Article Text
PDF
Abstract
Background Currently, there is no head-to-head comparison of novel pharmacological treatments for heart failure with reduced ejection fraction (HFrEF). A network meta-analysis aimed to compare effects of both conventional and alternative drug combinations on time to develop primary composite outcome of cardiovascular death or heart failure hospitalisation (PCO).
Methods Randomised controlled trials (RCTs) were identified from Medline, Scopus up to June 2021. The RCTs were included if comparing any single or combination of drugs, that is, ACE inhibitors (ACEI), angiotensin receptor blockers, beta-blockers (BB), mineralocorticoid receptor antagonists (MRA), ivabradine (IVA), angiotensin receptor blocker/neprilysin inhibitors (ARNI) and sodium-glucose cotransporter-2 inhibitors (SGLT2i), soluble guanylyl cyclase and omecamtiv mecarbil and reporting PCO. Data were extracted from Kaplan-Meier curves, individual patient data were generated. A mixed-effect Weibull regression was applied. Median time to PCO, HRs with 95% CI were estimated accordingly. Our findings suggested that ACEI+BB+MRA+SGLT2i, BB+MRA+ARNI, and ACEI+BB+MRA+IVA had lower probability of PCOs than the conventional triple therapy (ACEI+BB+MRA).
Results Median time to PCOs of ACEI+BB+MRA was 57.7 months whereas median times to those new combinations were longer than 57.7 months. In addition, the three new regimens had a significantly lower PCO risks than ACEI+BB+MRA, with the HRs (95% CI) of 0.51 (0.43 to 0.61), 0.55 (0.46 to 0.65) and 0.56 (0.47 to 0.67), accordingly.
Conclusion This study suggested that SGLT2i, ARNI and IVA in addition to ACEI+BB+MRA may be better in prolonging time to develop PCO in HFrEF patients.
- meta-analysis
- pharmacology, clinical
- heart failure, systolic
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Triple therapy with ACE inhibitors (ACEI)/angiotensin II receptor blockers (ARB), beta blocker (BB) and mineralocorticoid receptor antagonist (MRA) has long been used as the standard therapy for heart failure with reduced ejection fraction (HFrEF).
As angiotensin receptor blocker/neprilysin inhibitors (ARNI) and sodium-glucose co-transporter 2 inhibitors (SGLT2i) emerged as new alternatives for HFrEF, the recent guidelines suggest the quadruple combination of ARNI/ACEI/ARB+BB+MRA+SGLT2i to reduce mortality in HFrEF patients.
WHAT THIS STUDY ADDS
Our findings suggest that SGLT2i addition to the conventional triple therapy of ACEI+BB+MRA or ACEI replacement with ARNI could prolong median time to develop cardiovascular death or heart failure hospitalisation, with the median times of longer than 57.7 months compared with the conventional triple therapy at 57.7 months.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Our study supported the current guideline of quadruple combination of guideline-directed medical therapy, but further postguideline evaluation should be conducted using real world study.
Introduction
Approximately 64.3 million people are living with heart failure (HF) worldwide, of which approximately 50% are HFs with reduced ejection fraction (HFrEF).1 Throughout several decades, many drug therapies have been proven as effective treatments in reducing hospitalisations due to HFrEF or all-cause and cardiovascular mortality including ACE inhibitors (ACEI), angiotensin II receptor blockers (ARB), beta blockers (BB) and mineralocorticoid receptor antagonist (MRA).2–7 Recently, the treatment options for HFrEF increased with the addition of angiotensin receptor/neprilysin inhibitors (ARNI) and ivabradine (IVA).8 9 Most recently, sodium-glucose co-transporter 2 inhibitors (SGLT2i), soluble guanylyl cyclase stimulator (sGCS) and omecamtiv mecarbil (OMM) have emerged as new treatments for HFrEF.10–13 SGLT2i is now integrated as one of the four pillars of guideline-directed medical therapy (GDMT) among those most recent guidelines.14 15
With this many pharmacological regimens available to choose from, there became challenges for clinicians to select the appropriate drug combinations for their patients. There were several studies that looked at the effects of each drug when used in combinations with the conventional triple therapy, these studies were then pooled by applying a network meta-analysis (NMA). We have performed a scoping review on this topic and found that 10 out of 25 NMAs were relevant.
