In the United States, 30 million patients undergo noncardiac surgery every year and more than 1 million operations annually are complicated by adverse cardiovascular events, such as perioperative myocardial infarction or death from cardiac causes.1
Hence, a rapid and reliable predictor of perioperative and postoperative complications would be a valuable clinical addition simplifying the preoperative evaluation. B-type natriuretic peptide (BNP), a 32-amino-acid polypeptide, is a quantitative marker of cardiac wall stress that might be helpful in this setting.2,3 BNP is co-secreted with the inactive aminoterminal proBNP (NT-proBNP) from the cardiac ventricles in response to ventricular volume and pressure overload and regulates a wide array of physiological effects, including natriuresis, diuresis and vasodilatation.2,3 The main stimulus for the secretion of BNP is cardiac stress as determined by systolic and diastolic as well as valvular function.3 In addition, the level of BNP is significantly modified by renal failure and the BMI, caveats that need to be taken into account when interpreting BNP values.4,5
BNP has repeatedly been shown to be a strong indicator of prognosis in a variety of clinical settings, including patients with heart failure, pneumonia, coronary artery disease and heart transplantation.6–9 The recognition of any underlying cardiac dysfunction is of special clinical interest in orthopaedic patients, who are often limited by musculoskeletal pain and function. In these patients, the prevalence of cardiac diseases is typically underestimated due to a deceiving lack of exercise-induced symptoms. As BNP was recently found to predict postoperative complications, prolonged hospitalization and death in patients undergoing heart surgery,10,11 we hypothesized that BNP might also be helpful in the preoperative evaluation of patients undergoing low-to-intermediate-risk elective orthopaedic surgery.
Setting and study population
From May 2004 to October 2004, we prospectively recruited 270 consecutive patients scheduled to undergo elective orthopaedic surgery at the University Hospital Basel, Switzerland. All patients undergoing orthopaedic surgery were eligible for enrolment in the trial; there were no exclusion criteria. Before being accepted for surgery and study enrolment, all patients passed a routine preoperative cardiovascular evaluation according to current American College of Cardiology/American Heart Association (ACC/AHA) guidelines.12 The study was carried out according to the principles of the Declaration of Helsinki and approved by our local ethics committee. Written informed consent was obtained from all patients.
Blood for BNP measurement was collected from all patients during standard preoperative blood sampling. BNP was detected in ethylenediaminetetraacetate (EDTA) plasma using an extensively validated fluorescence immunoassay (Biosite Diagnostics, La Jolla, California, USA). The precision and sensitivity of this test kit has been previously described.13,14 All physicians directly involved in patient care and the patients were blinded to the BNP results. All treatment and monitoring decisions, including the need for evaluation of troponin I values, were left to the discretion of the treating physicians and followed standard clinical practice. All patients received low-molecular-weight heparin (dalteparin) for thromboprophylaxis according to guidelines. Patient enrolment was independent of the surgeon performing the operation and lasted for a prespecified period of 5 months. During the preoperative anaesthetic visit on the day before surgery, all patients were classified according to the Physical Status Classification System of the American Society of Anesthesiologists (ASA) by the responsible anaesthetist. A research assistant collected other data, including information about the medical history, the initial physical examination and results from other laboratory and diagnostic tests. We extracted procedure and intraoperative data from the finalized surgical notes and the anaesthesia records. Glomerular filtration rate was calculated using the abbreviated Modification of Diet in Renal Disease Study equation.15
Adverse cardiac events were assessed by two attending cardiologists not involved in the patients' care with the use of all medical records pertaining to each patient. Cardiac events included intraoperative and postoperative significant ST-segment changes and new Q-waves in two contiguous leads, as defined by the current AHA guidelines;16 significant arrhythmias, defined as any ventricular tachycardia, any ventricular fibrillation or new onset atrial fibrillation or flutter; myocardial necrosis as detected by the elevation of cardiac troponin I more than 2.0 μg l−1; and acute heart failure, defined by the rapid onset of signs and symptoms of fluid retention requiring treatment with nitrates or diuretics. The cardiac troponin I cut-off level of 2.0 μg l−1 was chosen according to the manufacturer's recommendations and offers a clinical sensitivity at 91.8% [95% confidence interval (CI) 85.4–96.0%], with a corresponding clinical specificity of 92.4% (95% CI 89.7–94.5%) for the diagnosis of an acute myocardial infarction.17 Cardiac troponin I was measured using the AxSYM immuno-91 analyser (Abbott Laboratories, Baar/Zug, Switzerland).
