF, LV end-systolic volume index (end-systolic volume/body surface area), LV end-diastolic volume index (end-diastolic volume/body surface area), mean systemic blood pressure, cardiac output, New York Heart Association functional class, final cardiac diagnosis, and previous cardiac procedures (coronary artery bypass surgery, angioplasty with/without stent deployment). For the final cardiac diagnosis, coronary artery disease was defined as an angiographic lesion of ⱖ50% occlusion involving 1, 2, or 3 major coronary vessels and no disease for normal coronary anatomy. Individual medication lists and select blood chemistry results at the time of catheterization, such as electrolytes, creatine kinase, troponin, and natriuretic peptide, were only entered into the database after January 2004 and therefore these variables were not included in the multivariate models. www.AJConline.org
Coronary Heart Disease/Inpatient Outcomes After Left Heart Catheterization
487
Table 1 Clinical characteristics of the entire patient population and as separated into the groups of preserved or reduced left ventricular ejection fraction Variable
Age (yrs) Body mass index (kg/m2) Women White African-American Asian Others Systemic hypertension Diabetes mellitus Cerebrovascular event Chronic renal insufficiency Anemia Chronic obstructive lung disease Cancer Other NYHA class I II III IV Mean LVEF (%) Mean LVEDP (mm Hg) Final cardiac diagnosis No occlusive coronary disease 1-vessel coronary disease 2-vessel coronary disease 3-vessel coronary disease Myocardial infarction Pulmonary hypertension Valvular heart disease
All Patients (n ⫽ 9,969)
LVEF
p Value*
ⱖ50% (n ⫽ 5,193)
⬍50% (n ⫽ 4,776)
60.64 ⫾ 3.37 29.85 ⫾ 7.14 36.46% 83.43% 14.29% 0.53% 1.49% 35.24% 24.53% 3.67% 7.32% 7.73% 11.84% 5.06% 14.38%
59.7 ⫾ 13.0 30.20 ⫾ 7.26 41.25% 83.73% 13.63% 0.69% 1.64% 31.64% 23.69% 3.02% 5.41% 7.72% 11.11% 5.05% 17.00%
61.2 ⫾ 13.6 29.41 ⫾ 6.97 30.80% 83.35% 14.61% 0.36% 1.38% 45.29% 29.71% 4.17% 6.64% 8.14% 12.04% 4.94% 15.91%
⬍0.0001 ⬍0.0001 ⬍0.0001 0.6150 0.1598 0.0207 0.2979 0.0002 0.0067 0.0021 0.0100 0.4348 0.1475 0.8120 0.0554
2.19% 26.04% 38.83% 32.95% 47.25 ⫾ 13.69 17.13 ⫾ 7.97
3.50% 36.88% 38.59% 20.99% 57.81 ⫾ 5.53 15.28 ⫾ 6.76
0.90% 16.25% 40.26% 42.57% 35.77 ⫾ 10.25 18.43 ⫾ 8.35
⬍0.0001 ⬍0.0001 0.0876 ⬍0.0001 ⬍0.0001 ⬍0.0012
4.75% 9.66% 14.19% 26.30% 18.88% 22.74% 3.48%
8.34% 12.48% 16.16% 24.13% 15.89% 17.25% 3.70%
1.03% 7.10% 12.65% 30.07% 19.68% 26.34% 2.85%
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0175
Data are presented as percentage of subjects or mean ⫾ 1 SD. * LVEF ⱖ50% versus ⬍50%. NYHA ⫽ New York Heart Association.
Statistical analyses were performed with SAS 8.2 (SAS Institute, Cary, North Carolina) and Excel (Microsoft, Redmond, Washington). The results are shown as mean ⫾ SD for continuous variables or percentages for categorical variables unless otherwise indicated. Clinical and laboratory characteristics were also compared between patients with an LVEF ⱖ50% and those with an LVEF ⬍50% using the Pearson chi-square test for categorical variables and unpaired Student’s t test for continuous variables. Primary outcomes of the study were all-cause in-hospital mortality and LOS. Repeat left heart catheterization and coronary angiographic data points were used only for LOS analysis. Predictive factors for in-hospital mortality and for LOS were identified using univariate analysis and tested with multivariate logistic regression models and multivariate Cox proportional hazard models. A p value ⬍0.05 with univariate analysis was used for entry into multivariate analysis and a p value ⬍0.05 for retention of a parameter in the multivariate model. The strength of predictive factors was determined using odds ratios with 95% confidence intervals in logistic regression models and hazard ratios with 95% confidence intervals in Cox proportional hazard models. LOS was defined as the time of admission to dis-
charge home or transfer to the cardiac surgery service or coronary intervention unit for a definitive cardiac procedure (angioplasty with/without stent deployment); LOS did not include duration of an intervention or the recovery from such. Because LVEF was found to be a powerful predictor of mortality and LVEF remains a prominent, clinically employed parameter, we reanalyzed the predictive factors for subgroups with a normal LVEF ⱖ50% (n ⫽ 5,193) and those with an LVEF ⬍50% (n ⫽ 4,776) using the same statistical models. Cox proportional hazard analysis revealed that LVEDP was the strongest predictor of LOS in the 2 subgroups; LVEDP was then studied in a stepwise fashion with 5-mm Hg increments. LVEF and LVEDP were powerful predictors of inpatient outcomes, yet independent of each other in their prognostication. Results From February 1999 to August 2005, 10,799 LHCs were performed in 9,420 patients. LVEDP measurements were available in 9,969 (of the 10,799) procedures, allowing inclusion in this study. Of the 9,969 studies, 60.9% were