Of them, six NMAs16–21 focused on alternative regimens with additional ARNI, SGLT2i, IVA, sGCS and OMM, whereas the other four NMAs focused on only ARNI,22 23 ARNI and IVA24, ARNI and SGLT2i.25 For those 6 NMAs considered more coverage alternative regimens, numbers of included randomised controlled trials (RCTs) varied from 6–75 due to type of control drugs and primary outcome of interests. For instance, number of all-cause mortalities ranged from 48 to 75; whereas smaller numbers from 6 to 1618 19 21 for the primary composite outcome (PCO) of cardiovascular death or HF hospitalisation (HFH). Relative treatment effects were estimated by pooling hazard ratios (HRs) of interventions relative to placebo,18 19 standard of care of GDMTs16 17 20 or both comparators.21 None of them considered time to PCO or survival time, except for Tromp et al,21 which included 75 relevant trials but used individual patient data from only 2 multicentre RCTs with 7376 patients. Since time to occurrence of PCO after receiving treatment was more clinically relevant and meaningful for clinicians to make decisions on the treatment of choice, this systematic review with NMA was, therefore, conducted to compare the time to develop PCO among all alternative regimens and GDMT in HFrEF patients.
Methods
A review protocol for this systematic review and NMA was registered on PROSPERO (registration number CRD42021262029). All steps were performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline.
Search strategy and study selection
A systematic search was performed on various electronic bibliographic databases including PubMed, Scopus, Cochrane Central Register of Controlled Trial (CENTRAL) and Web of Science from inception to June 2021 without language restriction. Search terms and strategies were constructed based on PICO and study design. Detailed search strategy and search term for each database were available in online supplemental table S1.
Supplemental material
Various synonyms and MeSH terms for HFrEF, pharmacotherapy and mortality were used to identify the relevant citations. Potential meta-analyses (MAs) or NMAs were identified and selected based on the inclusion criteria of MAs. The included RCTs of these selected MAs/NMAs were then chosen using specific eligibility criteria for individual RCTs. Furthermore, additional RCTs published after the most recent selected MAs that fitted the eligibility criteria were chosen. RCTs within MAs were chosen from inception to June 2021, while new RCTs were chosen from the latest search of the selected MAs (2018) to June 2021.
Eligibility criteria
Study selection was independently performed by two reviewers (TS and PC) as follows: First step: MAs were chosen if (1) including studies in only adult patients with HFrEF, (2) comparing any single or combination of drugs including ACEI, ARB, BB, MRA, SGLT2i, ARNI, OMM, IVA, sGCS and (3) reporting any outcome of interest defined in details below.
Second step: Individual RCTs included in previous MAs or additionally identified since the last search of MAs were chosen if (1) including studies in only adult patients with HFrEF, (2) comparing any single or combination of drugs including ACEI, ARB, BB, MRA, SGLT2i, ARNI, OMM, IVA, sGCS, (3) reporting any outcome of interest defined in details below and (4) containing sufficient data for pooling.
Intervention
Treatments of interest were all single or combination of drugs including ACEI, ARB, BB, MRA, SGLT2i, ARNI, OMM, IVA, sGCS regardless of dosages. The interested comparators were any combination of drugs including ACEI, ARB, BB, MRA, SGLT2i, ARNI, OMM, IVA, sGCS.
Outcomes
The primary outcome of interest was a time to develop PCO defined as a primary endpoint of cardiovascular death or heart failure hospitalisation. The secondary outcomes included cardiovascular death (CVD) and heart failure hospitalisation (HFH).
Data extraction
Two reviewers independently extracted the following data: publication year, country of origin, number of participants, study population covering sex, age, the New York Heart Association functional classification, mean left ventricular ejection fraction (LVEF), HF medications, comorbidities and outcome measures. For studies that did not report a SD for continous data, a calculated SD from the available data (eg, p values, t-values, CI or SEs) was done instead.
Data for pooling of dichotomous outcome, aggregated data between treatments and outcome were extracted if available, otherwise, summary statistic data along with 95% CI were extracted. For PCO, time-to-event data were extracted from Kaplan-Meier (KM) curves using a plot digitiser software (WebPlotDigitizer V.4.2), in which time and probability of failure (or event free) were extracted along with the number of interested events and corresponding patient at risk at each distinct time. In addition, summary statistical data of HRs along with 95% CIs were also extracted. Discrepancies between the two authors were solved by consensus with the author team.
Risk of bias assessment
The Cochrane Risk of Bias 2 (RoB 2.0) for RCTs was used in bias assessment, which consisted of five domains including randomisation process, deviation from intended interventions, missing outcome data, outcome measurements and selection of outcome reports. Studies were adjudged as high risk of bias if any domain was assessed as having a high risk of bias whereas they were judged as low risk of bias if all domains were low risks; otherwise, they were rated as some concerning risk of bias.