Further clinical endpoints included long-term cardiac events (significant arrhythmias, myocardial infarction, heart failure) and rehospitalizations and mortality that all were assessed by telephone 12 months after surgery. All telephone interviews were performed by a single trained researcher. In cases of remaining uncertainties regarding the health status, further hospitalizations and the respective discharge diagnoses, the referring physicians were contacted after the telephone interview. Public administrative databases were used to confirm the survival of those patients who could not be contacted by telephone. All information derived from contingent hospital readmission records or provided by the referring physician or outpatient clinic was reviewed and entered into the computer database.
The potential of preoperative BNP levels to predict short-term in-hospital cardiac events was assessed as the primary endpoint, whereas the utility of BNP to predict the long-term cardiac events, cardiac rehospitalization and cardiac mortality at 1 year of follow-up was considered as the secondary endpoint.
The statistical analyses were performed using the SPSS/PC (version 14.0; SPSS Inc., Chicago, Illinois, USA) and SAS 9.1.3 (SAS Inc., Cary, North Carolina, USA) software packages. All hypothesis testing was two-tailed and a P value of less than 0.05 was considered statistically significant. Discrete variables are expressed as counts (percentage) and continuous variables as means and SD or median and range or interquartile range (IQR), unless otherwise stated. The paucity of endpoint events observed throughout the study resulted in a small event group, rather scattered data in the no-event group and a skewed distribution; hence, all frequency comparisons were made using the Mann–Whitney U-test. The Spearman's rank correlation was used to perform correlation analyses. The Jonckheere–Terpstra test was used to test trend hypotheses between ordinal (i.e. ASA classes) and quantitative (i.e. BNP) variables. Nonparametric bootstrap was used to estimate 95% CIs for the areas under the receiver operating characteristic (ROC) curves (each simulation using 5000 bootstrap samples drawn from the original dataset). Cut-off points were calculated using Youden's J index. To assess the potential clinical relevance of BNP measurements, we used likelihood-ratio tests to determine whether logistic regression models that included measurements of BNP and ASA class provided a significantly better fit than logistic regression models including ASA class alone.18 The exact version of the SAS-procedure Logistic was used to estimate odds ratios (ORs) and their confidence limits. The predictor variables considered in this model included preoperative BNP values and ASA classes; first considered separately and then in combination.
We enrolled 270 consecutive orthopaedic patients. The patient flow through the study is depicted in Fig. 1 and detailed baseline characteristics of the study population are summarized in Table 1. The durations of surgery and anaesthesia as well as the type of surgery and the volume of crystalloids infused during surgery did not influence the need for intensive care unit treatment or rehospitalization. Patients undergoing spinal surgery were hospitalized longer than patients undergoing all other types of surgery (10 vs. 8 days, P = 0.04). The operations were performed by a total of 15 different surgeons, performing between one and 46 operations (median number of operations per surgeon = 10). Postoperative cardiac events occurred after surgeries performed by experienced surgeons (>25 study operations).
B-type natriuretic peptide levels in patients undergoing orthopaedic surgery
Preoperative BNP levels rose with increasing ASA physical class scores [ASA I: 19 pg ml−1 (14–42); ASA II 33 pg ml−1 (16–66); ASA III 69 pg ml−1 (30–190); P < 0.01]. There were no significant correlations between BNP levels and the duration of surgery (r = 0.09, P = 0.14), the duration of anaesthesia (r = 0.09, P = 0.13) or the amount of volume replaced intraoperatively (crystalloids: r = 0.08, P = 0.19; colloids: r = 0.09, P = 0.14). Similarly, preoperative BNP levels were comparable among patients receiving general anaesthesia and patients receiving regional anaesthesia (P = 0.51). However, as expected, we observed a moderate positive correlation between preoperative BNP levels and age (r = 0.469, P < 0.01) and a weak negative correlation between BNP and the estimated glomerular filtration rate (r = −0.12, P = 0.05).