488
The American Journal of Cardiology (www.AJConline.org)
Table 2 Multivariate models of in-hospital mortality in entire cohort of patients, in patients with heart failure and decreased ejection fraction, and in patients with preserved ejection fraction
All patients Age LVEDP Mean systemic blood pressure LVEF Chronic renal insufficiency White race LVEF ⬍50% Age LVEDP Mean systemic blood pressure Myocardial infarction LV end-systolic volume index Male gender 3-vessel coronary disease LVEF ⱖ50% Age Chronic renal insufficiency Pulmonary hypertension No obstructive coronary disease
Odds Ratio (95% CI)
p Value
1.062 (1.045–1.079) 1.433 (1.272–1.614) 0.952 (0.940–0.965) 0.111 (0.030–0.409) 1.990 (1.150–3.444) 0.563 (0.342–0.927)
⬍0.0001 ⬍0.0001 ⬍0.0001 0.0009 0.0140 0.0238
1.055 (1.034–1.077) 1.486 (1.290–1.712) 0.940 (0.925–0.956) 2.393 (1.335–4.291) 1.009 (1.002–1.016) 0.594 (0.371–0.950) 1.796 (1.031–3.126)
⬍0.0001 ⬍0.0001 ⬍0.0001 0.0034 0.0067 0.0298 0.0385
1.059 (1.028–1.090) 3.974 (1.716–9.204) 2.475 (1.221–5.015) 0.420 (0.188–0.934)
⬍0.0001 0.0013 0.0119 0.0335
CI ⫽ confidence interval.
performed in patients with significant but stable cardiac symptoms (elective), 31.5% in patients with relatively unstable symptoms (e.g., unstable angina, decompensated heart failure), and 7.6% in patients with an emergency condition (e.g., acute coronary syndromes, cardiogenic shock, ventilator support). Outcomes of the former 2 groups were statistically similar, with an in-hospital mortality of 15.3% in the latter group of patients with emergency conditions. Causes of death were designated as cardiac (e.g., dysrhythmia, shock) in 59.8%, pulmonic in 9.8%, infectious/septic in 7.6%, neurologic in 7.1%, hemorrhagic in 1.1%, other in 6.0%, and undetermined in 8.6%. Rate of major complications from LHC was low at ⬍0.2% and was statistically similar between groups. Clinical characteristics of the entire study population and subgroups defined by LVEFs ⱖ50% and ⬍50% are presented in Table 1. Patients with an LVEF ⱖ50% were slightly younger with a larger proportion of women. History of systemic hypertension, diabetes mellitus, stroke or transient ischemic attack, and/or chronic renal insufficiency was more common in patients with an LVEF ⬍50%. Patients with an LVEF ⬍50% also had a higher prevalence of 3-vessel coronary disease, myocardial infarction, and pulmonary hypertension. Valvular disorders, detected or verified by LHC, had a slightly higher prevalence in patients with an LVEF ⱖ50%. In-hospital all-cause mortality for the overall population undergoing LHC was 1.95% (184 of 9,420 patients died). As presented in Table 2, multivariate logistic regression analysis showed that increasing age, low LVEF, lower systemic blood pressure, and high LVEDP were statistically the strongest predictors of in-hospital mortality. Not unexpectedly, mortality rate was significantly higher in patients with an LVEF ⬍50% (2.16%) compared with
Table 3 Multivariate models of in-hospital length of stay in entire cohort of patients, in patients with heart failure and decreased ejection fraction, and in patients with preserved ejection fraction
All patients Age LVEDP LVEF Mean systemic blood pressure No occlusive coronary disease 3-vessel coronary disease Valvular heart disease Pulmonary hypertension Male gender Chronic renal insufficiency LVEF ⬍50% Age LVEDP Mean systemic blood pressure 3-vessel coronary disease Valvular heart disease Male gender No occlusive coronary disease 1-vessel coronary disease White race LVEF ⱖ50% Age LVEDP 3-vessel coronary disease Valvular heart disease Chronic renal insufficiency Pulmonary hypertension No occlusive coronary disease Mean systemic blood pressure Body mass index
Hazard Ratio (95% CI)
p Value
0.991 (0.990–0.993) 0.927 (0.914–0.941) 2.441 (2.074–2.874) 1.004 (1.003–1.006) 1.269 (1.145–1.407) 0.751 (0.713–0.792) 0.588 (0.518–0.666) 0.929 (0.875–0.986) 1.069 (1.022–1.117) 0.879 (0.797–0.969)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0027 0.0048 0.0177
0.992 (0.989–0.994) 0.906 (0.889–0.924) 1.006 (1.004–1.007) 0.735 (0.687–0.787) 0.684 (0.566–0.827) 1.116 (1.045–1.192) 1.111 (1.043–1.183) 1.158 (1.030–1.302) 0.916 (0.884–0.994)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0011 0.0011 0.0143 0.0354
0.991 (0.989–0.994) 0.949 (0.925–0.966) 0.757 (0.705–0.812) 0.531 (0.454–0.629) 0.773 (0.677–0.887) 0.868 (0.795–0.939) 1.188 (1.066–1.325) 1.002 (1.000–1.004) 1.002 (1.001–1.003)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0002 0.0008 0.0019 0.0143 0.0172
Abbreviation as in Table 2.
