It is estimated that 4% of the world's population undergo surgery every year, and that approximately 230 million operations are performed annually.1 In Europe, with a population of approximately 500 million people, at least 20 million surgical operations are performed annually2 with a crude in-hospital mortality rate of 4%.3 An ageing population might lead to an increase in these figures in the coming years. Preoperative identification of patients at risk of complications would enable a selective approach aimed at reducing the incidence of morbidity and mortality during the perioperative period. Patients at risk are most commonly identified by assigning a cardiovascular risk score, although this approach is subject to limitations and has scant predictive power in major noncardiac surgery.4
Evidence from nonsurgical groups suggests that biomarkers can improve stratification of patients according to their risk of mortality and cardiovascular events. The most promising potential biomarkers are troponins and natriuretic peptides. The large international VISION Study (Vascular Events in Noncardiac Surgery Patients Cohort Evaluation) found that measurement of troponin T after noncardiac surgery was strongly associated with 30-day mortality.5 Brain natriuretic peptide (BNP) and the amino terminal fraction of its propeptide (NT-proBNP) belong to a family of hormones involved in homeostasis and in the cardiovascular remodelling that occurs in the myocardium in response to increased stress, inflammation and ischaemia.6 Elevated values for these markers, with some caveats, indicate a cardiac origin. In contrast to atrial natriuretic peptide, BNP is not only secreted from the atria but also from the ventricles, and its circulating concentrations are raised in both symptomatic and asymptomatic patients with left ventricular dysfunction. Detecting increased preoperative and postoperative plasma levels has been shown to be effective in predicting perioperative morbidity and mortality.7
The aim of the present study was to determine the incidence of high levels of preoperative and postoperative NT-proBNP in adults with cardiovascular risk factors who were scheduled for major noncardiac surgery. We also investigated the relationship between these levels and cardiovascular complications, including all-cause mortality, during the first 30 days after surgery.
Materials and methods
We performed a prospective cohort study at Hospital Universitario de La Princesa, Madrid, Spain between 1 October 2011 and 1 July 2013. The study was approved on 22 September 2011 by an independent ethics committee (Comité Ético de Investigación Clínica del Hospital Universitario de La Princesa, Madrid, Spain; protocol number EC1835), and written consent was obtained from all participants.
The follow-up period comprised the 30 days after surgery. During this period, variables comprising the composite endpoint ’cardiovascular complications’ were recorded as follows: death from all causes, resuscitated cardiac arrest, stroke, exacerbation of congestive heart failure (CHF) and acute coronary syndrome. Data were gathered from medical records, and survival was confirmed via telephone calls.
The inclusion criteria were age more than 45 years, scheduled elective noncardiac surgery and a surgical risk estimated as intermediate or high according to the guidelines of the European Society of Cardiology (ESC) and the European Society of Anaesthesiology.2 Patients undergoing emergency surgery and surgical procedures classified as low risk were excluded from the study. To meet the inclusion criteria, patients had to meet at least one of the following cardiovascular risk factors:
- A history of ischaemic heart disease defined by the presence of one or more of the following in their medical records: a history of angina or heart attack (according to the definition of the ESC);8 a positive stress test result; myocardial ischaemia determined via a nuclear stress test or ultrasound stress test; or coronary artery stenosis of more than 50% of arterial lumen observed during coronary artery catheterisation.
- Peripheral arterial vascular disease as defined by one or more of the following criteria: a history of intermittent claudication (pain in the lower limbs when walking, which ceases within 10 min of rest); angiography or a Doppler study of the lower limbs showing arterial stenosis more than 70%; or an ankle-brachial index 0.90 or less in either leg when at rest.
- A history of stroke with focal neurological deficit lasting at least 1 week from the onset of symptoms.
- A history of hospitalisation because of CHF in the 3 years before the study.
- Scheduled surgery of the abdominal aorta.
- Age more than 70 years with at least one of the following comorbidities: CHF, diabetes mellitus and plasma creatinine at least 177 μmol l−1 (2 mg dl−1).