B-type natriuretic peptide as a marker for short-term in-hospital cardiac events
Overall, four (1.5%) patients had a postoperative cardiac event during the initial hospitalization. Two patients experienced postoperative angina pectoris; of these, one patient showed transient ST-segment changes (BNP 185 pg ml−1), whereas the other experienced a non-ST-segment elevation myocardial infarction (BNP 436 pg ml−1). The other two patients experiencing significant in-hospital cardiac events had acute heart failure (BNP 3958 pg ml−1) and tachycardiac atrial fibrillation (BNP 123 pg ml−1). The patient experiencing acute heart failure decompensation had previously been diagnosed with chronic, ischaemic heart failure and diabetes mellitus. Similarly, coronary artery disease was found in the patient experiencing postoperative ST-segment changes and the patient with a non-ST-segment elevation myocardial infarction. Tachycardiac atrial fibrillation occurred in a patient without previously known comorbidities. The patients' ages, the durations of surgery and anaesthesia, the type of surgery, the amount of intraoperative crystalloid and colloid volume as well as the intraoperative blood loss did not predict the occurrence of postoperative cardiac events. However, preoperative BNP levels were significantly higher in patients experiencing in-hospital cardiac events than in patients without such events [306 pg ml−1 (range 123–3958), mean 1156 pg ml−1 vs. 35 pg ml−1 (range 14–2074), mean 81 pg ml−1, P = 0.01].
To evaluate the potential of preoperative BNP levels to predict in-hospital cardiac events in patients undergoing orthopaedic surgery, a ROC analysis was performed. The area under the ROC curve (AUC) for the preoperative BNP value was 0.86 (95% CI 0.74–0.99; Fig. 2). The optimal predictive accuracy was achieved with a BNP threshold of 174 pg ml−1. When using this cut-off value, sensitivity and specificity were 75 and 90%, respectively. The positive and negative likelihood ratios were 7.67 and 0.27. In an exact logistic regression analysis including preoperative BNP values and ASA classes, both significantly predicted in-hospital cardiac events [OR 1.20, 95% CI 1.07–1.46, P < 0.01, for every increase in BNP by 100 pg ml−1; and OR 17.1, 95% CI 1.6–947, P = 0.01, for each additional level of ASA class, respectively]. Thus, BNP remains a statistically significant predictor even after additional inclusion of the ASA class into the model.
In a subgroup analysis including all patients with previously diagnosed cardiac disease, BNP retained a high predictive accuracy. The area under the curve for the prediction of in-hospital cardiac events in these patients was 0.85 (95% CI from 0.65 to <1). In a univariate analysis, the OR of 1.25 (95% CI 1.03–2.01) for the preoperative BNP value was similar to the one in the overall patient population. In contrast, the effect of ASA class could no longer be estimated with sufficient precision (OR 3.27, 95% CI 0.12–377), even before the inclusion of BNP values into the model.
B-type natriuretic peptide as a marker for long-term cardiac events
Overall 15 (5.5%) patients experienced a severe cardiac event during the follow-up period. The mean follow-up duration was 14.4 (SD 1.6) months. Six patients experienced transient ST-segment changes, six patients had a heart failure episode and atrial fibrillation occurred in two patients. A non-ST-segment elevation myocardial infarction occurred in only one patient. According to the ASA physical class score, five patients with long-term cardiac events were classified as ASA class II, nine patients were classified as class III and one patient was grouped as class IV. BNP levels were significantly higher in patients who experienced a cardiac event during the long-term follow-up period [148 pg ml−1 (IQR 30–479), mean 466 pg ml−1 vs. 35 pg ml−1 (IQR 17–76), mean 75 pg ml−1, P < 0.01] than in patients with an event-free survival. Similarly, preoperative BNP levels were significantly increased in the 25 patients who were rehospitalized during follow-up [62 pg ml−1 (IQR 29–140), mean 125 pg ml−1 vs. 34 pg ml−1 (IQR 16–77), mean 94 pg ml−1, P = 0.02]. Cardiac (nine cases), malignant (five cases) and infectious (three cases) diseases were the most common causes of rehospitalization.