that of patients with an LVEF ⱖ50% (0.89%, p ⬍0.0001). Predictors of inpatient death are presented in Table 2. For an LVEF ⬍50%, predictors of inpatient death included increasing age, female gender, lower systemic blood pressure, larger LV systolic volume, higher LVEDP, triple coronary artery occlusive disease, and myocardial infarction; of these, the strongest predictors of all-cause mortality were increased LVEDP and myocardial infarction. Higher mean systemic blood pressure was protective in patients with an LVEF ⬍50%. Predictors of inpatient death after LHC for patients with an LVEF ⱖ50% were increasing age, chronic renal insufficiency, and pulmonary hypertension, and no angiographic coronary artery disease lowered the risk. Table 3 presents the results of multivariate Cox proportional hazard analysis for LOS. Because survival concept was used in multivariate analysis (patients who expired could not be included in LOS analyses), longer LOS was associated with a smaller hazard ratio. Results show that, for the entire study population, increasing age, lower LVEF, higher LVEDP, more prevalent 3-vessel coronary occlusive disease, and more valvular disease were the most powerful predictors of LOS. For the 2 subgroups, LVEFs ⬍50% and ⬎50%, older age, higher LVEDP, greater prevalence of 3-vessel coronary
Coronary Heart Disease/Inpatient Outcomes After Left Heart Catheterization
489
Discussion
Figure 1. Effects of LVEDP on hospital LOS. Study patients are presented as all patients (A) and subgrouped into those with an LVEF ⬍50% (B) and those with an LVEF ⱖ50% (C). LVEDP was then splayed for each population and plotted against the medians, lower quartile, and upper quartile of LOS.
occlusive disease and valvular disease, and lower systemic blood pressure remained strong predictors of increased LOS after LHC (Table 3). However, for patients with an LVEF ⬍50%, male gender tended to favorably affect LOS, whereas LOS for patients with an LVEF ⬎50% was adversely affected by chronic renal disease and pulmonary hypertension. Not unexpectedly, absence of obstructive coronary lesions, or if the occlusive lesion was confined to a single coronary vessel, favorably affected LOS for patients with normal or decreased LV systolic function. Increased LVEDP was associated with prolonged LOS in patients with an LVEF ⬍50% or ⱖ50% (Table 3). Further analysis of LVEDP showed that increased LVEDP was closely associated with longer LOS when it was increased ⬎15 mm Hg (Figure 1). Median LOS was 2 days when LVEDP was ⱕ15 mm Hg, increased 1 day to 3 days for each increase of 10 mm Hg of LVEDP, and approached 10 days when LVEDP was ⬎35 mm Hg.
In this study, a sizable single-center cardiac catheterization database was utilized to determine inpatient outcomes of hospitalized patients referred for diagnostic LHC. Results of this study demonstrated that LVEF strongly predicted in-hospital mortality in patients undergoing LHC. Increasing age was found to be a consistent independent predictor of in-hospital mortality and LOS in patients with preserved or decreased LVEF. In addition to age, LVEDP, systemic blood pressure, and more extensive coronary artery disease negatively affected, whereas male gender favorably affected, inpatient mortality and LOS for patients with an LVEF ⬍50% undergoing LHC. Chronic renal insufficiency, in addition to age, is an independent predictor of increased in-hospital mortality and LOS in patients with preserved LVEF, whereas male gender had a favorable effect on these outcomes in this subgroup. Prognostic factors for mortality, hospitalization and rehospitalization rates, and duration of hospitalization have been reported for several major cardiovascular conditions and interventions with the formulation of a number of predictive models. Some predictors of inpatient mortality and LOS after LHC in the present study have also been noted as prognostic indicators for a number of these conditions, such as myocardial infarction and heart failure, and for major cardiovascular interventions; frequently encountered factors in common are age, LVEF, LVEDP, systemic blood pressure, extent of coronary artery disease, and chronic renal insufficiency.1–10 Despite the commonality of these particular prognosticators for adverse outcomes, patients with these cardiovascular conditions still often require LHC for more precise diagnostic information and formulation of the optimal intervention and treatment plan. There is little published information from sizable populations regarding in-hospital outcomes after major diagnostic testing, including the frequently performed “invasive” LHC. A report from the American College of Cardiology/ National Cardiovascular Data Registry of a large patient population undergoing coronary angiography found age, female gender, and extent of coronary artery disease to have the most unfavorable impact on in-hospital mortality,9 findings in keeping with the results of the present study. In contrast to the findings from the registry,9 race had a marginal impact on outcomes in the present study, perhaps related to a much smaller patient population. LOS data and information were not presented in the registry report. Despite the power of population size, the National Cardiovascular Data Registry is limited somewhat by the paucity of risk factors surveyed and submitted by 388 different hospitals nationwide. This study has the limitations inherent in retrospective analyses of database information in human disease; not all relevant data are available for analysis and a selection bias for LHC cannot be excluded. Because potentially predictive serologic markers (e.g., C-reactive protein, natriuretic peptide) were only recently available, these were not included in the analyses. Results cannot be transferred to outpatient left heart catheterization and coronary angiographic procedures and may not pertain to other medical centers, partic-
490
The American Journal of Cardiology (www.AJConline.org)
ularly nonuniversity centers of a different size and different patient populations.11 6.