Serum levels were determined in all patients 12 h before surgery and 24 h after surgery. We used Elecsys proBNP II, an in vitro immunoassay, to measure NT-proBNP concentrations in human serum (Cobas Roche). We defined normal NT-proBNP levels as less than 300 pg ml−1 (local reference values), and high levels as more than 300 pg ml−1. For high levels, we established two arbitrary divisions: values ranging between 300 and 1000 pg ml−1; and values more than 1000 pg ml−1.
Sample size was calculated using standard α and β values of 0.05 and 0.2 (power 0.8) and a two-sided test. With an estimated incidence of events of 5% under the null hypothesis, a sample of 292 patients was required to demonstrate a significant odds ratio of at least 2.25 with a maximum proportion of variance explained by other covariates, R2, of 0.25.
We used the data collected to carry out a descriptive analysis of the patients’ characteristics and of the primary outcome by recording measures of central tendency (mean) and dispersion (SD) for quantitative variables and frequencies and proportions for qualitative variables. Wherever necessary, comparisons of proportions were tested using the χ2 test or the Fisher exact test.
We used univariate binary logistic regression models to analyse the association between NT-proBNP levels and cardiovascular risk factors, cardiovascular complications and all-cause mortality after 30 days of follow-up. Cardiovascular risk factors were divided into three groups, according to the revised cardiac risk index (rCRI) score9 (rCRI of ≤1, 2 or ≥3). To control for the effect of potentially confounding variables, we used multivariate logistic regression models for variables with P < 0.10 in the univariate analysis. The confidence interval for all estimates was 95%. Values were considered statistically significant when P < 0.05.
All analyses were carried out at the Instituto de Investigación Sanitaria-Princesa, Madrid using Stata v.12.
We recruited 320 patients, of whom 16 were excluded owing to missing data during the follow-up (blood samples not extracted when required), leaving 304 for analysis. Their personal data are shown in Table 1.
The incidence of cardiovascular complications, including all-cause mortality, during the 30 days after surgery, was 7.8% (n = 25), although each of the patients may have been subject to more than one complication. The 30-day all-cause mortality rate was 4.3% (n = 13). The frequency of resuscitated cardiac arrest was 1% (n = 3), acute coronary syndrome 1% (n = 3), acute episodes of CHF 3.3% (n = 10) and stroke 0.3% (n = 1).
Preoperative NT-proBNP levels were within the normal range (<300 pg ml−1) in 157 patients (51.6%). The cardiovascular complications rate in this group was 1.3% (n = 2), with a mortality rate of 0.6% (n = 1). The remaining 147 patients (48.4%) had high preoperative NT-proBNP serum levels (>300 pg ml−1), with a cardiovascular complications rate of 15.6% (n = 23) and mortality rate of 8.2% (n = 12). NT-proBNP values ranged between 300 and 1000 pg ml−1 in 26% of patients (n = 79), with a cardiovascular complications rate of 10.1% (n = 8) and mortality rate of 3.8% (n = 3). NT-proBNP values more than 1000 pg ml−1 were detected in 22.4% of patients (n = 68), with a cardiovascular complications rate of 22.1% (n = 15) and a mortality rate of 13.2% (n = 9). Preoperative NT-proBNP serum values more than 300 pg ml−1 were statistically significantly associated with the primary outcome, 30-day cardiovascular complications rate (including all-cause mortality) (Table 2).
Postoperative NT-proBNP levels were within the normal range (<300 pg ml−1) in 150 of the patients (49.3%), with a cardiovascular complications rate of 1.3% (n = 2) and mortality rate of 0.7% (n = 1). High postoperative NT-proBNP serum levels (>300 pg ml−1) were detected in 154 patients (50.7%), who had a cardiovascular complications rate of 14.9% (n = 23) and mortality rate of 7.8% (n = 12). NT-proBNP levels of 300 to 1000 pg ml−1 were found in 26.0% of cases (n = 79) and cardiovascular complications in 5.1% (n = 4). The mortality rate was 3.8% (n = 3). NT-proBNP levels more than 1000 pg ml−1 were detected in 24.7% of cases (n = 75) and cardiovascular complications in 25.3% (n = 19). The mortality rate was 12.0% (n = 9). The risk of postoperative cardiovascular complications, including all-cause mortality, was significantly higher in patients with postoperative levels of NT-proBNP at least 1000 pg ml−1 (Table 2).