Statistical analysis
For the primary outcome of interest, that is, time to develop PCO, individual patient data were constructed for each RCT based on data extracted from the KM curve (ie, time, probability of survival, number of composite events along with patient at risk at each distinct time) using an algorithm by Wei and Royston.26 Then, these individual patient data were appended across all studies.
A one-stage NMA was applied using a mixed-effect parametric survival model with Weibull distribution and accelerated failure time.27 Survival time was regressed on treatment variables considering individual study as a random-effect and treatment as a fixed-effect. Median survival (or event-free) time for each treatment regimen and HRs comparing each treatment regimens along with 95% CIs were estimated accordingly. All analyses were performed using STATA V. 17.0. A p<0.05 was considered as statistically significant.
For secondary outcomes, that is, CVD and HFH, an NMA was performed by pooling lnHRs across studies using a consistency multivariate MA model with common between-study variance. Treatment comparisons in terms of HR along with their 95% CI were estimated afterwards. Because there is no loop in the network structure, inconsistency assumption could not be checked. Publication bias was assessed using a comparison-adjusted funnel plot.
Assessment of the confidence in NMA
The confidence in the primary outcome from the NMA was assessed using a web application tool of Confidence In Network Meta-Analysis (CINeMA).28 The following six domains were considered: within-study bias, reporting bias, indirectness, imprecision, heterogeneity and incoherence. Subsequently, the level of confidence in each comparison was rated as high, moderate, low or very low according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.
Results
Study selection
Two steps of searching were performed as follows: First, a systematic search of MAs was performed on PubMed, Scopus, Cochrane Central Register of Controlled Trial (CENTRAL) and Web of Science, which identified a total of 282 MAs after the removal of duplicates. Of those, 260 and 9 MAs were excluded in the screening and full-text review process respectively. This leaves only 13 MAs eligible for the study, which consisted of 247 RCTs in total. Second, a systematic search of individual RCTs was performed on the same databases from the latest search of the included MA to 2021 which retrieved additional 769 RCTs. After removal of duplicates, 50 RCTs were left for full-text review process after the screening of title/abstracts. Out of these, only 15 RCTs reported KM curves of the outcomes of interest (see figure 1). Of them, 11 RCTs reported PCO, 14 RCTs reported CVD and 14 RCTs reported HFH.
(A) PRISMA flow chart of a systematic search of meta-analyses. (B) PRISMA flow chart of a systematic search of RCTs schematic overview of study selection. PRISMA flow chart showing a systematic search of meta-analysis (A); and for randomised controlled trials (B). ACEI, ACE inhibitors; ARB, angiotensin II receptor blockers; ARNI, angiotensin receptor blocker/neprilysin inhibitor; BB, beta blockers; HFrEF, heart failure with reduced ejection fraction; IVA, ivabradine; MA, meta-analysis; MRA, mineralocorticoid receptor antagonist; NMA, network meta-analysis; OMM, omecamtiv mecarbil; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RCT, randomised controlled trials; sGCS, soluble guanylyl cyclase stimulator; SGLT2i, sodium-glucose co-transporter two inhibitors.
Study and patients characteristics
Characteristics of 15 RCTs with 51 280 patients are described in online supplemental table S2. Most patients were men with the mean age of 60–70 years and a mean LVEF was ranged from 23.0 to 37.7. The median follow-up period ranged from 15 to 48 months.
Risk of bias assessment
From 15 selected studies, only 2 (GALACTIC-HF, VIZZARDI), 1 (VIZZARDI) and 3 (PARADIGM-HF, PARALLEL-HF, ELTE II) had some concerns in the randomisation process, missing outcome data and selection bias of the reported results accordingly. Up to 10 studies had a low risk of bias in all 5 domains, which were considered as having an overall low risk of bias. On the other hand, the other five studies with some concerns in one or two domains were considered as having some concerns regarding the risk of bias overall (see online supplemental table S3).