Acknowledgment: The authors thank Kamal Jyodi, PhD, and Jennifer Santiangolo for their assistance in data collection. 7. 1. Moss AJ, Multicenter Postinfarction Research Group. Risk stratification and survival after myocardial infarction. N Engl J Med 1983;309: 331–336. 2. Boersma E, Pieper KS, Steyerberg EW, Wilcox RG, Chang WC, Lee KL, Akkerhuis KM, Harrington RA, Deckers JW, Armstrong PW, et al, and PURSUIT Investigators. Predictions of outcome in patients with acute coronary syndromes without persistent ST-segment elevation. Results from an international trial of 9461 patients. Circulation 2000;101:2557–2567. 3. Pocock SJ, Wang D, Pfeffer MA, Yusuf S, McMurray JJ, Swedberg KB, Ostergren J, Michelson EL, Pieper KS, Granger CB. Predictors of mortality and morbidity in patients with chronic heart failure. Eur Heart J 2006;27:65–75. 4. Nieminen MS, Brutsaert D, Dickstein K, Drexler H, Follath F, Harjola VP, Hochadel M, Komajda M, Lassus J, Lopez-Sendon JL, Ponikowski P, Tavazzi L, for the EuroHeart Survey Investigators, Heart Failure Association, European Society of Cardiology. EuroHeart Failure Survey II (EHFS II): a survey on hospitalized acute heart failure patients: description of population. Eur Heart J 2006;27:2725–2736. 5. Abraham WT, Adams KF, Fonarow GC, Costanzo MR, Berkowitz RL, LeJemtel TH, Cheng ML, Wynne J, ADHERE Study Group. In-hospital
8.
9.
10.
11.
mortality in patients with acute decompensated heart failure requiring intravenous vasoactive medications: an analysis from the Acute Decompensated Heart Failure National Registry (ADHERE). J Am Coll Cardiol 2005;46:57– 64. Hillege HL, Nitsch D, Pfeffer MA, Swedberg K, McMurray JJ, Yusuf S, Granger CB, Michelson EL, Ostergren J, Cornel JH, et al, CHARM Investigators. Renal function as a predictor of outcome in a broad spectrum of patients with heart failure. Circulation 2006;113:671– 678. Kini AS, Lee PC, Mitre CA, Kim MC, Kamran M, Duffy ME, Marmur JD, Sharma SK. Prediction of outcome after percutaneous coronary intervention for the acute coronary syndrome. Am J Med 2003;115: 708 –714. Hubacek J, Galbraith PD, Gao M, Humphries K, Graham MM, Knutson ML, Ghali WA, APPROACH Investigators. External validation of a percutaneous coronary intervention mortality prediction model in patients with acute coronary syndromes. Am Heart J 2006;151:308 – 315. Shaw LJ, Shaw RE, Bairey Merz CN, Brindis RG, Klein LW, Nallamothu B, Douglas PS, Krone RJ, McKay CR, Block PC, et al, for the American College of Cardiology–National Cardiovascular Data Registry Investigators. Impact of ethnicity and gender differences on angiographic coronary artery disease prevalence and in-hospital mortality in the American College of Cardiology–National Cardiovascular Data Registry. Circulation 2008;117:1787–1801. Rao SV, Ou FS, Wang TY, Roe MT, Brindis R, Rumsfeld JS, Peterson ED. Trends in the prevalence and outcomes of radial and femoral approaches to percutaneous coronary intervention. J Am Coll Cardiol 2008;1:379 –386. Shahian DM, Normand SLT. Comparison of “risk-adjusted” hospital outcomes. Circulation 2008;117:1955–1963.