Multivariate analysis revealed a significant relationship between the primary outcome and preoperative NT-proBNP values of 300 to 1000 pg ml−1 (P = 0.009; OR, 8.2) and more than 1000 pg ml−1 (P < 0.001; OR, 19.5). Preoperative NT-proBNP serum values more than 1000 pg ml−1 were an independent factor associated with 30-day all-cause mortality (P = 0.002; OR, 27.2) (Table 3). In the case of postoperative NT-proBNP, the multivariate analysis only showed statistically significant relationships between the incidence of the primary outcome and NT-proBNP values more than 1000 pg ml−1 (P < 0.001; OR, 22.8). Postoperative NT-proBNP serum values more than 1000 pg ml−1 were an independent factor associated with 30-day all-cause mortality (P = 0.003; OR, 24.4) (Table 3).
The univariate analysis of the rCRI score showed that only rCRI at least 3 was associated with the primary outcome (P = 0.004; OR, 4.7; 95%CI, 1.61 to 13.42).
In our series, the overall 30-day mortality was 4.3% (n = 13), which is consistent with the range reported in similar studies, such as the POISE Study,10 a systematic review and meta-analysis by Rodseth et al.11 and the European Surgical Outcomes Study by Pearse et al.3
We consider two findings to be especially significant. The first is the high incidence of patients with high NT-proBNP levels (>300 pg ml−1) before surgery (48.4%; n = 147) and after (50.7%; n = 154). Of special note is the detection of clearly pathological NT-proBNP values more than 1000 pg ml−1 in 22.4% (n = 68) of patients during the preoperative and in 24.7% (n = 75) during the postoperative period. This incidence can be partly explained by the fact that we selected patients with pre-existing cardiovascular risk factors. In the systematic review and meta-analysis published by Karthikeyan et al.12 in 2009 (3281 patients scheduled for noncardiac surgery), a preoperative increase in BNP or NT-proBNP levels was detected in 24.8% of patients, although the authors did not specify the threshold reference levels or whether the cohort studied was confined to patients with diseases or cardiovascular risk factors, as in our study, or whether it included all types of patients. The same limitation was present in the systematic review published by Rodseth et al.13 in 2014, which included 1921 patients and reported increases in postoperative levels of BNP or NT-proBNP in up to 76% of patients. In any event, if we assume that 30% of patients scheduled for noncardiac surgery have associated cardiovascular conditions,2 then around 50% of these patients will have high biomarker levels. This observation would appear to be relevant.
The second finding was that high levels of preoperative and postoperative NT-proBNP were associated with significantly higher incidences of cardiovascular complications and mortality, regardless of rCRI levels. The higher the biomarker levels, the greater the association (Tables 2 and 3). The negative predictive value of NT-proBNP also needs to be highlighted, as there is a clear association between normal biomarker values and a favourable outcome following surgery. In accordance with other series and meta-analyses,14 we found that the mortality rate was low (0.6%; n = 1) in patients with preoperative and postoperative values less than 300 pg ml−1.