Network meta-analysis
Primary composite outcome
A total of 11 RCTs (n=42 643) consisting of 8 interventions were included in the NMA of PCO (see figure 2). These treatment regimens included eight combinations, that is, ACEI+BB, ACEI+BB+ARB, ACEI+BB+MRA, BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+SGLT2i, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM. All RCTs had the conventional triple therapy (ACEI+BB+MRA) as the comparator, except one RCT which had ACEI+BB as the comparator. A mixed-effect Weibull regression was applied, and the failure curves plotting between cumulative probability of developing PCO and time after randomisation were drawn accordingly (see figure 3). The results showed that ACEI+BB+MRA+SGLT2i had the lowest failure probability of PCO occurrence, followed by BB+MRA+ARNI and ACEI+BB+MRA+IVA; whereas ACEI+BB had the highest probability of PCO occurrence, followed by ACEI+BB+ARB, ACEI+BB+MRA. The estimated median times to develop PCO for ACEI+BB and ACEI+BB+ARB were 37.2, 45.4, acrrodingly, and the median times to develop PCO for the conventional triple therapy of ACEI+BB+MRA was 57.7 months. However, the median times to develop PCO of the other drug combinations including ACEI+BB+MRA+SGLT2i, BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM were unable to be identified. In other words, their median times to PCO occurrence were longer than 57 months. The relative treatment effects estimated showed that all treatments were lowering risk of PCO relative to ACEI+BB, with the HRs (95% CI) of 0.51 (0.43 to 0.61), 0.55 (0.46 to 0.65) and 0.56 (0.47 to 0.67) for the first-three best treatments of ACEI+BB+MRA+SGLT2i, BB+MRA+ARNI and ACEI+BB+MRA+IVA (see table 1). In addition, the relative treatment effects of the alternative combinations of ACEI+BB+MRA+SGLT2i, BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM were significantly lowering risk of PCO occurrences when compared with the conventional triple therapy of ACEI+BB+MRA with the HRs (95% CI) of 0.75 (0.68 to 0.82), 0.80 (0.74 to 0.88), 0.82 (0.75 to 0.90), 0.91 (0.83 to 0.99) and 0.92 (0.86 to 0.99), respectively. Furthermore, comparing among the alternative regimens, ACEI+BB+MRA+IVA, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM had a higher PCO risk than BB+MRA+ARNI but only ACEI+BB+MRA+OMM had a significant result with HR of 1.15 (95% CI 1.03 to 1.28). On the other hand, ACEI+BB+MRA+SGLT2i had a 0.93-fold (95% CI 0.92 to 1.06) lower risk but this was not significant. Likewise, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM were at a higher PCO risk than ACEI+BB+MRA+IVA but none was significant. ACEI+BB+MRA+SGLT2i was 0.91-fold (95% CI 0.79 to 1.04) lower PCO risk but not significant as well. Finally, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM had a 1.21-fold (95% CI 1.06 to 1.38) and 1.24-fold (95% CI 1.10 to 1.39) higher risk of PCO than ACEI+BB+MRA+SGLT2i whereas both ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM were not significantly different. A comparison-adjusted funnel plot was constructed indicating symmetry of a funnel, although a few RCTs with high effect sizes but low precisions fell into the bottom (see online supplemental figure S1).
Network map generated by STATA V.17.0 shows the treatment regimens reporting primary composite outcome and Kaplan-Meier curve. The thickness of the lines represents the number of patients involved in the studies. ACEI, ACE inhibitors; ARB, angiotensin II receptor blockers; ARNI, angiotensin receptor blocker/neprilysin inhibitor; BB, beta blockers; IVA, ivabradine; MRA, mineralocorticoid receptor antagonist; OMM, omecamtiv mecarbil; sGCS, soluble guanylyl cyclase stimulator; SGLT2i, sodium-glucose co-transporter two inhibitors.
Cumulative probability curves of the primary composite outcome by treatment regimens. The curves generated by STATA V.17.0 using individual patient data show the probability of developing the primary composite outcome the cumulative probability of developing the primary composite outcome of cardiovascular death or heart failure hospitalisation of each treatment regimen. ACEI, ACE inhibitors; ARB, angiotensin II receptor blockers; ARNI, angiotensin receptor blocker/neprilysin inhibitor; BB, beta blockers; IVA, ivabradine; MRA, mineralocorticoid receptor antagonist; OMM, omecamtiv mecarbil; sGCS, soluble guanylyl cyclase stimulator; SGLT2i, sodium-glucose co-transporter two inhibitors.
