Introduction

Every year, 230 million adults undergo non-cardiac surgery worldwide (Weiser et al. 2008). In patients who are 45 years or older, 8% will suffer Myocardial Injury after Non-cardiac Surgery (MINS) (Botto et al. 2014) and 2% will die within 30 days (Devereaux et al. 2012). MINS is typically asymptomatic without the usual features of chest pain and electrocardiogram changes seen with myocardial infarction (Botto et al. 2014). MINS has prognostic importance up to a year after surgery (Puelacher et al. 2018).

The biomarker, B-type natriuretic peptide (BNP), has been identified as an important preoperative predictor of perioperative cardiovascular events (Rodseth et al. 2014). Despite this strong association, it remains unknown whether preoperative treatment to normalize or reduce NP (B-type natriuretic peptide and N-terminal pro-B-type natriuretic peptide) levels prior to surgery would result in improved perioperative outcomes. This is a novel approach that has not been tested in clinical trials involving surgical patients. Thus, a systematic review of non-surgical trials is necessary to establish whether this approach is safe and effective before it can be tested in a surgical population.

The objective of this systematic review of clinical trials was to determine whether, in adults, medical patients with cardiac failure, NP-directed medical therapy is able to decrease NP levels and whether this is associated with increased survival.

These data could then be used to inform preoperative protocols aimed at decreasing NPs prior to surgery, with the intention of improving perioperative cardiovascular outcomes.

Methods

Protocol and registration

The protocol was registered with PROSPERO (CRD42017051468). The Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) guidelines were adhered to (Moher et al. 2009).

Eligibility criteria

Clinical trials of adult medical patients who were randomized to either NP guided medical therapy or standard care were eligible. We included trials which used NPs to (i) guide medical therapy in non-surgical patients, (ii) up-titrate or modify medical therapy in the response to NP levels, or (iii) included exercise as part of cardiac rehabilitation in non-surgical patients. We required that the trials report the subsequent changes in NP levels. We excluded trials that (i) monitored natriuretic peptides for prognostic or diagnostic purposes, without a strategy to lower natriuretic peptide levels, (ii) reviews of natriuretic peptide or biomarker physiology, and (iii) trials reporting natriuretic peptides in patients with acute myocardial infarction, pulmonary hypertension, cardiac resynchronization therapy, and left ventricular devices.

Information sources, search, and study selection

Three searches were conducted using search terms “brain natriuretic peptide” AND “treatment,” “brain natriuretic peptide” AND “heart failure” and “brain natriuretic peptide” AND “exercise.” The following databases were accessed; MEDLINE (PubMed), CINAHL (EBSCO host), EMBASE (EBSCO host), ProQuest, Web of Science, and Cochrane database. No date filter was used. An example of the search is shown in Additional file 1. The initial search was conducted on 22 December 2016 and updated on the 4 March 2018.

Data collection process

Titles were screened for potential inclusion by CA and PG. The abstracts of the potential papers identified through the title search were then screened using inclusion and exclusion criteria by CA and PG. The full texts of potential trials were then extracted for detailed review and analysis. Reference lists were searched for additional papers that could be included in this review. Data extraction was done by one author (CA) and then checked by a co-author. When the required data was not presented in the publication, the authors were contacted for these data.

Data items

We extracted data on the NP reduction within the first six months of randomization and mortality at 6 months. The data items extracted for this review are shown in Additional file 1: Table S1.

Outcomes

The primary outcome for this review was to determine whether a NP-directed medical therapy protocol is effective in reducing NP levels at 6 months after initiation of therapy compared to standard care. The secondary outcome was to determine whether NP-directed medical therapy decreases mortality at 6 months and at the end of the trial. The safety outcomes of changes in medical therapy were evaluated. Specific medical treatment strategies are described.

Risk of bias in individual studies

Assessment of bias in the studies was conducted by CA and verified by BB following discussion. Each randomized trial was assessed using the Cochrane Collaboration risk of bias tool, assessing selection bias, concealment bias, performance bias, detection bias, attrition bias, and other bias. Studies were assessed as having a low, unclear, or high risk of bias.

Summary measures and synthesis of results

The statistical analyses were conducted using Review Manager Version 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Those trials which had data on NP levels within 6 months of therapy initiation were analyzed using standardized mean difference (SMD), and these data are presented as a forest plot. Mean and standard deviation (SD) of NP levels were used and those trials which reported NP levels as the median and interquartile range (IQR) were converted to mean and SD, using the formula proposed by Wan et al. (Wan et al. 2014). Reporting the SMD allowed for the inclusion of all trials, whether BNP or NT-proBNP, was used to monitor the medical therapeutic response. SMD addresses the difference in the effect size for an intervention when the units of measurement differ between trials e.g. use of BNP or NT-proBNP. The SMD is the difference between groups in the mean endpoint divided by the SD of the control group or pooled SD of the treatment and control groups (Hedges’ g) (Guyatt et al. 2015).