We hypothesized that certain clinical characteristics and laboratory findings can serve as independent prognostic factors for inpatient outcomes in hospitalized patients referred for left heart catheterization and coronary angiography (LHC). We conducted this retrospective, single-center study to determine the predictive factors for in-hospital mortality and length of stay (LOS) in hospitalized patients undergoing LHC. Methods From February 1, 1999, to August 31, 2005, all patients hospitalized at The Ohio State University Medical Center and referred for LHC were included in this study. Patients undergoing LHC as an outpatient procedure were not included in the data set. Approval for this study was obtained from our institutional review board for human subjects. The Health Insurance Portability and Accountability Act guidelines for patient confidentiality were applied to data collection, analysis, and reporting. All study patients underwent LHC consisting of left ventricular (LV) pressure measurements and LV and coronary artery angiography. During LHC, a fluid-filled 5Fr or 6Fr pigtail catheter was positioned in the mid-LV cavity. Transducers were balanced with the 0 level at the midaxil-
Division of Cardiovascular Medicine, The Ohio State University Medical Center, Columbus, Ohio. Manuscript received July 24, 2008; revised manuscript received and accepted October 7, 2008. *Corresponding author: Tel: 614-293-8963; fax: 614-293-5614. E-mail address: [email protected] (C.V. Leier). 0002-9149/09/$ – see front matter © 2009 Elsevier Inc. doi:10.1016/j.amjcard.2008.10.012
lary line with the patient in a fully supine position. Pressures were measured and recorded digitally before diagnostic LV and coronary angiography. LV angiography was performed in a 30° right anterior-oblique projection. LV end-diastolic and end-systolic endocardial contours were traced in the frames of maximal and minimal dimensions (volumes), from which LV ejection fraction (LVEF) was calculated using Innova (General Electric Marquette, Milwaukee, Wisconsin). Ectopic and immediate postectopic beats were excluded from analysis. Primary inpatient outcomes, specifically death and LOS, after LHC were obtained from the comprehensive cardiovascular database. We collected for analyses the major clinical and laboratory factors felt to be associated with in-hospital mortality or LOS, including gender, race, age, body mass index, co-morbid conditions, standard risk factors for atherosclerosis, LV end-diastolic pressure (LVEDP), LVEF, LV end-systolic volume index (end-systolic volume/body surface area), LV end-diastolic volume index (end-diastolic volume/body surface area), mean systemic blood pressure, cardiac output, New York Heart Association functional class, final cardiac diagnosis, and previous cardiac procedures (coronary artery bypass surgery, angioplasty with/without stent deployment). For the final cardiac diagnosis, coronary artery disease was defined as an angiographic lesion of ⱖ50% occlusion involving 1, 2, or 3 major coronary vessels and no disease for normal coronary anatomy. Individual medication lists and select blood chemistry results at the time of catheterization, such as electrolytes, creatine kinase, troponin, and natriuretic peptide, were only entered into the database after January 2004 and therefore these variables were not included in the multivariate models. www.AJConline.org
Coronary Heart Disease/Inpatient Outcomes After Left Heart Catheterization
487
Table 1 Clinical characteristics of the entire patient population and as separated into the groups of preserved or reduced left ventricular ejection fraction Variable
Age (yrs) Body mass index (kg/m2) Women White African-American Asian Others Systemic hypertension Diabetes mellitus Cerebrovascular event Chronic renal insufficiency Anemia Chronic obstructive lung disease Cancer Other NYHA class I II III IV Mean LVEF (%) Mean LVEDP (mm Hg) Final cardiac diagnosis No occlusive coronary disease 1-vessel coronary disease 2-vessel coronary disease 3-vessel coronary disease Myocardial infarction Pulmonary hypertension Valvular heart disease
All Patients (n ⫽ 9,969)
LVEF
p Value*
ⱖ50% (n ⫽ 5,193)
⬍50% (n ⫽ 4,776)
60.64 ⫾ 3.37 29.85 ⫾ 7.14 36.46% 83.43% 14.29% 0.53% 1.49% 35.24% 24.53% 3.67% 7.32% 7.73% 11.84% 5.06% 14.38%
59.7 ⫾ 13.0 30.20 ⫾ 7.26 41.25% 83.73% 13.63% 0.69% 1.64% 31.64% 23.69% 3.02% 5.41% 7.72% 11.11% 5.05% 17.00%
61.2 ⫾ 13.6 29.41 ⫾ 6.97 30.80% 83.35% 14.61% 0.36% 1.38% 45.29% 29.71% 4.17% 6.64% 8.14% 12.04% 4.94% 15.91%
⬍0.0001 ⬍0.0001 ⬍0.0001 0.6150 0.1598 0.0207 0.2979 0.0002 0.0067 0.0021 0.0100 0.4348 0.1475 0.8120 0.0554
2.19% 26.04% 38.83% 32.95% 47.25 ⫾ 13.69 17.13 ⫾ 7.97
3.50% 36.88% 38.59% 20.99% 57.81 ⫾ 5.53 15.28 ⫾ 6.76
0.90% 16.25% 40.26% 42.57% 35.77 ⫾ 10.25 18.43 ⫾ 8.35
⬍0.0001 ⬍0.0001 0.0876 ⬍0.0001 ⬍0.0001 ⬍0.0012
4.75% 9.66% 14.19% 26.30% 18.88% 22.74% 3.48%
8.34% 12.48% 16.16% 24.13% 15.89% 17.25% 3.70%
1.03% 7.10% 12.65% 30.07% 19.68% 26.34% 2.85%
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0175
Data are presented as percentage of subjects or mean ⫾ 1 SD. * LVEF ⱖ50% versus ⬍50%. NYHA ⫽ New York Heart Association.