The AUC determining the potential of BNP to predict long-term cardiac events was 0.71 (95% CI 0.57–0.84; Fig. 3). With an optimal, calculated BNP threshold of 145 pg ml−1, sensitivity and specificity were 53 and 90%. In an exact logistic regression analysis, preoperative BNP value and ASA class both significantly predicted long-term cardiac events (OR 1.21, 95% CI 1.06–1.47, P < 0.01 for every increase in BNP by 100 pg ml−1 and OR 6.7, 95% CI 2.2–24.0, P < 0.01 for each additional level of ASA class, respectively). Both predictors remained statistically significant when being considered in combination (OR 1.18, 95% CI 1.00–1.38, P = 0.05 for every increase in BNP by 100 pg ml−1 and OR 5.57, 95% CI 1.93–16.10, P < 0.01 for each additional level of ASA class).
The predictive accuracy of preoperative BNP levels in the patient subgroup with known cardiac diseases was modest (AUC 0.66) for the prediction of long-term cardiac events. ASA class failed to predict long-term cardiac events in patients with known cardiac diseases (OR 1.03, 95% CI 0.18–7.42, for each additional level of ASA class).
In the present study, we specifically examined the potential of preoperative BNP levels to predict short-term in-hospital and long-term cardiac events in 270 patients undergoing orthopaedic surgery. All patients had undergone and passed a preoperative risk assessment according to current ACC/AHA guidelines.12 Most importantly, we found a good predictive accuracy of preoperative BNP screening in orthopaedic patients, which appeared to be especially useful in patients with known cardiac diseases.
These results expand the findings of previous studies investigating the role of BNP in the preoperative risk stratification of patients undergoing noncardiac surgery. In a large study investigating 1590 patients undergoing various surgical procedures, BNP levels over 189 pg ml−1 were found to powerfully predict postoperative cardiac events.19 A comparable study selectively enrolling elderly patients reported NT-proBNP values above 201 pg ml−1 to be associated with perioperative cardiovascular complications.20 The diagnostic accuracy and the optimal cut-off value described in these studies were similar to our findings (AUC 0.86 vs. 0.84 vs. 0.86; cut-off 174 pg ml−1 vs. 189 pg ml−1 vs. NT-proBNP 201 pg ml−1). However, orthopaedic surgery constituted only 40 and 25% of the operations performed in these previous studies.
A recent meta-analysis evaluating the predictive value of BNP in noncardiac surgery also found elevated BNP values to powerfully predict all-cause and cardiac short-term mortality, with ORs reaching 9.2 and 23.8, respectively.21 Again, only a minority of patients studied in this meta-analysis underwent orthopaedic surgery. Furthermore, due to the diverse patient populations and various types of surgery, an optimal cut-off level of BNP could not be established by this meta-analysis. Of note, a small study examining 66 patients undergoing nonscheduled surgery for femoral neck fractures also confirmed the predictive accuracy of preoperative natriuretic peptide screening in the emergency setting. Preoperative NT-proBNP levels above 3984 pg ml−1 predicted perioperative cardiac events with an OR of 3.0.22
In the present study, preoperative BNP levels also predicted, in addition to in-hospital cardiac events, the occurrence of long-term cardiac complications. Similar results have recently been described in a follow-up study examining the 2-year outcome of 204 patients undergoing major noncardiac surgery.23 Interestingly, we observed decreased accuracy as well as lowered sensitivity for the prediction of long-term events compared with the prediction of in-hospital cardiac events. This observation suggests that BNP levels cannot be used as a single test predicting the long-term outcome but need to be considered in conjunction with all other clinical information. The low incidence of co-morbidities as well as the preserved functional status of our study population might at least have contributed to the reduced long-term test accuracy.