A meta-analysis of mortality within 6 months of the initiation of therapy, with subgroup analyses at 4 and 6 months was conducted. The results are reported as relative risk (RR), with 95% confidence intervals (CI), and presented as forest plots. Random effects models were used where the I2 statistic > 2s5% (representing significant heterogeneity), otherwise a fixed-effects model was used.

Risk of bias across studies

Risk of bias across studies was assessed with funnel plots for NP reduction and mortality.

Post hoc

After extracting and analyzing the data, it was noticed that the methodology used in the exercise trials differed significantly from the medical therapy trials. This difference was so substantial that we deemed it inappropriate to pool the two interventions. We therefore made a post hoc decision to separate the exercise studies from the medical therapy studies. These exercise study data are presented in the accompanying publication (CS Alphonsus et al. 2019).

Results

Study selection

Sixty-four full-text articles were reviewed for potential inclusion and 26 trials (presented in 27 publications) met the inclusion criteria. An additional eight trials were added from references (Fig. 1). Eighteen trials of medical therapy interventions were identified (reported in 19 publications; 1 trial was reported in 2 separate papers) (Maeder et al. 2013; Pfisterer et al. 2009) fulfilled the inclusion criteria, although only 14 trials provided data for this review’s outcomes. The 16 exercise trials were subsequently removed from this review, following the post hoc decision to present these trials in a separate paper (CS Alphonsus et. al. 2019).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

We evaluated previous systematic reviews identified in the search using the AMSTAR format (Additional file 1: Table S2).

Study characteristics of included studies

The characteristics of the included clinical trials are shown in Table 1. These trials included adult patients of 18 years and older. The majority of the trials examined outpatient NP-directed medical therapy, with follow-up of 15 months or more. In 10 out of 18 trials, patients were seen by a specialist at a clinic (Anguita et al. 2010; Berger et al. 2010; Eurlings et al. 2010; Januzzi et al. 2011; Jourdain et al. 2007; Lainchbury et al. 2009; Murdoch et al. 1999; Persson et al. 2010; Schou et al. 2013; Shah et al. 2011; Troughton et al. 2000). Most patients were randomized once heart failure was medically stabilized and 12 out of 18 trials enrolled patients with EF ≤ 50% (Beck-da-Silva et al. 2005; Felker et al. 2017; Januzzi et al. 2011; Jourdain et al. 2007; Karlstrom et al. 2011; Murdoch et al. 1999; Ozkara et al. 2007; Persson et al. 2010; Schou et al. 2013; Shah et al. 2011; Stienen et al. 2018; Troughton et al. 2000). The remainder of the trials combined patients with preserved and reduced ejection fraction (Anguita et al. 2010; Berger et al. 2010; Carubelli et al. 2016; Eurlings et al. 2010; Lainchbury et al. 2009; Maeder et al. 2013; Pfisterer et al. 2009).

Table 1 Characteristics of included clinical trials
Full size table

The conduct of the trial intervention arms is shown in Table 2. All trials randomized patients into NP-directed medical therapy or clinical/usual care. Two trials were three-arm trials, but for this analysis, only the interventional and usual care arms were included (Berger et al. 2010; Lainchbury et al. 2009). In the majority of trials, the NP threshold for inclusion was consistent across age and gender, with the exception of three trials, where the threshold was either age- or gender-specific (Karlstrom et al. 2011; Maeder et al. 2013; Persson et al. 2010; Pfisterer et al. 2009). Nine trials set population NP targets (Anguita et al. 2010; Berger et al. 2010; Carubelli et al. 2016; Felker et al. 2017; Januzzi et al. 2011; Jourdain et al. 2007; Lainchbury et al. 2009; Murdoch et al. 1999; Troughton et al. 2000), eight trials set individualized NP targets (Beck-da-Silva et al. 2005; Eurlings et al. 2010; Karlstrom et al. 2011; Maeder et al. 2013; Persson et al. 2010; Pfisterer et al. 2009; Schou et al. 2013; Shah et al. 2011; Stienen et al. 2018), and one trial had no set NP target, but directed medical therapy to reduce the NP level (Ozkara et al. 2007). The management of the standard care arms is shown in Additional file 1.

Table 2 The conduct of the natriuretic-peptide (NP)-directed clinical trials
Full size table

Two trials were stopped early (Felker et al. 2017; Januzzi et al. 2011), Felker et. al for the benefit, and Januzzi et al. for futility.