Statistical analyses were performed with SAS 8.2 (SAS Institute, Cary, North Carolina) and Excel (Microsoft, Redmond, Washington). The results are shown as mean ⫾ SD for continuous variables or percentages for categorical variables unless otherwise indicated. Clinical and laboratory characteristics were also compared between patients with an LVEF ⱖ50% and those with an LVEF ⬍50% using the Pearson chi-square test for categorical variables and unpaired Student’s t test for continuous variables. Primary outcomes of the study were all-cause in-hospital mortality and LOS. Repeat left heart catheterization and coronary angiographic data points were used only for LOS analysis. Predictive factors for in-hospital mortality and for LOS were identified using univariate analysis and tested with multivariate logistic regression models and multivariate Cox proportional hazard models. A p value ⬍0.05 with univariate analysis was used for entry into multivariate analysis and a p value ⬍0.05 for retention of a parameter in the multivariate model. The strength of predictive factors was determined using odds ratios with 95% confidence intervals in logistic regression models and hazard ratios with 95% confidence intervals in Cox proportional hazard models. LOS was defined as the time of admission to dis-
charge home or transfer to the cardiac surgery service or coronary intervention unit for a definitive cardiac procedure (angioplasty with/without stent deployment); LOS did not include duration of an intervention or the recovery from such. Because LVEF was found to be a powerful predictor of mortality and LVEF remains a prominent, clinically employed parameter, we reanalyzed the predictive factors for subgroups with a normal LVEF ⱖ50% (n ⫽ 5,193) and those with an LVEF ⬍50% (n ⫽ 4,776) using the same statistical models. Cox proportional hazard analysis revealed that LVEDP was the strongest predictor of LOS in the 2 subgroups; LVEDP was then studied in a stepwise fashion with 5-mm Hg increments. LVEF and LVEDP were powerful predictors of inpatient outcomes, yet independent of each other in their prognostication. Results From February 1999 to August 2005, 10,799 LHCs were performed in 9,420 patients. LVEDP measurements were available in 9,969 (of the 10,799) procedures, allowing inclusion in this study. Of the 9,969 studies, 60.9% were
488
The American Journal of Cardiology (www.AJConline.org)
Table 2 Multivariate models of in-hospital mortality in entire cohort of patients, in patients with heart failure and decreased ejection fraction, and in patients with preserved ejection fraction
All patients Age LVEDP Mean systemic blood pressure LVEF Chronic renal insufficiency White race LVEF ⬍50% Age LVEDP Mean systemic blood pressure Myocardial infarction LV end-systolic volume index Male gender 3-vessel coronary disease LVEF ⱖ50% Age Chronic renal insufficiency Pulmonary hypertension No obstructive coronary disease
Odds Ratio (95% CI)
p Value
1.062 (1.045–1.079) 1.433 (1.272–1.614) 0.952 (0.940–0.965) 0.111 (0.030–0.409) 1.990 (1.150–3.444) 0.563 (0.342–0.927)
⬍0.0001 ⬍0.0001 ⬍0.0001 0.0009 0.0140 0.0238
1.055 (1.034–1.077) 1.486 (1.290–1.712) 0.940 (0.925–0.956) 2.393 (1.335–4.291) 1.009 (1.002–1.016) 0.594 (0.371–0.950) 1.796 (1.031–3.126)
⬍0.0001 ⬍0.0001 ⬍0.0001 0.0034 0.0067 0.0298 0.0385
1.059 (1.028–1.090) 3.974 (1.716–9.204) 2.475 (1.221–5.015) 0.420 (0.188–0.934)
⬍0.0001 0.0013 0.0119 0.0335
CI ⫽ confidence interval.
performed in patients with significant but stable cardiac symptoms (elective), 31.5% in patients with relatively unstable symptoms (e.g., unstable angina, decompensated heart failure), and 7.6% in patients with an emergency condition (e.g., acute coronary syndromes, cardiogenic shock, ventilator support). Outcomes of the former 2 groups were statistically similar, with an in-hospital mortality of 15.3% in the latter group of patients with emergency conditions. Causes of death were designated as cardiac (e.g., dysrhythmia, shock) in 59.8%, pulmonic in 9.8%, infectious/septic in 7.6%, neurologic in 7.1%, hemorrhagic in 1.1%, other in 6.0%, and undetermined in 8.6%. Rate of major complications from LHC was low at ⬍0.2% and was statistically similar between groups. Clinical characteristics of the entire study population and subgroups defined by LVEFs ⱖ50% and ⬍50% are presented in Table 1. Patients with an LVEF ⱖ50% were slightly younger with a larger proportion of women. History of systemic hypertension, diabetes mellitus, stroke or transient ischemic attack, and/or chronic renal insufficiency was more common in patients with an LVEF ⬍50%. Patients with an LVEF ⬍50% also had a higher prevalence of 3-vessel coronary disease, myocardial infarction, and pulmonary hypertension. Valvular disorders, detected or verified by LHC, had a slightly higher prevalence in patients with an LVEF ⱖ50%. In-hospital all-cause mortality for the overall population undergoing LHC was 1.95% (184 of 9,420 patients died). As presented in Table 2, multivariate logistic regression analysis showed that increasing age, low LVEF, lower systemic blood pressure, and high LVEDP were statistically the strongest predictors of in-hospital mortality. Not unexpectedly, mortality rate was significantly higher in patients with an LVEF ⬍50% (2.16%) compared with