Relative treatment effects for primary composite outcome
Cardiovascular death
A total of 14 RCTs (n=48 128) reported relative treatment effects as HRs of 9 treatments regiomen on cardiovascular death. A two-stage NMA with consistency model was applied to pool relative treatment effects between any two drug combinations (see table 2). Results suggested that BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+SGLT2i and ACEI+BB+MRA+sGCS could lower cardiovascular death but not significant when compared with the conventional triple therapy of ACEI+BB+MRA with the HRs of 0.82 (95% CI 0.66 to 1.02), 0.91 (95% CI 0.74 to 1.13), 0.87 (95% CI 0.72 to 1.04), 0.93 (95% CI 0.75 to 1.16); whereas effect of ACEI+BB+MRA+OMM was very null. When compared with the conventional triple therapy of ACEI+BB+MRA, BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+SGLT2i, ACEI+BB+MRA+sGCS, ACEI+BB+MRA+OMM were better in lowering CVD but not significant with the HRs of 0.82 (95% CI 0.66 to 1.02), 0.91 (95% CI 0.74 to 1.13), 0.87 (95% CI 0.72 to 1.04), 0.93 (95% CI 0.75 to 1.16) and 1.01 (95% CI 0.82 to 1.24), respectively.
Relative treatment effects for cardiovascular death
Heart failure hospitalisation
A total of 14 RCTs (n=51 150) reported relative treatment effects as HRs of 10 interventions on HFH. A two-stage NMA with consistency model was used to pool relative treatment effects between the alternative and conventional regimens (see table 3). Results suggested that the comparative HRs of treatment BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+SGLT2i, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM could lower HFH when compared with the GDMT of ACEI+BB+MRA with the HRs of 0.81 (95% CI 0.73 to 0.90), 0.81 (95% CI 0.74 to 0.90), 0.75 (95% CI 0.68 to 0.82), 0.90 (95% CI 0.83 to 0.98) and 0.92 (95% CI 0.85 to 0.99), respectively.
Relative treatment effects for HF hospitalisation
Level of evidence
The CINeMA tool was used to grade the level of confidence of finings for PCO (see online supplemental table S4). As for the network map in figure 2, there was no closed loop comparison, thus inconsistency was unable to assess. Therefore, this item was considered non-applicable. Of 28 comparisons, 85.7% of them were graded as moderate to high levels of evidence.
Discussion
This systematic review with NMA was conducted to assess the comparative cardiovascular outcomes of pharmacological interventions for chronic HFrEF concerning the time to develop PCO. Our results showed that ACEI+BB+MRA+SGLT2i is the most effective treatment regimen for reducing PCO followed by BB+MRA+ARNI and ACEI+BB+MRA+IVA, in which all three regimens could significantly prolong PCO occurrence to longer than 57 months whereas all other regimens had time to PCO occurrence ranged from 37 to 57 months. In addition, ACEI+BB+MRA+SGLT2i, BB+MRA+ARNI, ACEI+BB+MRA+IVA, ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM could lower PCO occurrence of about 8%–25% when compared with ACEI+BB+MRA. Among the quadruple regimens, ACEI+BB+MRA+SGLT2i was the best in lowering PCO occurrence than ACEI+BB+MRA+sGCS and ACEI+BB+MRA+OMM, but it was not significant when compared with BB+MRA+ARNI. The analysis of HFH showed a corresponding trend with the primary outcome. However, regarding CVD, no quadruple regimen significantly lowered CVD when compared with the conventional triple therapy of ACEI+BB+MRA. Despite this, BB+MRA+ARNI and ACEI+BB+MRA+SGLT2i showed promising results with 18% and 13% risk reduction accordingly.
Recently, there have been several HFrEF trials studying the efficacy of novel and alternative drugs including SGLT2i, sGCS, OMM and IVA in reducing the primary endpoint of CVD or HFH. The DAPA-HF10 and EMPEROR-Reduced11 studies showed that dapagliflozin and empagliflozin (SGLT2i) in combination with baseline triple therapy (ACEI+BB+ARB) reduced the PCO. Therapy with vericiguat (ie, sGCS) also reduced the risk of the PCO occurrence as highlighted in the VICTORIA study29, OMM and IVA also reduced the incidence of the primary endpoint as suggested by the GALACTIC-HF study13 and SHIFT study.30 The latest European Society of Cardiology (ESC) guideline14 and American College of Cardiology/American Heart Association (ACC/AHA)15 recommends using ACEI/ARNI, BB, MRA and SGLT2i to reduce mortality for all patients with chronic HFrEF, while several trials also suggested the use of sGCS, OMM and IVA to reduce mortality. However, it is not yet clear which combination of these drugs yields the most effective result in reducing PCO, CVD or HFH.