Risk of bias within studies and across studies

The risk of bias of the included trials is shown in Additional file 1: Figures S1 and S2. The random sequence generation was unclear in half the trials, and blinding of patients and investigators was low. Many trials did not clearly document if outcome assessors were blinded. The funnel plots for SMD (Additional file 1: Figure S3), and 6-month mortality (Additional file 1: Figure S4) did not suggest publication bias.

Results of individual studies and synthesis of results

The efficacy of an NP-directed medical therapy in reducing NP levels within 6 months compared to standard care

Fourteen out of 18 medical therapy trials presented data on change in NP levels during the trial (Anguita et al. 2010; Carubelli et al. 2016; Eurlings et al. 2010; Felker et al. 2017; Januzzi et al. 2011; Jourdain et al. 2007; Karlstrom et al. 2011; Lainchbury et al. 2009; Maeder et al. 2013; Murdoch et al. 1999; Persson et al. 2010; Pfisterer et al. 2009; Schou et al. 2013; Shah et al. 2011; Stienen et al. 2018; Troughton et al. 2000), of which 7 out of 14 trials presented data on NP levels within the first 6 months of the trial (Anguita et al. 2010; Carubelli et al. 2016; Felker et al. 2017; Maeder et al. 2013; Pfisterer et al. 2009; Shah et al. 2011; Stienen et al. 2018). Three trials Shah, Carubelli, and Stienen were excluded as the data was reported at differing time points before 6 months: Stienen (mean 12 ± 10 days) (Stienen et al. 2018), Carubelli (mean 11 ± 9 days) (Carubelli et al. 2016) and Shah (4 months) (Shah et al. 2011). The overall point estimate of the four remaining trials was nonsignificant at 6 months of NP-directed medical therapy with low heterogeneity in the included trials (Fig. 2), (SMD − 0.04, 95% CI − 0.16, 0.07).

Fig. 2.
figure 2

Efficacy of natriuretic peptide-directed medical therapy versus control in reducing BNP-levels within 6 months

(Standardised mean difference in natriuretic peptide levels in NP-directed medical therapy clinical trials). SD, standard deviation; CI, confidence interval

Full size image

Reduction in NP levels and its association with mortality

Seventeen out of 18 studies reported mortality at trial completion (Anguita et al. 2010; Beck-da-Silva et al. 2005; Berger et al. 2010; Carubelli et al. 2016; Eurlings et al. 2010; Felker et al. 2017; Januzzi et al. 2011; Jourdain et al. 2007; Karlstrom et al. 2011; Lainchbury et al. 2009; Maeder et al. 2013; Ozkara et al. 2007; Persson et al. 2010; Pfisterer et al. 2009; Schou et al. 2013; Shah et al. 2011; Stienen et al. 2018; Troughton et al. 2000). After extracting the end of trial mortality data, it was deemed inappropriate to conduct a meta-analysis, as the duration of the trial follow up periods differed between the trials. It was therefore impossible to conduct a meta-analysis at a fixed long-term time point.

Eight of the 18 trials (Anguita et al. 2010; Eurlings et al. 2010; Felker et al. 2017; Lainchbury et al. 2009; Maeder et al. 2013; Pfisterer et al. 2009; Shah et al. 2011; Stienen et al. 2018) reported mortality within the first 6 months of the intervention. Two trials reported mortality at 4 months, and 6 trials reported mortality at 6 months. NP-directed medical therapy was associated with a reduction in mortality within the first 6 months of the intervention (RR 0.82, 95% CI 0.68–0.99). Subgroup analysis suggested little heterogeneity between the 4-month and 6-month outcomes (Fig. 3).

Fig. 3
figure 3

Effect of natriuretic peptide guided medical therapy versus control on mortality after 6 months. CI-confidence interval

Full size image

Adverse events in relation to the change in therapy

Seven out of 18 studies recorded adverse effects of medical therapy on electrolytes and the cardiovascular system (Eurlings et al. 2010; Januzzi et al. 2011; Jourdain et al. 2007; Maeder et al. 2013; Murdoch et al. 1999; Pfisterer et al. 2009; Shah et al. 2011; Troughton et al. 2000). These were deemed not serious and six of these studies showed no difference in the incidence of adverse effects of therapy between the intervention and control groups.

The specific treatment strategies used in the trials

The treatment strategies and efficacy of these treatments varied between the trials. The two trials that showed the most benefit associated with NP-directed medical therapy,(Carubelli et al. 2016) and (Shah et al. 2011), showed efficacy for diuretics (the former) and angiotensin-converting enzyme inhibitors and beta-blockers (the latter). The two studies that showed the greatest number of patients reaching target NP levels, (Lainchbury et al. 2009) and (Karlstrom et al. 2011), showed that a combination of therapies was effective, including diuretics, angiotensin-converting enzyme inhibitor/angiotensin II receptor blockers and beta-blockers.