The ability of NT-proBNP levels to predict morbidity and mortality in a perioperative setting can probably be explained by their sensitivity to alterations in ventricular function. Even small alterations in ventricular function brought on by transitional periods of ischaemia, volume overload or both15 lead to increases in plasma NT-proBNP levels, whereas higher levels of NT-proBNP are found in patients with poorer ventricular function or in patients with ischaemic heart disease. Other studies report the link between ‘silent heart disease’, which is only detectable via high biomarker levels, and a higher mortality rate among nonsurgical groups.7 In surgical cohorts, the higher rates of morbidity and mortality, such as those that were found with high biomarker levels in the present study, are probably associated with decreased tolerance to haemodynamic changes and with the pro-inflammatory response brought on by general anaesthesia and surgery.16 Several studies17 and meta-analyses12,13,16 have confirmed this link between high preoperative biomarker levels and mortality. During the postoperative period, published proBNP data suggest that it is the absolute postoperative value, rather than the increase between preoperative and postoperative pro-BNP levels, which is associated with a higher rate of morbidity and mortality.11 This observation has also been made in studies evaluating the postoperative predictive value of troponin.
There has been considerable diversity in the timing of the collection of NT-proBNP samples in previous studies, with a range from 24 days before surgery to 7 days after surgery,12,13,16 and in the pathological threshold values. We recorded NT-proBNP levels 12 h prior to surgery and 24 h after surgery, as we consider these to be the most representative values of the patient's perioperative clinical situation. Nonetheless, we believe that any therapeutic measures could still prove effective within a period of a few days or even weeks before or after surgery.
Our findings are consistent with the ESC/European Society of Anaesthesiology Guidelines2 published in 2014, which recommend measuring both preoperative and postoperative NT-proBNP levels for perioperative risk stratification in patients scheduled for noncardiac surgery with an alteration in functional status (less than four metabolic equivalents), or at least one factor on the rCRI in the case of vascular surgery, or two factors in the case of nonvascular surgery. Preoperative NT-proBNP measurement will improve risk stratification, whereas postoperative measurement can aid in the early diagnosis of clinically silent disease.
The present study is subject to a series of limitations. First, the mean age of the cohort was 73.6 ± 9.4 years, and 72% of our patients were aged more than 70 years. This may limit our ability to extrapolate our findings to other groups. Although an association between age and natriuretic peptide levels has been reported in healthy individuals, its influence on natriuretic peptides in CHF was negligible; therefore when natriuretic peptides are used for prognosis, there are no recommendations for indexing them by age.18 Second, the association between NT-proBNP blood levels and renal function is complex. Patients with chronic renal dysfunction tend to have higher systemic blood pressure and greater ventricular mass, both of which could lead to high ‘true’ NT-proBNP levels. However, NT-proBNP levels could also be elevated because of decreased renal filtration or decreased creatinine clearance.19 Our study includes patients with cardiovascular risk factors, including a creatinine value more than 177 μmol l−1. Serum creatinine more than 177 μmol l−1 was detected in 28 cases of the 304 patients in our series (9.2%); this finding may decrease the specificity of NT-proBNP in predicting cardiac complications and mortality. We conducted a sensitivity analysis in which patients with serum creatinine at least 177 μmol l−1 were excluded and found a nonsignificant variation in the primary outcome (7.2%) and in the effect on the results. Therefore we concluded that the results of our study were not affected by serum creatinine values. Finally, the cut-off point for the relationship between biomarker plasma levels and the appearance of complications remains open to debate. The range of laboratory values we used is common in clinical analysis, namely, an NT-proBNP reference threshold of 300 pg ml−1. We established two arbitrary strata above this level: 300 to 1000 pg ml−1 and more than 1000 pg ml−1. The 1000 pg ml−1 cut-off was chosen taking into account that levels more than 1000 pg ml−1 have been associated with severe heart failure and poor prognosis.20 The optimal cut-off point may well require further research, although we think it unlikely that this limitation could have influenced our main findings.
In conclusion, in a series of patients with cardiac risk factors who were scheduled for noncardiac surgery, we found that nearly 50% had high blood levels of preoperative and postoperative NT-proBNP and that these were associated with a greater incidence of cardiovascular complications, including all-cause mortality. The high prevalence of patients with these characteristics highlights the need to design clinical trials that evaluate strategies aimed at reducing perioperative morbidity and mortality rates in noncardiac surgery. Such strategies can be introduced before surgery (preconditioning), during surgery (haemodynamic monitoring and stress reduction strategies) and after surgery (admission to the critical care unit, enhanced recovery programme).