The comparative efficacy of ACEI+BB+MRA+SGLT2i and BB+MRA+ARNI is still not clear. It is possible that ACEI+BB+MRA+SGLT2i could reduce the risk of PCO and HFH more while BB+MRA+ARNI could reduce the risk of CVD more. Nevertheless, the superiority of ACEI+BB+MRA+SGLT2i over BB+MRA+ARNI in PCO reduction and vice versa in CVD risk reduction were both not significant, which was also corresponded with findings from previous NMAs.16 18 19 The comparative analysis of BB+MRA+ARNI treatment and the GDMT of ACEI+BB+MRA pooling results from the PARADIGM study8 and PARALLEL-HF study31 showed that ARNI is significantly superior to ACEI with significant risk reduction in all outcomes, suggesting the replacement of ACEI with ARNI in the conventional triple therapy. Furthermore, there was also the possible use of ARNI instead of ACEI in the treatment of ACEI+BB+MRA+SGLT2i which contributed to around 10% of the patients in the DAPA-HF study10 and 20% in the EMPEROR-reduced study.11 Since ACEI+BB+MRA+SGLT2i and BB+MRA+ARNI are the two most effective treatments according to the NMAs concerning all endpoints and ARNI is more effective than ACEI, ARNI+BB+MRA+SGLT2i might be the most effective combination. Findings from a large-scale NMA by Tromp et al21 showed that ACEI/ARNI+BB+MRA+SGLT2i could lower PCO occurrence of 64% (HR 0.36 (95% CI 0.29 to 0.46)), whereas ACEI+BB+MRA could reduce PCO risk of 42% (HR 0.58 (95% CI 0.47 to 0.71)); 95% CIs of these two relative treatment effects were not overlapped, which could imply that the two regimens were statistically significant. Our findings are in line with Tromp et al21 in terms of proving that the addition of SGLT2i and replacement of ACEI with ARNI in the conventional triple therapy of ACEI+BB+MRA is currently the best combination for the treatment of HFrEF.
The superiority of SGLT2i over other new generation drugs might be explained by a number of proposed mechanisms including systemic effects and myocardial effects.32 SGLT2i was shown to promote diuresis by causing a reduction in interstitial fluid more than intravascular fluid compared with a loop diuretic33, thus reducing neurohormonal stimulation triggered by intravascular volume loss when using loop diuretics.32 Other systemic effects included sympathetic nervous system inhibition29, glucose control and weight loss.30 32 Furthermore, SGLT2i also has direct effects on myocardium as it improved cardiac energy metabolism and prevented adverse cardiac remodelling.32
Our study has several strengths. First, we primarily aimed to estimate and compare time to PCO instead of HRs as other previous NMAs because it could provide information on time to PCO occurrence after receiving treatments. The time to PCO data were extracted from KM curves and converted to individual patient data for all included studies,9 31 34–40 the cumulative probability curve of PCO was then constructed. As a result, we were able to predict time to PCO for the combination of ACEI+BB, ACEI+BB+ARB and ACEI+BB+MRA. We focused on more recent treatment regimens by considering all alternative regimens with an addition of SGLT2i, ARNI, IVA, sGCS and OMM. We compared the combinations of these drugs with the conventional triple therapy instead of comparing with placebo. Thus, our study differs from previous NMAs.
However, there are some limitations that could not be avoided. First, this review included only studies reporting KM curves of time to PCO, CVD or HFH; so only recent or large trials which reported this data type of any outcome were eligible. This then excluded some notable trials, for example, the CIBIS study41 (ACEI vs ACEI+BB), ELITE-I study42 (ACEI vs ARB) and COPERNICUS study43 (ACEI vs ACEI+BB). Moreover, since this study aimed to compare the efficacy of different drug combinations the old trials which compared drug combinations with placebo were not included.
In summary, our findings suggested that adding SGLT2i to the conventional triple therapy of ACEI+BB+MRA may be the best treatment regimen in reducing the PCO with an even more promising results if ARNI is used instead of ACEI. Our findings are in line with and support the current ESC14 and ACC/AHA guideline15 suggestion which recommends ACEI/ARNI, BB, MRA and SGLT2i for the management of all chronic HFrEF patients.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
TS and PC contributed equally.
Contributors TS and PC were the guarantors of overall content and responsible for leading the project, performing systematic search, data extraction, data analysis. TY and TI was responsible for supervising clinical contents including search selection, data interpretation and discussion. AThakkinstian was responsible for supervising statistical analysis and ensure that all analytical data are correct. ATansawet and PN were responsible for statistical analysis.
Funding This study has been granted by the National Research Council of Thailand (NRCT: N42A640323).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.