We observed both a low rate of in-hospital and long-term cardiac events in this unselected cohort of patients undergoing orthopaedic surgery. This observation readily indicates a high standard of perioperative patient management and is a strong argument against any extension of the stepwise preoperative cardiac assessment plan outlined in the ACC/AHA screening guidelines.12 However, due to their additive predictive information, BNP levels may be helpful in cases of diagnostic uncertainty. We believe that the enrolment of consecutive patients, the stringent blinding of all treating physicians and the long observational period to include the prediction of long-term cardiac events are advantages of our study.
However, in addition to a low event rate, several other limitations of our study need to be mentioned. First, we conducted a single centre study on orthopaedic surgical patients. Further multicentre studies are, therefore, warranted. However, as baseline characteristics, co-morbidities24,25 and results closely resembled those observed in previous work,19 our findings also seem applicable to other settings. In addition, the predictive potential of BNP established by examining patients undergoing cardiac10,11 or major noncardiac21,26 surgery also appears to extend to patients undergoing intermediate-risk orthopaedic surgery. Second, Youden's J index has not been directly validated for the small event numbers observed in this trial. However, bearing in mind that over 90% of event-free patients presented with BNP levels below the calculated cut-off value of 174 pg ml−1, that prior small-scale studies were in agreement with the general literature11,22 and that our findings mirror previous results,19,20 we consider BNP values above 160–180 pg ml−1 to be good predictors of in-hospital cardiac complications.
In conclusion, preoperative BNP levels can predict short-term and long-term cardiac events in patients undergoing low-to-intermediate-risk orthopaedic surgery. Owing to the simplicity and widespread availability of BNP testing, this biomarker used in addition to an ACC/AHA guideline-based preoperative risk assessment might be a useful tool in the preoperative evaluation of patients undergoing orthopaedic surgery.
The present study was supported by grants from the Swiss National Science Foundation, the Swiss Heart Foundation and the Novartis Foundation to CM. We also thank Dr Miodrag Filipovic of the Department of Anesthesia at the University Hospital Basel for his valuable input during the design of this study as well as Professor Walter Dick and Professor Bernhard Jeanneret of the Department of Orthopedic Surgery at the University Hospital Basel for their help during the patient enrolment process as well as their valuable comments during manuscript writing.
1 Maddox TM. Preoperative cardiovascular evaluation for noncardiac surgery. Mt Sinai J Med 2005; 72:185–192.
2 Silver MA. The natriuretic peptide system: kidney and cardiovascular effects. Curr Opin Nephrol Hypertens 2006; 15:14–21.
3 Mueller C, Breidthardt T, Laule-Kilian K, et al
. The integration of BNP and NT-proBNP into clinical medicine. Swiss Med Wkly 2007; 137:4–12.
4 McCullough PA, Duc P, Omland T, et al
. B-type natriuretic peptide
and renal function in the diagnosis of heart failure: an analysis from the breathing not properly multinational study. Am J Kidney Dis 2003; 41:571–579.
5 Daniels LB, Clopton P, Bhalla V, et al
. How obesity affects the cut-points for B-type natriuretic peptide
in the diagnosis of acute heart failure. Results from the breathing not properly multinational study. Am Heart J 2006; 151:999–1005.
6 Daniels LB, Maisel AS. Natriuretic peptides. J Am Coll Cardiol 2007; 50:2357–2368.
7 Park MH, Scott RL, Uber PA, et al
. Usefulness of B-type natriuretic peptide
levels in predicting hemodynamic perturbations after heart transplantation despite preserved left ventricular systolic function. Am J Cardiol 2002; 90:1326–1329.
8 Christ-Crain M, Breidthardt T, Stolz D, et al.
Use of B-type natriuretic peptide
in the risk stratification of community-acquired pneumonia. J Intern Med
9 Bibbins-Domingo K, Gupta R, Na B, et al
. N-terminal fragment of the prohormone brain-type natriuretic peptide (NT-proBNP), cardiovascular events, and mortality in patients with stable coronary heart disease. JAMA 2007; 297:169–176.