Table 3 Multivariate models of in-hospital length of stay in entire cohort of patients, in patients with heart failure and decreased ejection fraction, and in patients with preserved ejection fraction
All patients Age LVEDP LVEF Mean systemic blood pressure No occlusive coronary disease 3-vessel coronary disease Valvular heart disease Pulmonary hypertension Male gender Chronic renal insufficiency LVEF ⬍50% Age LVEDP Mean systemic blood pressure 3-vessel coronary disease Valvular heart disease Male gender No occlusive coronary disease 1-vessel coronary disease White race LVEF ⱖ50% Age LVEDP 3-vessel coronary disease Valvular heart disease Chronic renal insufficiency Pulmonary hypertension No occlusive coronary disease Mean systemic blood pressure Body mass index
Hazard Ratio (95% CI)
p Value
0.991 (0.990–0.993) 0.927 (0.914–0.941) 2.441 (2.074–2.874) 1.004 (1.003–1.006) 1.269 (1.145–1.407) 0.751 (0.713–0.792) 0.588 (0.518–0.666) 0.929 (0.875–0.986) 1.069 (1.022–1.117) 0.879 (0.797–0.969)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0027 0.0048 0.0177
0.992 (0.989–0.994) 0.906 (0.889–0.924) 1.006 (1.004–1.007) 0.735 (0.687–0.787) 0.684 (0.566–0.827) 1.116 (1.045–1.192) 1.111 (1.043–1.183) 1.158 (1.030–1.302) 0.916 (0.884–0.994)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0011 0.0011 0.0143 0.0354
0.991 (0.989–0.994) 0.949 (0.925–0.966) 0.757 (0.705–0.812) 0.531 (0.454–0.629) 0.773 (0.677–0.887) 0.868 (0.795–0.939) 1.188 (1.066–1.325) 1.002 (1.000–1.004) 1.002 (1.001–1.003)
⬍0.0001 ⬍0.0001 ⬍0.0001 ⬍0.0001 0.0002 0.0008 0.0019 0.0143 0.0172
Abbreviation as in Table 2.
that of patients with an LVEF ⱖ50% (0.89%, p ⬍0.0001). Predictors of inpatient death are presented in Table 2. For an LVEF ⬍50%, predictors of inpatient death included increasing age, female gender, lower systemic blood pressure, larger LV systolic volume, higher LVEDP, triple coronary artery occlusive disease, and myocardial infarction; of these, the strongest predictors of all-cause mortality were increased LVEDP and myocardial infarction. Higher mean systemic blood pressure was protective in patients with an LVEF ⬍50%. Predictors of inpatient death after LHC for patients with an LVEF ⱖ50% were increasing age, chronic renal insufficiency, and pulmonary hypertension, and no angiographic coronary artery disease lowered the risk. Table 3 presents the results of multivariate Cox proportional hazard analysis for LOS. Because survival concept was used in multivariate analysis (patients who expired could not be included in LOS analyses), longer LOS was associated with a smaller hazard ratio. Results show that, for the entire study population, increasing age, lower LVEF, higher LVEDP, more prevalent 3-vessel coronary occlusive disease, and more valvular disease were the most powerful predictors of LOS. For the 2 subgroups, LVEFs ⬍50% and ⬎50%, older age, higher LVEDP, greater prevalence of 3-vessel coronary
Coronary Heart Disease/Inpatient Outcomes After Left Heart Catheterization
489
Discussion
Figure 1. Effects of LVEDP on hospital LOS. Study patients are presented as all patients (A) and subgrouped into those with an LVEF ⬍50% (B) and those with an LVEF ⱖ50% (C). LVEDP was then splayed for each population and plotted against the medians, lower quartile, and upper quartile of LOS.
occlusive disease and valvular disease, and lower systemic blood pressure remained strong predictors of increased LOS after LHC (Table 3). However, for patients with an LVEF ⬍50%, male gender tended to favorably affect LOS, whereas LOS for patients with an LVEF ⬎50% was adversely affected by chronic renal disease and pulmonary hypertension. Not unexpectedly, absence of obstructive coronary lesions, or if the occlusive lesion was confined to a single coronary vessel, favorably affected LOS for patients with normal or decreased LV systolic function. Increased LVEDP was associated with prolonged LOS in patients with an LVEF ⬍50% or ⱖ50% (Table 3). Further analysis of LVEDP showed that increased LVEDP was closely associated with longer LOS when it was increased ⬎15 mm Hg (Figure 1). Median LOS was 2 days when LVEDP was ⱕ15 mm Hg, increased 1 day to 3 days for each increase of 10 mm Hg of LVEDP, and approached 10 days when LVEDP was ⬎35 mm Hg.