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflicts of interest: none.
1. Weiser TG, Rogenbogen SE, Thompson KD, et al. An estimation of the global volume of surgery: a modelling strategy based on available data. Lancet
2. Kristensen SD, Knuuti J, Saraste A, et al. 2014 ESC/ESA Guidelines on noncardiac surgery: cardiovascular assessment and management: The Joint Task Force on noncardiac surgery: cardiovascular assessment and management of the European Society of Cardiology (ESC) and the European Society of Anaesthesiology (ESA). Eur J Anaesthesiol
3. Pearse RM, Moreno RP, Bauer P, et al. Mortality after surgery in Europe: a 7-day cohort study. Lancet
4. Ford MK, Beattie WS, Wijeysundera DN. Systematic review: prediction of perioperative cardiac complications and mortality by the revised cardiac risk index. Ann Intern Med
5. Devereaux PJ, Chan MT, Alonso-Coello P, et al. Association between postoperative troponin levels and 30-day mortality among patients undergoing noncardiac surgery. JAMA
6. Rodseth RN. B type natriuretic peptide. A diagnostic breakthrough in perioperative cardiac risk assessment? Anaesthesia
7. Biccard BM, Deveraux PJ, Rodseth RN. Cardiac biomarkers in the prediction of risk in the noncardiac surgery setting. Anaesthesia
8. Hamm CW, Bassand JP, Agewall S, et al. ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: The Task Force for the management of acute coronary syndromes (ACS) in patients presenting without persistent ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J
9. Lee TH, Marcantonio ER, Mangione CM, et al. Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery. Circulation
10. Devereaux PJ, Yang H, Yusuf S, et al. The effect of extended-release metoprolol succinate in patients undergoing noncardiac surgery (POISE trial): a randomised controlled trial. Lancet
11. Rodseth RN, Biccard BM, Chu R, et al. Postoperative B-type natriurtic peptide for prediction of major cardiac events in patients undergoing noncardiac surgery. Anesthesiology
12. Karthikeyan G, Moncur RA, Levine O, et al. Is a preoperative brain natriuretic peptide or N-terminal pro-B-type natriuretic peptide measurement an independent predictor of adverse cardiovascular outcomes within 30 days of noncardiac surgery? A systematic review and meta-analysis of observational studies. J Am Coll Cardiol
13. Rodseth RN, Bicard BM, Le Manach Y, et al. The prognostic value of preoperative and postoperative B-type natriuretic peptides (BNP and NT proBNP) in patients having noncardiac surgery. A systematic review and individual patient data meta-analysis. J Am Coll Cardiol
14. Bolliger D, Seeberg MD, Filipovic M. Preoperative cardiac risk assessment in noncardiac surgery. Are natriuretic peptides the magic bullet? J Am Coll Cardiol
15. Struthers A, Lang C. The potential to improve primary prevention in the future by using BNP/N-BNP as an indicator of silent pancardiac target organ damage: BNP/N-BNP could become for the heart what microalbuminuria is for the kidney. European Heart Journal
16. Ryding A, Kumar S, Worthington AM, Burgess D. Prognostic value of brain natriuretic peptide in noncardiac surgery. A meta-analysis. Anesthesiology
17. Goei D, Flu WJ, Boersma E, et al. The interrelationship between preoperative anemia and N-terminal pro-B-type natriuretic peptide: the effect on predicting postoperative cardiac outcome in vascular surgery patients. Anesth Analg
18. Hogenhuis J, Voors AA, Jaarsma T, et al. Influence of age on natriuretic peptides in patients with chronic heart failure: a comparison between ANP/NT-ANP and BNP/NT-proBNP. Eur J Heart Fail
19. Daniels LB, Maisel AS. Natriuretic peptides. J Am Coll Cardiol
20. Maisel AS, Mueller C, Adams K, et al. State of the art: using natriuretic peptide levels in clinical practice. Eur J Heart Fail