10 Hutfless R, Kazanegra R, Madani M, et al
. Utility of B-type natriuretic peptide
in predicting postoperative complications
and outcomes in patients undergoing heart surgery. J Am Coll Cardiol 2004; 43:1873–1879.
11 Eliasdottir SB, Klemenzson G, Torfason B, Valsson F. Brain natriuretic peptide is a good predictor for outcome
in cardiac surgery. Acta Anaesthesiol Scand 2008; 52:182–187.
12 Eagle KA, Berger PB, Calkins H, et al
. ACC/AHA guideline update for perioperative cardiovascular evaluation for noncardiac surgery: executive summary. A report of the American College of Cardiology/American Heart Association task force on practice guidelines (committee to update the 1996 guidelines on perioperative cardiovascular evaluation for noncardiac surgery). Anesth Analg 2002; 94:1052–1064.
13 Cabanes L, Richaud-Thiriez B, Fulla Y, et al
. Brain natriuretic peptide blood levels in the differential diagnosis of dyspnea. Chest 2001; 120:2047–2050.
14 Cheng V, Kazanagra R, Garcia A, et al
. A rapid bedside test for B-type peptide predicts treatment outcomes in patients admitted for decompensated heart failure: a pilot study. J Am Coll Cardiol 2001; 37:386–391.
15 Levey AS, Bosch JP, Lewis JB, et al
. 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.
16 Thygesen K, Alpert JS, White HD, et al
. Universal definition of myocardial infarction. Circulation 2007; 116:2634–2653.
17 Apple FS, Maturen AJ, Mullins RE, et al
. Multicenter clinical and analytical evaluation of the AxSYM troponin-I immunoassay to assist in the diagnosis of myocardial infarction. Clin Chem 1999; 45:206–212.
18 Harbarth S, Holeckova K, Froidevaux C, et al
. Diagnostic value of procalcitonin, interleukin-6, and interleukin-8 in critically ill patients admitted with suspected sepsis. Am J Respir Crit Care Med 2001; 164:396–402.
19 Dernellis J, Panaretou M. Assessment of cardiac risk before noncardiac surgery: brain natriuretic peptide in 1590 patients. Heart 2006; 92:1645–1650.
20 Yun KH, Jeong MH, Oh SK, et al
. Preoperative plasma N-terminal pro-brain natriuretic peptide concentration and perioperative cardiovascular risk in elderly patients. Circ J 2008; 72:195–199.
21 Ryding AD, Kumar S, Worthington AM, Burgess D. Prognostic value of brain natriuretic peptide in noncardiac surgery: a meta-analysis. Anesthesiology 2009; 111:311–319.
22 Oscarsson A, Fredrikson M, Sorliden M, et al
. N-terminal fragment of pro-B-type natriuretic peptide
is a predictor of cardiac events in high-risk patients undergoing acute hip fracture surgery. Br J Anaesth 2009; 103:206–212.
23 Cuthbertson BH, Amiri AR, Croal BL, et al
. Utility of B-type natriuretic peptide
in predicting medium-term mortality in patients undergoing major noncardiac surgery. Am J Cardiol 2007; 100:1310–1313.
24 Cram P, Vaughan-Sarrazin MS, Wolf B, et al
. A comparison of total hip and knee replacement in specialty and general hospitals. J Bone Joint Surg Am 2007; 89:1675–1684.
25 Meding JB, Wing JT, Keating EM, Ritter MA. Total knee arthroplasty for isolated patellofemoral arthritis in younger patients. Clin Orthop Relat Res 2007; 464:78–82.
26 Rodseth RN, Padayachee L, Biccard BM. A meta-analysis of the utility of preoperative brain natriuretic peptide in predicting early and intermediate-term mortality and major adverse cardiac events in vascular surgical patients. Anaesthesia 2008; 63:1226–1233.