In this study, a sizable single-center cardiac catheterization database was utilized to determine inpatient outcomes of hospitalized patients referred for diagnostic LHC. Results of this study demonstrated that LVEF strongly predicted in-hospital mortality in patients undergoing LHC. Increasing age was found to be a consistent independent predictor of in-hospital mortality and LOS in patients with preserved or decreased LVEF. In addition to age, LVEDP, systemic blood pressure, and more extensive coronary artery disease negatively affected, whereas male gender favorably affected, inpatient mortality and LOS for patients with an LVEF ⬍50% undergoing LHC. Chronic renal insufficiency, in addition to age, is an independent predictor of increased in-hospital mortality and LOS in patients with preserved LVEF, whereas male gender had a favorable effect on these outcomes in this subgroup. Prognostic factors for mortality, hospitalization and rehospitalization rates, and duration of hospitalization have been reported for several major cardiovascular conditions and interventions with the formulation of a number of predictive models. Some predictors of inpatient mortality and LOS after LHC in the present study have also been noted as prognostic indicators for a number of these conditions, such as myocardial infarction and heart failure, and for major cardiovascular interventions; frequently encountered factors in common are age, LVEF, LVEDP, systemic blood pressure, extent of coronary artery disease, and chronic renal insufficiency.1–10 Despite the commonality of these particular prognosticators for adverse outcomes, patients with these cardiovascular conditions still often require LHC for more precise diagnostic information and formulation of the optimal intervention and treatment plan. There is little published information from sizable populations regarding in-hospital outcomes after major diagnostic testing, including the frequently performed “invasive” LHC. A report from the American College of Cardiology/ National Cardiovascular Data Registry of a large patient population undergoing coronary angiography found age, female gender, and extent of coronary artery disease to have the most unfavorable impact on in-hospital mortality,9 findings in keeping with the results of the present study. In contrast to the findings from the registry,9 race had a marginal impact on outcomes in the present study, perhaps related to a much smaller patient population. LOS data and information were not presented in the registry report. Despite the power of population size, the National Cardiovascular Data Registry is limited somewhat by the paucity of risk factors surveyed and submitted by 388 different hospitals nationwide. This study has the limitations inherent in retrospective analyses of database information in human disease; not all relevant data are available for analysis and a selection bias for LHC cannot be excluded. Because potentially predictive serologic markers (e.g., C-reactive protein, natriuretic peptide) were only recently available, these were not included in the analyses. Results cannot be transferred to outpatient left heart catheterization and coronary angiographic procedures and may not pertain to other medical centers, partic-
490
The American Journal of Cardiology (www.AJConline.org)
ularly nonuniversity centers of a different size and different patient populations.11 6.
Acknowledgment: The authors thank Kamal Jyodi, PhD, and Jennifer Santiangolo for their assistance in data collection. 7. 1. Moss AJ, Multicenter Postinfarction Research Group. Risk stratification and survival after myocardial infarction. N Engl J Med 1983;309: 331–336. 2. Boersma E, Pieper KS, Steyerberg EW, Wilcox RG, Chang WC, Lee KL, Akkerhuis KM, Harrington RA, Deckers JW, Armstrong PW, et al, and PURSUIT Investigators. Predictions of outcome in patients with acute coronary syndromes without persistent ST-segment elevation. Results from an international trial of 9461 patients. Circulation 2000;101:2557–2567. 3. Pocock SJ, Wang D, Pfeffer MA, Yusuf S, McMurray JJ, Swedberg KB, Ostergren J, Michelson EL, Pieper KS, Granger CB. Predictors of mortality and morbidity in patients with chronic heart failure. Eur Heart J 2006;27:65–75. 4. Nieminen MS, Brutsaert D, Dickstein K, Drexler H, Follath F, Harjola VP, Hochadel M, Komajda M, Lassus J, Lopez-Sendon JL, Ponikowski P, Tavazzi L, for the EuroHeart Survey Investigators, Heart Failure Association, European Society of Cardiology. EuroHeart Failure Survey II (EHFS II): a survey on hospitalized acute heart failure patients: description of population. Eur Heart J 2006;27:2725–2736. 5. Abraham WT, Adams KF, Fonarow GC, Costanzo MR, Berkowitz RL, LeJemtel TH, Cheng ML, Wynne J, ADHERE Study Group. In-hospital
8.
9.
10.
11.
mortality in patients with acute decompensated heart failure requiring intravenous vasoactive medications: an analysis from the Acute Decompensated Heart Failure National Registry (ADHERE). J Am Coll Cardiol 2005;46:57– 64. Hillege HL, Nitsch D, Pfeffer MA, Swedberg K, McMurray JJ, Yusuf S, Granger CB, Michelson EL, Ostergren J, Cornel JH, et al, CHARM Investigators. Renal function as a predictor of outcome in a broad spectrum of patients with heart failure. Circulation 2006;113:671– 678. Kini AS, Lee PC, Mitre CA, Kim MC, Kamran M, Duffy ME, Marmur JD, Sharma SK. Prediction of outcome after percutaneous coronary intervention for the acute coronary syndrome. Am J Med 2003;115: 708 –714. Hubacek J, Galbraith PD, Gao M, Humphries K, Graham MM, Knutson ML, Ghali WA, APPROACH Investigators. External validation of a percutaneous coronary intervention mortality prediction model in patients with acute coronary syndromes. Am Heart J 2006;151:308 – 315. Shaw LJ, Shaw RE, Bairey Merz CN, Brindis RG, Klein LW, Nallamothu B, Douglas PS, Krone RJ, McKay CR, Block PC, et al, for the American College of Cardiology–National Cardiovascular Data Registry Investigators. Impact of ethnicity and gender differences on angiographic coronary artery disease prevalence and in-hospital mortality in the American College of Cardiology–National Cardiovascular Data Registry. Circulation 2008;117:1787–1801. Rao SV, Ou FS, Wang TY, Roe MT, Brindis R, Rumsfeld JS, Peterson ED. Trends in the prevalence and outcomes of radial and femoral approaches to percutaneous coronary intervention. J Am Coll Cardiol 2008;1:379 –386. Shahian DM, Normand SLT. Comparison of “risk-adjusted” hospital outcomes. Circulation 2008;117:1955–1963.