Discussion

The principal findings of this systematic review are that NP-directed medical therapy does not significantly reduce NP levels at 6 months after initiation of NP-directed medical therapy. However, NP-directed medical therapy may be associated with decreased mortality in the short term, and there is little heterogeneity for this finding.

Strengths

The strength of this review is that we evaluated the efficacy of NP-directed medical therapy from clinical trials, on the biochemical response of patient NP levels, and the clinically relevant outcome of mortality. The methodology of this systematic review and meta-analysis is robust.

Findings in relation to other studies

There is an important fundamental difference between this meta-analysis, and the other two meta-analyses that were published after the protocol for our meta-analysis was registered (Khan et al. 2018; Pufulete et al. 2017). The primary outcome of our meta-analysis was to evaluate if it was possible to decrease NP levels with NP-directed therapy, while the primary outcome of the other two meta-analyses was to determine if NP-directed medical therapy was associated with a survival benefit. Evaluation of a potential survival benefit was a secondary outcome in our meta-analysis. Our primary interest was to determine whether perioperative physicians could possibly decrease NP levels prior to elective surgery in patients with high NP levels (and thereby potentially improve the risk profile of poor surgical candidates). Both these meta-analyses also had point estimates favoring survival benefit with NP-directed therapy in the long term. The importance of our meta-analysis is that (i) a reduction in NP levels is not necessarily essential to demonstrate a survival benefit with NP-directed medical therapy, and (ii) that this survival benefit may be seen earlier than what has been previously documented. The utility of NP-directed medical therapy in preoperative surgical patients is unknown, as there are currently no surgical trials in this field. This meta-analysis suggests that there is potential utility in this approach in surgical patients.

Preoperative risk stratification of high-risk patients is advocated by international guidelines, the most recent being the Canadian Cardiovascular Society Guidelines on Perioperative Cardiac Risk Assessment and Management for Patients Who Undergo Noncardiac Surgery (Duceppe et al. 2017). Screening for natriuretic peptides is a key component of risk stratification (Duceppe et al. 2017).

Our meta-analysis suggests that there may be a further benefit to the reduction of NP levels prior to surgery. The survival benefit seen with NP-directed medical therapy in this meta-analysis may suggest that an intensification of medical therapy is warranted in patients with marked physiological derangement reflected by a markedly elevated NP level. These high NP levels may reflect some reversibility in volume status and myocyte ischemia which is responsive to further medical therapy. Indeed, the trials which demonstrated the greatest number of patients reaching a target NP level included a combination of therapies which would have had both volume and ischemia efficacy (Karlstrom et al. 2011; Lainchbury et al. 2009). The importance of this systematic review is the following. Firstly, these findings suggest that there is potential to improve survival for an elective surgical population through NP-directed medical therapy. Secondly, the perioperative period is a powerful modifier of risk, and decreasing this risk, has the potential to change morbidity and mortality up to a year after surgery (Puelacher et al. 2018).

Limitations

We were unable to obtain data from all the included trials for the SMD, the patients reaching the target NP and the time to NP reduction analysis. This is because most trials did not publish these end-points, nor was this included as part of the trial protocols. It is possible that if we had a larger sample which included data from all trials, then we may have shown an association between NP-directed medical therapy and a reduction in NP levels. However, it appears from this meta-analysis, that it is the intensification of medical therapy, rather than the reduction in NP levels, which may be important for short-term survival.

The non-parametric data for the SMD analysis was transformed to mean and standard deviation to facilitate analysis and caution should be taken when interpreting these results. The range of starting NP level on randomization in the intervention groups is large and could dramatically influence responsiveness to NP-directed therapy. However, despite these differences in the pre-intervention NP levels, mortality decreased in the NP-directed therapy arm, with little heterogeneity. It could be argued, however, that this early mortality (i.e., at 6 months of therapy initiation) signal is fragile. If a random-effects meta-analysis is conducted, then one cannot demonstrate a survival benefit associated with NP-directed medical therapy (RR 0.88, 95% 0.75–1.04, p = 0.14). Similarly, a sensitivity analysis which excludes all trials with a high risk of bias is not associated with a survival benefit (RR 0.84, 95% CI 0.61–1.15, p = 0.27) in a random-effects model. The survival benefit demonstrated in this meta-analysis therefore should be considered “hypothesis-generating” at best. It was not possible to control for the effect of age or renal function on NP for this analysis.

Finally, the included trials had very different implemented protocols, and thus it is not possible to identify a preferred medical management plan based on these data.

Future research

This systematic review provides support for a clinical trial of preoperative NP-directed medical therapy in high-risk elective surgical patients.

Conclusion

NP-directed medical therapy does not necessarily decrease NP levels, but it may be associated with a survival benefit. There may be a place for preoperative NP-directed medical therapy in high-risk surgical patients.